H4.3 top features

Top feature 0 in H4.3: (feature 4296)

TOP ACTIVATIONS
MAX = 2.483

in
Token in
Feature activation+0.000
Orlando
Token Orlando
Feature activation+0.000
,
Token,
Feature activation+0.000
Fla
Token Fla
Feature activation+0.000
.,
Token.,
Feature activation+0.000
requires
Token requires
Feature activation+2.483
making
Token making
Feature activation+0.681
Muslim
Token Muslim
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
more
Token more
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ending
Tokenending
Feature activation+0.000
parental
Token parental
Feature activation+0.000
leave
Token leave
Feature activation+0.000
is
Token is
Feature activation+1.653
really
Token really
Feature activation+1.998
something
Token something
Feature activation+2.440
that
Token that
Feature activation+0.382
requires
Token requires
Feature activation+0.334
Congress
Token Congress
Feature activation+0.000
to
Token to
Feature activation+0.000
act
Token act
Feature activation+0.000
"
Token "
Feature activation+0.000
Rem
TokenRem
Feature activation+0.000
oving
Tokenoving
Feature activation+0.033
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
anyone
Token anyone
Feature activation+0.146
oving
Tokenoving
Feature activation+0.033
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
anyone
Token anyone
Feature activation+0.146
close
Token close
Feature activation+0.000
to
Token to
Feature activation+0.000
Rem
TokenRem
Feature activation+0.000
oving
Tokenoving
Feature activation+0.033
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
anyone
Token anyone
Feature activation+0.146
close
Token close
Feature activation+0.000
s
Tokens
Feature activation+0.000
existing
Token existing
Feature activation+0.000
coverage
Token coverage
Feature activation+0.000
area
Token area
Feature activation+0.000
is
Token is
Feature activation+2.091
not
Token not
Feature activation+2.320
allowed
Token allowed
Feature activation+1.002
.
Token.
Feature activation+0.000
Chatt
Token Chatt
Feature activation+0.000
ano
Tokenano
Feature activation+0.000
og
Tokenog
Feature activation+0.000
cotton
Token cotton
Feature activation+0.000
seed
Tokenseed
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.530
soil
Token soil
Feature activation+0.000
provides
Token provides
Feature activation+2.320
protein
Token protein
Feature activation+0.000
which
Token which
Feature activation+0.000
counter
Token counter
Feature activation+0.000
acts
Tokenacts
Feature activation+0.000
the
Token the
Feature activation+0.000
that
Token that
Feature activation+0.000
blocking
Token blocking
Feature activation+0.000
certain
Token certain
Feature activation+0.000
content
Token content
Feature activation+0.000
may
Token may
Feature activation+2.031
work
Token work
Feature activation+2.315
to
Token to
Feature activation+1.534
temporarily
Token temporarily
Feature activation+1.423
disrupt
Token disrupt
Feature activation+1.184
groups
Token groups
Feature activation+0.000
,
Token,
Feature activation+0.000
iping
Tokeniping
Feature activation+0.000
on
Token on
Feature activation+0.545
B
Token B
Feature activation+0.000
umble
Tokenumble
Feature activation+0.000
won
Token won
Feature activation+1.337
't
Token't
Feature activation+2.306
just
Token just
Feature activation+2.083
land
Token land
Feature activation+1.040
you
Token you
Feature activation+0.000
with
Token with
Feature activation+0.000
a
Token a
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
office
Token office
Feature activation+0.000
said
Token said
Feature activation+0.000
,
Token,
Feature activation+0.000
is
Token is
Feature activation+2.292
equivalent
Token equivalent
Feature activation+1.939
to
Token to
Feature activation+1.567
taking
Token taking
Feature activation+0.000
out
Token out
Feature activation+0.000
1
Token 1
Feature activation+0.220
her
Token her
Feature activation+0.000
office
Token office
Feature activation+0.000
at
Token at
Feature activation+0.000
Northwestern
Token Northwestern
Feature activation+0.000
University
Token University
Feature activation+0.000
is
Token is
Feature activation+2.245
Jacqu
Token Jacqu
Feature activation+0.000
eline
Tokeneline
Feature activation+0.000
Stevens
Token Stevens
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
anyone
Token anyone
Feature activation+0.146
close
Token close
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
.
Token.
Feature activation+0.000
"
Token "
Feature activation+0.000
Rem
TokenRem
Feature activation+0.000
oving
Tokenoving
Feature activation+0.033
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
,
Token,
Feature activation+0.000
said
Token said
Feature activation+0.000
dragging
Token dragging
Feature activation+1.123
the
Token the
Feature activation+0.000
shark
Token shark
Feature activation+0.000
was
Token was
Feature activation+2.218
âĢ
Token âĢ
Feature activation+0.947
ľ
Tokenľ
Feature activation+0.174
un
Tokenun
Feature activation+1.244
necessary
Tokennecessary
Feature activation+2.112
and
Token and
Feature activation+0.981
crushing
Token crushing
Feature activation+0.132
and
Token and
Feature activation+0.954
sn
Token sn
Feature activation+0.000
orting
Tokenorting
Feature activation+0.105
opioids
Token opioids
Feature activation+0.000
was
Token was
Feature activation+2.214
the
Token the
Feature activation+2.146
most
Token most
Feature activation+1.269
popular
Token popular
Feature activation+0.728
way
Token way
Feature activation+0.976
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
missile
Token missile
Feature activation+0.000
defence
Token defence
Feature activation+0.000
system
Token system
Feature activation+0.000
was
Token was
Feature activation+2.148
not
Token not
Feature activation+2.174
a
Token a
Feature activation+2.153
move
Token move
Feature activation+1.419
aimed
Token aimed
Feature activation+0.000
at
Token at
Feature activation+0.000
Russia
Token Russia
Feature activation+0.000
Transfer
Token Transfer
Feature activation+0.000
ring
Tokenring
Feature activation+0.000
a
Token a
Feature activation+0.000
Character
Token Character
Feature activation+0.000
would
Token would
Feature activation+1.997
allow
Token allow
Feature activation+2.165
the
Token the
Feature activation+0.000
player
Token player
Feature activation+0.000
to
Token to
Feature activation+0.000
keep
Token keep
Feature activation+0.000
their
Token their
Feature activation+0.000
the
Token the
Feature activation+0.000
natural
Token natural
Feature activation+0.000
laws
Token laws
Feature activation+0.000
of
Token of
Feature activation+0.000
man
Token man
Feature activation+0.000
is
Token is
Feature activation+2.160
another
Token another
Feature activation+1.814
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Howell
Token Howell
Feature activation+0.000
missile
Token missile
Feature activation+0.000
defence
Token defence
Feature activation+0.000
system
Token system
Feature activation+0.000
was
Token was
Feature activation+2.148
not
Token not
Feature activation+2.174
a
Token a
Feature activation+2.153
move
Token move
Feature activation+1.419
aimed
Token aimed
Feature activation+0.000
at
Token at
Feature activation+0.000
Russia
Token Russia
Feature activation+0.000
.
Token.
Feature activation+0.000
the
Token the
Feature activation+0.000
group
Token group
Feature activation+0.000
says
Token says
Feature activation+0.000
,
Token,
Feature activation+0.000
could
Token could
Feature activation+1.886
lead
Token lead
Feature activation+2.152
to
Token to
Feature activation+1.466
serious
Token serious
Feature activation+0.000
eye
Token eye
Feature activation+0.000
damage
Token damage
Feature activation+0.000
.
Token.
Feature activation+0.000

Top DFA by src position
MAX = 2.948

,
Token,
Feature activation-0.011
Top resid features:
told
Token told
Feature activation-0.021
Top resid features:
Yahoo
Token Yahoo
Feature activation-0.003
Top resid features:
News
Token News
Feature activation+0.000
Top resid features:
that
Token that
Feature activation-0.026
Top resid features:
stopping
Token stopping
Feature activation+2.948
Top resid features:
future
Token future
Feature activation+0.019
Top resid features:
atrocities
Token atrocities
Feature activation+0.003
Top resid features:
like
Token like
Feature activation+0.016
Top resid features:
the
Token the
Feature activation-0.008
Top resid features:
mass
Token mass
Feature activation-0.002
Top resid features:
Ċ
TokenĊ
Feature activation-0.004
Top resid features:
Ċ
TokenĊ
Feature activation-0.004
Top resid features:
âĢ
TokenâĢ
Feature activation-0.011
Top resid features:
ľ
Tokenľ
Feature activation-0.013
Top resid features:
Ext
TokenExt
Feature activation+0.008
Top resid features:
ending
Tokenending
Feature activation+2.933
Top resid features:
parental
Token parental
Feature activation-0.004
Top resid features:
leave
Token leave
Feature activation-0.006
Top resid features:
is
Token is
Feature activation-0.012
Top resid features:
really
Token really
Feature activation+0.055
Top resid features:
something
Token something
Feature activation-0.038
Top resid features:
mer
Tokenmer
Feature activation-0.001
Top resid features:
said
Token said
Feature activation+0.063
Top resid features:
.
Token.
Feature activation-0.001
Top resid features:
"
Token "
Feature activation+0.016
Top resid features:
Rem
TokenRem
Feature activation-0.013
Top resid features:
oving
Tokenoving
Feature activation+2.826
Top resid features:
them
Token them
Feature activation-0.002
Top resid features:
would
Token would
Feature activation+0.095
Top resid features:
have
Token have
Feature activation-0.070
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
mer
Tokenmer
Feature activation-0.002
Top resid features:
said
Token said
Feature activation+0.042
Top resid features:
.
Token.
Feature activation-0.001
Top resid features:
"
Token "
Feature activation+0.012
Top resid features:
Rem
TokenRem
Feature activation+0.001
Top resid features:
oving
Tokenoving
Feature activation+2.844
Top resid features:
them
Token them
Feature activation+0.023
Top resid features:
would
Token would
Feature activation+0.012
Top resid features:
have
Token have
Feature activation+0.038
Top resid features:
been
Token been
Feature activation+0.064
Top resid features:
a
Token a
Feature activation-0.131
Top resid features:
mer
Tokenmer
Feature activation-0.001
Top resid features:
said
Token said
Feature activation+0.051
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
"
Token "
Feature activation+0.013
Top resid features:
Rem
TokenRem
Feature activation-0.009
Top resid features:
oving
Tokenoving
Feature activation+2.800
Top resid features:
them
Token them
Feature activation+0.005
Top resid features:
would
Token would
Feature activation+0.010
Top resid features:
have
Token have
Feature activation+0.066
Top resid features:
been
Token been
Feature activation-0.082
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
broadband
Token broadband
Feature activation-0.000
Top resid features:
remains
Token remains
Feature activation-0.005
Top resid features:
legal
Token legal
Feature activation-0.004
Top resid features:
âĢĶ
TokenâĢĶ
Feature activation+0.007
Top resid features:
but
Tokenbut
Feature activation+0.038
Top resid features:
expanding
Token expanding
Feature activation+2.786
Top resid features:
it
Token it
Feature activation+0.016
Top resid features:
beyond
Token beyond
Feature activation+0.038
Top resid features:
the
Token the
Feature activation+0.011
Top resid features:
provider
Token provider
Feature activation-0.002
Top resid features:
âĢ
TokenâĢ
Feature activation+0.014
Top resid features:
in
Token in
Feature activation+0.007
Top resid features:
your
Token your
Feature activation+0.018
Top resid features:
marijuana
Token marijuana
Feature activation-0.003
Top resid features:
garden
Token garden
Feature activation+0.011
Top resid features:
,
Token,
Feature activation+0.037
Top resid features:
adding
Token adding
Feature activation+2.853
Top resid features:
gran
Token gran
Feature activation+0.001
Top resid features:
ulated
Tokenulated
Feature activation+0.002
Top resid features:
products
Token products
Feature activation+0.002
Top resid features:
made
Token made
Feature activation+0.025
Top resid features:
from
Token from
Feature activation+0.017
Top resid features:
Extrem
Token Extrem
Feature activation+0.001
Top resid features:
ism
Tokenism
Feature activation-0.003
Top resid features:
researchers
Token researchers
Feature activation+0.001
Top resid features:
said
Token said
Feature activation+0.018
Top resid features:
that
Token that
Feature activation+0.010
Top resid features:
blocking
Token blocking
Feature activation+2.764
Top resid features:
certain
Token certain
Feature activation-0.022
Top resid features:
content
Token content
Feature activation+0.004
Top resid features:
may
Token may
Feature activation+0.068
Top resid features:
work
Token work
Feature activation-0.094
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
that
Token that
Feature activation-0.001
Top resid features:
moment
Token moment
Feature activation+0.001
Top resid features:
.
Token.
Feature activation-0.001
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.028
Top resid features:
Sw
TokenSw
Feature activation-0.013
Top resid features:
iping
Tokeniping
Feature activation+2.845
Top resid features:
on
Token on
Feature activation+0.042
Top resid features:
B
Token B
Feature activation+0.001
Top resid features:
umble
Tokenumble
Feature activation+0.008
Top resid features:
won
Token won
Feature activation+0.060
Top resid features:
't
Token't
Feature activation-0.116
Top resid features:
potency
Token potency
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.009
Top resid features:
Ċ
TokenĊ
Feature activation-0.003
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
Rem
TokenRem
Feature activation-0.011
Top resid features:
oving
Tokenoving
Feature activation+2.767
Top resid features:
just
Token just
Feature activation+0.007
Top resid features:
one
Token one
Feature activation-0.018
Top resid features:
ton
Token ton
Feature activation-0.006
Top resid features:
of
Token of
Feature activation+0.001
Top resid features:
diesel
Token diesel
Feature activation-0.004
Top resid features:
ATION
TokenATION
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.001
Top resid features:
Ċ
TokenĊ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation-0.009
Top resid features:
Jo
TokenJo
Feature activation+0.013
Top resid features:
ining
Tokenining
Feature activation+2.831
Top resid features:
us
Token us
Feature activation+0.024
Top resid features:
now
Token now
Feature activation+0.006
Top resid features:
from
Token from
Feature activation+0.045
Top resid features:
her
Token her
Feature activation+0.015
Top resid features:
office
Token office
Feature activation-0.010
Top resid features:
mer
Tokenmer
Feature activation-0.001
Top resid features:
said
Token said
Feature activation+0.036
Top resid features:
.
Token.
Feature activation-0.002
Top resid features:
"
Token "
Feature activation+0.004
Top resid features:
Rem
TokenRem
Feature activation-0.002
Top resid features:
oving
Tokenoving
Feature activation+2.670
Top resid features:
them
Token them
Feature activation+0.005
Top resid features:
would
Token would
Feature activation+0.035
Top resid features:
have
Token have
Feature activation+0.036
Top resid features:
been
Token been
Feature activation+0.003
Top resid features:
a
Token a
Feature activation+0.056
Top resid features:
mer
Tokenmer
Feature activation-0.002
Top resid features:
said
Token said
Feature activation+0.055
Top resid features:
.
Token.
Feature activation-0.002
Top resid features:
"
Token "
Feature activation+0.017
Top resid features:
Rem
TokenRem
Feature activation-0.004
Top resid features:
oving
Tokenoving
Feature activation+2.833
Top resid features:
them
Token them
Feature activation-0.085
Top resid features:
would
Token would
Feature activation-0.090
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
Fishing
Token Fishing
Feature activation+0.001
Top resid features:
in
Token in
Feature activation-0.001
Top resid features:
Jupiter
Token Jupiter
Feature activation+0.001
Top resid features:
,
Token,
Feature activation+0.001
Top resid features:
said
Token said
Feature activation+0.072
Top resid features:
dragging
Token dragging
Feature activation+2.884
Top resid features:
the
Token the
Feature activation+0.053
Top resid features:
shark
Token shark
Feature activation-0.084
Top resid features:
was
Token was
Feature activation-0.219
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
car
Token car
Feature activation-0.005
Top resid features:
accidents
Token accidents
Feature activation-0.010
Top resid features:
than
Token than
Feature activation+0.023
Top resid features:
overdoses
Token overdoses
Feature activation-0.013
Top resid features:
,
Token,
Feature activation+0.011
Top resid features:
crushing
Token crushing
Feature activation+2.706
Top resid features:
and
Token and
Feature activation-0.029
Top resid features:
sn
Token sn
Feature activation-0.008
Top resid features:
orting
Tokenorting
Feature activation-0.001
Top resid features:
opioids
Token opioids
Feature activation-0.022
Top resid features:
was
Token was
Feature activation-0.107
Top resid features:
W
Token W
Feature activation-0.001
Top resid features:
am
Tokenam
Feature activation-0.001
Top resid features:
men
Tokenmen
Feature activation+0.001
Top resid features:
said
Token said
Feature activation+0.033
Top resid features:
that
Token that
Feature activation-0.010
Top resid features:
joining
Token joining
Feature activation+2.800
Top resid features:
the
Token the
Feature activation+0.009
Top resid features:
missile
Token missile
Feature activation+0.001
Top resid features:
defence
Token defence
Feature activation-0.003
Top resid features:
system
Token system
Feature activation-0.018
Top resid features:
was
Token was
Feature activation-0.017
Top resid features:
new
Token new
Feature activation-0.003
Top resid features:
factions
Token factions
Feature activation-0.001
Top resid features:
only
Token only
Feature activation-0.004
Top resid features:
.
Token.
Feature activation+0.020
Top resid features:
Transfer
Token Transfer
Feature activation+0.057
Top resid features:
ring
Tokenring
Feature activation+2.582
Top resid features:
a
Token a
Feature activation-0.033
Top resid features:
Character
Token Character
Feature activation-0.070
Top resid features:
would
Token would
Feature activation+0.086
Top resid features:
allow
Token allow
Feature activation-0.032
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
wrote
Token wrote
Feature activation+0.007
Top resid features:
.
Token.
Feature activation+0.015
Top resid features:
âĢ
Token âĢ
Feature activation+0.004
Top resid features:
ľ
Tokenľ
Feature activation-0.010
Top resid features:
Decl
TokenDecl
Feature activation-0.001
Top resid features:
aring
Tokenaring
Feature activation+2.686
Top resid features:
that
Token that
Feature activation+0.010
Top resid features:
homosexual
Token homosexual
Feature activation-0.005
Top resid features:
acts
Token acts
Feature activation+0.002
Top resid features:
violate
Token violate
Feature activation-0.038
Top resid features:
the
Token the
Feature activation+0.003
Top resid features:
W
Token W
Feature activation-0.001
Top resid features:
am
Tokenam
Feature activation-0.001
Top resid features:
men
Tokenmen
Feature activation+0.001
Top resid features:
said
Token said
Feature activation+0.029
Top resid features:
that
Token that
Feature activation+0.008
Top resid features:
joining
Token joining
Feature activation+2.777
Top resid features:
the
Token the
Feature activation+0.005
Top resid features:
missile
Token missile
Feature activation+0.001
Top resid features:
defence
Token defence
Feature activation-0.003
Top resid features:
system
Token system
Feature activation-0.022
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
being
Token being
Feature activation-0.001
Top resid features:
sold
Token sold
Feature activation-0.000
Top resid features:
online
Token online
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.008
Top resid features:
Bu
Token Bu
Feature activation+0.015
Top resid features:
ying
Tokenying
Feature activation+2.397
Top resid features:
inadequate
Token inadequate
Feature activation-0.038
Top resid features:
glasses
Token glasses
Feature activation-0.016
Top resid features:
,
Token,
Feature activation+0.032
Top resid features:
the
Token the
Feature activation+0.007
Top resid features:
group
Token group
Feature activation+0.009
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.05

Head 2: 0.03

Head 3: 0.52

Head 4: 0.04

Head 5: 0.03

Head 6: 0.08

Head 7: 0.08

Head 8: 0.03

Head 9: 0.05

Head 10: 0.02

Head 11: 0.06

Positive logits

quickShipAvailable1.86

FLAG1.61

��1.56

olean1.51

futile1.49

frowned1.46

counterproductive1.45

oggles1.45

aza1.42

unwanted1.40

disguise1.40

undesirable1.40

deterrent1.39

hind1.39

Allows1.37

chore1.36

pointless1.35

discouraged1.34

accent1.33

clipboard1.32

Negative logits

Jur-1.65

Zi-1.40

Zy-1.38

Hai-1.33

Kirin-1.32

arsen-1.31

assi-1.30

ahar-1.30

mol-1.29

Bridgewater-1.29

kar-1.27

Jal-1.27

ebted-1.27

capital-1.26

Lumin-1.26

vir-1.25

hail-1.25

Ragnar-1.25

Antar-1.23

Riy-1.23

INTERVAL 2.234 - 2.483
CONTAINS 0.001%

that
Token that
Feature activation+0.000
blocking
Token blocking
Feature activation+0.000
certain
Token certain
Feature activation+0.000
content
Token content
Feature activation+0.000
may
Token may
Feature activation+2.031
work
Token work
Feature activation+2.315
to
Token to
Feature activation+1.534
temporarily
Token temporarily
Feature activation+1.423
disrupt
Token disrupt
Feature activation+1.184
groups
Token groups
Feature activation+0.000
,
Token,
Feature activation+0.000
Rem
TokenRem
Feature activation+0.000
oving
Tokenoving
Feature activation+0.033
them
Token them
Feature activation+0.000
would
Token would
Feature activation+2.230
have
Token have
Feature activation+2.415
been
Token been
Feature activation+2.364
a
Token a
Feature activation+2.407
priority
Token priority
Feature activation+2.234
for
Token for
Feature activation+0.194
anyone
Token anyone
Feature activation+0.146
close
Token close
Feature activation+0.000
in
Token in
Feature activation+0.000
Orlando
Token Orlando
Feature activation+0.000
,
Token,
Feature activation+0.000
Fla
Token Fla
Feature activation+0.000
.,
Token.,
Feature activation+0.000
requires
Token requires
Feature activation+2.483
making
Token making
Feature activation+0.681
Muslim
Token Muslim
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
more
Token more
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ending
Tokenending
Feature activation+0.000
parental
Token parental
Feature activation+0.000
leave
Token leave
Feature activation+0.000
is
Token is
Feature activation+1.653
really
Token really
Feature activation+1.998
something
Token something
Feature activation+2.440
that
Token that
Feature activation+0.382
requires
Token requires
Feature activation+0.334
Congress
Token Congress
Feature activation+0.000
to
Token to
Feature activation+0.000
act
Token act
Feature activation+0.000
s
Tokens
Feature activation+0.000
existing
Token existing
Feature activation+0.000
coverage
Token coverage
Feature activation+0.000
area
Token area
Feature activation+0.000
is
Token is
Feature activation+2.091
not
Token not
Feature activation+2.320
allowed
Token allowed
Feature activation+1.002
.
Token.
Feature activation+0.000
Chatt
Token Chatt
Feature activation+0.000
ano
Tokenano
Feature activation+0.000
og
Tokenog
Feature activation+0.000

INTERVAL 1.986 - 2.234
CONTAINS 0.005%

in
Token in
Feature activation+0.000
between
Token between
Feature activation+0.000
the
Token the
Feature activation+0.000
teeth
Token teeth
Feature activation+0.000
is
Token is
Feature activation+2.021
a
Token a
Feature activation+2.055
tedious
Token tedious
Feature activation+2.056
task
Token task
Feature activation+2.098
.
Token.
Feature activation+0.360
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
existing
Token existing
Feature activation+0.000
coverage
Token coverage
Feature activation+0.000
area
Token area
Feature activation+0.000
is
Token is
Feature activation+2.091
not
Token not
Feature activation+2.320
allowed
Token allowed
Feature activation+1.002
.
Token.
Feature activation+0.000
Chatt
Token Chatt
Feature activation+0.000
ano
Tokenano
Feature activation+0.000
bone
Token bone
Feature activation+0.000
wasn
Token wasn
Feature activation+0.918
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.247
enough
Token enough
Feature activation+2.081
.
Token.
Feature activation+0.081
On
Token On
Feature activation+0.000
Thursday
Token Thursday
Feature activation+0.000
afternoon
Token afternoon
Feature activation+0.000
,
Token,
Feature activation+0.000
joining
Token joining
Feature activation+0.110
the
Token the
Feature activation+0.000
missile
Token missile
Feature activation+0.000
defence
Token defence
Feature activation+0.000
system
Token system
Feature activation+0.000
was
Token was
Feature activation+2.148
not
Token not
Feature activation+2.174
a
Token a
Feature activation+2.153
move
Token move
Feature activation+1.419
aimed
Token aimed
Feature activation+0.000
at
Token at
Feature activation+0.000
oving
Tokenoving
Feature activation+0.000
me
Token me
Feature activation+0.000
from
Token from
Feature activation+0.000
this
Token this
Feature activation+0.000
location
Token location
Feature activation+0.000
will
Token will
Feature activation+1.988
rob
Token rob
Feature activation+1.797
my
Token my
Feature activation+0.000
message
Token message
Feature activation+0.000
of
Token of
Feature activation+0.000
contextual
Token contextual
Feature activation+0.000

INTERVAL 1.738 - 1.986
CONTAINS 0.011%

and
Token and
Feature activation+0.000
moving
Token moving
Feature activation+0.000
through
Token through
Feature activation+0.000
tight
Token tight
Feature activation+0.000
spaces
Token spaces
Feature activation+0.000
can
Token can
Feature activation+1.830
be
Token be
Feature activation+2.025
like
Token like
Feature activation+1.711
trying
Token trying
Feature activation+0.000
to
Token to
Feature activation+0.000
navigate
Token navigate
Feature activation+0.000
ting
Tokenting
Feature activation+0.000
individual
Token individual
Feature activation+0.000
ism
Tokenism
Feature activation+0.000
does
Token does
Feature activation+1.514
not
Token not
Feature activation+1.505
mean
Token mean
Feature activation+1.846
that
Token that
Feature activation+0.208
everyone
Token everyone
Feature activation+0.000
must
Token must
Feature activation+0.000
be
Token be
Feature activation+0.000
identical
Token identical
Feature activation+0.000
the
Token the
Feature activation+0.000
two
Token two
Feature activation+0.000
films
Token films
Feature activation+0.000
would
Token would
Feature activation+1.758
be
Token be
Feature activation+1.877
problematic
Token problematic
Feature activation+1.829
and
Token and
Feature activation+0.034
highly
Token highly
Feature activation+0.000
unlikely
Token unlikely
Feature activation+0.590
to
Token to
Feature activation+0.194
happen
Token happen
Feature activation+0.095
in
Token in
Feature activation+0.000
front
Token front
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
FCC
Token FCC
Feature activation+0.000
is
Token is
Feature activation+1.867
essential
Token essential
Feature activation+1.920
for
Token for
Feature activation+0.831
them
Token them
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
.
Token.
Feature activation+0.000
Getting
Token Getting
Feature activation+0.000
the
Token the
Feature activation+0.000
work
Token work
Feature activation+0.000
done
Token done
Feature activation+0.000
becomes
Token becomes
Feature activation+1.755
more
Token more
Feature activation+1.712
important
Token important
Feature activation+1.841
than
Token than
Feature activation+0.906
yourself
Token yourself
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 1.490 - 1.738
CONTAINS 0.018%

phosph
Token phosph
Feature activation+0.000
at
Tokenat
Feature activation+0.000
ase
Tokenase
Feature activation+0.000
inhibitor
Token inhibitor
Feature activation+0.000
transm
Token transm
Feature activation+0.834
its
Tokenits
Feature activation+1.643
signals
Token signals
Feature activation+0.210
for
Token for
Feature activation+0.000
cell
Token cell
Feature activation+0.000
division
Token division
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+1.102
cars
Token cars
Feature activation+0.453
is
Token is
Feature activation+1.961
more
Token more
Feature activation+1.620
dangerous
Token dangerous
Feature activation+1.204
than
Token than
Feature activation+0.123
traffic
Token traffic
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
1
Token1
Feature activation+0.000
is
Token is
Feature activation+1.522
NOT
Token NOT
Feature activation+1.419
a
Token a
Feature activation+1.498
good
Token good
Feature activation+1.721
idea
Token idea
Feature activation+1.636
.
Token.
Feature activation+0.047
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
For
TokenFor
Feature activation+0.000
those
Token those
Feature activation+0.000
so
Token so
Feature activation+0.000
upgrading
Token upgrading
Feature activation+0.107
early
Token early
Feature activation+0.338
is
Token is
Feature activation+1.583
not
Token not
Feature activation+1.577
a
Token a
Feature activation+1.603
way
Token way
Feature activation+1.376
out
Token out
Feature activation+0.214
.
Token.
Feature activation+0.000
minim
Token minim
Feature activation+0.000
ise
Tokenise
Feature activation+0.000
the
Token the
Feature activation+0.000
US
Token US
Feature activation+0.000
Open
Token Open
Feature activation+0.000
Cup
Token Cup
Feature activation+0.000
could
Token could
Feature activation+1.876
help
Token help
Feature activation+1.721
him
Token him
Feature activation+0.000
earn
Token earn
Feature activation+0.000
a
Token a
Feature activation+0.000
pro
Token pro
Feature activation+0.000
contract
Token contract
Feature activation+0.000

INTERVAL 1.241 - 1.490
CONTAINS 0.020%

emotions
Token emotions
Feature activation+0.000
,
Token,
Feature activation+0.000
playing
Token playing
Feature activation+0.000
against
Token against
Feature activation+0.000
Ireland
Token Ireland
Feature activation+0.000
is
Token is
Feature activation+1.477
not
Token not
Feature activation+1.506
an
Token an
Feature activation+1.184
em
Token em
Feature activation+0.000
to
Tokento
Feature activation+0.000
ional
Tokenional
Feature activation+0.000
onwards
Token onwards
Feature activation+0.000
into
Token into
Feature activation+0.000
a
Token a
Feature activation+0.000
hex
Token hex
Feature activation+0.000
editor
Token editor
Feature activation+0.000
reveals
Token reveals
Feature activation+1.266
this
Token this
Feature activation+0.000
image
Token image
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.000
confirms
Token confirms
Feature activation+0.000
,
Token,
Feature activation+0.000
spotting
Token spotting
Feature activation+0.000
such
Token such
Feature activation+0.000
errors
Token errors
Feature activation+0.000
can
Token can
Feature activation+1.106
easily
Token easily
Feature activation+1.285
be
Token be
Feature activation+1.394
done
Token done
Feature activation+0.915
by
Token by
Feature activation+0.310
a
Token a
Feature activation+0.008
beginner
Token beginner
Feature activation+0.000
25
Token 25
Feature activation+0.000
years
Token years
Feature activation+0.000
.
Token.
Feature activation+0.000
Talking
Token Talking
Feature activation+0.000
is
Token is
Feature activation+0.920
not
Token not
Feature activation+1.333
the
Token the
Feature activation+1.160
answer
Token answer
Feature activation+1.425
!"
Token!"
Feature activation+0.000
Trump
Token Trump
Feature activation+0.000
tweeted
Token tweeted
Feature activation+0.000
that
Token that
Feature activation+0.000
banning
Token banning
Feature activation+0.000
the
Token the
Feature activation+0.000
Confederate
Token Confederate
Feature activation+0.000
flag
Token flag
Feature activation+0.000
is
Token is
Feature activation+1.479
anti
Token anti
Feature activation+1.410
-
Token-
Feature activation+0.000
American
TokenAmerican
Feature activation+0.000
.
Token.
Feature activation+0.000
Nonetheless
Token Nonetheless
Feature activation+0.000

INTERVAL 0.993 - 1.241
CONTAINS 0.021%

over
Token over
Feature activation+0.000
the
Token the
Feature activation+0.000
internet
Token internet
Feature activation+0.351
will
Token will
Feature activation+1.431
make
Token make
Feature activation+1.543
up
Token up
Feature activation+1.130
for
Token for
Feature activation+0.000
that
Token that
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Adding
Token Adding
Feature activation+0.000
another
Token another
Feature activation+0.000
thing
Token thing
Feature activation+0.000
on
Token on
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+1.120
just
Token just
Feature activation+1.349
crazy
Token crazy
Feature activation+1.593
.
Token.
Feature activation+0.041
It
Token It
Feature activation+0.000
's
Token's
Feature activation+0.000
vel
Tokenvel
Feature activation+0.000
ing
Tokening
Feature activation+0.000
wet
Token wet
Feature activation+0.000
compost
Token compost
Feature activation+0.000
is
Token is
Feature activation+0.292
a
Token a
Feature activation+1.041
chore
Token chore
Feature activation+1.111
.
Token.
Feature activation+0.000
(
Token (
Feature activation+0.000
Well
TokenWell
Feature activation+0.000
,
Token,
Feature activation+0.000
feeling
Token feeling
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.000
feels
Token feels
Feature activation+1.021
like
Token like
Feature activation+1.009
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.056
is
Token is
Feature activation+1.480
not
Token not
Feature activation+1.689
push
Token push
Feature activation+1.620
-
Token-
Feature activation+0.000
button
Tokenbutton
Feature activation+0.000
simple
Token simple
Feature activation+1.082
.
Token.
Feature activation+0.285
It
Token It
Feature activation+0.000
takes
Token takes
Feature activation+0.000
a
Token a
Feature activation+0.000
long
Token long
Feature activation+0.000

INTERVAL 0.745 - 0.993
CONTAINS 0.021%

in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
city
Token city
Feature activation+0.000
is
Token is
Feature activation+1.586
a
Token a
Feature activation+1.588
top
Token top
Feature activation+0.917
priority
Token priority
Feature activation+0.650
.
Token.
Feature activation+0.000
Authorities
Token Authorities
Feature activation+0.000
say
Token say
Feature activation+0.000
they
Token they
Feature activation+0.000
work
Token work
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
mix
Token mix
Feature activation+0.000
could
Token could
Feature activation+1.393
make
Token make
Feature activation+0.781
your
Token your
Feature activation+0.000
kitchen
Token kitchen
Feature activation+0.000
even
Token even
Feature activation+0.000
smarter
Token smarter
Feature activation+0.000
.
Token.
Feature activation+0.000
reasonably
Token reasonably
Feature activation+0.000
secure
Token secure
Feature activation+0.000
local
Token local
Feature activation+0.000
economy
Token economy
Feature activation+0.000
can
Token can
Feature activation+0.918
still
Token still
Feature activation+0.839
despair
Token despair
Feature activation+0.066
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
future
Token future
Feature activation+0.000
.
Token.
Feature activation+0.000
.
Token.
Feature activation+0.000
Summon
Token Summon
Feature activation+0.000
ing
Tokening
Feature activation+0.000
would
Token would
Feature activation+0.414
eventually
Token eventually
Feature activation+0.747
become
Token become
Feature activation+0.983
more
Token more
Feature activation+0.543
orche
Token orche
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
NEW
TokenNEW
Feature activation+0.000
DEL
Token DEL
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+0.833
not
Token not
Feature activation+1.067
necessarily
Token necessarily
Feature activation+1.045
equate
Token equate
Feature activation+1.061
to
Token to
Feature activation+0.803
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000
a
Tokena
Feature activation+0.000
lot
Token lot
Feature activation+0.000
of
Token of
Feature activation+0.000

INTERVAL 0.497 - 0.745
CONTAINS 0.034%

cing
Tokencing
Feature activation+0.000
issues
Token issues
Feature activation+0.000
shouldn
Token shouldn
Feature activation+0.554
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.504
be
Token be
Feature activation+0.966
as
Token as
Feature activation+0.268
big
Token big
Feature activation+0.000
a
Token a
Feature activation+0.096
problem
Token problem
Feature activation+0.000
t
Tokent
Feature activation+1.944
be
Token be
Feature activation+2.139
much
Token much
Feature activation+1.605
of
Token of
Feature activation+0.285
a
Token a
Feature activation+1.292
stretch
Token stretch
Feature activation+0.498
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
For
TokenFor
Feature activation+0.000
affected
Token affected
Feature activation+0.000
volcanic
Token volcanic
Feature activation+0.000
activity
Token activity
Feature activation+0.000
of
Token of
Feature activation+0.000
this
Token this
Feature activation+0.000
location
Token location
Feature activation+0.000
is
Token is
Feature activation+0.565
therefore
Token therefore
Feature activation+0.571
crucial
Token crucial
Feature activation+0.768
for
Token for
Feature activation+0.000
predicting
Token predicting
Feature activation+0.000
future
Token future
Feature activation+0.000
ifting
Tokenifting
Feature activation+0.000
of
Token of
Feature activation+0.000
white
Token white
Feature activation+0.000
has
Token has
Feature activation+0.192
completely
Token completely
Feature activation+0.360
filled
Token filled
Feature activation+0.653
in
Token in
Feature activation+0.000
our
Token our
Feature activation+0.000
tracks
Token tracks
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
backing
Token backing
Feature activation+0.000
of
Token of
Feature activation+0.000
African
Token African
Feature activation+0.000
nations
Token nations
Feature activation+0.000
,
Token,
Feature activation+0.000
will
Token will
Feature activation+0.666
fall
Token fall
Feature activation+0.608
under
Token under
Feature activation+0.000
the
Token the
Feature activation+0.000
command
Token command
Feature activation+0.000
of
Token of
Feature activation+0.000

INTERVAL 0.248 - 0.497
CONTAINS 0.069%

neighbours
Token neighbours
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
conf
Tokenconf
Feature activation+0.016
irms
Tokenirms
Feature activation+0.253
beyond
Token beyond
Feature activation+0.000
any
Token any
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
Qatar
Token Qatar
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
work
Token work
Feature activation+0.000
was
Token was
Feature activation+1.166
no
Token no
Feature activation+1.279
small
Token small
Feature activation+1.383
task
Token task
Feature activation+1.180
.
Token.
Feature activation+0.250
"
Token "
Feature activation+0.000
It
TokenIt
Feature activation+0.000
was
Token was
Feature activation+0.000
so
Token so
Feature activation+0.000
intense
Token intense
Feature activation+0.000
understand
Token understand
Feature activation+0.000
that
Token that
Feature activation+0.000
file
Token file
Feature activation+0.000
sharing
Token sharing
Feature activation+0.000
is
Token is
Feature activation+0.000
bad
Token bad
Feature activation+0.344
,
Token,
Feature activation+0.067
and
Token and
Feature activation+0.000
this
Token this
Feature activation+0.000
is
Token is
Feature activation+0.000
a
Token a
Feature activation+0.000
problems
Token problems
Feature activation+0.000
.
Token.
Feature activation+0.000
And
Token And
Feature activation+0.000
prohibition
Token prohibition
Feature activation+0.000
may
Token may
Feature activation+0.000
deter
Token deter
Feature activation+0.317
people
Token people
Feature activation+0.000
with
Token with
Feature activation+0.000
drug
Token drug
Feature activation+0.000
use
Token use
Feature activation+0.000
disorders
Token disorders
Feature activation+0.000
looked
Token looked
Feature activation+0.000
up
Token up
Feature activation+0.000
to
Token to
Feature activation+0.000
coming
Token coming
Feature activation+0.000
up
Token up
Feature activation+0.000
and
Token and
Feature activation+0.285
even
Token even
Feature activation+0.000
playing
Token playing
Feature activation+0.000
against
Token against
Feature activation+0.000
.
Token.
Feature activation+0.000
So
Token So
Feature activation+0.000

INTERVAL 0.000 - 0.248
CONTAINS 99.799%

Military
Token Military
Feature activation+0.000
Journal
Token Journal
Feature activation+0.000
,
Token,
Feature activation+0.000
E
Token E
Feature activation+0.000
wing
Tokenwing
Feature activation+0.000
mentions
Token mentions
Feature activation+0.000
that
Token that
Feature activation+0.000
George
Token George
Feature activation+0.000
Washington
Token Washington
Feature activation+0.000
himself
Token himself
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
th
Tokenth
Feature activation+0.000
overall
Token overall
Feature activation+0.000
pick
Token pick
Feature activation+0.000
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
It
TokenIt
Feature activation+0.000
has
Token has
Feature activation+0.000
been
Token been
Feature activation+0.000
a
Token a
Feature activation+0.000
roller
Token roller
Feature activation+0.000
co
Tokenco
Feature activation+0.000
room
Token room
Feature activation+0.000
she
Token she
Feature activation+0.000
was
Token was
Feature activation+0.000
in
Token in
Feature activation+0.000
.
Token.
Feature activation+0.000
She
Token She
Feature activation+0.000
stood
Token stood
Feature activation+0.000
up
Token up
Feature activation+0.000
and
Token and
Feature activation+0.000
d
Token d
Feature activation+0.000
usted
Tokenusted
Feature activation+0.000
Mech
Token Mech
Feature activation+0.000
War
TokenWar
Feature activation+0.000
rior
Tokenrior
Feature activation+0.000
Online
Token Online
Feature activation+0.000
,
Token,
Feature activation+0.000
as
Token as
Feature activation+0.000
getting
Token getting
Feature activation+0.000
blown
Token blown
Feature activation+0.000
to
Token to
Feature activation+0.505
pieces
Token pieces
Feature activation+0.102
shortly
Token shortly
Feature activation+0.000
("
Token ("
Feature activation+0.000
Give
TokenGive
Feature activation+0.000
us
Token us
Feature activation+0.000
back
Token back
Feature activation+0.000
our
Token our
Feature activation+0.000
country
Token country
Feature activation+0.000
,"
Token,"
Feature activation+0.000
he
Token he
Feature activation+0.000
bell
Token bell
Feature activation+0.000
owed
Tokenowed
Feature activation+0.000
to
Token to
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 1 in H4.3: (feature 20664)

TOP ACTIVATIONS
MAX = 1.855

.
Token.
Feature activation+1.007
Pat
Token Pat
Feature activation+0.000
more
Tokenmore
Feature activation+0.000
t
Token t
Feature activation+1.078
uss
Tokenuss
Feature activation+0.111
les
Tokenles
Feature activation+1.855
with
Token with
Feature activation+1.548
the
Token the
Feature activation+0.321
new
Token new
Feature activation+0.000
mixer
Token mixer
Feature activation+0.000
,
Token,
Feature activation+0.000
very
Token very
Feature activation+0.000
day
Token day
Feature activation+0.000
,
Token,
Feature activation+0.624
never
Token never
Feature activation+1.346
really
Token really
Feature activation+0.613
liked
Token liked
Feature activation+1.728
Mole
Token Mole
Feature activation+0.000
very
Token very
Feature activation+0.000
much
Token much
Feature activation+0.000
because
Token because
Feature activation+0.141
he
Token he
Feature activation+0.187
to
Token to
Feature activation+0.000
prevent
Token prevent
Feature activation+0.000
his
Token his
Feature activation+0.000
mother
Token mother
Feature activation+0.000
from
Token from
Feature activation+1.441
being
Token being
Feature activation+1.639
killed
Token killed
Feature activation+1.198
,
Token,
Feature activation+0.070
only
Token only
Feature activation+0.000
to
Token to
Feature activation+0.000
come
Token come
Feature activation+0.000
of
Token of
Feature activation+1.106
the
Token the
Feature activation+0.300
Green
Token Green
Feature activation+0.000
party
Token party
Feature activation+0.000
,
Token,
Feature activation+0.033
said
Token said
Feature activation+1.557
:
Token:
Feature activation+0.676
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
shameless
Token shameless
Feature activation+0.000
Pat
Token Pat
Feature activation+0.000
more
Tokenmore
Feature activation+0.000
t
Token t
Feature activation+1.078
uss
Tokenuss
Feature activation+0.111
les
Tokenles
Feature activation+1.855
with
Token with
Feature activation+1.548
the
Token the
Feature activation+0.321
new
Token new
Feature activation+0.000
mixer
Token mixer
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
to
Token to
Feature activation+0.427
be
Token be
Feature activation+0.149
identified
Token identified
Feature activation+0.235
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
said
Token said
Feature activation+1.540
:
Token:
Feature activation+0.118
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
first
Token first
Feature activation+0.000
orie
Tokenorie
Feature activation+0.000
Kelly
Token Kelly
Feature activation+0.500
We
Token We
Feature activation+0.000
er
Tokener
Feature activation+0.000
,
Token,
Feature activation+1.115
took
Token took
Feature activation+1.517
to
Token to
Feature activation+0.000
Facebook
Token Facebook
Feature activation+0.288
to
Token to
Feature activation+0.000
confirm
Token confirm
Feature activation+0.665
that
Token that
Feature activation+0.643
interview
Token interview
Feature activation+0.000
for
Token for
Feature activation+0.000
10
Token 10
Feature activation+0.000
days
Token days
Feature activation+0.000
,
Token,
Feature activation+0.088
showed
Token showed
Feature activation+1.496
only
Token only
Feature activation+0.000
one
Token one
Feature activation+0.000
line
Token line
Feature activation+0.000
,"
Token,"
Feature activation+0.000
he
Token he
Feature activation+0.000
known
Token known
Feature activation+0.000
sighting
Token sighting
Feature activation+0.000
of
Token of
Feature activation+0.000
Ms
Token Ms
Feature activation+0.000
Con
Token Con
Feature activation+0.658
lon
Tokenlon
Feature activation+1.486
was
Token was
Feature activation+0.000
in
Token in
Feature activation+0.000
January
Token January
Feature activation+0.000
2015
Token 2015
Feature activation+0.000
.
Token.
Feature activation+0.000
br
Tokenbr
Feature activation+0.000
ina
Tokenina
Feature activation+0.000
Fulton
Token Fulton
Feature activation+0.317
,
Token,
Feature activation+1.042
will
Token will
Feature activation+1.120
testify
Token testify
Feature activation+1.483
.
Token.
Feature activation+0.217
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
Russian
TokenRussian
Feature activation+0.000
school
Token school
Feature activation+0.000
girl
Tokengirl
Feature activation+0.000
reference
Token reference
Feature activation+0.000
to
Token to
Feature activation+0.000
Mrs
Token Mrs
Feature activation+0.000
Kir
Token Kir
Feature activation+0.909
ch
Tokench
Feature activation+1.245
ner
Tokenner
Feature activation+1.476
and
Token and
Feature activation+0.898
her
Token her
Feature activation+1.013
late
Token late
Feature activation+0.365
husband
Token husband
Feature activation+0.000
,
Token,
Feature activation+0.000
wife
Token wife
Feature activation+0.000
,
Token,
Feature activation+0.896
Vik
Token Vik
Feature activation+0.000
ki
Tokenki
Feature activation+0.000
,
Token,
Feature activation+0.014
died
Token died
Feature activation+1.473
unexpectedly
Token unexpectedly
Feature activation+0.100
of
Token of
Feature activation+0.000
natural
Token natural
Feature activation+0.000
causes
Token causes
Feature activation+0.000
at
Token at
Feature activation+0.000
Rachel
Token Rachel
Feature activation+0.000
Not
Token Not
Feature activation+0.000
ley
Tokenley
Feature activation+0.387
âĢ
TokenâĢ
Feature activation+0.286
Ļ
TokenĻ
Feature activation+0.519
s
Tokens
Feature activation+1.453
NDP
Token NDP
Feature activation+0.256
.
Token.
Feature activation+0.040
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
PC
TokenPC
Feature activation+0.000
time
Token time
Feature activation+0.000
to
Token to
Feature activation+0.000
prevent
Token prevent
Feature activation+0.000
his
Token his
Feature activation+0.000
mother
Token mother
Feature activation+0.000
from
Token from
Feature activation+1.441
being
Token being
Feature activation+1.639
killed
Token killed
Feature activation+1.198
,
Token,
Feature activation+0.070
only
Token only
Feature activation+0.000
to
Token to
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
His
TokenHis
Feature activation+0.000
mother
Token mother
Feature activation+0.000
,
Token,
Feature activation+0.954
who
Token who
Feature activation+1.270
has
Token has
Feature activation+1.403
drinking
Token drinking
Feature activation+0.552
problems
Token problems
Feature activation+0.000
up
Token up
Feature activation+0.000
to
Token to
Feature activation+0.000
this
Token this
Feature activation+0.000
to
Token to
Feature activation+0.000
prevent
Token prevent
Feature activation+0.000
his
Token his
Feature activation+0.000
wife
Token wife
Feature activation+0.000
from
Token from
Feature activation+1.359
recovering
Token recovering
Feature activation+1.390
marital
Token marital
Feature activation+0.000
assets
Token assets
Feature activation+0.000
,
Token,
Feature activation+0.214
and
Token and
Feature activation+0.000
therefore
Token therefore
Feature activation+0.000
X
Token X
Feature activation+0.000
il
Tokenil
Feature activation+0.000
ai
Tokenai
Feature activation+0.000
's
Token's
Feature activation+0.000
wife
Token wife
Feature activation+0.000
arrested
Token arrested
Feature activation+1.389
in
Token in
Feature activation+0.496
Brit
Token Brit
Feature activation+0.000
on
Tokenon
Feature activation+0.000
death
Token death
Feature activation+0.000
Replay
Token Replay
Feature activation+0.000
us
Tokenus
Feature activation+0.000
finds
Token finds
Feature activation+0.000
his
Token his
Feature activation+0.000
mother
Token mother
Feature activation+0.000
still
Token still
Feature activation+0.929
alive
Token alive
Feature activation+1.372
.
Token.
Feature activation+0.410
She
Token She
Feature activation+0.000
is
Token is
Feature activation+0.804
over
Token over
Feature activation+0.693
joy
Tokenjoy
Feature activation+0.270
Ċ
TokenĊ
Feature activation+0.000
Car
TokenCar
Feature activation+0.000
oline
Tokenoline
Feature activation+0.000
Lucas
Token Lucas
Feature activation+0.690
,
Token,
Feature activation+1.257
the
Token the
Feature activation+1.367
co
Token co
Feature activation+0.465
-
Token-
Feature activation+0.475
leader
Tokenleader
Feature activation+1.217
of
Token of
Feature activation+1.106
the
Token the
Feature activation+0.300
my
Token my
Feature activation+0.000
mum
Token mum
Feature activation+0.000
was
Token was
Feature activation+0.900
termin
Token termin
Feature activation+0.003
ally
Tokenally
Feature activation+0.292
ill
Token ill
Feature activation+1.362
and
Token and
Feature activation+0.376
I
Token I
Feature activation+0.000
think
Token think
Feature activation+0.000
she
Token she
Feature activation+0.132
saw
Token saw
Feature activation+0.144

Top DFA by src position
MAX = 1.930

economics
Token economics
Feature activation-0.004
Top resid features:
.
Token.
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation+0.004
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
When
TokenWhen
Feature activation-0.049
Top resid features:
Mrs
Token Mrs
Feature activation+1.852
Top resid features:
.
Token.
Feature activation-0.001
Top resid features:
Pat
Token Pat
Feature activation+0.020
Top resid features:
more
Tokenmore
Feature activation-0.021
Top resid features:
t
Token t
Feature activation-0.002
Top resid features:
uss
Tokenuss
Feature activation+0.116
Top resid features:
luck
Token luck
Feature activation+0.005
Top resid features:
.
Token.
Feature activation+0.006
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation+0.002
Top resid features:
His
TokenHis
Feature activation-0.022
Top resid features:
mother
Token mother
Feature activation+1.694
Top resid features:
,
Token,
Feature activation+0.023
Top resid features:
who
Token who
Feature activation+0.007
Top resid features:
has
Token has
Feature activation+0.003
Top resid features:
drinking
Token drinking
Feature activation+0.003
Top resid features:
problems
Token problems
Feature activation-0.001
Top resid features:
in
Token in
Feature activation+0.001
Top resid features:
time
Token time
Feature activation+0.001
Top resid features:
to
Token to
Feature activation-0.001
Top resid features:
prevent
Token prevent
Feature activation-0.020
Top resid features:
his
Token his
Feature activation-0.051
Top resid features:
mother
Token mother
Feature activation+1.930
Top resid features:
from
Token from
Feature activation-0.008
Top resid features:
being
Token being
Feature activation-0.075
Top resid features:
killed
Token killed
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
only
Token only
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.002
Top resid features:
Ŀ
TokenĿ
Feature activation+0.001
Top resid features:
Ċ
TokenĊ
Feature activation+0.002
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Car
TokenCar
Feature activation+0.004
Top resid features:
oline
Tokenoline
Feature activation+1.729
Top resid features:
Lucas
Token Lucas
Feature activation-0.028
Top resid features:
,
Token,
Feature activation+0.025
Top resid features:
the
Token the
Feature activation+0.004
Top resid features:
co
Token co
Feature activation+0.007
Top resid features:
-
Token-
Feature activation+0.002
Top resid features:
economics
Token economics
Feature activation-0.002
Top resid features:
.
Token.
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation+0.006
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
When
TokenWhen
Feature activation-0.040
Top resid features:
Mrs
Token Mrs
Feature activation+1.547
Top resid features:
.
Token.
Feature activation+0.020
Top resid features:
Pat
Token Pat
Feature activation+0.033
Top resid features:
more
Tokenmore
Feature activation+0.003
Top resid features:
t
Token t
Feature activation-0.013
Top resid features:
uss
Tokenuss
Feature activation+0.001
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.016
Top resid features:
introduced
Token introduced
Feature activation-0.010
Top resid features:
to
Token to
Feature activation+0.003
Top resid features:
him
Token him
Feature activation+0.023
Top resid features:
.
Token.
Feature activation+0.017
Top resid features:
She
Token She
Feature activation+1.741
Top resid features:
did
Token did
Feature activation-0.023
Top resid features:
not
Token not
Feature activation+0.020
Top resid features:
want
Token want
Feature activation+0.035
Top resid features:
to
Token to
Feature activation+0.004
Top resid features:
be
Token be
Feature activation-0.001
Top resid features:
Mont
Token Mont
Feature activation+0.003
Top resid features:
y
Tokeny
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ļ
TokenĻ
Feature activation+0.002
Top resid features:
s
Tokens
Feature activation-0.006
Top resid features:
mother
Token mother
Feature activation+1.533
Top resid features:
,
Token,
Feature activation+0.018
Top resid features:
Mar
Token Mar
Feature activation+0.000
Top resid features:
j
Tokenj
Feature activation+0.001
Top resid features:
orie
Tokenorie
Feature activation+0.064
Top resid features:
Kelly
Token Kelly
Feature activation-0.010
Top resid features:
what
Token what
Feature activation+0.003
Top resid features:
he
Token he
Feature activation-0.001
Top resid features:
said
Token said
Feature activation-0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
"
Token "
Feature activation+0.006
Top resid features:
She
TokenShe
Feature activation+1.666
Top resid features:
took
Token took
Feature activation-0.049
Top resid features:
my
Token my
Feature activation+0.011
Top resid features:
interview
Token interview
Feature activation+0.007
Top resid features:
for
Token for
Feature activation+0.019
Top resid features:
10
Token 10
Feature activation-0.000
Top resid features:
the
Token the
Feature activation-0.012
Top resid features:
last
Token last
Feature activation-0.013
Top resid features:
known
Token known
Feature activation-0.002
Top resid features:
sighting
Token sighting
Feature activation+0.005
Top resid features:
of
Token of
Feature activation-0.017
Top resid features:
Ms
Token Ms
Feature activation+1.682
Top resid features:
Con
Token Con
Feature activation-0.088
Top resid features:
lon
Tokenlon
Feature activation-0.057
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
January
Token January
Feature activation+0.000
Top resid features:
whether
Token whether
Feature activation-0.019
Top resid features:
or
Token or
Feature activation-0.006
Top resid features:
not
Token not
Feature activation-0.001
Top resid features:
Trayvon
Token Trayvon
Feature activation+0.030
Top resid features:
's
Token's
Feature activation+0.009
Top resid features:
mother
Token mother
Feature activation+1.443
Top resid features:
,
Token,
Feature activation+0.022
Top resid features:
Sy
Token Sy
Feature activation-0.002
Top resid features:
br
Tokenbr
Feature activation+0.001
Top resid features:
ina
Tokenina
Feature activation+0.073
Top resid features:
Fulton
Token Fulton
Feature activation+0.003
Top resid features:
were
Token were
Feature activation-0.007
Top resid features:
on
Token on
Feature activation+0.008
Top resid features:
in
Token in
Feature activation+0.002
Top resid features:
reference
Token reference
Feature activation+0.004
Top resid features:
to
Token to
Feature activation-0.040
Top resid features:
Mrs
Token Mrs
Feature activation+1.790
Top resid features:
Kir
Token Kir
Feature activation+0.079
Top resid features:
ch
Tokench
Feature activation-0.131
Top resid features:
ner
Tokenner
Feature activation-0.124
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
her
Token her
Feature activation+0.000
Top resid features:
blow
Token blow
Feature activation-0.012
Top resid features:
in
Token in
Feature activation-0.003
Top resid features:
2009
Token 2009
Feature activation-0.012
Top resid features:
when
Token when
Feature activation-0.023
Top resid features:
his
Token his
Feature activation-0.137
Top resid features:
wife
Token wife
Feature activation+1.849
Top resid features:
,
Token,
Feature activation+0.022
Top resid features:
Vik
Token Vik
Feature activation-0.004
Top resid features:
ki
Tokenki
Feature activation+0.023
Top resid features:
,
Token,
Feature activation-0.037
Top resid features:
died
Token died
Feature activation-0.029
Top resid features:
,
Token,
Feature activation+0.004
Top resid features:
which
Token which
Feature activation-0.003
Top resid features:
it
Token it
Feature activation+0.004
Top resid features:
lost
Token lost
Feature activation-0.029
Top resid features:
to
Token to
Feature activation-0.077
Top resid features:
Rachel
Token Rachel
Feature activation+1.736
Top resid features:
Not
Token Not
Feature activation-0.001
Top resid features:
ley
Tokenley
Feature activation+0.032
Top resid features:
âĢ
TokenâĢ
Feature activation-0.054
Top resid features:
Ļ
TokenĻ
Feature activation+0.037
Top resid features:
s
Tokens
Feature activation-0.052
Top resid features:
in
Token in
Feature activation+0.004
Top resid features:
time
Token time
Feature activation+0.000
Top resid features:
to
Token to
Feature activation-0.000
Top resid features:
prevent
Token prevent
Feature activation+0.002
Top resid features:
his
Token his
Feature activation-0.055
Top resid features:
mother
Token mother
Feature activation+1.737
Top resid features:
from
Token from
Feature activation-0.104
Top resid features:
being
Token being
Feature activation+0.000
Top resid features:
killed
Token killed
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
only
Token only
Feature activation+0.000
Top resid features:
luck
Token luck
Feature activation+0.006
Top resid features:
.
Token.
Feature activation+0.009
Top resid features:
Ċ
TokenĊ
Feature activation+0.002
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
His
TokenHis
Feature activation-0.036
Top resid features:
mother
Token mother
Feature activation+1.789
Top resid features:
,
Token,
Feature activation-0.004
Top resid features:
who
Token who
Feature activation-0.121
Top resid features:
has
Token has
Feature activation-0.131
Top resid features:
drinking
Token drinking
Feature activation+0.000
Top resid features:
problems
Token problems
Feature activation+0.000
Top resid features:
malicious
Token malicious
Feature activation+0.004
Top resid features:
attempt
Token attempt
Feature activation+0.003
Top resid features:
to
Token to
Feature activation+0.005
Top resid features:
prevent
Token prevent
Feature activation+0.008
Top resid features:
his
Token his
Feature activation-0.062
Top resid features:
wife
Token wife
Feature activation+1.716
Top resid features:
from
Token from
Feature activation+0.024
Top resid features:
recovering
Token recovering
Feature activation-0.129
Top resid features:
marital
Token marital
Feature activation+0.000
Top resid features:
assets
Token assets
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
Bo
Token Bo
Feature activation-0.012
Top resid features:
X
Token X
Feature activation+0.009
Top resid features:
il
Tokenil
Feature activation+0.011
Top resid features:
ai
Tokenai
Feature activation+0.001
Top resid features:
's
Token's
Feature activation-0.045
Top resid features:
wife
Token wife
Feature activation+1.726
Top resid features:
arrested
Token arrested
Feature activation-0.101
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
Brit
Token Brit
Feature activation+0.000
Top resid features:
on
Tokenon
Feature activation+0.000
Top resid features:
death
Token death
Feature activation+0.000
Top resid features:
ac
Tokenac
Feature activation+0.002
Top resid features:
ot
Tokenot
Feature activation+0.003
Top resid features:
us
Tokenus
Feature activation+0.005
Top resid features:
finds
Token finds
Feature activation-0.025
Top resid features:
his
Token his
Feature activation-0.040
Top resid features:
mother
Token mother
Feature activation+1.761
Top resid features:
still
Token still
Feature activation-0.081
Top resid features:
alive
Token alive
Feature activation-0.090
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
She
Token She
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.003
Top resid features:
Ŀ
TokenĿ
Feature activation+0.001
Top resid features:
Ċ
TokenĊ
Feature activation+0.007
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Car
TokenCar
Feature activation+0.011
Top resid features:
oline
Tokenoline
Feature activation+1.684
Top resid features:
Lucas
Token Lucas
Feature activation-0.041
Top resid features:
,
Token,
Feature activation-0.008
Top resid features:
the
Token the
Feature activation-0.105
Top resid features:
co
Token co
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
lots
Token lots
Feature activation-0.004
Top resid features:
of
Token of
Feature activation-0.001
Top resid features:
money
Token money
Feature activation+0.006
Top resid features:
but
Token but
Feature activation+0.005
Top resid features:
my
Token my
Feature activation-0.037
Top resid features:
mum
Token mum
Feature activation+1.569
Top resid features:
was
Token was
Feature activation+0.006
Top resid features:
termin
Token termin
Feature activation-0.020
Top resid features:
ally
Tokenally
Feature activation-0.019
Top resid features:
ill
Token ill
Feature activation+0.012
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.03

Head 2: 0.03

Head 3: 0.51

Head 4: 0.04

Head 5: 0.03

Head 6: 0.06

Head 7: 0.10

Head 8: 0.03

Head 9: 0.04

Head 10: 0.02

Head 11: 0.09

Positive logits

herself3.28

husband2.02

breasts1.87

boyfriend1.78

childbirth1.78

maid1.77

maternity1.76

menstrual1.74

pregnant1.73

maiden1.71

ova1.71

womb1.62

pregnancy1.61

Lou1.61

breastfeeding1.59

miscar1.59

dow1.59

Christina1.54

hers1.53

boobs1.53

Negative logits

rall-1.31

raltar-1.24

imus-1.22

hindsight-1.22

don-1.21

ENS-1.19

-1.18

scrimmage-1.17

actory-1.16

Canaver-1.13

Players-1.13

erection-1.12

looph-1.10

ETA-1.09

EStreamFrame-1.08

click-1.08

projects-1.08

linemen-1.07

guiActiveUnfocused-1.05

PayPal-1.05

INTERVAL 1.670 - 1.855
CONTAINS 0.000%

very
Token very
Feature activation+0.000
day
Token day
Feature activation+0.000
,
Token,
Feature activation+0.624
never
Token never
Feature activation+1.346
really
Token really
Feature activation+0.613
liked
Token liked
Feature activation+1.728
Mole
Token Mole
Feature activation+0.000
very
Token very
Feature activation+0.000
much
Token much
Feature activation+0.000
because
Token because
Feature activation+0.141
he
Token he
Feature activation+0.187
.
Token.
Feature activation+1.007
Pat
Token Pat
Feature activation+0.000
more
Tokenmore
Feature activation+0.000
t
Token t
Feature activation+1.078
uss
Tokenuss
Feature activation+0.111
les
Tokenles
Feature activation+1.855
with
Token with
Feature activation+1.548
the
Token the
Feature activation+0.321
new
Token new
Feature activation+0.000
mixer
Token mixer
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 1.484 - 1.670
CONTAINS 0.001%

to
Token to
Feature activation+0.427
be
Token be
Feature activation+0.149
identified
Token identified
Feature activation+0.235
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
said
Token said
Feature activation+1.540
:
Token:
Feature activation+0.118
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
first
Token first
Feature activation+0.000
orie
Tokenorie
Feature activation+0.000
Kelly
Token Kelly
Feature activation+0.500
We
Token We
Feature activation+0.000
er
Tokener
Feature activation+0.000
,
Token,
Feature activation+1.115
took
Token took
Feature activation+1.517
to
Token to
Feature activation+0.000
Facebook
Token Facebook
Feature activation+0.288
to
Token to
Feature activation+0.000
confirm
Token confirm
Feature activation+0.665
that
Token that
Feature activation+0.643
known
Token known
Feature activation+0.000
sighting
Token sighting
Feature activation+0.000
of
Token of
Feature activation+0.000
Ms
Token Ms
Feature activation+0.000
Con
Token Con
Feature activation+0.658
lon
Tokenlon
Feature activation+1.486
was
Token was
Feature activation+0.000
in
Token in
Feature activation+0.000
January
Token January
Feature activation+0.000
2015
Token 2015
Feature activation+0.000
.
Token.
Feature activation+0.000
of
Token of
Feature activation+1.106
the
Token the
Feature activation+0.300
Green
Token Green
Feature activation+0.000
party
Token party
Feature activation+0.000
,
Token,
Feature activation+0.033
said
Token said
Feature activation+1.557
:
Token:
Feature activation+0.676
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
shameless
Token shameless
Feature activation+0.000
Pat
Token Pat
Feature activation+0.000
more
Tokenmore
Feature activation+0.000
t
Token t
Feature activation+1.078
uss
Tokenuss
Feature activation+0.111
les
Tokenles
Feature activation+1.855
with
Token with
Feature activation+1.548
the
Token the
Feature activation+0.321
new
Token new
Feature activation+0.000
mixer
Token mixer
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000

INTERVAL 1.299 - 1.484
CONTAINS 0.002%

with
Token with
Feature activation+0.000
Michel
Token Michel
Feature activation+0.000
Barn
Token Barn
Feature activation+0.000
ier
Tokenier
Feature activation+0.000
,
Token,
Feature activation+0.000
refused
Token refused
Feature activation+1.313
to
Token to
Feature activation+0.756
engage
Token engage
Feature activation+0.326
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
civil
Token civil
Feature activation+0.000
wife
Token wife
Feature activation+0.000
,
Token,
Feature activation+0.896
Vik
Token Vik
Feature activation+0.000
ki
Tokenki
Feature activation+0.000
,
Token,
Feature activation+0.014
died
Token died
Feature activation+1.473
unexpectedly
Token unexpectedly
Feature activation+0.100
of
Token of
Feature activation+0.000
natural
Token natural
Feature activation+0.000
causes
Token causes
Feature activation+0.000
at
Token at
Feature activation+0.000
up
Token up
Feature activation+0.000
drunk
Token drunk
Feature activation+0.000
at
Token at
Feature activation+0.000
his
Token his
Feature activation+0.000
mother
Token mother
Feature activation+0.000
's
Token's
Feature activation+1.309
funeral
Token funeral
Feature activation+0.366
last
Token last
Feature activation+0.000
year
Token year
Feature activation+0.000
,
Token,
Feature activation+0.218
before
Token before
Feature activation+0.000
X
Token X
Feature activation+0.000
il
Tokenil
Feature activation+0.000
ai
Tokenai
Feature activation+0.000
's
Token's
Feature activation+0.000
wife
Token wife
Feature activation+0.000
arrested
Token arrested
Feature activation+1.389
in
Token in
Feature activation+0.496
Brit
Token Brit
Feature activation+0.000
on
Tokenon
Feature activation+0.000
death
Token death
Feature activation+0.000
Replay
Token Replay
Feature activation+0.000
wife
Token wife
Feature activation+0.000
,
Token,
Feature activation+0.942
El
Token El
Feature activation+0.000
sie
Tokensie
Feature activation+0.000
,
Token,
Feature activation+0.248
came
Token came
Feature activation+1.325
to
Token to
Feature activation+1.082
visit
Token visit
Feature activation+1.032
him
Token him
Feature activation+0.000
at
Token at
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 1.113 - 1.299
CONTAINS 0.006%

with
Token with
Feature activation+0.000
Bo
Token Bo
Feature activation+0.000
's
Token's
Feature activation+0.000
wife
Token wife
Feature activation+0.000
--
Token --
Feature activation+1.003
an
Token an
Feature activation+1.232
allegation
Token allegation
Feature activation+0.055
that
Token that
Feature activation+0.000
appears
Token appears
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
His
TokenHis
Feature activation+0.000
mother
Token mother
Feature activation+0.000
,
Token,
Feature activation+0.954
who
Token who
Feature activation+1.270
has
Token has
Feature activation+1.403
drinking
Token drinking
Feature activation+0.552
problems
Token problems
Feature activation+0.000
up
Token up
Feature activation+0.000
to
Token to
Feature activation+0.000
Sy
Token Sy
Feature activation+0.798
br
Tokenbr
Feature activation+0.000
ina
Tokenina
Feature activation+0.000
Fulton
Token Fulton
Feature activation+0.317
,
Token,
Feature activation+1.042
will
Token will
Feature activation+1.120
testify
Token testify
Feature activation+1.483
.
Token.
Feature activation+0.217
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
Russian
TokenRussian
Feature activation+0.000
school
Token school
Feature activation+0.000
ist
Tokenist
Feature activation+0.000
)
Token)
Feature activation+0.785
challenged
Token challenged
Feature activation+0.890
R
Token R
Feature activation+0.000
hea
Tokenhea
Feature activation+0.064
(
Token (
Feature activation+1.230
Ter
TokenTer
Feature activation+0.351
i
Tokeni
Feature activation+0.000
Hat
Token Hat
Feature activation+0.053
cher
Tokencher
Feature activation+0.206
)
Token)
Feature activation+0.567
is
Token is
Feature activation+0.804
over
Token over
Feature activation+0.693
joy
Tokenjoy
Feature activation+0.270
ed
Tokened
Feature activation+0.961
to
Token to
Feature activation+1.070
learn
Token learn
Feature activation+1.174
that
Token that
Feature activation+1.286
he
Token he
Feature activation+0.000
is
Token is
Feature activation+0.000
well
Token well
Feature activation+0.000
studied
Token studied
Feature activation+0.000

INTERVAL 0.928 - 1.113
CONTAINS 0.015%

Pot
TokenPot
Feature activation+0.000
ip
Tokenip
Feature activation+0.000
har
Tokenhar
Feature activation+0.000
's
Token's
Feature activation+0.000
wife
Token wife
Feature activation+0.000
attempted
Token attempted
Feature activation+1.006
to
Token to
Feature activation+1.028
sed
Token sed
Feature activation+0.151
uce
Tokenuce
Feature activation+0.712
Joseph
Token Joseph
Feature activation+0.000
in
Token in
Feature activation+0.000
,
Token,
Feature activation+0.195
Marc
Token Marc
Feature activation+0.000
ia
Tokenia
Feature activation+0.000
Grant
Token Grant
Feature activation+0.798
,
Token,
Feature activation+0.626
principal
Token principal
Feature activation+0.988
at
Token at
Feature activation+0.121
the
Token the
Feature activation+0.000
Resurrection
Token Resurrection
Feature activation+0.000
Christian
Token Christian
Feature activation+0.000
Academy
Token Academy
Feature activation+0.000
committed
Token committed
Feature activation+0.000
suicide
Token suicide
Feature activation+0.000
with
Token with
Feature activation+0.000
his
Token his
Feature activation+0.000
wife
Token wife
Feature activation+0.000
,
Token,
Feature activation+0.928
Eva
Token Eva
Feature activation+0.000
Braun
Token Braun
Feature activation+0.000
.
Token.
Feature activation+0.111
Their
Token Their
Feature activation+0.000
corpses
Token corpses
Feature activation+0.000
to
Token to
Feature activation+0.000
ensure
Token ensure
Feature activation+0.000
him
Token him
Feature activation+0.000
that
Token that
Feature activation+0.000
Mary
Token Mary
Feature activation+0.000
has
Token has
Feature activation+0.950
not
Token not
Feature activation+0.892
been
Token been
Feature activation+1.171
cheating
Token cheating
Feature activation+1.017
but
Token but
Feature activation+0.000
that
Token that
Feature activation+0.000
current
Token current
Feature activation+0.000
Fed
Token Fed
Feature activation+0.000
Chair
Token Chair
Feature activation+0.000
Janet
Token Janet
Feature activation+0.000
L
Token L
Feature activation+0.000
.
Token.
Feature activation+1.088
Y
Token Y
Feature activation+0.601
ellen
Tokenellen
Feature activation+0.286
.
Token.
Feature activation+0.044
In
Token In
Feature activation+0.000
many
Token many
Feature activation+0.000

INTERVAL 0.742 - 0.928
CONTAINS 0.029%

of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
woman
Token woman
Feature activation+0.000
after
Token after
Feature activation+0.000
murdering
Token murdering
Feature activation+0.000
her
Token her
Feature activation+0.783
.
Token.
Feature activation+0.202
There
Token There
Feature activation+0.000
were
Token were
Feature activation+0.000
38
Token 38
Feature activation+0.000
minor
Token minor
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
We
TokenWe
Feature activation+0.000
iner
Tokeniner
Feature activation+0.000
's
Token's
Feature activation+0.000
wife
Token wife
Feature activation+0.000
,
Token,
Feature activation+0.836
Huma
Token Huma
Feature activation+0.000
Abedin
Token Abedin
Feature activation+0.000
,
Token,
Feature activation+0.497
is
Token is
Feature activation+1.164
a
Token a
Feature activation+1.312
s
Tokens
Feature activation+0.000
Claire
Token Claire
Feature activation+0.000
McC
Token McC
Feature activation+0.000
ask
Tokenask
Feature activation+0.000
ill
Tokenill
Feature activation+0.613
,
Token,
Feature activation+0.926
whom
Token whom
Feature activation+1.137
the
Token the
Feature activation+0.076
president
Token president
Feature activation+0.000
will
Token will
Feature activation+0.000
likely
Token likely
Feature activation+0.000
also
Token also
Feature activation+0.000
include
Token include
Feature activation+0.000
Laura
Token Laura
Feature activation+0.000
Po
Token Po
Feature activation+0.000
it
Tokenit
Feature activation+0.397
ras
Tokenras
Feature activation+0.756
,
Token,
Feature activation+0.849
the
Token the
Feature activation+1.169
document
Token document
Feature activation+0.000
arian
Tokenarian
Feature activation+0.023
who
Token who
Feature activation+0.000
.,
Token.,
Feature activation+0.000
whose
Token whose
Feature activation+0.000
wife
Token wife
Feature activation+0.000
Susan
Token Susan
Feature activation+0.000
was
Token was
Feature activation+0.311
killed
Token killed
Feature activation+0.816
in
Token in
Feature activation+0.196
August
Token August
Feature activation+0.000
2010
Token 2010
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 0.557 - 0.742
CONTAINS 0.056%

Sus
TokenSus
Feature activation+0.000
ie
Tokenie
Feature activation+0.000
Le
Token Le
Feature activation+0.000
afe
Tokenafe
Feature activation+0.000
,
Token,
Feature activation+0.419
from
Token from
Feature activation+0.612
the
Token the
Feature activation+0.044
Christian
Token Christian
Feature activation+0.000
campaign
Token campaign
Feature activation+0.000
group
Token group
Feature activation+0.000
Reform
Token Reform
Feature activation+0.000
(
Token(
Feature activation+0.000
CNN
TokenCNN
Feature activation+0.000
)
Token)
Feature activation+0.000
Michelle
Token Michelle
Feature activation+0.000
Carter
Token Carter
Feature activation+0.000
was
Token was
Feature activation+0.694
found
Token found
Feature activation+0.756
guilty
Token guilty
Feature activation+0.548
Friday
Token Friday
Feature activation+0.000
of
Token of
Feature activation+0.343
involuntary
Token involuntary
Feature activation+0.000
sh
Token sh
Feature activation+0.000
acked
Tokenacked
Feature activation+0.000
up
Token up
Feature activation+0.000
)
Token)
Feature activation+0.128
is
Token is
Feature activation+0.822
still
Token still
Feature activation+0.691
a
Token a
Feature activation+0.659
parental
Token parental
Feature activation+0.000
figure
Token figure
Feature activation+0.000
to
Token to
Feature activation+0.000
me
Token me
Feature activation+0.000
opt
Tokenopt
Feature activation+0.000
imal
Tokenimal
Feature activation+0.000
control
Token control
Feature activation+0.173
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.428
speeches
Token speeches
Feature activation+0.613
in
Token in
Feature activation+0.000
2011
Token 2011
Feature activation+0.000
and
Token and
Feature activation+0.000
2012
Token 2012
Feature activation+0.000
,
Token,
Feature activation+0.000
about
Token about
Feature activation+0.000
Hillary
Token Hillary
Feature activation+0.000
Clinton
Token Clinton
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.050
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.595
email
Token email
Feature activation+0.351
practices
Token practices
Feature activation+0.000
.
Token.
Feature activation+0.000
(
Token (
Feature activation+0.000
Cl
TokenCl
Feature activation+0.000

INTERVAL 0.371 - 0.557
CONTAINS 0.100%

Palin
Token Palin
Feature activation+0.000
had
Token had
Feature activation+0.308
little
Token little
Feature activation+0.075
to
Token to
Feature activation+0.000
say
Token say
Feature activation+0.479
about
Token about
Feature activation+0.423
policy
Token policy
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
her
Token her
Feature activation+0.281
speech
Token speech
Feature activation+0.161
two
Token two
Feature activation+0.000
women
Token women
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.330
who
Token who
Feature activation+0.175
go
Token go
Feature activation+0.582
by
Token by
Feature activation+0.387
the
Token the
Feature activation+0.000
single
Token single
Feature activation+0.000
names
Token names
Feature activation+0.000
of
Token of
Feature activation+0.000
Melissa
Token Melissa
Feature activation+0.000
she
Token she
Feature activation+0.000
went
Token went
Feature activation+0.077
to
Token to
Feature activation+0.000
friends
Token friends
Feature activation+0.000
and
Token and
Feature activation+0.000
told
Token told
Feature activation+0.496
people
Token people
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Hill
Token Hill
Feature activation+0.000
writing
Token writing
Feature activation+0.463
for
Token for
Feature activation+0.000
In
Token In
Feature activation+0.000
verse
Tokenverse
Feature activation+0.000
,
Token,
Feature activation+0.000
said
Token said
Feature activation+0.467
it
Token it
Feature activation+0.000
best
Token best
Feature activation+0.000
early
Token early
Feature activation+0.000
last
Token last
Feature activation+0.000
month
Token month
Feature activation+0.000
ĺ
Tokenĺ
Feature activation+0.000
Next
TokenNext
Feature activation+0.000
Erin
Token Erin
Feature activation+0.000
Andrews
Token Andrews
Feature activation+0.198
âĢ
TokenâĢ
Feature activation+0.078
Ļ
TokenĻ
Feature activation+0.409
series
Token series
Feature activation+0.000
.
Token.
Feature activation+0.000
That
Token That
Feature activation+0.000
post
Token post
Feature activation+0.000
was
Token was
Feature activation+0.000

INTERVAL 0.186 - 0.371
CONTAINS 0.187%

,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
girls
Token girls
Feature activation+0.000
were
Token were
Feature activation+0.455
required
Token required
Feature activation+0.235
to
Token to
Feature activation+0.307
administer
Token administer
Feature activation+0.305
"
Token "
Feature activation+0.000
top
Tokentop
Feature activation+0.000
less
Tokenless
Feature activation+0.000
or
Token or
Feature activation+0.000
recording
Token recording
Feature activation+0.000
Clement
Token Clement
Feature activation+0.000
i
Tokeni
Feature activation+0.000
having
Token having
Feature activation+0.147
sex
Token sex
Feature activation+0.072
with
Token with
Feature activation+0.208
another
Token another
Feature activation+0.000
male
Token male
Feature activation+0.000
and
Token and
Feature activation+0.000
broadcasting
Token broadcasting
Feature activation+0.000
the
Token the
Feature activation+0.000
"
Token "
Feature activation+0.000
meaning
Tokenmeaning
Feature activation+0.000
less
Tokenless
Feature activation+0.000
"
Token"
Feature activation+0.000
and
Token and
Feature activation+0.000
said
Token said
Feature activation+0.219
that
Token that
Feature activation+0.000
they
Token they
Feature activation+0.000
would
Token would
Feature activation+0.000
not
Token not
Feature activation+0.000
prevent
Token prevent
Feature activation+0.000
,"
Token,"
Feature activation+0.000
warned
Token warned
Feature activation+0.000
city
Token city
Feature activation+0.000
resident
Token resident
Feature activation+0.000
Melanie
Token Melanie
Feature activation+0.000
Roberts
Token Roberts
Feature activation+0.279
.
Token.
Feature activation+0.147
But
TokenBut
Feature activation+0.000
local
Token local
Feature activation+0.000
attorney
Token attorney
Feature activation+0.000
Jeff
Token Jeff
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ms
Token Ms
Feature activation+0.000
.
Token.
Feature activation+0.000
Nau
Token Nau
Feature activation+0.000
ert
Tokenert
Feature activation+0.067
said
Token said
Feature activation+0.279
.
Token.
Feature activation+0.052
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
Val
TokenVal
Feature activation+0.000
encia
Tokenencia
Feature activation+0.000
agree
Token agree
Feature activation+0.000

INTERVAL 0.000 - 0.186
CONTAINS 99.605%

to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
do
Token do
Feature activation+0.000
with
Token with
Feature activation+0.000
this
Token this
Feature activation+0.000
:
Token:
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Cong
TokenCong
Feature activation+0.000
rats
Tokenrats
Feature activation+0.000
to
Token to
Feature activation+0.000
surprising
Token surprising
Feature activation+0.000
that
Token that
Feature activation+0.000
Cantor
Token Cantor
Feature activation+0.000
did
Token did
Feature activation+0.000
not
Token not
Feature activation+0.000
want
Token want
Feature activation+0.000
to
Token to
Feature activation+0.000
face
Token face
Feature activation+0.000
the
Token the
Feature activation+0.000
"
Token "
Feature activation+0.000
m
Tokenm
Feature activation+0.000
So
Token So
Feature activation+0.000
beginning
Token beginning
Feature activation+0.000
next
Token next
Feature activation+0.000
week
Token week
Feature activation+0.000
,
Token,
Feature activation+0.000
my
Token my
Feature activation+0.000
blog
Token blog
Feature activation+0.000
will
Token will
Feature activation+0.000
begin
Token begin
Feature activation+0.000
to
Token to
Feature activation+0.000
include
Token include
Feature activation+0.000
not
Token not
Feature activation+0.000
know
Token know
Feature activation+0.000
,
Token,
Feature activation+0.000
Abraham
Token Abraham
Feature activation+0.000
ob
Token ob
Feature activation+0.000
eyed
Tokeneyed
Feature activation+0.000
when
Token when
Feature activation+0.000
the
Token the
Feature activation+0.000
Lord
Token Lord
Feature activation+0.000
asked
Token asked
Feature activation+0.000
him
Token him
Feature activation+0.000
I
Token I
Feature activation+0.000
was
Token was
Feature activation+0.000
3
Token 3
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
I
Token I
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
ve
Tokenve
Feature activation+0.000
never
Token never
Feature activation+0.000
encountered
Token encountered
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 2 in H4.3: (feature 8949)

TOP ACTIVATIONS
MAX = 4.074

THERE
Token THERE
Feature activation+0.000
'
Token'
Feature activation+0.829
S
TokenS
Feature activation+0.000
NOT
Token NOT
Feature activation+0.000
GO
Token GO
Feature activation+1.433
ING
TokenING
Feature activation+3.710
TO
Token TO
Feature activation+1.918
BE
Token BE
Feature activation+2.221
A
Token A
Feature activation+2.847
LOT
Token LOT
Feature activation+1.620
D
Token D
Feature activation+0.000
applying
Token applying
Feature activation+0.000
themselves
Token themselves
Feature activation+0.000
worked
Token worked
Feature activation+0.000
with
Token with
Feature activation+0.000
there
Token there
Feature activation+0.000
being
Token being
Feature activation+3.650
a
Token a
Feature activation+0.471
number
Token number
Feature activation+0.000
of
Token of
Feature activation+0.000
notable
Token notable
Feature activation+0.000
abs
Token abs
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
helpful
Token helpful
Feature activation+0.000
for
Token for
Feature activation+0.000
there
Token there
Feature activation+0.000
to
Token to
Feature activation+4.074
be
Token be
Feature activation+3.619
fl
Token fl
Feature activation+1.453
ot
Tokenot
Feature activation+0.000
illas
Tokenillas
Feature activation+0.000
that
Token that
Feature activation+0.000
try
Token try
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
has
Token has
Feature activation+0.851
never
Token never
Feature activation+3.612
been
Token been
Feature activation+1.683
anyone
Token anyone
Feature activation+0.000
like
Token like
Feature activation+0.000
M
Token M
Feature activation+0.000
eryl
Tokeneryl
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
was
Token was
Feature activation+0.000
,
Token,
Feature activation+1.886
however
Token however
Feature activation+0.912
,
Token,
Feature activation+3.585
a
Token a
Feature activation+0.000
very
Token very
Feature activation+0.000
pronounced
Token pronounced
Feature activation+0.089
ethos
Token ethos
Feature activation+0.000
of
Token of
Feature activation+0.000
-
Token-
Feature activation+0.000
wise
Tokenwise
Feature activation+0.000
we
Token we
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
there
Token there
Feature activation+0.000
'd
Token'd
Feature activation+3.545
be
Token be
Feature activation+1.829
a
Token a
Feature activation+0.370
difference
Token difference
Feature activation+0.000
.
Token.
Feature activation+0.000
After
Token After
Feature activation+0.000
tomorrow
Token tomorrow
Feature activation+0.000
night
Token night
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
would
Token would
Feature activation+2.419
've
Token've
Feature activation+3.485
been
Token been
Feature activation+1.820
a
Token a
Feature activation+1.103
celebr
Token celebr
Feature activation+0.815
atory
Tokenatory
Feature activation+0.000
riot
Token riot
Feature activation+0.219
an
Token an
Feature activation+0.000
era
Token era
Feature activation+0.000
when
Token when
Feature activation+0.000
there
Token there
Feature activation+0.000
was
Token was
Feature activation+0.900
indeed
Token indeed
Feature activation+3.484
some
Token some
Feature activation+0.509
prestige
Token prestige
Feature activation+0.000
involved
Token involved
Feature activation+0.000
in
Token in
Feature activation+0.000
participation
Token participation
Feature activation+0.000
all
Token all
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
certainly
Token certainly
Feature activation+0.815
does
Token does
Feature activation+2.280
exist
Token exist
Feature activation+3.473
much
Token much
Feature activation+0.154
for
Token for
Feature activation+0.000
Anarch
Token Anarch
Feature activation+0.000
ists
Tokenists
Feature activation+0.000
to
Token to
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
despite
Tokendespite
Feature activation+0.000
there
Token there
Feature activation+0.000
being
Token being
Feature activation+3.393
examples
Token examples
Feature activation+0.136
of
Token of
Feature activation+0.739
serious
Token serious
Feature activation+0.000
bullying
Token bullying
Feature activation+0.000
and
Token and
Feature activation+0.000
five
Token five
Feature activation+0.000
million
Token million
Feature activation+0.000
people
Token people
Feature activation+0.000
would
Token would
Feature activation+0.000
there
Token there
Feature activation+0.000
not
Token not
Feature activation+3.349
be
Token be
Feature activation+1.437
enough
Token enough
Feature activation+0.611
fans
Token fans
Feature activation+0.000
to
Token to
Feature activation+0.000
fill
Token fill
Feature activation+0.000
this
Token this
Feature activation+0.000
election
Token election
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+0.675
no
Token no
Feature activation+0.060
longer
Token longer
Feature activation+3.348
a
Token a
Feature activation+2.639
non
Token non
Feature activation+0.983
-
Token-
Feature activation+0.000
partisan
Tokenpartisan
Feature activation+0.000
space
Token space
Feature activation+0.000
chips
Token chips
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
will
Token will
Feature activation+1.086
likely
Token likely
Feature activation+2.928
need
Token need
Feature activation+3.321
to
Token to
Feature activation+2.746
include
Token include
Feature activation+0.382
an
Token an
Feature activation+0.000
ins
Token ins
Feature activation+0.000
ulator
Tokenulator
Feature activation+0.000
an
Token an
Feature activation+0.000
'
Token '
Feature activation+0.000
X
TokenX
Feature activation+0.000
'
Token'
Feature activation+0.000
there
Token there
Feature activation+0.000
will
Token will
Feature activation+3.310
be
Token be
Feature activation+1.688
a
Token a
Feature activation+0.476
pin
Token pin
Feature activation+0.000
(
Token (
Feature activation+0.000
and
Tokenand
Feature activation+0.000
said
Token said
Feature activation+0.000
that
Token that
Feature activation+0.000
they
Token they
Feature activation+0.000
showed
Token showed
Feature activation+0.000
there
Token there
Feature activation+0.000
had
Token had
Feature activation+3.306
been
Token been
Feature activation+0.000
a
Token a
Feature activation+0.095
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
huge
Tokenhuge
Feature activation+1.152
Ċ
TokenĊ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
used
Token used
Feature activation+2.536
to
Token to
Feature activation+3.253
be
Token be
Feature activation+1.926
a
Token a
Feature activation+1.568
thing
Token thing
Feature activation+1.112
called
Token called
Feature activation+0.000
the
Token the
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.074
currently
Token currently
Feature activation+3.226
a
Token a
Feature activation+0.248
21
Token 21
Feature activation+0.000
-
Token-
Feature activation+0.000
mile
Tokenmile
Feature activation+0.000
(
Token (
Feature activation+0.000
for
Token for
Feature activation+0.000
Leave
Token Leave
Feature activation+0.000
to
Token to
Feature activation+0.000
claim
Token claim
Feature activation+0.000
there
Token there
Feature activation+0.000
'd
Token'd
Feature activation+3.197
be
Token be
Feature activation+2.075
£
Token £
Feature activation+1.027
350
Token350
Feature activation+0.000
M
TokenM
Feature activation+0.000
a
Token a
Feature activation+0.000
the
Token the
Feature activation+0.000
researchers
Token researchers
Feature activation+0.000
said
Token said
Feature activation+0.000
there
Token there
Feature activation+0.000
would
Token would
Feature activation+2.819
need
Token need
Feature activation+3.189
to
Token to
Feature activation+2.528
be
Token be
Feature activation+1.210
a
Token a
Feature activation+0.024
switch
Token switch
Feature activation+0.000
from
Token from
Feature activation+0.000
express
Token express
Feature activation+0.000
apology
Token apology
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
will
Token will
Feature activation+1.452
likely
Token likely
Feature activation+3.184
be
Token be
Feature activation+0.454
moral
Token moral
Feature activation+0.000
equival
Token equival
Feature activation+0.000
ence
Tokenence
Feature activation+0.000
like
Token like
Feature activation+0.000

Top DFA by src position
MAX = 5.280

FR
Token FR
Feature activation+0.004
Top resid features:
ID
TokenID
Feature activation+0.003
Top resid features:
AY
TokenAY
Feature activation+0.012
Top resid features:
.
Token.
Feature activation-0.025
Top resid features:
SO
Token SO
Feature activation+0.034
Top resid features:
THERE
Token THERE
Feature activation+4.307
Top resid features:
'
Token'
Feature activation-0.039
Top resid features:
S
TokenS
Feature activation-0.047
Top resid features:
NOT
Token NOT
Feature activation+0.058
Top resid features:
GO
Token GO
Feature activation-0.125
Top resid features:
ING
TokenING
Feature activation-0.031
Top resid features:
not
Token not
Feature activation-0.003
Top resid features:
applying
Token applying
Feature activation-0.014
Top resid features:
themselves
Token themselves
Feature activation-0.006
Top resid features:
worked
Token worked
Feature activation-0.017
Top resid features:
with
Token with
Feature activation-0.069
Top resid features:
there
Token there
Feature activation+4.505
Top resid features:
being
Token being
Feature activation-0.295
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
number
Token number
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
notable
Token notable
Feature activation+0.000
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.177
Top resid features:
helpful
Token helpful
Feature activation-0.182
Top resid features:
for
Token for
Feature activation-0.345
Top resid features:
there
Token there
Feature activation+5.280
Top resid features:
to
Token to
Feature activation-0.320
Top resid features:
be
Token be
Feature activation-0.325
Top resid features:
fl
Token fl
Feature activation+0.000
Top resid features:
ot
Tokenot
Feature activation+0.000
Top resid features:
illas
Tokenillas
Feature activation+0.000
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.045
Top resid features:
Ċ
TokenĊ
Feature activation-0.045
Top resid features:
Ċ
TokenĊ
Feature activation-0.018
Top resid features:
There
TokenThere
Feature activation+4.854
Top resid features:
has
Token has
Feature activation-0.362
Top resid features:
never
Token never
Feature activation-0.460
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
anyone
Token anyone
Feature activation+0.000
Top resid features:
like
Token like
Feature activation+0.000
Top resid features:
free
Token free
Feature activation+0.000
Top resid features:
exchange
Token exchange
Feature activation-0.002
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
ideas
Token ideas
Feature activation-0.005
Top resid features:
.
Token.
Feature activation-0.028
Top resid features:
There
Token There
Feature activation+4.306
Top resid features:
was
Token was
Feature activation-0.336
Top resid features:
,
Token,
Feature activation+0.077
Top resid features:
however
Token however
Feature activation+0.205
Top resid features:
,
Token,
Feature activation-0.184
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
damage
Token damage
Feature activation-0.003
Top resid features:
-
Token-
Feature activation-0.011
Top resid features:
wise
Tokenwise
Feature activation-0.018
Top resid features:
we
Token we
Feature activation+0.033
Top resid features:
doubt
Token doubt
Feature activation-0.102
Top resid features:
there
Token there
Feature activation+4.448
Top resid features:
'd
Token'd
Feature activation-0.298
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
difference
Token difference
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
National
Token National
Feature activation+0.013
Top resid features:
Championship
Token Championship
Feature activation-0.007
Top resid features:
tomorrow
Token tomorrow
Feature activation+0.007
Top resid features:
night
Token night
Feature activation+0.002
Top resid features:
,
Token,
Feature activation-0.010
Top resid features:
there
Token there
Feature activation+4.486
Top resid features:
would
Token would
Feature activation-0.480
Top resid features:
've
Token've
Feature activation-0.220
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
celebr
Token celebr
Feature activation+0.000
Top resid features:
came
Token came
Feature activation-0.002
Top resid features:
from
Token from
Feature activation-0.008
Top resid features:
an
Token an
Feature activation-0.028
Top resid features:
era
Token era
Feature activation+0.010
Top resid features:
when
Token when
Feature activation-0.069
Top resid features:
there
Token there
Feature activation+4.688
Top resid features:
was
Token was
Feature activation-0.312
Top resid features:
indeed
Token indeed
Feature activation-0.262
Top resid features:
some
Token some
Feature activation+0.000
Top resid features:
prestige
Token prestige
Feature activation+0.000
Top resid features:
involved
Token involved
Feature activation+0.000
Top resid features:
tendencies
Token tendencies
Feature activation-0.008
Top resid features:
.
Token.
Feature activation-0.074
Top resid features:
After
Token After
Feature activation+0.077
Top resid features:
all
Token all
Feature activation-0.031
Top resid features:
,
Token,
Feature activation-0.020
Top resid features:
there
Token there
Feature activation+4.481
Top resid features:
certainly
Token certainly
Feature activation-0.014
Top resid features:
does
Token does
Feature activation-0.294
Top resid features:
exist
Token exist
Feature activation-0.198
Top resid features:
much
Token much
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
them
Token them
Feature activation-0.007
Top resid features:
,
Token,
Feature activation-0.031
Top resid features:
âĢ
Token âĢ
Feature activation-0.030
Top resid features:
ľ
Tokenľ
Feature activation-0.019
Top resid features:
despite
Tokendespite
Feature activation-0.020
Top resid features:
there
Token there
Feature activation+4.419
Top resid features:
being
Token being
Feature activation-0.541
Top resid features:
examples
Token examples
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
serious
Token serious
Feature activation+0.000
Top resid features:
bullying
Token bullying
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.002
Top resid features:
five
Token five
Feature activation-0.008
Top resid features:
million
Token million
Feature activation-0.093
Top resid features:
people
Token people
Feature activation-0.050
Top resid features:
would
Token would
Feature activation+0.038
Top resid features:
there
Token there
Feature activation+4.195
Top resid features:
not
Token not
Feature activation-0.280
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
enough
Token enough
Feature activation+0.000
Top resid features:
fans
Token fans
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
is
Token is
Feature activation-0.023
Top resid features:
that
Token that
Feature activation-0.020
Top resid features:
in
Token in
Feature activation+0.006
Top resid features:
this
Token this
Feature activation-0.010
Top resid features:
election
Token election
Feature activation-0.015
Top resid features:
there
Token there
Feature activation+4.412
Top resid features:
is
Token is
Feature activation-0.319
Top resid features:
no
Token no
Feature activation-0.168
Top resid features:
longer
Token longer
Feature activation-0.092
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
non
Token non
Feature activation+0.000
Top resid features:
flow
Token flow
Feature activation-0.011
Top resid features:
of
Token of
Feature activation-0.014
Top resid features:
the
Token the
Feature activation-0.007
Top resid features:
chips
Token chips
Feature activation-0.013
Top resid features:
,
Token,
Feature activation-0.005
Top resid features:
there
Token there
Feature activation+4.231
Top resid features:
will
Token will
Feature activation-0.101
Top resid features:
likely
Token likely
Feature activation-0.147
Top resid features:
need
Token need
Feature activation-0.201
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
include
Token include
Feature activation+0.000
Top resid features:
is
Token is
Feature activation-0.070
Top resid features:
an
Token an
Feature activation-0.064
Top resid features:
'
Token '
Feature activation+0.042
Top resid features:
X
TokenX
Feature activation+0.019
Top resid features:
'
Token'
Feature activation+0.006
Top resid features:
there
Token there
Feature activation+4.465
Top resid features:
will
Token will
Feature activation-0.497
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
pin
Token pin
Feature activation+0.000
Top resid features:
(
Token (
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.001
Top resid features:
said
Token said
Feature activation-0.043
Top resid features:
that
Token that
Feature activation-0.008
Top resid features:
they
Token they
Feature activation-0.054
Top resid features:
showed
Token showed
Feature activation-0.278
Top resid features:
there
Token there
Feature activation+4.681
Top resid features:
had
Token had
Feature activation-0.516
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
Ŀ
TokenĿ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation-0.004
Top resid features:
Ċ
TokenĊ
Feature activation-0.007
Top resid features:
âĢ
TokenâĢ
Feature activation-0.005
Top resid features:
ľ
Tokenľ
Feature activation-0.033
Top resid features:
There
TokenThere
Feature activation+4.308
Top resid features:
used
Token used
Feature activation-0.445
Top resid features:
to
Token to
Feature activation+0.013
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
thing
Token thing
Feature activation+0.000
Top resid features:
1980
Token 1980
Feature activation-0.000
Top resid features:
s
Tokens
Feature activation-0.004
Top resid features:
.
Token.
Feature activation-0.036
Top resid features:
Ċ
TokenĊ
Feature activation-0.025
Top resid features:
Ċ
TokenĊ
Feature activation-0.021
Top resid features:
There
TokenThere
Feature activation+4.437
Top resid features:
is
Token is
Feature activation-0.296
Top resid features:
currently
Token currently
Feature activation-0.270
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
21
Token 21
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
'
Token'
Feature activation+0.000
Top resid features:
for
Token for
Feature activation-0.082
Top resid features:
Leave
Token Leave
Feature activation-0.005
Top resid features:
to
Token to
Feature activation+0.005
Top resid features:
claim
Token claim
Feature activation-0.101
Top resid features:
there
Token there
Feature activation+3.934
Top resid features:
'd
Token'd
Feature activation-0.148
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
£
Token £
Feature activation+0.000
Top resid features:
350
Token350
Feature activation+0.000
Top resid features:
M
TokenM
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.001
Top resid features:
But
TokenBut
Feature activation+0.047
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
researchers
Token researchers
Feature activation-0.087
Top resid features:
said
Token said
Feature activation-0.087
Top resid features:
there
Token there
Feature activation+4.229
Top resid features:
would
Token would
Feature activation-0.281
Top resid features:
need
Token need
Feature activation-0.172
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
without
Token without
Feature activation+0.035
Top resid features:
an
Token an
Feature activation-0.005
Top resid features:
express
Token express
Feature activation+0.019
Top resid features:
apology
Token apology
Feature activation-0.005
Top resid features:
,
Token,
Feature activation+0.018
Top resid features:
there
Token there
Feature activation+4.117
Top resid features:
will
Token will
Feature activation-0.453
Top resid features:
likely
Token likely
Feature activation-0.156
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
moral
Token moral
Feature activation+0.000
Top resid features:
equival
Token equival
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.07

Head 1: 0.05

Head 2: 0.02

Head 3: 0.50

Head 4: 0.02

Head 5: 0.05

Head 6: 0.06

Head 7: 0.06

Head 8: 0.02

Head 9: 0.03

Head 10: 0.01

Head 11: 0.10

Positive logits

unanim1.33

ogeneity1.32

disagreement1.22

outcry1.16

differences1.16

universal1.15

icable1.12

dL1.11

murm1.11

dispute1.11

ovan1.09

unity1.09

heterogeneity1.08

fam1.08

attering1.07

everal1.07

somew1.07

plenty1.06

differing1.06

zero1.05

Negative logits

Cola-1.24

quote-1.21

iddles-1.13

Hed-1.13

leased-1.11

haul-1.11

soDeliveryDate-1.08

Cmd-1.08

Released-1.07

sted-1.05

gypt-1.04

prone-1.04

prone-1.03

merce-1.02

CEO-1.02

rants-1.02

poses-1.00

iller-0.99

oriented-0.99

umbered-0.98

INTERVAL 3.667 - 4.074
CONTAINS 0.000%

THERE
Token THERE
Feature activation+0.000
'
Token'
Feature activation+0.829
S
TokenS
Feature activation+0.000
NOT
Token NOT
Feature activation+0.000
GO
Token GO
Feature activation+1.433
ING
TokenING
Feature activation+3.710
TO
Token TO
Feature activation+1.918
BE
Token BE
Feature activation+2.221
A
Token A
Feature activation+2.847
LOT
Token LOT
Feature activation+1.620
D
Token D
Feature activation+0.000

INTERVAL 3.259 - 3.667
CONTAINS 0.002%

an
Token an
Feature activation+0.000
'
Token '
Feature activation+0.000
X
TokenX
Feature activation+0.000
'
Token'
Feature activation+0.000
there
Token there
Feature activation+0.000
will
Token will
Feature activation+3.310
be
Token be
Feature activation+1.688
a
Token a
Feature activation+0.476
pin
Token pin
Feature activation+0.000
(
Token (
Feature activation+0.000
and
Tokenand
Feature activation+0.000
five
Token five
Feature activation+0.000
million
Token million
Feature activation+0.000
people
Token people
Feature activation+0.000
would
Token would
Feature activation+0.000
there
Token there
Feature activation+0.000
not
Token not
Feature activation+3.349
be
Token be
Feature activation+1.437
enough
Token enough
Feature activation+0.611
fans
Token fans
Feature activation+0.000
to
Token to
Feature activation+0.000
fill
Token fill
Feature activation+0.000
tomorrow
Token tomorrow
Feature activation+0.000
night
Token night
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
would
Token would
Feature activation+2.419
've
Token've
Feature activation+3.485
been
Token been
Feature activation+1.820
a
Token a
Feature activation+1.103
celebr
Token celebr
Feature activation+0.815
atory
Tokenatory
Feature activation+0.000
riot
Token riot
Feature activation+0.219
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
was
Token was
Feature activation+0.000
,
Token,
Feature activation+1.886
however
Token however
Feature activation+0.912
,
Token,
Feature activation+3.585
a
Token a
Feature activation+0.000
very
Token very
Feature activation+0.000
pronounced
Token pronounced
Feature activation+0.089
ethos
Token ethos
Feature activation+0.000
of
Token of
Feature activation+0.000
an
Token an
Feature activation+0.000
era
Token era
Feature activation+0.000
when
Token when
Feature activation+0.000
there
Token there
Feature activation+0.000
was
Token was
Feature activation+0.900
indeed
Token indeed
Feature activation+3.484
some
Token some
Feature activation+0.509
prestige
Token prestige
Feature activation+0.000
involved
Token involved
Feature activation+0.000
in
Token in
Feature activation+0.000
participation
Token participation
Feature activation+0.000

INTERVAL 2.852 - 3.259
CONTAINS 0.003%

it
Token it
Feature activation+0.000
.
Token.
Feature activation+0.000
I
Token I
Feature activation+0.000
expect
Token expect
Feature activation+0.000
there
Token there
Feature activation+0.000
to
Token to
Feature activation+3.050
be
Token be
Feature activation+2.846
better
Token better
Feature activation+0.328
reception
Token reception
Feature activation+0.202
this
Token this
Feature activation+0.000
year
Token year
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Not
TokenNot
Feature activation+0.000
only
Token only
Feature activation+0.000
is
Token is
Feature activation+0.000
there
Token there
Feature activation+3.087
pressure
Token pressure
Feature activation+0.000
to
Token to
Feature activation+0.000
build
Token build
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
basis
Token basis
Feature activation+0.000
for
Token for
Feature activation+0.000
holding
Token holding
Feature activation+0.000
that
Token that
Feature activation+0.000
there
Token there
Feature activation+0.000
should
Token should
Feature activation+3.086
be
Token be
Feature activation+1.857
property
Token property
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
content
Token content
Feature activation+0.000
called
Token called
Feature activation+0.000
in
Token in
Feature activation+0.000
June
Token June
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
seemed
Token seemed
Feature activation+2.997
little
Token little
Feature activation+0.283
reason
Token reason
Feature activation+0.000
to
Token to
Feature activation+0.000
believe
Token believe
Feature activation+0.401
it
Token it
Feature activation+0.000
.
Token.
Feature activation+0.000
People
Token People
Feature activation+0.000
understood
Token understood
Feature activation+0.000
that
Token that
Feature activation+0.000
there
Token there
Feature activation+0.000
could
Token could
Feature activation+3.177
be
Token be
Feature activation+1.211
no
Token no
Feature activation+0.000
normal
Token normal
Feature activation+0.000
izing
Tokenizing
Feature activation+0.000
of
Token of
Feature activation+0.000

INTERVAL 2.444 - 2.852
CONTAINS 0.005%

the
Token the
Feature activation+0.000
role
Token role
Feature activation+0.000
,
Token,
Feature activation+0.000
so
Token so
Feature activation+0.000
there
Token there
Feature activation+0.000
must
Token must
Feature activation+2.552
be
Token be
Feature activation+0.000
a
Token a
Feature activation+0.000
large
Token large
Feature activation+0.960
pool
Token pool
Feature activation+0.000
of
Token of
Feature activation+0.000
when
Token when
Feature activation+0.000
it
Token it
Feature activation+0.000
seemed
Token seemed
Feature activation+0.000
there
Token there
Feature activation+0.000
couldn
Token couldn
Feature activation+2.706
't
Token't
Feature activation+2.689
be
Token be
Feature activation+0.000
any
Token any
Feature activation+0.000
more
Token more
Feature activation+0.000
good
Token good
Feature activation+0.000
news
Token news
Feature activation+0.298
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
needs
Token needs
Feature activation+2.161
to
Token to
Feature activation+2.487
be
Token be
Feature activation+0.914
concrete
Token concrete
Feature activation+0.830
reasons
Token reasons
Feature activation+0.000
why
Token why
Feature activation+0.000
a
Token a
Feature activation+0.000
salaries
Token salaries
Feature activation+0.000
there
Token there
Feature activation+0.000
doesn
Token doesn
Feature activation+1.094
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+2.749
appear
Token appear
Feature activation+2.697
to
Token to
Feature activation+2.107
be
Token be
Feature activation+2.225
any
Token any
Feature activation+1.168
rhy
Token rhy
Feature activation+0.000
think
Token think
Feature activation+0.000
there
Token there
Feature activation+0.000
won
Token won
Feature activation+2.919
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+2.642
be
Token be
Feature activation+0.807
room
Token room
Feature activation+0.000
for
Token for
Feature activation+0.000
them
Token them
Feature activation+0.000
in
Token in
Feature activation+0.000

INTERVAL 2.037 - 2.444
CONTAINS 0.007%

say
Token say
Feature activation+0.000
he
Token he
Feature activation+0.000
had
Token had
Feature activation+0.000
heard
Token heard
Feature activation+0.000
there
Token there
Feature activation+0.000
was
Token was
Feature activation+2.112
some
Token some
Feature activation+1.046
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
con
Tokencon
Feature activation+0.000
cern
Tokencern
Feature activation+0.621
stating
Token stating
Feature activation+0.000
,
Token,
Feature activation+0.000
"
Token "
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.000
clearly
Token clearly
Feature activation+2.164
an
Token an
Feature activation+0.116
unprecedented
Token unprecedented
Feature activation+0.215
desire
Token desire
Feature activation+0.000
for
Token for
Feature activation+0.069
alternatives
Token alternatives
Feature activation+0.000
fraught
Token fraught
Feature activation+0.000
history
Token history
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
has
Token has
Feature activation+2.231
been
Token been
Feature activation+0.000
historical
Token historical
Feature activation+0.024
animosity
Token animosity
Feature activation+0.668
between
Token between
Feature activation+0.000
the
Token the
Feature activation+0.000
.
Token.
Feature activation+0.000
"
Token "
Feature activation+0.000
I
TokenI
Feature activation+0.000
wish
Token wish
Feature activation+0.000
there
Token there
Feature activation+0.000
was
Token was
Feature activation+2.158
more
Token more
Feature activation+0.521
courage
Token courage
Feature activation+0.300
to
Token to
Feature activation+0.000
find
Token find
Feature activation+0.069
new
Token new
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
does
Token does
Feature activation+1.570
appear
Token appear
Feature activation+2.118
to
Token to
Feature activation+1.023
be
Token be
Feature activation+0.571
a
Token a
Feature activation+0.922
hereditary
Token hereditary
Feature activation+0.000
risk
Token risk
Feature activation+0.000

INTERVAL 1.630 - 2.037
CONTAINS 0.010%

there
Token there
Feature activation+0.000
doesn
Token doesn
Feature activation+0.571
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.825
seem
Token seem
Feature activation+1.789
to
Token to
Feature activation+1.255
be
Token be
Feature activation+0.047
much
Token much
Feature activation+0.603
point
Token point
Feature activation+0.000
.
Token.
Feature activation+0.000
donors
Token donors
Feature activation+0.000
.
Token.
Feature activation+0.000
But
Token But
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+0.432
not
Token not
Feature activation+1.765
enough
Token enough
Feature activation+0.841
donor
Token donor
Feature activation+0.000
milk
Token milk
Feature activation+0.000
for
Token for
Feature activation+0.000
that
Token that
Feature activation+0.000
combine
Token combine
Feature activation+0.000
sugar
Token sugar
Feature activation+0.000
and
Token and
Feature activation+0.000
fructose
Token fructose
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+1.799
more
Token more
Feature activation+0.820
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.499
risk
Token risk
Feature activation+0.000
for
Token for
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
used
Token used
Feature activation+2.536
to
Token to
Feature activation+3.253
be
Token be
Feature activation+1.926
a
Token a
Feature activation+1.568
thing
Token thing
Feature activation+1.112
called
Token called
Feature activation+0.000
the
Token the
Feature activation+0.000
Arab
Token Arab
Feature activation+0.000
whole
Token whole
Feature activation+0.000
time
Token time
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
was
Token was
Feature activation+0.000
never
Token never
Feature activation+1.856
a
Token a
Feature activation+0.492
question
Token question
Feature activation+0.000
about
Token about
Feature activation+0.000
leaving
Token leaving
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 1.222 - 1.630
CONTAINS 0.013%

happen
Token happen
Feature activation+0.000
there
Token there
Feature activation+0.000
are
Token are
Feature activation+1.588
going
Token going
Feature activation+0.033
to
Token to
Feature activation+1.490
be
Token be
Feature activation+1.342
more
Token more
Feature activation+0.931
environmental
Token environmental
Feature activation+0.000
implications
Token implications
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
there
Token there
Feature activation+0.000
has
Token has
Feature activation+2.854
always
Token always
Feature activation+1.904
been
Token been
Feature activation+1.664
a
Token a
Feature activation+2.004
minority
Token minority
Feature activation+1.337
of
Token of
Feature activation+1.555
spirited
Token spirited
Feature activation+0.000
radicals
Token radicals
Feature activation+0.000
,
Token,
Feature activation+0.000
including
Token including
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
was
Token was
Feature activation+0.000
too
Token too
Feature activation+1.393
much
Token much
Feature activation+0.000
money
Token money
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
how
Token how
Feature activation+0.000
many
Token many
Feature activation+0.000
total
Token total
Feature activation+0.000
items
Token items
Feature activation+0.000
there
Token there
Feature activation+0.000
are
Token are
Feature activation+1.272
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.109
dock
Token dock
Feature activation+0.000
.
Token.
Feature activation+0.000
You
Token You
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.770
Ļ
TokenĻ
Feature activation+0.741
s
Tokens
Feature activation+0.475
no
Token no
Feature activation+0.000
such
Token such
Feature activation+0.104
thing
Token thing
Feature activation+1.365
as
Token as
Feature activation+0.000
a
Token a
Feature activation+0.000
bisexual
Token bisexual
Feature activation+0.000
man
Token man
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 0.815 - 1.222
CONTAINS 0.024%

're
Token're
Feature activation+0.000
stressing
Token stressing
Feature activation+0.000
is
Token is
Feature activation+0.000
perhaps
Token perhaps
Feature activation+0.000
there
Token there
Feature activation+0.000
was
Token was
Feature activation+0.832
additional
Token additional
Feature activation+0.000
displeasure
Token displeasure
Feature activation+0.000
because
Token because
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
and
Token and
Feature activation+0.000
that
Token that
Feature activation+0.000
there
Token there
Feature activation+0.000
simply
Token simply
Feature activation+0.974
âĢ
Token âĢ
Feature activation+0.074
ľ
Tokenľ
Feature activation+0.000
was
Tokenwas
Feature activation+0.841
n
Tokenn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.148
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.589
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
always
Token always
Feature activation+0.889
a
Token a
Feature activation+0.604
question
Token question
Feature activation+0.000
of
Token of
Feature activation+0.544
whether
Token whether
Feature activation+0.000
you
Token you
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
are
Token are
Feature activation+0.058
approximately
Token approximately
Feature activation+1.066
5
Token 5
Feature activation+0.358
85
Token85
Feature activation+0.000
permanent
Token permanent
Feature activation+0.000
road
Token road
Feature activation+0.000
blocks
Tokenblocks
Feature activation+0.000
Minnesota
Token Minnesota
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
may
Token may
Feature activation+1.806
not
Token not
Feature activation+2.627
be
Token be
Feature activation+0.962
a
Token a
Feature activation+0.493
single
Token single
Feature activation+0.667
insurer
Token insurer
Feature activation+0.000
offering
Token offering
Feature activation+0.000
M
Token M
Feature activation+0.000

INTERVAL 0.407 - 0.815
CONTAINS 0.040%

challenging
Token challenging
Feature activation+0.000
collision
Token collision
Feature activation+0.000
in
Token in
Feature activation+0.000
that
Token that
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+0.644
a
Token a
Feature activation+0.000
considerable
Token considerable
Feature activation+0.000
lack
Token lack
Feature activation+0.084
of
Token of
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
What
TokenWhat
Feature activation+0.000
if
Token if
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+0.539
something
Token something
Feature activation+0.639
there
Token there
Feature activation+0.000
?
Token?
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
asked
Token asked
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.825
seem
Token seem
Feature activation+1.789
to
Token to
Feature activation+1.255
be
Token be
Feature activation+0.047
much
Token much
Feature activation+0.603
point
Token point
Feature activation+0.000
.
Token.
Feature activation+0.000
This
Token This
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
feeling
Token feeling
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
is
Token is
Feature activation+0.000
but
Token but
Feature activation+1.784
one
Token one
Feature activation+0.587
safe
Token safe
Feature activation+0.000
rule
Token rule
Feature activation+0.000
in
Token in
Feature activation+0.000
dealing
Token dealing
Feature activation+0.000
with
Token with
Feature activation+0.000
There
TokenThere
Feature activation+0.000
are
Token are
Feature activation+0.000
many
Token many
Feature activation+0.000
bones
Token bones
Feature activation+0.000
of
Token of
Feature activation+0.000
contention
Token contention
Feature activation+0.415
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
these
Token these
Feature activation+0.000
are
Token are
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 0.000 - 0.407
CONTAINS 99.897%

this
Token this
Feature activation+0.000
point
Token point
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
trend
Token trend
Feature activation+0.000
is
Token is
Feature activation+0.000
clear
Token clear
Feature activation+0.000
:
Token:
Feature activation+0.000
Papers
Token Papers
Feature activation+0.000
that
Token that
Feature activation+0.000
should
Token should
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
stupid
Token stupid
Feature activation+0.000
Republicans
Token Republicans
Feature activation+0.000
)
Token)
Feature activation+0.000
and
Token and
Feature activation+0.000
will
Token will
Feature activation+0.000
soar
Token soar
Feature activation+0.000
to
Token to
Feature activation+0.000
victory
Token victory
Feature activation+0.000
on
Token on
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
Education
Token Education
Feature activation+0.000
Funding
Token Funding
Feature activation+0.000
Agency
Token Agency
Feature activation+0.000
-
Token -
Feature activation+0.000
from
Token from
Feature activation+0.000
which
Token which
Feature activation+0.000
the
Token the
Feature activation+0.000
school
Token school
Feature activation+0.000
door
Token door
Feature activation+0.000
around
Token around
Feature activation+0.000
midnight
Token midnight
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
we
Token we
Feature activation+0.000
were
Token were
Feature activation+0.000
all
Token all
Feature activation+0.000
sob
Token sob
Feature activation+0.000
bing
Tokenbing
Feature activation+0.000
"
Token"
Feature activation+0.000
from
Token from
Feature activation+0.000
establishing
Token establishing
Feature activation+0.000
higher
Token higher
Feature activation+0.000
minimum
Token minimum
Feature activation+0.000
wages
Token wages
Feature activation+0.000
(
Token (
Feature activation+0.000
not
Tokennot
Feature activation+0.000
relating
Token relating
Feature activation+0.000
specifically
Token specifically
Feature activation+0.000
to
Token to
Feature activation+0.000
LGBT
Token LGBT
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 3 in H4.3: (feature 16677)

TOP ACTIVATIONS
MAX = 2.957

and
Token and
Feature activation+0.000
turned
Token turned
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.226
over
Token over
Feature activation+1.954
to
Token to
Feature activation+2.957
a
Token a
Feature activation+2.066
jury
Token jury
Feature activation+0.000
to
Token to
Feature activation+0.000
determine
Token determine
Feature activation+0.000
the
Token the
Feature activation+0.000
Wednesday
Token Wednesday
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.000
should
Token should
Feature activation+2.127
be
Token be
Feature activation+2.742
put
Token put
Feature activation+2.921
to
Token to
Feature activation+1.677
the
Token the
Feature activation+1.748
public
Token public
Feature activation+0.846
through
Token through
Feature activation+0.682
a
Token a
Feature activation+0.005
conflict
Token conflict
Feature activation+0.000
and
Token and
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686
how
Token how
Feature activation+0.000
much
Token much
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.000
has
Token has
Feature activation+2.007
impacted
Token impacted
Feature activation+2.759
my
Token my
Feature activation+0.105
life
Token life
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
on
Token on
Feature activation+0.000
Wednesday
Token Wednesday
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.000
should
Token should
Feature activation+2.127
be
Token be
Feature activation+2.742
put
Token put
Feature activation+2.921
to
Token to
Feature activation+1.677
the
Token the
Feature activation+1.748
public
Token public
Feature activation+0.846
through
Token through
Feature activation+0.682
not
Token not
Feature activation+0.000
known
Token known
Feature activation+0.000
whether
Token whether
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.092
involved
Token involved
Feature activation+2.671
the
Token the
Feature activation+0.202
jury
Token jury
Feature activation+0.000
itself
Token itself
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
reson
Token reson
Feature activation+2.291
ates
Tokenates
Feature activation+2.652
strongly
Token strongly
Feature activation+1.494
with
Token with
Feature activation+1.836
scientists
Token scientists
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686
any
Token any
Feature activation+0.000
ceasefire
Token ceasefire
Feature activation+0.000
were
Token were
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
conflict
Token conflict
Feature activation+0.000
and
Token and
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
ľ
Tokenľ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
not
Token not
Feature activation+1.858
only
Token only
Feature activation+1.799
affects
Token affects
Feature activation+2.385
overseas
Token overseas
Feature activation+0.000
talent
Token talent
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
smart
Token smart
Feature activation+0.000
about
Token about
Feature activation+0.000
how
Token how
Feature activation+0.000
this
Token this
Feature activation+0.000
issue
Token issue
Feature activation+0.619
is
Token is
Feature activation+2.353
framed
Token framed
Feature activation+1.929
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
he
Token he
Feature activation+0.000
Instead
Token Instead
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
sensitive
Token sensitive
Feature activation+0.000
issue
Token issue
Feature activation+0.457
was
Token was
Feature activation+2.337
âĢ
Token âĢ
Feature activation+1.783
ľ
Tokenľ
Feature activation+0.000
being
Tokenbeing
Feature activation+2.262
discussed
Token discussed
Feature activation+1.786
on
Token on
Feature activation+1.173
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686
any
Token any
Feature activation+0.000
ceasefire
Token ceasefire
Feature activation+0.000
were
Token were
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
reson
Token reson
Feature activation+2.291
ates
Tokenates
Feature activation+2.652
strongly
Token strongly
Feature activation+1.494
with
Token with
Feature activation+1.836
scientists
Token scientists
Feature activation+0.000
,
Token,
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686
any
Token any
Feature activation+0.000
ceasefire
Token ceasefire
Feature activation+0.000
sensitive
Token sensitive
Feature activation+0.000
issue
Token issue
Feature activation+0.457
was
Token was
Feature activation+2.337
âĢ
Token âĢ
Feature activation+1.783
ľ
Tokenľ
Feature activation+0.000
being
Tokenbeing
Feature activation+2.262
discussed
Token discussed
Feature activation+1.786
on
Token on
Feature activation+1.173
a
Token a
Feature activation+0.497
separate
Token separate
Feature activation+0.254
track
Token track
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
At
TokenAt
Feature activation+0.000
issue
Token issue
Feature activation+0.179
is
Token is
Feature activation+2.249
how
Token how
Feature activation+0.897
closely
Token closely
Feature activation+0.000
the
Token the
Feature activation+0.000
U
Token U
Feature activation+0.000
.
Token.
Feature activation+0.000
now
Token now
Feature activation+0.000
a
Token a
Feature activation+0.000
mainstream
Token mainstream
Feature activation+0.000
issue
Token issue
Feature activation+0.000
,
Token,
Feature activation+0.000
one
Token one
Feature activation+2.234
which
Token which
Feature activation+2.212
is
Token is
Feature activation+0.097
supported
Token supported
Feature activation+0.304
by
Token by
Feature activation+0.000
a
Token a
Feature activation+0.000
a
Token a
Feature activation+0.000
mainstream
Token mainstream
Feature activation+0.000
issue
Token issue
Feature activation+0.000
,
Token,
Feature activation+0.000
one
Token one
Feature activation+2.234
which
Token which
Feature activation+2.212
is
Token is
Feature activation+0.097
supported
Token supported
Feature activation+0.304
by
Token by
Feature activation+0.000
a
Token a
Feature activation+0.000
majority
Token majority
Feature activation+0.000
answer
Token answer
Feature activation+0.000
on
Token on
Feature activation+0.000
key
Token key
Feature activation+0.000
issues
Token issues
Feature activation+0.000
that
Token that
Feature activation+0.825
are
Token are
Feature activation+2.210
in
Token in
Feature activation+2.151
one
Token one
Feature activation+0.254
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000

Top DFA by src position
MAX = 4.034

Amendment
Token Amendment
Feature activation+0.022
Top resid features:
rights
Token rights
Feature activation+0.008
Top resid features:
and
Token and
Feature activation-0.008
Top resid features:
turned
Token turned
Feature activation-0.023
Top resid features:
the
Token the
Feature activation+0.058
Top resid features:
issue
Token issue
Feature activation+3.553
Top resid features:
over
Token over
Feature activation+0.265
Top resid features:
to
Token to
Feature activation+0.221
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
jury
Token jury
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
Abbott
Token Abbott
Feature activation+0.003
Top resid features:
said
Token said
Feature activation-0.019
Top resid features:
on
Token on
Feature activation+0.014
Top resid features:
Wednesday
Token Wednesday
Feature activation+0.018
Top resid features:
the
Token the
Feature activation+0.082
Top resid features:
issue
Token issue
Feature activation+3.697
Top resid features:
should
Token should
Feature activation+0.059
Top resid features:
be
Token be
Feature activation+0.080
Top resid features:
put
Token put
Feature activation+0.029
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
for
Token for
Feature activation-0.012
Top resid features:
the
Token the
Feature activation+0.017
Top resid features:
conflict
Token conflict
Feature activation+0.063
Top resid features:
and
Token and
Feature activation+0.037
Top resid features:
an
Token an
Feature activation+0.078
Top resid features:
issue
Token issue
Feature activation+3.415
Top resid features:
that
Token that
Feature activation+0.073
Top resid features:
would
Token would
Feature activation+0.083
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
reflecting
Token reflecting
Feature activation-0.023
Top resid features:
on
Token on
Feature activation+0.012
Top resid features:
how
Token how
Feature activation-0.011
Top resid features:
much
Token much
Feature activation-0.154
Top resid features:
the
Token the
Feature activation+0.060
Top resid features:
issue
Token issue
Feature activation+4.034
Top resid features:
has
Token has
Feature activation+0.069
Top resid features:
impacted
Token impacted
Feature activation-0.050
Top resid features:
my
Token my
Feature activation+0.000
Top resid features:
life
Token life
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Abbott
Token Abbott
Feature activation-0.000
Top resid features:
said
Token said
Feature activation-0.041
Top resid features:
on
Token on
Feature activation+0.034
Top resid features:
Wednesday
Token Wednesday
Feature activation+0.037
Top resid features:
the
Token the
Feature activation+0.084
Top resid features:
issue
Token issue
Feature activation+3.614
Top resid features:
should
Token should
Feature activation+0.193
Top resid features:
be
Token be
Feature activation-0.092
Top resid features:
put
Token put
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
was
Token was
Feature activation-0.001
Top resid features:
not
Token not
Feature activation-0.009
Top resid features:
known
Token known
Feature activation-0.044
Top resid features:
whether
Token whether
Feature activation-0.150
Top resid features:
the
Token the
Feature activation+0.006
Top resid features:
issue
Token issue
Feature activation+3.953
Top resid features:
involved
Token involved
Feature activation-0.070
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
jury
Token jury
Feature activation+0.000
Top resid features:
itself
Token itself
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
scientists
Token scientists
Feature activation-0.004
Top resid features:
.
Token.
Feature activation-0.004
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.014
Top resid features:
This
TokenThis
Feature activation+0.012
Top resid features:
issue
Token issue
Feature activation+3.651
Top resid features:
reson
Token reson
Feature activation+0.064
Top resid features:
ates
Tokenates
Feature activation-0.195
Top resid features:
strongly
Token strongly
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
scientists
Token scientists
Feature activation+0.000
Top resid features:
for
Token for
Feature activation-0.002
Top resid features:
the
Token the
Feature activation+0.013
Top resid features:
conflict
Token conflict
Feature activation+0.075
Top resid features:
and
Token and
Feature activation+0.024
Top resid features:
an
Token an
Feature activation+0.043
Top resid features:
issue
Token issue
Feature activation+3.059
Top resid features:
that
Token that
Feature activation+0.050
Top resid features:
would
Token would
Feature activation+0.009
Top resid features:
have
Token have
Feature activation-0.021
Top resid features:
to
Token to
Feature activation+0.107
Top resid features:
be
Token be
Feature activation+0.164
Top resid features:
for
Token for
Feature activation-0.052
Top resid features:
the
Token the
Feature activation+0.032
Top resid features:
conflict
Token conflict
Feature activation+0.031
Top resid features:
and
Token and
Feature activation+0.045
Top resid features:
an
Token an
Feature activation+0.171
Top resid features:
issue
Token issue
Feature activation+3.182
Top resid features:
that
Token that
Feature activation+0.058
Top resid features:
would
Token would
Feature activation+0.000
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.007
Top resid features:
Ċ
TokenĊ
Feature activation+0.001
Top resid features:
âĢ
TokenâĢ
Feature activation+0.004
Top resid features:
ľ
Tokenľ
Feature activation-0.003
Top resid features:
This
TokenThis
Feature activation+0.039
Top resid features:
issue
Token issue
Feature activation+3.765
Top resid features:
not
Token not
Feature activation+0.156
Top resid features:
only
Token only
Feature activation-0.340
Top resid features:
affects
Token affects
Feature activation-0.122
Top resid features:
overseas
Token overseas
Feature activation+0.000
Top resid features:
talent
Token talent
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.001
Top resid features:
smart
Token smart
Feature activation+0.020
Top resid features:
about
Token about
Feature activation+0.004
Top resid features:
how
Token how
Feature activation-0.187
Top resid features:
this
Token this
Feature activation+0.075
Top resid features:
issue
Token issue
Feature activation+3.624
Top resid features:
is
Token is
Feature activation-0.023
Top resid features:
framed
Token framed
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ŀ
TokenĿ
Feature activation+0.000
Top resid features:
.
Token.
Feature activation-0.009
Top resid features:
Instead
Token Instead
Feature activation-0.061
Top resid features:
,
Token,
Feature activation+0.050
Top resid features:
the
Token the
Feature activation+0.042
Top resid features:
sensitive
Token sensitive
Feature activation+0.564
Top resid features:
issue
Token issue
Feature activation+2.675
Top resid features:
was
Token was
Feature activation+0.036
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
being
Tokenbeing
Feature activation+0.000
Top resid features:
discussed
Token discussed
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.009
Top resid features:
the
Token the
Feature activation+0.011
Top resid features:
conflict
Token conflict
Feature activation+0.070
Top resid features:
and
Token and
Feature activation+0.030
Top resid features:
an
Token an
Feature activation+0.063
Top resid features:
issue
Token issue
Feature activation+2.763
Top resid features:
that
Token that
Feature activation+0.122
Top resid features:
would
Token would
Feature activation+0.015
Top resid features:
have
Token have
Feature activation-0.030
Top resid features:
to
Token to
Feature activation+0.364
Top resid features:
be
Token be
Feature activation-0.018
Top resid features:
scientists
Token scientists
Feature activation-0.004
Top resid features:
.
Token.
Feature activation-0.007
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.026
Top resid features:
This
TokenThis
Feature activation+0.001
Top resid features:
issue
Token issue
Feature activation+3.321
Top resid features:
reson
Token reson
Feature activation-0.039
Top resid features:
ates
Tokenates
Feature activation+0.000
Top resid features:
strongly
Token strongly
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
scientists
Token scientists
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.021
Top resid features:
conflict
Token conflict
Feature activation+0.078
Top resid features:
and
Token and
Feature activation+0.031
Top resid features:
an
Token an
Feature activation+0.067
Top resid features:
issue
Token issue
Feature activation+2.871
Top resid features:
that
Token that
Feature activation+0.084
Top resid features:
would
Token would
Feature activation+0.015
Top resid features:
have
Token have
Feature activation+0.226
Top resid features:
to
Token to
Feature activation-0.033
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.010
Top resid features:
Instead
Token Instead
Feature activation+0.007
Top resid features:
,
Token,
Feature activation+0.034
Top resid features:
the
Token the
Feature activation+0.042
Top resid features:
sensitive
Token sensitive
Feature activation+0.682
Top resid features:
issue
Token issue
Feature activation+2.410
Top resid features:
was
Token was
Feature activation+0.029
Top resid features:
âĢ
Token âĢ
Feature activation+0.013
Top resid features:
ľ
Tokenľ
Feature activation+0.220
Top resid features:
being
Tokenbeing
Feature activation-0.093
Top resid features:
discussed
Token discussed
Feature activation+0.000
Top resid features:
extremists
Token extremists
Feature activation+0.007
Top resid features:
.
Token.
Feature activation-0.004
Top resid features:
Ċ
TokenĊ
Feature activation+0.019
Top resid features:
Ċ
TokenĊ
Feature activation+0.006
Top resid features:
At
TokenAt
Feature activation+0.061
Top resid features:
issue
Token issue
Feature activation+3.326
Top resid features:
is
Token is
Feature activation-0.010
Top resid features:
how
Token how
Feature activation+0.000
Top resid features:
closely
Token closely
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
U
Token U
Feature activation+0.000
Top resid features:
reform
Token reform
Feature activation+0.024
Top resid features:
is
Token is
Feature activation+0.014
Top resid features:
now
Token now
Feature activation+0.055
Top resid features:
a
Token a
Feature activation+0.034
Top resid features:
mainstream
Token mainstream
Feature activation+0.035
Top resid features:
issue
Token issue
Feature activation+2.904
Top resid features:
,
Token,
Feature activation+0.080
Top resid features:
one
Token one
Feature activation+0.093
Top resid features:
which
Token which
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
supported
Token supported
Feature activation+0.000
Top resid features:
reform
Token reform
Feature activation+0.030
Top resid features:
is
Token is
Feature activation+0.009
Top resid features:
now
Token now
Feature activation+0.046
Top resid features:
a
Token a
Feature activation+0.054
Top resid features:
mainstream
Token mainstream
Feature activation+0.028
Top resid features:
issue
Token issue
Feature activation+3.252
Top resid features:
,
Token,
Feature activation+0.040
Top resid features:
one
Token one
Feature activation-0.066
Top resid features:
which
Token which
Feature activation-0.112
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
supported
Token supported
Feature activation+0.000
Top resid features:
refuses
Token refuses
Feature activation-0.014
Top resid features:
to
Token to
Feature activation-0.016
Top resid features:
answer
Token answer
Feature activation+0.023
Top resid features:
on
Token on
Feature activation+0.031
Top resid features:
key
Token key
Feature activation-0.077
Top resid features:
issues
Token issues
Feature activation+3.383
Top resid features:
that
Token that
Feature activation-0.071
Top resid features:
are
Token are
Feature activation+0.025
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
one
Token one
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.04

Head 2: 0.02

Head 3: 0.48

Head 4: 0.03

Head 5: 0.02

Head 6: 0.06

Head 7: 0.11

Head 8: 0.02

Head 9: 0.07

Head 10: 0.02

Head 11: 0.09

Positive logits

unsolved1.67

unresolved1.60

intrigue1.29

isson1.29

aceutical1.28

AIDS1.26

enegger1.24

trivia1.22

affecting1.22

channelAvailability1.17

alities1.16

concern1.14

concerns1.13

ality1.12

severity1.12

imaginable1.12

moot1.12

thorn1.11

Merit1.11

Sever1.11

Negative logits

adra-1.37

ramids-1.34

spons-1.29

react-1.25

onga-1.22

certific-1.17

atron-1.16

rift-1.14

stationed-1.13

bah-1.12

hiba-1.12

azy-1.12

elta-1.11

assisted-1.09

psey-1.09

ogun-1.08

tz-1.08

wrists-1.07

inclined-1.07

indo-1.07

INTERVAL 2.661 - 2.957
CONTAINS 0.001%

and
Token and
Feature activation+0.000
turned
Token turned
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.226
over
Token over
Feature activation+1.954
to
Token to
Feature activation+2.957
a
Token a
Feature activation+2.066
jury
Token jury
Feature activation+0.000
to
Token to
Feature activation+0.000
determine
Token determine
Feature activation+0.000
the
Token the
Feature activation+0.000
on
Token on
Feature activation+0.000
Wednesday
Token Wednesday
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.000
should
Token should
Feature activation+2.127
be
Token be
Feature activation+2.742
put
Token put
Feature activation+2.921
to
Token to
Feature activation+1.677
the
Token the
Feature activation+1.748
public
Token public
Feature activation+0.846
through
Token through
Feature activation+0.682
how
Token how
Feature activation+0.000
much
Token much
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.000
has
Token has
Feature activation+2.007
impacted
Token impacted
Feature activation+2.759
my
Token my
Feature activation+0.105
life
Token life
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
not
Token not
Feature activation+0.000
known
Token known
Feature activation+0.000
whether
Token whether
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.092
involved
Token involved
Feature activation+2.671
the
Token the
Feature activation+0.202
jury
Token jury
Feature activation+0.000
itself
Token itself
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
conflict
Token conflict
Feature activation+0.000
and
Token and
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686

INTERVAL 2.366 - 2.661
CONTAINS 0.000%

Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
reson
Token reson
Feature activation+2.291
ates
Tokenates
Feature activation+2.652
strongly
Token strongly
Feature activation+1.494
with
Token with
Feature activation+1.836
scientists
Token scientists
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
not
Token not
Feature activation+1.858
only
Token only
Feature activation+1.799
affects
Token affects
Feature activation+2.385
overseas
Token overseas
Feature activation+0.000
talent
Token talent
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
the
Token the
Feature activation+0.000
conflict
Token conflict
Feature activation+0.000
and
Token and
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.397
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
that
Token that
Feature activation+2.465
would
Token would
Feature activation+2.761
have
Token have
Feature activation+1.916
to
Token to
Feature activation+2.269
be
Token be
Feature activation+2.326
tackled
Token tackled
Feature activation+2.540
if
Token if
Feature activation+0.686
any
Token any
Feature activation+0.000
ceasefire
Token ceasefire
Feature activation+0.000
were
Token were
Feature activation+0.000
to
Token to
Feature activation+0.000

INTERVAL 2.070 - 2.366
CONTAINS 0.002%

Obamacare
Token Obamacare
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
issue
Token issue
Feature activation+0.063
here
Token here
Feature activation+1.096
is
Token is
Feature activation+2.192
the
Token the
Feature activation+1.677
role
Token role
Feature activation+0.631
and
Token and
Feature activation+0.000
institutional
Token institutional
Feature activation+0.000
integrity
Token integrity
Feature activation+0.000
answer
Token answer
Feature activation+0.000
on
Token on
Feature activation+0.000
key
Token key
Feature activation+0.000
issues
Token issues
Feature activation+0.000
that
Token that
Feature activation+0.825
are
Token are
Feature activation+2.210
in
Token in
Feature activation+2.151
one
Token one
Feature activation+0.254
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
on
Token on
Feature activation+0.000
key
Token key
Feature activation+0.000
issues
Token issues
Feature activation+0.000
that
Token that
Feature activation+0.825
are
Token are
Feature activation+2.210
in
Token in
Feature activation+2.151
one
Token one
Feature activation+0.254
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
proclaimed
Token proclaimed
Feature activation+0.000
a
Token a
Feature activation+0.000
mainstream
Token mainstream
Feature activation+0.000
issue
Token issue
Feature activation+0.000
,
Token,
Feature activation+0.000
one
Token one
Feature activation+2.234
which
Token which
Feature activation+2.212
is
Token is
Feature activation+0.097
supported
Token supported
Feature activation+0.304
by
Token by
Feature activation+0.000
a
Token a
Feature activation+0.000
majority
Token majority
Feature activation+0.000
Instead
Token Instead
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
sensitive
Token sensitive
Feature activation+0.000
issue
Token issue
Feature activation+0.457
was
Token was
Feature activation+2.337
âĢ
Token âĢ
Feature activation+1.783
ľ
Tokenľ
Feature activation+0.000
being
Tokenbeing
Feature activation+2.262
discussed
Token discussed
Feature activation+1.786
on
Token on
Feature activation+1.173

INTERVAL 1.774 - 2.070
CONTAINS 0.002%

the
Token the
Feature activation+0.000
math
Token math
Feature activation+0.000
on
Token on
Feature activation+0.000
issues
Token issues
Feature activation+0.000
such
Token such
Feature activation+1.230
as
Token as
Feature activation+1.873
the
Token the
Feature activation+0.996
cost
Token cost
Feature activation+0.000
-
Token-
Feature activation+0.000
effect
Tokeneffect
Feature activation+0.000
iveness
Tokeniveness
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
This
TokenThis
Feature activation+0.000
issue
Token issue
Feature activation+0.000
not
Token not
Feature activation+1.858
only
Token only
Feature activation+1.799
affects
Token affects
Feature activation+2.385
overseas
Token overseas
Feature activation+0.000
talent
Token talent
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Uber
Token Uber
Feature activation+0.000
issue
Token issue
Feature activation+0.000
still
Token still
Feature activation+1.262
might
Token might
Feature activation+2.014
work
Token work
Feature activation+2.021
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
Liberals
Token Liberals
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
turned
Token turned
Feature activation+0.000
the
Token the
Feature activation+0.000
issue
Token issue
Feature activation+0.226
over
Token over
Feature activation+1.954
to
Token to
Feature activation+2.957
a
Token a
Feature activation+2.066
jury
Token jury
Feature activation+0.000
to
Token to
Feature activation+0.000
determine
Token determine
Feature activation+0.000
the
Token the
Feature activation+0.000
amount
Token amount
Feature activation+0.000
highlighted
Token highlighted
Feature activation+0.941
in
Token in
Feature activation+0.522
the
Token the
Feature activation+0.121
report
Token report
Feature activation+0.000
were
Token were
Feature activation+1.671
in
Token in
Feature activation+1.788
the
Token the
Feature activation+0.928
past
Token past
Feature activation+0.454
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000

INTERVAL 1.479 - 1.774
CONTAINS 0.003%

Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
matter
Token matter
Feature activation+0.302
has
Token has
Feature activation+1.336
been
Token been
Feature activation+1.710
fought
Token fought
Feature activation+1.383
in
Token in
Feature activation+0.326
New
Token New
Feature activation+0.000
Jersey
Token Jersey
Feature activation+0.000
's
Token's
Feature activation+0.000
issue
Token issue
Feature activation+0.000
,
Token,
Feature activation+1.572
we
Token we
Feature activation+0.410
believe
Token believe
Feature activation+0.000
,
Token,
Feature activation+0.000
is
Token is
Feature activation+1.688
a
Token a
Feature activation+1.832
red
Token red
Feature activation+0.000
her
Token her
Feature activation+0.000
ring
Tokenring
Feature activation+0.000
in
Token in
Feature activation+0.000
,
Token,
Feature activation+0.000
saying
Token saying
Feature activation+0.000
the
Token the
Feature activation+0.000
matter
Token matter
Feature activation+0.397
will
Token will
Feature activation+1.117
be
Token be
Feature activation+1.629
given
Token given
Feature activation+1.259
proper
Token proper
Feature activation+0.000
consideration
Token consideration
Feature activation+0.579
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
pertinent
Token pertinent
Feature activation+0.000
and
Token and
Feature activation+0.000
an
Token an
Feature activation+0.000
important
Token important
Feature activation+0.000
issue
Token issue
Feature activation+0.000
involving
Token involving
Feature activation+1.491
public
Token public
Feature activation+0.000
interest
Token interest
Feature activation+0.000
through
Token through
Feature activation+0.429
his
Token his
Feature activation+0.000
RT
Token RT
Feature activation+0.000
immigration
Token immigration
Feature activation+0.000
status
Token status
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
topic
Token topic
Feature activation+0.232
that
Token that
Feature activation+1.609
prompted
Token prompted
Feature activation+1.590
banter
Token banter
Feature activation+0.000
between
Token between
Feature activation+0.000
attorneys
Token attorneys
Feature activation+0.000
and
Token and
Feature activation+0.000

INTERVAL 1.183 - 1.479
CONTAINS 0.004%

investigations
Token investigations
Feature activation+0.000
into
Token into
Feature activation+0.000
issues
Token issues
Feature activation+0.000
relating
Token relating
Feature activation+0.916
to
Token to
Feature activation+0.909
those
Token those
Feature activation+1.286
lies
Token lies
Feature activation+0.000
and
Token and
Feature activation+0.000
that
Token that
Feature activation+0.000
he
Token he
Feature activation+0.000
resign
Token resign
Feature activation+0.000
,
Token,
Feature activation+0.000
to
Token to
Feature activation+0.000
describe
Token describe
Feature activation+0.000
an
Token an
Feature activation+0.000
issue
Token issue
Feature activation+0.000
that
Token that
Feature activation+1.380
unfortunately
Token unfortunately
Feature activation+0.808
occurs
Token occurs
Feature activation+2.018
all
Token all
Feature activation+0.000
over
Token over
Feature activation+0.000
the
Token the
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
issues
Token issues
Feature activation+0.000
which
Token which
Feature activation+0.499
were
Token were
Feature activation+1.201
not
Token not
Feature activation+1.279
communicated
Token communicated
Feature activation+1.222
to
Token to
Feature activation+0.176
last
Token last
Feature activation+0.654
night
Token night
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
supplies
Token supplies
Feature activation+0.000
is
Token is
Feature activation+0.522
far
Token far
Feature activation+1.065
-
Token-
Feature activation+0.000
reaching
Tokenreaching
Feature activation+2.087
across
Token across
Feature activation+1.475
the
Token the
Feature activation+0.042
U
Token U
Feature activation+0.000
.
Token.
Feature activation+0.000
S
TokenS
Feature activation+0.000
.
Token.
Feature activation+0.000
of
Token of
Feature activation+0.589
personal
Token personal
Feature activation+0.000
and
Token and
Feature activation+0.000
sexual
Token sexual
Feature activation+0.000
behaviour
Token behaviour
Feature activation+0.000
are
Token are
Feature activation+1.311
handled
Token handled
Feature activation+1.214
by
Token by
Feature activation+0.510
society
Token society
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 0.887 - 1.183
CONTAINS 0.007%

to
Token to
Feature activation+0.000
know
Token know
Feature activation+0.000
why
Token why
Feature activation+0.000
the
Token the
Feature activation+0.000
topic
Token topic
Feature activation+0.392
is
Token is
Feature activation+1.091
important
Token important
Feature activation+1.523
to
Token to
Feature activation+0.955
them
Token them
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
but
Token but
Feature activation+0.000
the
Token the
Feature activation+0.000
topic
Token topic
Feature activation+0.393
of
Token of
Feature activation+0.473
racism
Token racism
Feature activation+0.000
has
Token has
Feature activation+1.083
arrived
Token arrived
Feature activation+1.061
per
Token per
Feature activation+0.135
force
Token force
Feature activation+0.099
in
Token in
Feature activation+0.455
the
Token the
Feature activation+0.000
the
Token the
Feature activation+0.000
Nile
Token Nile
Feature activation+0.000
water
Token water
Feature activation+0.000
issue
Token issue
Feature activation+0.000
has
Token has
Feature activation+0.579
paramount
Token paramount
Feature activation+0.933
importance
Token importance
Feature activation+0.670
in
Token in
Feature activation+0.000
all
Token all
Feature activation+0.000
of
Token of
Feature activation+0.000
our
Token our
Feature activation+0.000
strong
Token strong
Feature activation+0.000
authority
Token authority
Feature activation+0.000
on
Token on
Feature activation+0.000
issues
Token issues
Feature activation+0.000
such
Token such
Feature activation+1.224
the
Token the
Feature activation+1.170
disconnect
Token disconnect
Feature activation+0.000
ion
Tokenion
Feature activation+0.000
of
Token of
Feature activation+0.000
Tel
Token Tel
Feature activation+0.000
stra
Tokenstra
Feature activation+0.000
broad
Token broad
Feature activation+0.000
range
Token range
Feature activation+0.000
of
Token of
Feature activation+0.000
topics
Token topics
Feature activation+0.000
,
Token,
Feature activation+0.000
include
Token include
Feature activation+1.051
sports
Token sports
Feature activation+0.000
and
Token and
Feature activation+0.000
entertainment
Token entertainment
Feature activation+0.023
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000

INTERVAL 0.591 - 0.887
CONTAINS 0.010%

to
Token to
Feature activation+0.000
find
Token find
Feature activation+0.000
subject
Token subject
Feature activation+0.312
matter
Token matter
Feature activation+0.716
in
Token in
Feature activation+0.644
a
Token a
Feature activation+0.797
wide
Token wide
Feature activation+0.000
variety
Token variety
Feature activation+0.000
of
Token of
Feature activation+0.000
places
Token places
Feature activation+0.000
,
Token,
Feature activation+0.000
brought
Token brought
Feature activation+0.000
into
Token into
Feature activation+0.084
the
Token the
Feature activation+0.000
arena
Token arena
Feature activation+0.000
are
Token are
Feature activation+0.598
central
Token central
Feature activation+0.767
to
Token to
Feature activation+0.047
the
Token the
Feature activation+0.000
nation
Token nation
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
,
Token,
Feature activation+0.000
saying
Token saying
Feature activation+0.000
all
Token all
Feature activation+0.000
issues
Token issues
Feature activation+0.000
were
Token were
Feature activation+0.774
considered
Token considered
Feature activation+0.783
patient
Token patient
Feature activation+0.046
confidentiality
Token confidentiality
Feature activation+0.000
issues
Token issues
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
around
Token around
Feature activation+0.000
different
Token different
Feature activation+0.000
themes
Token themes
Feature activation+0.000
,
Token,
Feature activation+0.000
such
Token such
Feature activation+0.585
as
Token as
Feature activation+0.778
Open
Token Open
Feature activation+0.000
Education
Token Education
Feature activation+0.000
,
Token,
Feature activation+0.000
Open
Token Open
Feature activation+0.000
Access
Token Access
Feature activation+0.000
envoy
Token envoy
Feature activation+0.000
asserted
Token asserted
Feature activation+0.000
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
problem
Token problem
Feature activation+0.000
is
Token is
Feature activation+0.733
much
Token much
Feature activation+0.923
deeper
Token deeper
Feature activation+1.080
and
Token and
Feature activation+0.324
much
Token much
Feature activation+0.117
more
Token more
Feature activation+0.000

INTERVAL 0.296 - 0.591
CONTAINS 0.019%

basic
Token basic
Feature activation+0.000
items
Token items
Feature activation+0.000
or
Token or
Feature activation+0.000
school
Token school
Feature activation+0.000
supplies
Token supplies
Feature activation+0.000
is
Token is
Feature activation+0.522
far
Token far
Feature activation+1.065
-
Token-
Feature activation+0.000
reaching
Tokenreaching
Feature activation+2.087
across
Token across
Feature activation+1.475
the
Token the
Feature activation+0.042
marriage
Token marriage
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
matter
Token matter
Feature activation+0.302
has
Token has
Feature activation+1.336
been
Token been
Feature activation+1.710
fought
Token fought
Feature activation+1.383
in
Token in
Feature activation+0.326
New
Token New
Feature activation+0.000
.
Token.
Feature activation+0.000
And
Token And
Feature activation+0.000
there
Token there
Feature activation+0.000
are
Token are
Feature activation+0.000
issues
Token issues
Feature activation+0.000
around
Token around
Feature activation+0.437
how
Token how
Feature activation+0.000
to
Token to
Feature activation+0.000
reward
Token reward
Feature activation+0.000
and
Token and
Feature activation+0.000
recognize
Token recognize
Feature activation+0.000
huge
Token huge
Feature activation+0.000
systemic
Token systemic
Feature activation+0.000
issue
Token issue
Feature activation+0.000
for
Token for
Feature activation+0.184
Scotland
Token Scotland
Feature activation+0.000
about
Token about
Feature activation+0.346
making
Token making
Feature activation+0.000
sure
Token sure
Feature activation+0.000
at
Token at
Feature activation+0.000
every
Token every
Feature activation+0.000
stage
Token stage
Feature activation+0.000
footed
Tokenfooted
Feature activation+0.000
on
Token on
Feature activation+0.000
another
Token another
Feature activation+0.000
issue
Token issue
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.356
pur
Token pur
Feature activation+0.000
vey
Tokenvey
Feature activation+0.000
ors
Tokenors
Feature activation+0.000
of
Token of
Feature activation+0.000
Hudson
Token Hudson
Feature activation+0.000

INTERVAL 0.000 - 0.296
CONTAINS 99.953%

we
Token we
Feature activation+0.000
give
Token give
Feature activation+0.000
ourselves
Token ourselves
Feature activation+0.000
the
Token the
Feature activation+0.000
best
Token best
Feature activation+0.000
shot
Token shot
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
they
Token they
Feature activation+0.000
created
Token created
Feature activation+0.000
a
Token a
Feature activation+0.000
worker
Token worker
Feature activation+0.000
thread
Token thread
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
middle
Token middle
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
block
Token block
Feature activation+0.000
23
Token 23
Feature activation+0.000
,
Token,
Feature activation+0.000
would
Token would
Feature activation+0.000
be
Token be
Feature activation+0.000
far
Token far
Feature activation+0.000
cheaper
Token cheaper
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
team
Token team
Feature activation+0.000
he
Token he
Feature activation+0.000
chooses
Token chooses
Feature activation+0.000
alon
Tokenalon
Feature activation+0.000
|
Token |
Feature activation+0.000
History
Token History
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
At
TokenAt
Feature activation+0.000
the
Token the
Feature activation+0.000
end
Token end
Feature activation+0.000
of
Token of
Feature activation+0.000
his
Token his
Feature activation+0.000
History
Token History
Feature activation+0.000
first
Token first
Feature activation+0.000
one
Token one
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
one
Token one
Feature activation+0.000
I
Token I
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
d
Tokend
Feature activation+0.000
recommend
Token recommend
Feature activation+0.000
anyone
Token anyone
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 4 in H4.3: (feature 4757)

TOP ACTIVATIONS
MAX = 3.055

being
Token being
Feature activation+0.000
an
Token an
Feature activation+0.000
older
Token older
Feature activation+0.000
mom
Token mom
Feature activation+0.139
has
Token has
Feature activation+2.922
had
Token had
Feature activation+3.055
many
Token many
Feature activation+0.300
advantages
Token advantages
Feature activation+1.292
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
But
Token But
Feature activation+0.000
being
Token being
Feature activation+0.000
an
Token an
Feature activation+0.000
older
Token older
Feature activation+0.000
mom
Token mom
Feature activation+0.139
has
Token has
Feature activation+2.922
had
Token had
Feature activation+3.055
many
Token many
Feature activation+0.300
advantages
Token advantages
Feature activation+1.292
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
that
Token that
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000
a
Tokena
Feature activation+0.000
understand
Token understand
Feature activation+0.000
that
Token that
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000
standpoint
Token standpoint
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.166
Ļ
TokenĻ
Feature activation+0.000
is
Token is
Feature activation+2.607
not
Token not
Feature activation+2.777
a
Token a
Feature activation+2.880
qualification
Token qualification
Feature activation+1.960
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
job
Token job
Feature activation+0.000
,"
Token,"
Feature activation+0.000
to
Token to
Feature activation+0.206
wear
Token wear
Feature activation+0.509
what
Token what
Feature activation+0.094
one
Token one
Feature activation+0.000
wants
Token wants
Feature activation+0.000
is
Token is
Feature activation+2.880
âĢ
Token âĢ
Feature activation+0.648
ľ
Tokenľ
Feature activation+0.000
a
Tokena
Feature activation+2.516
basic
Token basic
Feature activation+0.453
human
Token human
Feature activation+0.000
and
Token and
Feature activation+0.000
understand
Token understand
Feature activation+0.000
that
Token that
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
dead
Tokendead
Feature activation+0.000
',
Token',
Feature activation+0.298
they
Token they
Feature activation+0.000
say
Token say
Feature activation+0.000
,
Token,
Feature activation+0.000
means
Token means
Feature activation+2.797
losing
Token losing
Feature activation+0.000
touch
Token touch
Feature activation+0.000
with
Token with
Feature activation+0.000
the
Token the
Feature activation+0.000
ground
Token ground
Feature activation+0.000
diversity
Token diversity
Feature activation+0.000
standpoint
Token standpoint
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.166
Ļ
TokenĻ
Feature activation+0.000
is
Token is
Feature activation+2.607
not
Token not
Feature activation+2.777
a
Token a
Feature activation+2.880
qualification
Token qualification
Feature activation+1.960
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
job
Token job
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000
a
Tokena
Feature activation+0.000
lot
Token lot
Feature activation+0.000
Iraq
Token Iraq
Feature activation+0.000
or
Token or
Feature activation+0.000
Afghanistan
Token Afghanistan
Feature activation+0.000
made
Token made
Feature activation+2.617
no
Token no
Feature activation+1.984
difference
Token difference
Feature activation+2.752
in
Token in
Feature activation+0.000
rates
Token rates
Feature activation+0.000
of
Token of
Feature activation+0.000
violent
Token violent
Feature activation+0.000
crime
Token crime
Feature activation+0.000
sent
Token sent
Feature activation+0.010
to
Token to
Feature activation+0.000
Iraq
Token Iraq
Feature activation+0.000
or
Token or
Feature activation+0.000
Afghanistan
Token Afghanistan
Feature activation+0.000
made
Token made
Feature activation+2.617
no
Token no
Feature activation+1.984
difference
Token difference
Feature activation+2.752
in
Token in
Feature activation+0.000
rates
Token rates
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
diversity
Token diversity
Feature activation+0.000
standpoint
Token standpoint
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.166
Ļ
TokenĻ
Feature activation+0.000
is
Token is
Feature activation+2.607
not
Token not
Feature activation+2.777
a
Token a
Feature activation+2.880
qualification
Token qualification
Feature activation+1.960
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
deployed
Token deployed
Feature activation+0.000
in
Token in
Feature activation+0.000
itself
Token itself
Feature activation+1.365
wasn
Token wasn
Feature activation+2.329
't
Token't
Feature activation+2.554
a
Token a
Feature activation+2.598
risk
Token risk
Feature activation+2.410
factor
Token factor
Feature activation+0.000
for
Token for
Feature activation+0.345
violent
Token violent
Feature activation+0.000
offenses
Token offenses
Feature activation+0.000
this
Token this
Feature activation+0.000
experience
Token experience
Feature activation+0.000
so
Token so
Feature activation+0.000
young
Token young
Feature activation+0.000
will
Token will
Feature activation+1.596
help
Token help
Feature activation+2.590
me
Token me
Feature activation+0.000
moving
Token moving
Feature activation+0.000
to
Token to
Feature activation+0.000
my
Token my
Feature activation+0.000
prime
Token prime
Feature activation+0.000
,
Token,
Feature activation+0.000
what
Token what
Feature activation+0.000
being
Token being
Feature activation+0.000
a
Token a
Feature activation+1.496
shareholder
Token shareholder
Feature activation+0.000
makes
Token makes
Feature activation+2.560
you
Token you
Feature activation+0.910
:
Token:
Feature activation+0.000
a
Token a
Feature activation+0.000
part
Token part
Feature activation+0.000
-
Token-
Feature activation+0.000
Being
TokenBeing
Feature activation+0.000
deployed
Token deployed
Feature activation+0.000
in
Token in
Feature activation+0.000
itself
Token itself
Feature activation+1.365
wasn
Token wasn
Feature activation+2.329
't
Token't
Feature activation+2.554
a
Token a
Feature activation+2.598
risk
Token risk
Feature activation+2.410
factor
Token factor
Feature activation+0.000
for
Token for
Feature activation+0.345
violent
Token violent
Feature activation+0.000
one
Token one
Feature activation+0.000
wants
Token wants
Feature activation+0.000
is
Token is
Feature activation+2.880
âĢ
Token âĢ
Feature activation+0.648
ľ
Tokenľ
Feature activation+0.000
a
Tokena
Feature activation+2.516
basic
Token basic
Feature activation+0.453
human
Token human
Feature activation+0.000
right
Token right
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
horrible
Token horrible
Feature activation+0.053
physical
Token physical
Feature activation+0.000
and
Token and
Feature activation+0.000
emotional
Token emotional
Feature activation+0.000
pain
Token pain
Feature activation+0.346
could
Token could
Feature activation+2.431
happen
Token happen
Feature activation+2.175
to
Token to
Feature activation+0.742
me
Token me
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
or
Token or
Feature activation+0.000
path
Token path
Feature activation+0.000
was
Token was
Feature activation+2.266
significantly
Token significantly
Feature activation+2.288
associated
Token associated
Feature activation+2.089
with
Token with
Feature activation+2.422
riding
Token riding
Feature activation+0.000
for
Token for
Feature activation+0.000
women
Token women
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000

Top DFA by src position
MAX = 4.167

<|endoftext|>
Token<|endoftext|>
Feature activation+0.035
Top resid features:
to
Token to
Feature activation-0.007
Top resid features:
judgement
Token judgement
Feature activation-0.013
Top resid features:
.
Token.
Feature activation-0.002
Top resid features:
But
Token But
Feature activation+0.025
Top resid features:
being
Token being
Feature activation+3.629
Top resid features:
an
Token an
Feature activation+0.109
Top resid features:
older
Token older
Feature activation-0.003
Top resid features:
mom
Token mom
Feature activation+0.025
Top resid features:
has
Token has
Feature activation+0.467
Top resid features:
had
Token had
Feature activation-0.094
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation+0.060
Top resid features:
to
Token to
Feature activation-0.008
Top resid features:
judgement
Token judgement
Feature activation-0.033
Top resid features:
.
Token.
Feature activation-0.037
Top resid features:
But
Token But
Feature activation+0.014
Top resid features:
being
Token being
Feature activation+3.968
Top resid features:
an
Token an
Feature activation+0.142
Top resid features:
older
Token older
Feature activation-0.027
Top resid features:
mom
Token mom
Feature activation+0.071
Top resid features:
has
Token has
Feature activation-0.110
Top resid features:
had
Token had
Feature activation+0.000
Top resid features:
Brass
Token Brass
Feature activation-0.004
Top resid features:
)
Token)
Feature activation+0.015
Top resid features:
and
Token and
Feature activation+0.019
Top resid features:
understand
Token understand
Feature activation+0.006
Top resid features:
that
Token that
Feature activation+0.022
Top resid features:
being
Token being
Feature activation+3.722
Top resid features:
small
Token small
Feature activation+0.114
Top resid features:
does
Token does
Feature activation+0.194
Top resid features:
not
Token not
Feature activation+0.241
Top resid features:
necessarily
Token necessarily
Feature activation-0.248
Top resid features:
equate
Token equate
Feature activation+0.000
Top resid features:
Brass
Token Brass
Feature activation-0.008
Top resid features:
)
Token)
Feature activation+0.012
Top resid features:
and
Token and
Feature activation+0.019
Top resid features:
understand
Token understand
Feature activation+0.004
Top resid features:
that
Token that
Feature activation+0.027
Top resid features:
being
Token being
Feature activation+3.678
Top resid features:
small
Token small
Feature activation+0.180
Top resid features:
does
Token does
Feature activation+0.214
Top resid features:
not
Token not
Feature activation-0.112
Top resid features:
necessarily
Token necessarily
Feature activation+0.000
Top resid features:
equate
Token equate
Feature activation+0.000
Top resid features:
rejected
Token rejected
Feature activation-0.031
Top resid features:
that
Token that
Feature activation-0.017
Top resid features:
rationale
Token rationale
Feature activation-0.002
Top resid features:
.
Token.
Feature activation+0.016
Top resid features:
"
Token "
Feature activation-0.017
Top resid features:
Being
TokenBeing
Feature activation+4.044
Top resid features:
a
Token a
Feature activation+0.030
Top resid features:
âĢ
Token âĢ
Feature activation+0.003
Top resid features:
ĺ
Tokenĺ
Feature activation+0.000
Top resid features:
home
Tokenhome
Feature activation+0.014
Top resid features:
run
Token run
Feature activation-0.036
Top resid features:
Say
Token Say
Feature activation+0.061
Top resid features:
re
Tokenre
Feature activation+0.004
Top resid features:
said
Token said
Feature activation+0.031
Top resid features:
,
Token,
Feature activation+0.005
Top resid features:
because
Token because
Feature activation+0.071
Top resid features:
being
Token being
Feature activation+3.686
Top resid features:
able
Token able
Feature activation+0.137
Top resid features:
to
Token to
Feature activation+0.056
Top resid features:
wear
Token wear
Feature activation-0.005
Top resid features:
what
Token what
Feature activation-0.030
Top resid features:
one
Token one
Feature activation+0.017
Top resid features:
Brass
Token Brass
Feature activation-0.008
Top resid features:
)
Token)
Feature activation-0.007
Top resid features:
and
Token and
Feature activation+0.023
Top resid features:
understand
Token understand
Feature activation-0.004
Top resid features:
that
Token that
Feature activation+0.009
Top resid features:
being
Token being
Feature activation+3.910
Top resid features:
small
Token small
Feature activation+0.170
Top resid features:
does
Token does
Feature activation-0.088
Top resid features:
not
Token not
Feature activation+0.000
Top resid features:
necessarily
Token necessarily
Feature activation+0.000
Top resid features:
equate
Token equate
Feature activation+0.000
Top resid features:
work
Token work
Feature activation-0.002
Top resid features:
on
Token on
Feature activation+0.005
Top resid features:
July
Token July
Feature activation+0.015
Top resid features:
28
Token 28
Feature activation+0.021
Top resid features:
.
Token.
Feature activation-0.003
Top resid features:
Being
Token Being
Feature activation+3.895
Top resid features:
'
Token '
Feature activation+0.074
Top resid features:
dead
Tokendead
Feature activation+0.051
Top resid features:
',
Token',
Feature activation+0.124
Top resid features:
they
Token they
Feature activation+0.055
Top resid features:
say
Token say
Feature activation+0.013
Top resid features:
rejected
Token rejected
Feature activation-0.051
Top resid features:
that
Token that
Feature activation-0.044
Top resid features:
rationale
Token rationale
Feature activation-0.001
Top resid features:
.
Token.
Feature activation-0.019
Top resid features:
"
Token "
Feature activation-0.025
Top resid features:
Being
TokenBeing
Feature activation+4.167
Top resid features:
a
Token a
Feature activation+0.051
Top resid features:
âĢ
Token âĢ
Feature activation+0.012
Top resid features:
ĺ
Tokenĺ
Feature activation-0.005
Top resid features:
home
Tokenhome
Feature activation+0.016
Top resid features:
run
Token run
Feature activation-0.028
Top resid features:
Brass
Token Brass
Feature activation-0.003
Top resid features:
)
Token)
Feature activation+0.016
Top resid features:
and
Token and
Feature activation+0.014
Top resid features:
understand
Token understand
Feature activation+0.001
Top resid features:
that
Token that
Feature activation+0.023
Top resid features:
being
Token being
Feature activation+3.689
Top resid features:
small
Token small
Feature activation+0.112
Top resid features:
does
Token does
Feature activation+0.035
Top resid features:
not
Token not
Feature activation+0.071
Top resid features:
necessarily
Token necessarily
Feature activation+0.051
Top resid features:
equate
Token equate
Feature activation-0.143
Top resid features:
,
Token,
Feature activation+0.005
Top resid features:
however
Token however
Feature activation-0.025
Top resid features:
,
Token,
Feature activation+0.007
Top resid features:
that
Token that
Feature activation+0.006
Top resid features:
merely
Token merely
Feature activation+0.069
Top resid features:
being
Token being
Feature activation+3.657
Top resid features:
sent
Token sent
Feature activation+0.016
Top resid features:
to
Token to
Feature activation-0.004
Top resid features:
Iraq
Token Iraq
Feature activation+0.014
Top resid features:
or
Token or
Feature activation+0.029
Top resid features:
Afghanistan
Token Afghanistan
Feature activation-0.031
Top resid features:
,
Token,
Feature activation-0.011
Top resid features:
however
Token however
Feature activation-0.009
Top resid features:
,
Token,
Feature activation+0.002
Top resid features:
that
Token that
Feature activation-0.049
Top resid features:
merely
Token merely
Feature activation+0.064
Top resid features:
being
Token being
Feature activation+4.102
Top resid features:
sent
Token sent
Feature activation+0.067
Top resid features:
to
Token to
Feature activation-0.005
Top resid features:
Iraq
Token Iraq
Feature activation+0.028
Top resid features:
or
Token or
Feature activation+0.007
Top resid features:
Afghanistan
Token Afghanistan
Feature activation-0.118
Top resid features:
rejected
Token rejected
Feature activation-0.077
Top resid features:
that
Token that
Feature activation-0.045
Top resid features:
rationale
Token rationale
Feature activation-0.017
Top resid features:
.
Token.
Feature activation-0.121
Top resid features:
"
Token "
Feature activation-0.016
Top resid features:
Being
TokenBeing
Feature activation+4.109
Top resid features:
a
Token a
Feature activation+0.093
Top resid features:
âĢ
Token âĢ
Feature activation+0.016
Top resid features:
ĺ
Tokenĺ
Feature activation-0.006
Top resid features:
home
Tokenhome
Feature activation+0.027
Top resid features:
run
Token run
Feature activation-0.038
Top resid features:
ing
Tokening
Feature activation-0.002
Top resid features:
.
Token.
Feature activation-0.001
Top resid features:
Ċ
TokenĊ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation-0.002
Top resid features:
"
Token"
Feature activation-0.003
Top resid features:
Being
TokenBeing
Feature activation+3.774
Top resid features:
deployed
Token deployed
Feature activation+0.119
Top resid features:
in
Token in
Feature activation+0.026
Top resid features:
itself
Token itself
Feature activation+0.046
Top resid features:
wasn
Token wasn
Feature activation-0.039
Top resid features:
't
Token't
Feature activation+0.084
Top resid features:
and
Token and
Feature activation+0.004
Top resid features:
downs
Token downs
Feature activation-0.004
Top resid features:
and
Token and
Feature activation-0.001
Top resid features:
everything
Token everything
Feature activation-0.009
Top resid features:
.
Token.
Feature activation+0.003
Top resid features:
Being
Token Being
Feature activation+3.719
Top resid features:
able
Token able
Feature activation+0.055
Top resid features:
to
Token to
Feature activation+0.016
Top resid features:
get
Token get
Feature activation-0.003
Top resid features:
this
Token this
Feature activation-0.043
Top resid features:
experience
Token experience
Feature activation-0.031
Top resid features:
,
Token,
Feature activation+0.024
Top resid features:
after
Token after
Feature activation+0.052
Top resid features:
all
Token all
Feature activation-0.065
Top resid features:
,
Token,
Feature activation+0.029
Top resid features:
what
Token what
Feature activation-0.225
Top resid features:
being
Token being
Feature activation+3.823
Top resid features:
a
Token a
Feature activation+0.100
Top resid features:
shareholder
Token shareholder
Feature activation+0.034
Top resid features:
makes
Token makes
Feature activation-0.219
Top resid features:
you
Token you
Feature activation+0.000
Top resid features:
:
Token:
Feature activation+0.000
Top resid features:
ing
Tokening
Feature activation-0.002
Top resid features:
.
Token.
Feature activation-0.008
Top resid features:
Ċ
TokenĊ
Feature activation-0.014
Top resid features:
Ċ
TokenĊ
Feature activation-0.011
Top resid features:
"
Token"
Feature activation-0.011
Top resid features:
Being
TokenBeing
Feature activation+3.786
Top resid features:
deployed
Token deployed
Feature activation+0.105
Top resid features:
in
Token in
Feature activation+0.035
Top resid features:
itself
Token itself
Feature activation+0.095
Top resid features:
wasn
Token wasn
Feature activation+0.014
Top resid features:
't
Token't
Feature activation-0.294
Top resid features:
Say
Token Say
Feature activation+0.050
Top resid features:
re
Tokenre
Feature activation+0.005
Top resid features:
said
Token said
Feature activation+0.037
Top resid features:
,
Token,
Feature activation+0.010
Top resid features:
because
Token because
Feature activation+0.068
Top resid features:
being
Token being
Feature activation+3.525
Top resid features:
able
Token able
Feature activation+0.134
Top resid features:
to
Token to
Feature activation+0.017
Top resid features:
wear
Token wear
Feature activation-0.008
Top resid features:
what
Token what
Feature activation+0.010
Top resid features:
one
Token one
Feature activation+0.015
Top resid features:
a
Token a
Feature activation+0.001
Top resid features:
ton
Token ton
Feature activation+0.001
Top resid features:
of
Token of
Feature activation-0.010
Top resid features:
bricks
Token bricks
Feature activation-0.009
Top resid features:
.
Token.
Feature activation-0.003
Top resid features:
Being
Token Being
Feature activation+3.644
Top resid features:
in
Token in
Feature activation+0.100
Top resid features:
horrible
Token horrible
Feature activation-0.081
Top resid features:
physical
Token physical
Feature activation+0.002
Top resid features:
and
Token and
Feature activation+0.081
Top resid features:
emotional
Token emotional
Feature activation-0.015
Top resid features:
detail
Token detail
Feature activation-0.034
Top resid features:
,
Token,
Feature activation+0.001
Top resid features:
they
Token they
Feature activation+0.000
Top resid features:
found
Token found
Feature activation-0.014
Top resid features:
that
Token that
Feature activation+0.013
Top resid features:
being
Token being
Feature activation+3.427
Top resid features:
within
Token within
Feature activation+0.052
Top resid features:
half
Token half
Feature activation+0.005
Top resid features:
a
Token a
Feature activation+0.006
Top resid features:
mile
Token mile
Feature activation-0.012
Top resid features:
of
Token of
Feature activation-0.024
Top resid features:

Decoder Weights Distribution

Head 0: 0.04

Head 1: 0.05

Head 2: 0.03

Head 3: 0.48

Head 4: 0.04

Head 5: 0.06

Head 6: 0.07

Head 7: 0.07

Head 8: 0.03

Head 9: 0.04

Head 10: 0.02

Head 11: 0.07

Positive logits

��1.60

Cause1.32

Discrimination1.23

clipboard1.23

UV1.21

impover1.21

ELY1.19

ife1.18

CTV1.17

ModLoader1.16

Marketplace1.14

hind1.14

Helper1.14

attendant1.12

ead1.11

Planet1.11

fright1.11

IER1.11

Orwell1.09

��1.09

Negative logits

ortium-1.33

Krug-1.33

enium-1.32

ologue-1.29

-1.29

-1.28

perse-1.28

sidx-1.26

bombed-1.26

executed-1.24

volt-1.24

theoret-1.23

vati-1.23

lambda-1.22

ascus-1.21

olate-1.21

chlor-1.19

disbanded-1.19

qt-1.18

arsen-1.17

INTERVAL 2.749 - 3.055
CONTAINS 0.001%

being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000
a
Tokena
Feature activation+0.000
lot
Token lot
Feature activation+0.000
But
Token But
Feature activation+0.000
being
Token being
Feature activation+0.000
an
Token an
Feature activation+0.000
older
Token older
Feature activation+0.000
mom
Token mom
Feature activation+0.139
has
Token has
Feature activation+2.922
had
Token had
Feature activation+3.055
many
Token many
Feature activation+0.300
advantages
Token advantages
Feature activation+1.292
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
to
Token to
Feature activation+0.206
wear
Token wear
Feature activation+0.509
what
Token what
Feature activation+0.094
one
Token one
Feature activation+0.000
wants
Token wants
Feature activation+0.000
is
Token is
Feature activation+2.880
âĢ
Token âĢ
Feature activation+0.648
ľ
Tokenľ
Feature activation+0.000
a
Tokena
Feature activation+2.516
basic
Token basic
Feature activation+0.453
human
Token human
Feature activation+0.000
and
Token and
Feature activation+0.000
understand
Token understand
Feature activation+0.000
that
Token that
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
understand
Token understand
Feature activation+0.000
that
Token that
Feature activation+0.000
being
Token being
Feature activation+0.000
small
Token small
Feature activation+0.000
does
Token does
Feature activation+2.819
not
Token not
Feature activation+2.904
necessarily
Token necessarily
Feature activation+2.909
equate
Token equate
Feature activation+2.757
to
Token to
Feature activation+2.165
weakness
Token weakness
Feature activation+0.000
...
Token...
Feature activation+0.000

INTERVAL 2.444 - 2.749
CONTAINS 0.001%

deployed
Token deployed
Feature activation+0.000
in
Token in
Feature activation+0.000
itself
Token itself
Feature activation+1.365
wasn
Token wasn
Feature activation+2.329
't
Token't
Feature activation+2.554
a
Token a
Feature activation+2.598
risk
Token risk
Feature activation+2.410
factor
Token factor
Feature activation+0.000
for
Token for
Feature activation+0.345
violent
Token violent
Feature activation+0.000
offenses
Token offenses
Feature activation+0.000
one
Token one
Feature activation+0.000
wants
Token wants
Feature activation+0.000
is
Token is
Feature activation+2.880
âĢ
Token âĢ
Feature activation+0.648
ľ
Tokenľ
Feature activation+0.000
a
Tokena
Feature activation+2.516
basic
Token basic
Feature activation+0.453
human
Token human
Feature activation+0.000
right
Token right
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
sent
Token sent
Feature activation+0.010
to
Token to
Feature activation+0.000
Iraq
Token Iraq
Feature activation+0.000
or
Token or
Feature activation+0.000
Afghanistan
Token Afghanistan
Feature activation+0.000
made
Token made
Feature activation+2.617
no
Token no
Feature activation+1.984
difference
Token difference
Feature activation+2.752
in
Token in
Feature activation+0.000
rates
Token rates
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
diversity
Token diversity
Feature activation+0.000
standpoint
Token standpoint
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.166
Ļ
TokenĻ
Feature activation+0.000
is
Token is
Feature activation+2.607
not
Token not
Feature activation+2.777
a
Token a
Feature activation+2.880
qualification
Token qualification
Feature activation+1.960
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
Being
TokenBeing
Feature activation+0.000
deployed
Token deployed
Feature activation+0.000
in
Token in
Feature activation+0.000
itself
Token itself
Feature activation+1.365
wasn
Token wasn
Feature activation+2.329
't
Token't
Feature activation+2.554
a
Token a
Feature activation+2.598
risk
Token risk
Feature activation+2.410
factor
Token factor
Feature activation+0.000
for
Token for
Feature activation+0.345
violent
Token violent
Feature activation+0.000

INTERVAL 2.138 - 2.444
CONTAINS 0.002%

or
Token or
Feature activation+0.000
expelled
Token expelled
Feature activation+0.000
from
Token from
Feature activation+0.149
school
Token school
Feature activation+0.000
made
Token made
Feature activation+2.253
a
Token a
Feature activation+2.374
student
Token student
Feature activation+0.022
nearly
Token nearly
Feature activation+0.000
three
Token three
Feature activation+0.000
times
Token times
Feature activation+0.000
more
Token more
Feature activation+0.000
out
Token out
Feature activation+0.000
of
Token of
Feature activation+1.447
Sunday
Token Sunday
Feature activation+0.000
's
Token's
Feature activation+0.000
game
Token game
Feature activation+0.000
will
Token will
Feature activation+2.320
affect
Token affect
Feature activation+2.055
him
Token him
Feature activation+0.000
moving
Token moving
Feature activation+0.000
forward
Token forward
Feature activation+0.000
.
Token.
Feature activation+0.000
in
Token in
Feature activation+0.000
time
Token time
Feature activation+0.000
.
Token.
Feature activation+0.000
Being
Token Being
Feature activation+0.000
mindful
Token mindful
Feature activation+0.000
is
Token is
Feature activation+2.174
simply
Token simply
Feature activation+2.272
being
Token being
Feature activation+1.362
aware
Token aware
Feature activation+0.699
of
Token of
Feature activation+1.384
everything
Token everything
Feature activation+0.000
suspended
Token suspended
Feature activation+0.000
or
Token or
Feature activation+0.000
expelled
Token expelled
Feature activation+0.000
from
Token from
Feature activation+0.149
school
Token school
Feature activation+0.000
made
Token made
Feature activation+2.253
a
Token a
Feature activation+2.374
student
Token student
Feature activation+0.022
nearly
Token nearly
Feature activation+0.000
three
Token three
Feature activation+0.000
times
Token times
Feature activation+0.000
or
Token or
Feature activation+0.000
path
Token path
Feature activation+0.000
was
Token was
Feature activation+2.266
significantly
Token significantly
Feature activation+2.288
associated
Token associated
Feature activation+2.089
with
Token with
Feature activation+2.422
riding
Token riding
Feature activation+0.000
for
Token for
Feature activation+0.000
women
Token women
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000

INTERVAL 1.833 - 2.138
CONTAINS 0.001%

of
Token of
Feature activation+1.447
Sunday
Token Sunday
Feature activation+0.000
's
Token's
Feature activation+0.000
game
Token game
Feature activation+0.000
will
Token will
Feature activation+2.320
affect
Token affect
Feature activation+2.055
him
Token him
Feature activation+0.000
moving
Token moving
Feature activation+0.000
forward
Token forward
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
that
Token that
Feature activation+0.000
amount
Token amount
Feature activation+0.000
of
Token of
Feature activation+0.000
time
Token time
Feature activation+0.000
comes
Token comes
Feature activation+1.878
with
Token with
Feature activation+1.916
its
Token its
Feature activation+0.440
own
Token own
Feature activation+0.034
challenges
Token challenges
Feature activation+1.322
.
Token.
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
trail
Token trail
Feature activation+0.000
or
Token or
Feature activation+0.000
path
Token path
Feature activation+0.000
was
Token was
Feature activation+2.266
significantly
Token significantly
Feature activation+2.288
associated
Token associated
Feature activation+2.089
with
Token with
Feature activation+2.422
riding
Token riding
Feature activation+0.000
for
Token for
Feature activation+0.000
women
Token women
Feature activation+0.000
,
Token,
Feature activation+0.000
instincts
Token instincts
Feature activation+0.000
.
Token.
Feature activation+0.000
Being
Token Being
Feature activation+0.000
as
Token as
Feature activation+0.000
smart
Token smart
Feature activation+0.139
as
Token as
Feature activation+2.022
he
Token he
Feature activation+0.214
was
Token was
Feature activation+0.000
he
Token he
Feature activation+1.299
had
Token had
Feature activation+1.531
a
Token a
Feature activation+0.364
to
Token to
Feature activation+0.000
Iraq
Token Iraq
Feature activation+0.000
or
Token or
Feature activation+0.000
Afghanistan
Token Afghanistan
Feature activation+0.000
made
Token made
Feature activation+2.617
no
Token no
Feature activation+1.984
difference
Token difference
Feature activation+2.752
in
Token in
Feature activation+0.000
rates
Token rates
Feature activation+0.000
of
Token of
Feature activation+0.000
violent
Token violent
Feature activation+0.000

INTERVAL 1.527 - 1.833
CONTAINS 0.001%

But
TokenBut
Feature activation+0.000
being
Token being
Feature activation+0.000
unable
Token unable
Feature activation+0.000
to
Token to
Feature activation+0.000
win
Token win
Feature activation+0.000
is
Token is
Feature activation+1.670
not
Token not
Feature activation+2.134
equivalent
Token equivalent
Feature activation+2.048
to
Token to
Feature activation+1.753
surrender
Token surrender
Feature activation+0.000
ing
Tokening
Feature activation+0.000
living
Token living
Feature activation+0.000
there
Token there
Feature activation+0.000
day
Token day
Feature activation+0.000
to
Token to
Feature activation+0.000
day
Token day
Feature activation+0.000
is
Token is
Feature activation+1.814
spectacular
Token spectacular
Feature activation+1.908
.
Token.
Feature activation+0.000
That
Token That
Feature activation+0.000
helped
Token helped
Feature activation+0.000
me
Token me
Feature activation+0.000
And
Token And
Feature activation+0.000
also
Token also
Feature activation+0.000
being
Token being
Feature activation+0.000
Christmas
Token Christmas
Feature activation+0.000
,
Token,
Feature activation+0.064
lets
Token lets
Feature activation+1.582
talk
Token talk
Feature activation+0.000
about
Token about
Feature activation+0.000
presents
Token presents
Feature activation+0.000
.
Token.
Feature activation+0.000
Furthermore
Token Furthermore
Feature activation+0.000
me
Token me
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.226
probably
Token probably
Feature activation+0.000
will
Token will
Feature activation+1.753
happen
Token happen
Feature activation+1.534
to
Token to
Feature activation+0.000
me
Token me
Feature activation+0.000
,
Token,
Feature activation+0.000
if
Token if
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.226
probably
Token probably
Feature activation+0.000
will
Token will
Feature activation+1.753
happen
Token happen
Feature activation+1.534
to
Token to
Feature activation+0.000
me
Token me
Feature activation+0.000
,
Token,
Feature activation+0.000
if
Token if
Feature activation+0.000
only
Token only
Feature activation+0.000

INTERVAL 1.222 - 1.527
CONTAINS 0.001%

as
Token as
Feature activation+0.000
smart
Token smart
Feature activation+0.139
as
Token as
Feature activation+2.022
he
Token he
Feature activation+0.214
was
Token was
Feature activation+0.000
he
Token he
Feature activation+1.299
had
Token had
Feature activation+1.531
a
Token a
Feature activation+0.364
remarkable
Token remarkable
Feature activation+1.210
ability
Token ability
Feature activation+0.638
to
Token to
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
I
Token I
Feature activation+0.000
think
Token think
Feature activation+0.000
being
Token being
Feature activation+0.000
encouraged
Token encouraged
Feature activation+1.314
to
Token to
Feature activation+0.279
be
Token be
Feature activation+0.375
disenfranch
Token disenfranch
Feature activation+1.601
ised
Tokenised
Feature activation+0.780
on
Token on
Feature activation+0.336
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
Being
TokenBeing
Feature activation+0.000
deployed
Token deployed
Feature activation+0.000
in
Token in
Feature activation+0.000
itself
Token itself
Feature activation+1.365
wasn
Token wasn
Feature activation+2.329
't
Token't
Feature activation+2.554
a
Token a
Feature activation+2.598
risk
Token risk
Feature activation+2.410
factor
Token factor
Feature activation+0.000
.
Token.
Feature activation+0.000
But
Token But
Feature activation+0.000
living
Token living
Feature activation+0.000
in
Token in
Feature activation+0.000
it
Token it
Feature activation+0.000
is
Token is
Feature activation+1.483
quite
Token quite
Feature activation+1.047
good
Token good
Feature activation+1.493
fun
Token fun
Feature activation+1.055
.
Token.
Feature activation+0.000
And
Token And
Feature activation+0.000
time
Token time
Feature activation+0.000
comes
Token comes
Feature activation+1.878
with
Token with
Feature activation+1.916
its
Token its
Feature activation+0.440
own
Token own
Feature activation+0.034
challenges
Token challenges
Feature activation+1.322
.
Token.
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
We
TokenWe
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000

INTERVAL 0.916 - 1.222
CONTAINS 0.002%

he
Token he
Feature activation+0.214
was
Token was
Feature activation+0.000
he
Token he
Feature activation+1.299
had
Token had
Feature activation+1.531
a
Token a
Feature activation+0.364
remarkable
Token remarkable
Feature activation+1.210
ability
Token ability
Feature activation+0.638
to
Token to
Feature activation+0.000
pull
Token pull
Feature activation+0.000
people
Token people
Feature activation+0.000
together
Token together
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.000
feels
Token feels
Feature activation+0.946
like
Token like
Feature activation+1.153
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000
,'
Token,'
Feature activation+0.000
sites
Token sites
Feature activation+0.000
being
Token being
Feature activation+0.000
so
Token so
Feature activation+0.000
far
Token far
Feature activation+0.000
apart
Token apart
Feature activation+0.000
causes
Token causes
Feature activation+1.152
high
Token high
Feature activation+0.000
rotation
Token rotation
Feature activation+0.000
times
Token times
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
being
Token being
Feature activation+0.000
an
Token an
Feature activation+0.000
independent
Token independent
Feature activation+0.000
âĢĶ
TokenâĢĶ
Feature activation+0.013
has
Tokenhas
Feature activation+0.000
given
Token given
Feature activation+0.951
New
Token New
Feature activation+0.000
York
Token York
Feature activation+0.000
Mayor
Token Mayor
Feature activation+0.000
Michael
Token Michael
Feature activation+0.000
Bloomberg
Token Bloomberg
Feature activation+0.000
in
Token in
Feature activation+0.000
it
Token it
Feature activation+0.000
is
Token is
Feature activation+1.483
quite
Token quite
Feature activation+1.047
good
Token good
Feature activation+1.493
fun
Token fun
Feature activation+1.055
.
Token.
Feature activation+0.000
And
Token And
Feature activation+0.000
you
Token you
Feature activation+0.000
always
Token always
Feature activation+0.000
find
Token find
Feature activation+0.000

INTERVAL 0.611 - 0.916
CONTAINS 0.002%

the
Token the
Feature activation+0.000
United
Token United
Feature activation+0.055
States
Token States
Feature activation+0.000
is
Token is
Feature activation+1.989
safer
Token safer
Feature activation+1.038
than
Token than
Feature activation+0.797
being
Token being
Feature activation+0.000
a
Token a
Feature activation+0.000
trash
Token trash
Feature activation+0.000
collector
Token collector
Feature activation+0.000
.
Token.
Feature activation+0.000
having
Token having
Feature activation+0.000
only
Token only
Feature activation+0.000
depression
Token depression
Feature activation+0.000
was
Token was
Feature activation+0.760
associated
Token associated
Feature activation+0.553
with
Token with
Feature activation+0.884
future
Token future
Feature activation+0.000
dementia
Token dementia
Feature activation+0.000
diagnosis
Token diagnosis
Feature activation+0.000
but
Token but
Feature activation+0.000
a
Token a
Feature activation+0.000
was
Token was
Feature activation+0.000
he
Token he
Feature activation+1.299
had
Token had
Feature activation+1.531
a
Token a
Feature activation+0.364
remarkable
Token remarkable
Feature activation+1.210
ability
Token ability
Feature activation+0.638
to
Token to
Feature activation+0.000
pull
Token pull
Feature activation+0.000
people
Token people
Feature activation+0.000
together
Token together
Feature activation+0.000
.
Token.
Feature activation+0.000
just
Token just
Feature activation+0.000
before
Token before
Feature activation+0.000
the
Token the
Feature activation+0.000
rye
Token rye
Feature activation+0.000
,
Token,
Feature activation+0.000
is
Token is
Feature activation+0.671
also
Token also
Feature activation+0.352
sort
Token sort
Feature activation+0.079
of
Token of
Feature activation+0.365
keeping
Token keeping
Feature activation+0.000
with
Token with
Feature activation+0.000
what
Token what
Feature activation+0.000
being
Token being
Feature activation+0.000
a
Token a
Feature activation+1.496
shareholder
Token shareholder
Feature activation+0.000
makes
Token makes
Feature activation+2.560
you
Token you
Feature activation+0.910
:
Token:
Feature activation+0.000
a
Token a
Feature activation+0.000
part
Token part
Feature activation+0.000
-
Token-
Feature activation+0.000
owner
Tokenowner
Feature activation+0.000

INTERVAL 0.305 - 0.611
CONTAINS 0.006%

from
Token from
Feature activation+0.000
slides
Token slides
Feature activation+0.000
is
Token is
Feature activation+0.000
"
Token "
Feature activation+0.000
very
Tokenvery
Feature activation+0.040
hard
Token hard
Feature activation+0.409
for
Token for
Feature activation+0.000
a
Token a
Feature activation+0.000
child
Token child
Feature activation+0.000
,"
Token,"
Feature activation+0.000
said
Token said
Feature activation+0.000
understand
Token understand
Feature activation+0.000
the
Token the
Feature activation+0.000
feelings
Token feelings
Feature activation+0.000
of
Token of
Feature activation+0.000
being
Token being
Feature activation+0.000
that
Token that
Feature activation+0.330
age
Token age
Feature activation+0.000
and
Token and
Feature activation+0.000
being
Token being
Feature activation+0.120
the
Token the
Feature activation+0.358
target
Token target
Feature activation+0.000
Association
Token Association
Feature activation+0.000
is
Token is
Feature activation+0.000
either
Token either
Feature activation+0.000
consciously
Token consciously
Feature activation+0.000
lying
Token lying
Feature activation+0.000
or
Token or
Feature activation+0.330
has
Token has
Feature activation+0.000
absolutely
Token absolutely
Feature activation+0.000
failed
Token failed
Feature activation+0.000
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
being
Token being
Feature activation+0.000
for
Token for
Feature activation+0.245
the
Token the
Feature activation+0.000
seizures
Token seizures
Feature activation+0.000
is
Token is
Feature activation+0.000
just
Token just
Feature activation+0.561
plain
Token plain
Feature activation+0.000
stress
Token stress
Feature activation+0.235
,
Token,
Feature activation+0.000
no
Token no
Feature activation+0.000
rest
Token rest
Feature activation+0.000
as
Token as
Feature activation+2.022
he
Token he
Feature activation+0.214
was
Token was
Feature activation+0.000
he
Token he
Feature activation+1.299
had
Token had
Feature activation+1.531
a
Token a
Feature activation+0.364
remarkable
Token remarkable
Feature activation+1.210
ability
Token ability
Feature activation+0.638
to
Token to
Feature activation+0.000
pull
Token pull
Feature activation+0.000
people
Token people
Feature activation+0.000

INTERVAL 0.000 - 0.305
CONTAINS 99.983%

:
Token:
Feature activation+0.000
Out
Token Out
Feature activation+0.000
k
Tokenk
Feature activation+0.000
ast
Tokenast
Feature activation+0.000
.
Token.
Feature activation+0.000
It
Token It
Feature activation+0.000
's
Token's
Feature activation+0.000
one
Token one
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
dick
Token dick
Feature activation+0.000
inside
Token inside
Feature activation+0.000
her
Token her
Feature activation+0.000
without
Token without
Feature activation+0.000
her
Token her
Feature activation+0.000
consent
Token consent
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
same
Token same
Feature activation+0.000
rape
Token rape
Feature activation+0.000
culture
Token culture
Feature activation+0.000
allowed
Token allowed
Feature activation+0.000
to
Token to
Feature activation+0.000
both
Token both
Feature activation+0.000
Sanders
Token Sanders
Feature activation+0.000
and
Token and
Feature activation+0.000
Hillary
Token Hillary
Feature activation+0.000
during
Token during
Feature activation+0.000
the
Token the
Feature activation+0.000
primary
Token primary
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
extremism
Token extremism
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
stated
Token stated
Feature activation+0.000
rationale
Token rationale
Feature activation+0.000
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
U
Token U
Feature activation+0.000
.
Token.
Feature activation+0.000
S
TokenS
Feature activation+0.000
.-
Token.-
Feature activation+0.000
as
Token as
Feature activation+0.000
Barrett
Token Barrett
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
es
Token es
Feature activation+0.000
oph
Tokenoph
Feature activation+0.000
agus
Tokenagus
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
prec
Token prec
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 5 in H4.3: (feature 13831)

TOP ACTIVATIONS
MAX = 2.583

it
Token it
Feature activation+0.000
could
Token could
Feature activation+2.144
âĢ
TokenâĢ
Feature activation+0.009
Ļ
TokenĻ
Feature activation+0.846
ve
Tokenve
Feature activation+0.000
been
Token been
Feature activation+2.583
avoided
Token avoided
Feature activation+2.285
.
Token.
Feature activation+0.000
Or
Token Or
Feature activation+0.000
if
Token if
Feature activation+0.000
she
Token she
Feature activation+0.000
four
Token four
Feature activation+0.000
of
Token of
Feature activation+0.000
us
Token us
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+2.209
gone
Token gone
Feature activation+2.573
quite
Token quite
Feature activation+0.578
quickly
Token quickly
Feature activation+0.664
.
Token.
Feature activation+0.000
We
Token We
Feature activation+0.000
play
Token play
Feature activation+0.000
and
Token and
Feature activation+0.000
waited
Token waited
Feature activation+0.000
for
Token for
Feature activation+0.000
it
Token it
Feature activation+0.000
to
Token to
Feature activation+1.829
come
Token come
Feature activation+2.570
by
Token by
Feature activation+0.400
so
Token so
Feature activation+0.000
we
Token we
Feature activation+0.000
could
Token could
Feature activation+0.000
see
Token see
Feature activation+0.000
have
Token have
Feature activation+0.837
had
Token had
Feature activation+0.818
something
Token something
Feature activation+0.557
to
Token to
Feature activation+0.700
do
Token do
Feature activation+2.476
with
Token with
Feature activation+2.541
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
that
Token that
Feature activation+0.319
Bart
Token Bart
Feature activation+0.000
lett
Tokenlett
Feature activation+0.000
may
Token may
Feature activation+0.072
have
Token have
Feature activation+0.837
had
Token had
Feature activation+0.818
something
Token something
Feature activation+0.557
to
Token to
Feature activation+0.700
do
Token do
Feature activation+2.476
with
Token with
Feature activation+2.541
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
that
Token that
Feature activation+0.319
Bart
Token Bart
Feature activation+0.000
pretty
Token pretty
Feature activation+0.000
bad
Token bad
Feature activation+0.000
for
Token for
Feature activation+0.000
it
Token it
Feature activation+0.000
to
Token to
Feature activation+1.638
be
Token be
Feature activation+2.466
impossible
Token impossible
Feature activation+0.272
to
Token to
Feature activation+0.000
improve
Token improve
Feature activation+0.000
anything
Token anything
Feature activation+0.000
else
Token else
Feature activation+0.000
really
Token really
Feature activation+0.659
had
Token had
Feature activation+0.271
nothing
Token nothing
Feature activation+0.581
to
Token to
Feature activation+0.838
do
Token do
Feature activation+2.258
with
Token with
Feature activation+2.438
the
Token the
Feature activation+0.000
question
Token question
Feature activation+0.000
whatsoever
Token whatsoever
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
and
Token and
Feature activation+0.000
most
Token most
Feature activation+0.000
upsetting
Token upsetting
Feature activation+1.955
,
Token,
Feature activation+0.000
still
Token still
Feature activation+0.178
relevant
Token relevant
Feature activation+2.411
to
Token to
Feature activation+0.867
the
Token the
Feature activation+0.000
world
Token world
Feature activation+0.000
today
Token today
Feature activation+0.000
.
Token.
Feature activation+0.000
allow
Token allow
Feature activation+0.000
[
Token [
Feature activation+0.000
it
Tokenit
Feature activation+0.000
]
Token]
Feature activation+1.172
to
Token to
Feature activation+2.217
be
Token be
Feature activation+2.367
submerged
Token submerged
Feature activation+0.744
by
Token by
Feature activation+0.224
a
Token a
Feature activation+0.000
wave
Token wave
Feature activation+0.000
of
Token of
Feature activation+0.000
it
Token it
Feature activation+0.000
had
Token had
Feature activation+0.383
something
Token something
Feature activation+0.423
to
Token to
Feature activation+0.140
do
Token do
Feature activation+2.071
with
Token with
Feature activation+2.367
Easter
Token Easter
Feature activation+0.000
,
Token,
Feature activation+0.000
obviously
Token obviously
Feature activation+0.000
âĢ¦
TokenâĢ¦
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
could
Token could
Feature activation+2.144
âĢ
TokenâĢ
Feature activation+0.009
Ļ
TokenĻ
Feature activation+0.846
ve
Tokenve
Feature activation+0.000
been
Token been
Feature activation+2.583
avoided
Token avoided
Feature activation+2.285
.
Token.
Feature activation+0.000
Or
Token Or
Feature activation+0.000
if
Token if
Feature activation+0.000
she
Token she
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
it
Token it
Feature activation+0.986
didn
Token didn
Feature activation+1.755
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.319
work
Token work
Feature activation+2.273
is
Token is
Feature activation+0.000
because
Token because
Feature activation+0.079
after
Token after
Feature activation+0.000
an
Token an
Feature activation+0.000
illicit
Token illicit
Feature activation+0.000
It
TokenIt
Feature activation+0.000
really
Token really
Feature activation+0.659
had
Token had
Feature activation+0.271
nothing
Token nothing
Feature activation+0.581
to
Token to
Feature activation+0.838
do
Token do
Feature activation+2.258
with
Token with
Feature activation+2.438
the
Token the
Feature activation+0.000
question
Token question
Feature activation+0.000
whatsoever
Token whatsoever
Feature activation+0.000
.
Token.
Feature activation+0.000
it
Token it
Feature activation+0.000
didn
Token didn
Feature activation+1.244
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.564
get
Token get
Feature activation+2.255
published
Token published
Feature activation+1.849
until
Token until
Feature activation+0.000
1967
Token 1967
Feature activation+0.000
.
Token.
Feature activation+0.000
This
Token This
Feature activation+0.000
not
Token not
Feature activation+0.000
allow
Token allow
Feature activation+0.000
[
Token [
Feature activation+0.000
it
Tokenit
Feature activation+0.000
]
Token]
Feature activation+1.172
to
Token to
Feature activation+2.217
be
Token be
Feature activation+2.367
submerged
Token submerged
Feature activation+0.744
by
Token by
Feature activation+0.224
a
Token a
Feature activation+0.000
wave
Token wave
Feature activation+0.000
the
Token the
Feature activation+0.000
four
Token four
Feature activation+0.000
of
Token of
Feature activation+0.000
us
Token us
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+2.209
gone
Token gone
Feature activation+2.573
quite
Token quite
Feature activation+0.578
quickly
Token quickly
Feature activation+0.664
.
Token.
Feature activation+0.000
We
Token We
Feature activation+0.000
If
TokenIf
Feature activation+0.000
it
Token it
Feature activation+0.000
turns
Token turns
Feature activation+0.000
out
Token out
Feature activation+1.033
to
Token to
Feature activation+1.646
be
Token be
Feature activation+2.195
Jeb
Token Jeb
Feature activation+0.000
versus
Token versus
Feature activation+0.000
Hillary
Token Hillary
Feature activation+0.000
we
Token we
Feature activation+0.000
would
Token would
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
where
Token where
Feature activation+0.000
it
Token it
Feature activation+0.165
starts
Token starts
Feature activation+2.189
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
more
Token more
Feature activation+0.000
society
Token society
Feature activation+0.000
in
Token in
Feature activation+0.000
,
Token,
Feature activation+0.000
it
Token it
Feature activation+0.000
had
Token had
Feature activation+0.375
nothing
Token nothing
Feature activation+0.355
to
Token to
Feature activation+0.628
do
Token do
Feature activation+2.158
with
Token with
Feature activation+1.958
today
Token today
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.764
come
Token come
Feature activation+1.252
as
Token as
Feature activation+1.010
a
Token a
Feature activation+0.602
surprise
Token surprise
Feature activation+2.156
as
Token as
Feature activation+0.266
nobody
Token nobody
Feature activation+0.000
knows
Token knows
Feature activation+0.000
the
Token the
Feature activation+0.000
character
Token character
Feature activation+0.000

Top DFA by src position
MAX = 3.427

t
Tokent
Feature activation-0.002
Top resid features:
broken
Token broken
Feature activation+0.000
Top resid features:
up
Token up
Feature activation+0.032
Top resid features:
with
Token with
Feature activation+0.006
Top resid features:
me
Token me
Feature activation+0.004
Top resid features:
it
Token it
Feature activation+3.234
Top resid features:
could
Token could
Feature activation+0.014
Top resid features:
âĢ
TokenâĢ
Feature activation-0.062
Top resid features:
Ļ
TokenĻ
Feature activation-0.007
Top resid features:
ve
Tokenve
Feature activation+0.046
Top resid features:
been
Token been
Feature activation-0.166
Top resid features:
with
Token with
Feature activation+0.007
Top resid features:
the
Token the
Feature activation+0.006
Top resid features:
four
Token four
Feature activation-0.022
Top resid features:
of
Token of
Feature activation+0.005
Top resid features:
us
Token us
Feature activation+0.002
Top resid features:
it
Token it
Feature activation+3.427
Top resid features:
has
Token has
Feature activation-0.108
Top resid features:
gone
Token gone
Feature activation-0.145
Top resid features:
quite
Token quite
Feature activation+0.000
Top resid features:
quickly
Token quickly
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
pulled
Token pulled
Feature activation+0.010
Top resid features:
over
Token over
Feature activation+0.024
Top resid features:
and
Token and
Feature activation+0.013
Top resid features:
waited
Token waited
Feature activation-0.031
Top resid features:
for
Token for
Feature activation+0.049
Top resid features:
it
Token it
Feature activation+3.358
Top resid features:
to
Token to
Feature activation-0.012
Top resid features:
come
Token come
Feature activation-0.191
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
so
Token so
Feature activation+0.000
Top resid features:
we
Token we
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.004
Top resid features:
pear
Token pear
Feature activation-0.000
Top resid features:
industry
Token industry
Feature activation-0.002
Top resid features:
.
Token.
Feature activation-0.010
Top resid features:
(
Token (
Feature activation-0.031
Top resid features:
It
TokenIt
Feature activation+3.130
Top resid features:
may
Token may
Feature activation-0.004
Top resid features:
have
Token have
Feature activation+0.019
Top resid features:
had
Token had
Feature activation+0.025
Top resid features:
something
Token something
Feature activation+0.061
Top resid features:
to
Token to
Feature activation+0.025
Top resid features:
the
Token the
Feature activation+0.004
Top resid features:
pear
Token pear
Feature activation-0.000
Top resid features:
industry
Token industry
Feature activation-0.003
Top resid features:
.
Token.
Feature activation-0.008
Top resid features:
(
Token (
Feature activation-0.030
Top resid features:
It
TokenIt
Feature activation+2.938
Top resid features:
may
Token may
Feature activation+0.009
Top resid features:
have
Token have
Feature activation+0.034
Top resid features:
had
Token had
Feature activation+0.059
Top resid features:
something
Token something
Feature activation+0.114
Top resid features:
to
Token to
Feature activation-0.046
Top resid features:
them
Token them
Feature activation+0.035
Top resid features:
up
Token up
Feature activation+0.037
Top resid features:
pretty
Token pretty
Feature activation+0.002
Top resid features:
bad
Token bad
Feature activation-0.025
Top resid features:
for
Token for
Feature activation+0.034
Top resid features:
it
Token it
Feature activation+2.999
Top resid features:
to
Token to
Feature activation+0.098
Top resid features:
be
Token be
Feature activation-0.160
Top resid features:
impossible
Token impossible
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
improve
Token improve
Feature activation+0.000
Top resid features:
told
Token told
Feature activation-0.001
Top resid features:
me
Token me
Feature activation-0.000
Top resid features:
.
Token.
Feature activation-0.013
Top resid features:
âĢ
Token âĢ
Feature activation-0.011
Top resid features:
ľ
Tokenľ
Feature activation+0.004
Top resid features:
It
TokenIt
Feature activation+2.983
Top resid features:
really
Token really
Feature activation-0.017
Top resid features:
had
Token had
Feature activation+0.052
Top resid features:
nothing
Token nothing
Feature activation+0.045
Top resid features:
to
Token to
Feature activation+0.024
Top resid features:
do
Token do
Feature activation-0.043
Top resid features:
the
Token the
Feature activation-0.000
Top resid features:
time
Token time
Feature activation-0.004
Top resid features:
.
Token.
Feature activation-0.012
Top resid features:
I
Token I
Feature activation-0.008
Top resid features:
found
Token found
Feature activation+0.067
Top resid features:
it
Token it
Feature activation+2.808
Top resid features:
fascinating
Token fascinating
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.009
Top resid features:
gruesome
Token gruesome
Feature activation+0.008
Top resid features:
,
Token,
Feature activation-0.008
Top resid features:
and
Token and
Feature activation+0.020
Top resid features:
We
TokenWe
Feature activation+0.003
Top resid features:
will
Token will
Feature activation-0.006
Top resid features:
not
Token not
Feature activation-0.017
Top resid features:
allow
Token allow
Feature activation+0.099
Top resid features:
[
Token [
Feature activation+0.023
Top resid features:
it
Tokenit
Feature activation+2.857
Top resid features:
]
Token]
Feature activation+0.050
Top resid features:
to
Token to
Feature activation+0.056
Top resid features:
be
Token be
Feature activation-0.063
Top resid features:
submerged
Token submerged
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
I
Token I
Feature activation-0.021
Top resid features:
âĢ
TokenâĢ
Feature activation-0.046
Top resid features:
Ļ
TokenĻ
Feature activation+0.000
Top resid features:
m
Tokenm
Feature activation+0.022
Top resid features:
sure
Token sure
Feature activation+0.005
Top resid features:
it
Token it
Feature activation+2.907
Top resid features:
had
Token had
Feature activation+0.031
Top resid features:
something
Token something
Feature activation+0.106
Top resid features:
to
Token to
Feature activation+0.043
Top resid features:
do
Token do
Feature activation-0.042
Top resid features:
with
Token with
Feature activation+0.025
Top resid features:
t
Tokent
Feature activation+0.006
Top resid features:
broken
Token broken
Feature activation+0.015
Top resid features:
up
Token up
Feature activation+0.027
Top resid features:
with
Token with
Feature activation+0.014
Top resid features:
me
Token me
Feature activation+0.003
Top resid features:
it
Token it
Feature activation+2.825
Top resid features:
could
Token could
Feature activation+0.050
Top resid features:
âĢ
TokenâĢ
Feature activation-0.079
Top resid features:
Ļ
TokenĻ
Feature activation-0.007
Top resid features:
ve
Tokenve
Feature activation-0.002
Top resid features:
been
Token been
Feature activation+0.043
Top resid features:
.
Token.
Feature activation-0.011
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
The
TokenThe
Feature activation+0.009
Top resid features:
reason
Token reason
Feature activation+0.072
Top resid features:
it
Token it
Feature activation+2.996
Top resid features:
didn
Token didn
Feature activation+0.027
Top resid features:
âĢ
TokenâĢ
Feature activation-0.060
Top resid features:
Ļ
TokenĻ
Feature activation-0.015
Top resid features:
t
Tokent
Feature activation+0.010
Top resid features:
work
Token work
Feature activation-0.148
Top resid features:
told
Token told
Feature activation+0.001
Top resid features:
me
Token me
Feature activation-0.003
Top resid features:
.
Token.
Feature activation-0.008
Top resid features:
âĢ
Token âĢ
Feature activation-0.004
Top resid features:
ľ
Tokenľ
Feature activation+0.003
Top resid features:
It
TokenIt
Feature activation+2.743
Top resid features:
really
Token really
Feature activation-0.038
Top resid features:
had
Token had
Feature activation+0.066
Top resid features:
nothing
Token nothing
Feature activation+0.076
Top resid features:
to
Token to
Feature activation+0.035
Top resid features:
do
Token do
Feature activation+0.041
Top resid features:
being
Token being
Feature activation+0.001
Top resid features:
Stalin
Token Stalin
Feature activation+0.000
Top resid features:
ist
Tokenist
Feature activation-0.005
Top resid features:
Russia
Token Russia
Feature activation+0.000
Top resid features:
)
Token)
Feature activation+0.005
Top resid features:
it
Token it
Feature activation+2.979
Top resid features:
didn
Token didn
Feature activation+0.075
Top resid features:
âĢ
TokenâĢ
Feature activation-0.065
Top resid features:
Ļ
TokenĻ
Feature activation-0.013
Top resid features:
t
Tokent
Feature activation+0.019
Top resid features:
get
Token get
Feature activation-0.058
Top resid features:
We
TokenWe
Feature activation+0.010
Top resid features:
will
Token will
Feature activation-0.010
Top resid features:
not
Token not
Feature activation-0.010
Top resid features:
allow
Token allow
Feature activation+0.062
Top resid features:
[
Token [
Feature activation+0.029
Top resid features:
it
Tokenit
Feature activation+2.897
Top resid features:
]
Token]
Feature activation-0.080
Top resid features:
to
Token to
Feature activation-0.052
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
submerged
Token submerged
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.018
Top resid features:
the
Token the
Feature activation+0.010
Top resid features:
four
Token four
Feature activation-0.036
Top resid features:
of
Token of
Feature activation+0.011
Top resid features:
us
Token us
Feature activation-0.007
Top resid features:
it
Token it
Feature activation+3.025
Top resid features:
has
Token has
Feature activation-0.185
Top resid features:
gone
Token gone
Feature activation+0.000
Top resid features:
quite
Token quite
Feature activation+0.000
Top resid features:
quickly
Token quickly
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.013
Top resid features:
ľ
Tokenľ
Feature activation-0.003
Top resid features:
If
TokenIf
Feature activation+0.067
Top resid features:
it
Token it
Feature activation+2.861
Top resid features:
turns
Token turns
Feature activation+0.012
Top resid features:
out
Token out
Feature activation+0.014
Top resid features:
to
Token to
Feature activation+0.055
Top resid features:
be
Token be
Feature activation-0.137
Top resid features:
Jeb
Token Jeb
Feature activation+0.000
Top resid features:
That
Token That
Feature activation+0.004
Top resid features:
âĢ
TokenâĢ
Feature activation-0.039
Top resid features:
Ļ
TokenĻ
Feature activation-0.005
Top resid features:
s
Tokens
Feature activation-0.002
Top resid features:
where
Token where
Feature activation+0.138
Top resid features:
it
Token it
Feature activation+2.880
Top resid features:
starts
Token starts
Feature activation-0.112
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
The
Token The
Feature activation+0.000
Top resid features:
more
Token more
Feature activation+0.000
Top resid features:
society
Token society
Feature activation+0.000
Top resid features:
-
Token-
Feature activation-0.000
Top resid features:
trans
Tokentrans
Feature activation+0.004
Top resid features:
forming
Tokenforming
Feature activation-0.002
Top resid features:
tool
Token tool
Feature activation+0.003
Top resid features:
,
Token,
Feature activation+0.019
Top resid features:
it
Token it
Feature activation+2.506
Top resid features:
had
Token had
Feature activation+0.063
Top resid features:
nothing
Token nothing
Feature activation+0.147
Top resid features:
to
Token to
Feature activation+0.023
Top resid features:
do
Token do
Feature activation+0.036
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
be
Token be
Feature activation-0.005
Top resid features:
a
Token a
Feature activation+0.003
Top resid features:
massive
Token massive
Feature activation-0.003
Top resid features:
hit
Token hit
Feature activation-0.003
Top resid features:
.
Token.
Feature activation-0.018
Top resid features:
It
Token It
Feature activation+2.687
Top resid features:
really
Token really
Feature activation-0.013
Top resid features:
shouldn
Token shouldn
Feature activation+0.055
Top resid features:
âĢ
TokenâĢ
Feature activation-0.047
Top resid features:
Ļ
TokenĻ
Feature activation+0.001
Top resid features:
t
Tokent
Feature activation+0.036
Top resid features:

Decoder Weights Distribution

Head 0: 0.05

Head 1: 0.04

Head 2: 0.02

Head 3: 0.48

Head 4: 0.04

Head 5: 0.05

Head 6: 0.05

Head 7: 0.09

Head 8: 0.02

Head 9: 0.06

Head 10: 0.01

Head 11: 0.09

Positive logits

raining2.22

easier1.40

aspir1.34

impossible1.32

uphill1.31

costs1.25

cheaper1.24

depressing1.22

feasible1.16

easy1.15

incumbent1.14

unbeliev1.11

aggrav1.11

advisable1.11

gross1.11

accidents1.11

sequels1.10

interstate1.09

economical1.08

overload1.07

Negative logits

arta-1.52

76561-1.41

isse-1.32

otos-1.29

dor-1.28

raid-1.26

kan-1.25

tor-1.24

anton-1.23

CARE-1.22

mire-1.21

ickr-1.20

row-1.17

inka-1.17

andum-1.17

CHA-1.16

-1.15

FFER-1.15

cussion-1.15

avage-1.15

INTERVAL 2.325 - 2.583
CONTAINS 0.001%

allow
Token allow
Feature activation+0.000
[
Token [
Feature activation+0.000
it
Tokenit
Feature activation+0.000
]
Token]
Feature activation+1.172
to
Token to
Feature activation+2.217
be
Token be
Feature activation+2.367
submerged
Token submerged
Feature activation+0.744
by
Token by
Feature activation+0.224
a
Token a
Feature activation+0.000
wave
Token wave
Feature activation+0.000
of
Token of
Feature activation+0.000
pretty
Token pretty
Feature activation+0.000
bad
Token bad
Feature activation+0.000
for
Token for
Feature activation+0.000
it
Token it
Feature activation+0.000
to
Token to
Feature activation+1.638
be
Token be
Feature activation+2.466
impossible
Token impossible
Feature activation+0.272
to
Token to
Feature activation+0.000
improve
Token improve
Feature activation+0.000
anything
Token anything
Feature activation+0.000
else
Token else
Feature activation+0.000
have
Token have
Feature activation+0.837
had
Token had
Feature activation+0.818
something
Token something
Feature activation+0.557
to
Token to
Feature activation+0.700
do
Token do
Feature activation+2.476
with
Token with
Feature activation+2.541
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
that
Token that
Feature activation+0.319
Bart
Token Bart
Feature activation+0.000
lett
Tokenlett
Feature activation+0.000
it
Token it
Feature activation+0.000
could
Token could
Feature activation+2.144
âĢ
TokenâĢ
Feature activation+0.009
Ļ
TokenĻ
Feature activation+0.846
ve
Tokenve
Feature activation+0.000
been
Token been
Feature activation+2.583
avoided
Token avoided
Feature activation+2.285
.
Token.
Feature activation+0.000
Or
Token Or
Feature activation+0.000
if
Token if
Feature activation+0.000
she
Token she
Feature activation+0.000
and
Token and
Feature activation+0.000
most
Token most
Feature activation+0.000
upsetting
Token upsetting
Feature activation+1.955
,
Token,
Feature activation+0.000
still
Token still
Feature activation+0.178
relevant
Token relevant
Feature activation+2.411
to
Token to
Feature activation+0.867
the
Token the
Feature activation+0.000
world
Token world
Feature activation+0.000
today
Token today
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 2.067 - 2.325
CONTAINS 0.002%

it
Token it
Feature activation+0.986
didn
Token didn
Feature activation+1.755
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.319
work
Token work
Feature activation+2.273
is
Token is
Feature activation+0.000
because
Token because
Feature activation+0.079
after
Token after
Feature activation+0.000
an
Token an
Feature activation+0.000
illicit
Token illicit
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
where
Token where
Feature activation+0.000
it
Token it
Feature activation+0.165
starts
Token starts
Feature activation+2.189
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
more
Token more
Feature activation+0.000
society
Token society
Feature activation+0.000
in
Token in
Feature activation+0.000
it
Token it
Feature activation+0.200
doesn
Token doesn
Feature activation+1.041
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.172
t
Tokent
Feature activation+1.178
put
Token put
Feature activation+2.084
a
Token a
Feature activation+0.181
foot
Token foot
Feature activation+0.000
wrong
Token wrong
Feature activation+0.000
,
Token,
Feature activation+0.010
because
Token because
Feature activation+0.000
it
Token it
Feature activation+0.000
didn
Token didn
Feature activation+1.244
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.564
get
Token get
Feature activation+2.255
published
Token published
Feature activation+1.849
until
Token until
Feature activation+0.000
1967
Token 1967
Feature activation+0.000
.
Token.
Feature activation+0.000
This
Token This
Feature activation+0.000
,
Token,
Feature activation+0.000
it
Token it
Feature activation+0.000
had
Token had
Feature activation+0.375
nothing
Token nothing
Feature activation+0.355
to
Token to
Feature activation+0.628
do
Token do
Feature activation+2.158
with
Token with
Feature activation+1.958
today
Token today
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000

INTERVAL 1.808 - 2.067
CONTAINS 0.006%

over
Token over
Feature activation+0.000
and
Token and
Feature activation+0.000
waited
Token waited
Feature activation+0.000
for
Token for
Feature activation+0.000
it
Token it
Feature activation+0.000
to
Token to
Feature activation+1.829
come
Token come
Feature activation+2.570
by
Token by
Feature activation+0.400
so
Token so
Feature activation+0.000
we
Token we
Feature activation+0.000
could
Token could
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
It
TokenIt
Feature activation+0.000
seems
Token seems
Feature activation+0.000
to
Token to
Feature activation+0.236
have
Token have
Feature activation+0.998
done
Token done
Feature activation+2.047
the
Token the
Feature activation+0.041
trick
Token trick
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
now
Token now
Feature activation+0.000
Why
TokenWhy
Feature activation+0.000
it
Token it
Feature activation+0.000
Doesn
Token Doesn
Feature activation+1.769
't
Token't
Feature activation+1.986
Make
Token Make
Feature activation+1.643
Sense
Token Sense
Feature activation+1.963
:
Token:
Feature activation+0.004
Ċ
TokenĊ
Feature activation+0.404
Ċ
TokenĊ
Feature activation+0.033
Jay
TokenJay
Feature activation+0.000
Gr
Token Gr
Feature activation+0.000
it
Token it
Feature activation+0.182
wouldn
Token wouldn
Feature activation+1.823
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.865
hit
Token hit
Feature activation+1.910
the
Token the
Feature activation+0.000
planet
Token planet
Feature activation+0.000
till
Token till
Feature activation+0.000
March
Token March
Feature activation+0.000
16
Token 16
Feature activation+0.000
certainly
Token certainly
Feature activation+0.517
doesn
Token doesn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.015
warrant
Token warrant
Feature activation+1.870
such
Token such
Feature activation+0.000
extended
Token extended
Feature activation+0.000
disruption
Token disruption
Feature activation+0.000
of
Token of
Feature activation+0.000
parliament
Token parliament
Feature activation+0.000

INTERVAL 1.550 - 1.808
CONTAINS 0.010%

it
Token it
Feature activation+0.195
just
Token just
Feature activation+0.601
well
Token well
Feature activation+0.553
,
Token,
Feature activation+0.000
colon
Token colon
Feature activation+0.205
izes
Tokenizes
Feature activation+1.675
North
Token North
Feature activation+0.000
Africa
Token Africa
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
I
TokenI
Feature activation+0.000
can
Token can
Feature activation+0.000
understand
Token understand
Feature activation+0.000
what
Token what
Feature activation+0.000
it
Token it
Feature activation+0.000
means
Token means
Feature activation+1.675
to
Token to
Feature activation+0.878
be
Token be
Feature activation+0.000
Indian
Token Indian
Feature activation+0.000
,
Token,
Feature activation+0.000
because
Token because
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.076
s
Tokens
Feature activation+0.014
darn
Token darn
Feature activation+0.000
ed
Tokened
Feature activation+0.000
peculiar
Token peculiar
Feature activation+1.555
to
Token to
Feature activation+1.455
say
Token say
Feature activation+1.292
the
Token the
Feature activation+0.000
least
Token least
Feature activation+0.201
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
reason
Token reason
Feature activation+0.000
it
Token it
Feature activation+0.986
didn
Token didn
Feature activation+1.755
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.319
work
Token work
Feature activation+2.273
is
Token is
Feature activation+0.000
accept
Token accept
Feature activation+0.000
that
Token that
Feature activation+0.000
it
Token it
Feature activation+0.028
has
Token has
Feature activation+1.162
to
Token to
Feature activation+1.179
be
Token be
Feature activation+1.758
this
Token this
Feature activation+0.549
way
Token way
Feature activation+0.403
.
Token.
Feature activation+0.000
Phen
Token Phen
Feature activation+0.000
omen
Tokenomen
Feature activation+0.000

INTERVAL 1.292 - 1.550
CONTAINS 0.022%

a
Token a
Feature activation+0.000
reserved
Token reserved
Feature activation+0.000
word
Token word
Feature activation+0.000
it
Token it
Feature activation+0.000
may
Token may
Feature activation+0.908
be
Token be
Feature activation+1.367
used
Token used
Feature activation+0.653
as
Token as
Feature activation+0.333
a
Token a
Feature activation+0.556
function
Token function
Feature activation+0.000
argument
Token argument
Feature activation+0.000
White
Token White
Feature activation+0.000
House
Token House
Feature activation+0.000
for
Token for
Feature activation+0.000
making
Token making
Feature activation+0.000
it
Token it
Feature activation+0.000
harder
Token harder
Feature activation+1.459
on
Token on
Feature activation+0.000
veterans
Token veterans
Feature activation+0.000
and
Token and
Feature activation+0.000
playing
Token playing
Feature activation+0.000
politics
Token politics
Feature activation+0.000
some
Token some
Feature activation+0.000
suggesting
Token suggesting
Feature activation+0.000
it
Token it
Feature activation+0.000
needs
Token needs
Feature activation+0.334
to
Token to
Feature activation+0.815
be
Token be
Feature activation+1.408
ner
Token ner
Feature activation+0.402
fed
Tokenfed
Feature activation+0.443
/
Token/
Feature activation+0.000
rem
Tokenrem
Feature activation+0.000
oved
Tokenoved
Feature activation+0.103
it
Token it
Feature activation+0.000
's
Token's
Feature activation+0.000
been
Token been
Feature activation+0.630
at
Token at
Feature activation+0.549
the
Token the
Feature activation+0.346
forefront
Token forefront
Feature activation+1.318
of
Token of
Feature activation+1.888
this
Token this
Feature activation+0.000
league
Token league
Feature activation+0.000
and
Token and
Feature activation+0.000
I
Token I
Feature activation+0.000
it
Token it
Feature activation+0.000
isn
Token isn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.415
legal
Token legal
Feature activation+1.486
currently
Token currently
Feature activation+0.943
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
board
Token board
Feature activation+0.000
was
Token was
Feature activation+0.000

INTERVAL 1.033 - 1.292
CONTAINS 0.044%

it
Token it
Feature activation+0.000
doesn
Token doesn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+1.173
get
Token get
Feature activation+1.212
in
Token in
Feature activation+0.005
the
Token the
Feature activation+0.000
way
Token way
Feature activation+0.000
.
Token.
Feature activation+0.000
For
Token For
Feature activation+0.000
it
Token it
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
done
Token done
Feature activation+0.823
to
Token to
Feature activation+1.081
you
Token you
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
3
Token 3
Feature activation+0.000
eth
Token eth
Feature activation+0.369
ically
Tokenically
Feature activation+0.232
right
Token right
Feature activation+0.691
but
Token but
Feature activation+0.000
will
Token will
Feature activation+1.219
also
Token also
Feature activation+1.133
have
Token have
Feature activation+1.811
a
Token a
Feature activation+0.305
huge
Token huge
Feature activation+0.000
impact
Token impact
Feature activation+0.000
on
Token on
Feature activation+0.052
But
Token But
Feature activation+0.000
it
Token it
Feature activation+0.000
is
Token is
Feature activation+0.000
a
Token a
Feature activation+0.000
recurring
Token recurring
Feature activation+0.000
role
Token role
Feature activation+1.051
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
since
Token since
Feature activation+0.000
the
Token the
Feature activation+0.000
introduction
Token introduction
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
area
Token area
Feature activation+0.000
,
Token,
Feature activation+0.000
it
Token it
Feature activation+0.000
can
Token can
Feature activation+1.079
kick
Token kick
Feature activation+0.444
-
Token-
Feature activation+0.000
start
Tokenstart
Feature activation+1.064
the
Token the
Feature activation+0.000
area
Token area
Feature activation+0.000

INTERVAL 0.775 - 1.033
CONTAINS 0.077%

I
TokenI
Feature activation+0.000
assume
Token assume
Feature activation+0.000
it
Token it
Feature activation+0.019
's
Token's
Feature activation+0.066
going
Token going
Feature activation+1.196
to
Token to
Feature activation+0.908
stop
Token stop
Feature activation+1.595
,"
Token,"
Feature activation+0.000
he
Token he
Feature activation+0.000
said
Token said
Feature activation+0.000
.
Token.
Feature activation+0.000
easy
Token easy
Feature activation+0.370
leap
Token leap
Feature activation+0.000
to
Token to
Feature activation+0.000
say
Token say
Feature activation+0.000
it
Token it
Feature activation+0.000
would
Token would
Feature activation+0.857
make
Token make
Feature activation+0.773
getting
Token getting
Feature activation+0.000
at
Token at
Feature activation+0.000
customer
Token customer
Feature activation+0.000
data
Token data
Feature activation+0.000
returned
Token returned
Feature activation+0.000
it
Token it
Feature activation+0.000
to
Token to
Feature activation+0.546
the
Token the
Feature activation+0.210
United
Token United
Feature activation+0.000
States
Token States
Feature activation+0.798
.
Token.
Feature activation+0.000
Instead
Token Instead
Feature activation+0.000
,
Token,
Feature activation+0.000
it
Token it
Feature activation+0.000
sits
Token sits
Feature activation+0.000
And
Token And
Feature activation+0.000
it
Token it
Feature activation+0.000
was
Token was
Feature activation+0.000
sw
Token sw
Feature activation+0.000
illing
Tokenilling
Feature activation+0.471
over
Token over
Feature activation+0.944
the
Token the
Feature activation+0.000
poor
Token poor
Feature activation+0.000
man
Token man
Feature activation+0.000
who
Token who
Feature activation+0.000
,
Token,
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
On
TokenOn
Feature activation+0.000
its
Token its
Feature activation+0.000
face
Token face
Feature activation+1.023
,
Token,
Feature activation+0.276
the
Token the
Feature activation+0.000
chord
Token chord
Feature activation+0.000
progression
Token progression
Feature activation+0.000
to
Token to
Feature activation+0.000

INTERVAL 0.517 - 0.775
CONTAINS 0.123%

In
Token In
Feature activation+0.000
one
Token one
Feature activation+0.000
sense
Token sense
Feature activation+0.000
,
Token,
Feature activation+0.000
it
Token it
Feature activation+0.000
doesn
Token doesn
Feature activation+0.599
't
Token't
Feature activation+0.891
quite
Token quite
Feature activation+0.690
matter
Token matter
Feature activation+0.000
.
Token.
Feature activation+0.000
David
Token David
Feature activation+0.000
if
Token if
Feature activation+0.000
it
Token it
Feature activation+0.000
is
Token is
Feature activation+0.000
borne
Token borne
Feature activation+0.000
out
Token out
Feature activation+0.461
by
Token by
Feature activation+0.754
observations
Token observations
Feature activation+0.000
,
Token,
Feature activation+0.000
would
Token would
Feature activation+0.086
help
Token help
Feature activation+0.000
explain
Token explain
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
It
TokenIt
Feature activation+0.000
seems
Token seems
Feature activation+0.000
hard
Token hard
Feature activation+0.000
to
Token to
Feature activation+0.070
believe
Token believe
Feature activation+0.519
someone
Token someone
Feature activation+0.000
would
Token would
Feature activation+0.000
resort
Token resort
Feature activation+0.000
to
Token to
Feature activation+0.000
injecting
Token injecting
Feature activation+0.000
So
TokenSo
Feature activation+0.000
it
Token it
Feature activation+0.000
wasn
Token wasn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.552
surprising
Token surprising
Feature activation+0.843
to
Token to
Feature activation+0.983
hear
Token hear
Feature activation+0.873
McC
Token McC
Feature activation+0.000
oll
Tokenoll
Feature activation+0.000
must
Token must
Feature activation+0.000
use
Token use
Feature activation+0.000
if
Token if
Feature activation+0.000
it
Token it
Feature activation+0.000
wants
Token wants
Feature activation+0.000
to
Token to
Feature activation+0.539
throw
Token throw
Feature activation+0.061
a
Token a
Feature activation+0.000
full
Token full
Feature activation+0.000
light
Token light
Feature activation+0.000
on
Token on
Feature activation+0.000

INTERVAL 0.258 - 0.517
CONTAINS 0.212%

are
Token are
Feature activation+0.000
Catholics
Token Catholics
Feature activation+0.000
and
Token and
Feature activation+0.000
it
Token it
Feature activation+0.000
is
Token is
Feature activation+0.149
described
Token described
Feature activation+0.346
as
Token as
Feature activation+1.393
one
Token one
Feature activation+1.379
of
Token of
Feature activation+1.555
the
Token the
Feature activation+0.963
five
Token five
Feature activation+0.000
I
Token I
Feature activation+0.000
fear
Token fear
Feature activation+0.000
it
Token it
Feature activation+0.167
could
Token could
Feature activation+1.285
be
Token be
Feature activation+1.103
far
Token far
Feature activation+0.516
too
Token too
Feature activation+0.702
long
Token long
Feature activation+0.098
before
Token before
Feature activation+0.045
the
Token the
Feature activation+0.000
book
Token book
Feature activation+0.000
.
Token.
Feature activation+0.000
Some
Token Some
Feature activation+0.000
of
Token of
Feature activation+0.000
how
Token how
Feature activation+0.000
it
Token it
Feature activation+0.389
works
Token works
Feature activation+0.265
,
Token,
Feature activation+0.035
its
Token its
Feature activation+0.362
goals
Token goals
Feature activation+0.585
,
Token,
Feature activation+0.000
specifications
Token specifications
Feature activation+0.000
was
Token was
Feature activation+0.000
a
Token a
Feature activation+0.000
spur
Token spur
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
moment
Token moment
Feature activation+0.493
thing
Token thing
Feature activation+0.130
,"
Token,"
Feature activation+0.000
Ma
Token Ma
Feature activation+0.000
ior
Tokenior
Feature activation+0.000
ana
Tokenana
Feature activation+0.000
very
Token very
Feature activation+0.666
much
Token much
Feature activation+0.000
like
Token like
Feature activation+0.084
a
Token a
Feature activation+0.297
coming
Token coming
Feature activation+0.000
together
Token together
Feature activation+0.471
of
Token of
Feature activation+0.000
anybody
Token anybody
Feature activation+0.000
who
Token who
Feature activation+0.000
valued
Token valued
Feature activation+0.000
human
Token human
Feature activation+0.000

INTERVAL 0.000 - 0.258
CONTAINS 99.502%

to
Token to
Feature activation+0.000
consumers
Token consumers
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
A
TokenA
Feature activation+0.000
metaph
Token metaph
Feature activation+0.000
orical
Tokenorical
Feature activation+0.000
slap
Token slap
Feature activation+0.000
high
Token high
Feature activation+0.000
speed
Token speed
Feature activation+0.000
also
Token also
Feature activation+0.000
came
Token came
Feature activation+0.000
up
Token up
Feature activation+0.000
in
Token in
Feature activation+0.000
discussion
Token discussion
Feature activation+0.000
âĢĶ
TokenâĢĶ
Feature activation+0.000
the
Tokenthe
Feature activation+0.000
phenomenon
Token phenomenon
Feature activation+0.000
would
Token would
Feature activation+0.000
the
Token the
Feature activation+0.000
Gra
Token Gra
Feature activation+0.000
etz
Tokenetz
Feature activation+0.000
legacy
Token legacy
Feature activation+0.000
.
Token.
Feature activation+0.000
He
Token He
Feature activation+0.000
launched
Token launched
Feature activation+0.000
a
Token a
Feature activation+0.000
propaganda
Token propaganda
Feature activation+0.000
campaign
Token campaign
Feature activation+0.000
of
Token of
Feature activation+0.000
fibre
Token fibre
Feature activation+0.000
and
Token and
Feature activation+0.000
calcium
Token calcium
Feature activation+0.000
.
Token.
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
kan
Token kan
Feature activation+0.000
ji
Tokenji
Feature activation+0.000
with
Token with
Feature activation+0.000
a
Token a
Feature activation+0.000
different
Token different
Feature activation+0.000
street
Token street
Feature activation+0.000
and
Token and
Feature activation+0.000
not
Token not
Feature activation+0.000
be
Token be
Feature activation+0.000
made
Token made
Feature activation+0.000
fun
Token fun
Feature activation+0.000
of
Token of
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 6 in H4.3: (feature 9039)

TOP ACTIVATIONS
MAX = 2.358

a
Token a
Feature activation+0.000
Tus
Token Tus
Feature activation+0.000
can
Tokencan
Feature activation+0.000
setting
Token setting
Feature activation+0.000
,
Token,
Feature activation+0.000
was
Token was
Feature activation+2.358
surrounded
Token surrounded
Feature activation+2.285
by
Token by
Feature activation+1.030
manic
Token manic
Feature activation+0.000
ured
Tokenured
Feature activation+0.000
hed
Token hed
Feature activation+0.000
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
GM
Token GM
Feature activation+0.000
plant
Token plant
Feature activation+0.076
was
Token was
Feature activation+1.688
firing
Token firing
Feature activation+2.294
750
Token 750
Feature activation+0.000
people
Token people
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Tus
Token Tus
Feature activation+0.000
can
Tokencan
Feature activation+0.000
setting
Token setting
Feature activation+0.000
,
Token,
Feature activation+0.000
was
Token was
Feature activation+2.358
surrounded
Token surrounded
Feature activation+2.285
by
Token by
Feature activation+1.030
manic
Token manic
Feature activation+0.000
ured
Tokenured
Feature activation+0.000
hed
Token hed
Feature activation+0.000
ges
Tokenges
Feature activation+0.460
the
Token the
Feature activation+0.000
reasons
Token reasons
Feature activation+0.000
that
Token that
Feature activation+0.000
plant
Token plant
Feature activation+0.323
got
Token got
Feature activation+1.837
shut
Token shut
Feature activation+2.276
down
Token down
Feature activation+2.178
was
Token was
Feature activation+0.000
$
Token $
Feature activation+0.671
4
Token4
Feature activation+0.000
gasoline
Token gasoline
Feature activation+0.000
reasons
Token reasons
Feature activation+0.000
that
Token that
Feature activation+0.000
plant
Token plant
Feature activation+0.323
got
Token got
Feature activation+1.837
shut
Token shut
Feature activation+2.276
down
Token down
Feature activation+2.178
was
Token was
Feature activation+0.000
$
Token $
Feature activation+0.671
4
Token4
Feature activation+0.000
gasoline
Token gasoline
Feature activation+0.000
.
Token.
Feature activation+0.000
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
field
Token field
Feature activation+0.000
hospital
Token hospital
Feature activation+0.000
,
Token,
Feature activation+0.845
which
Token which
Feature activation+2.150
G
Token G
Feature activation+0.000
ij
Tokenij
Feature activation+0.000
s
Tokens
Feature activation+0.000
expects
Token expects
Feature activation+0.000
to
Token to
Feature activation+0.208
reveal
Token reveal
Feature activation+0.000
a
Token a
Feature activation+0.000
hacking
Token hacking
Feature activation+0.000
operations
Token operations
Feature activation+0.000
center
Token center
Feature activation+0.000
called
Token called
Feature activation+2.142
the
Token the
Feature activation+0.652
Center
Token Center
Feature activation+0.000
for
Token for
Feature activation+0.045
Cyber
Token Cyber
Feature activation+0.000
Intelligence
Token Intelligence
Feature activation+0.011
after
Token after
Feature activation+0.000
the
Token the
Feature activation+0.000
medical
Token medical
Feature activation+0.000
facility
Token facility
Feature activation+0.000
was
Token was
Feature activation+2.096
hit
Token hit
Feature activation+2.135
by
Token by
Feature activation+1.607
bombs
Token bombs
Feature activation+0.025
containing
Token containing
Feature activation+0.000
toxic
Token toxic
Feature activation+0.000
gas
Token gas
Feature activation+0.000
centre
Token centre
Feature activation+0.000
in
Token in
Feature activation+1.393
Tap
Token Tap
Feature activation+0.000
ach
Tokenach
Feature activation+0.000
ula
Tokenula
Feature activation+0.000
is
Token is
Feature activation+2.128
the
Token the
Feature activation+1.879
largest
Token largest
Feature activation+1.715
in
Token in
Feature activation+1.681
Latin
Token Latin
Feature activation+0.000
America
Token America
Feature activation+0.000
of
Token of
Feature activation+0.000
seven
Token seven
Feature activation+0.000
hospitals
Token hospitals
Feature activation+0.561
are
Token are
Feature activation+1.667
still
Token still
Feature activation+1.878
in
Token in
Feature activation+2.110
operation
Token operation
Feature activation+1.672
serving
Token serving
Feature activation+1.840
about
Token about
Feature activation+0.298
250
Token 250
Feature activation+0.388
,
Token,
Feature activation+0.000
outlet
Token outlet
Feature activation+0.000
mall
Token mall
Feature activation+0.000
,
Token,
Feature activation+1.435
scheduled
Token scheduled
Feature activation+1.788
to
Token to
Feature activation+1.326
open
Token open
Feature activation+2.110
next
Token next
Feature activation+0.863
year
Token year
Feature activation+0.491
,
Token,
Feature activation+0.539
and
Token and
Feature activation+0.597
a
Token a
Feature activation+0.289
breath
Token breath
Feature activation+0.000
after
Token after
Feature activation+0.000
the
Token the
Feature activation+0.000
medical
Token medical
Feature activation+0.000
facility
Token facility
Feature activation+0.000
was
Token was
Feature activation+2.096
hit
Token hit
Feature activation+2.135
by
Token by
Feature activation+1.607
bombs
Token bombs
Feature activation+0.025
containing
Token containing
Feature activation+0.000
toxic
Token toxic
Feature activation+0.000
?
Token?
Feature activation+0.000
Gy
Token Gy
Feature activation+0.000
ms
Tokenms
Feature activation+0.589
are
Token are
Feature activation+1.341
wonderful
Token wonderful
Feature activation+1.920
for
Token for
Feature activation+2.090
that
Token that
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.131
unfortunately
Token unfortunately
Feature activation+0.000
they
Token they
Feature activation+1.356
where
Token where
Feature activation+1.286
they
Token they
Feature activation+0.000
lived
Token lived
Feature activation+0.000
has
Token has
Feature activation+1.648
stood
Token stood
Feature activation+1.921
empty
Token empty
Feature activation+2.049
and
Token and
Feature activation+0.804
crumbling
Token crumbling
Feature activation+1.158
,
Token,
Feature activation+0.756
with
Token with
Feature activation+1.585
nobody
Token nobody
Feature activation+0.000
he
Token he
Feature activation+0.000
'll
Token'll
Feature activation+0.000
keep
Token keep
Feature activation+0.000
that
Token that
Feature activation+0.000
plant
Token plant
Feature activation+0.314
open
Token open
Feature activation+2.018
,"
Token,"
Feature activation+0.000
Ryan
Token Ryan
Feature activation+0.000
said
Token said
Feature activation+0.000
during
Token during
Feature activation+0.000
a
Token a
Feature activation+0.000
oks
Tokenoks
Feature activation+0.000
'
Token'
Feature activation+0.000
hotel
Token hotel
Feature activation+0.000
,
Token,
Feature activation+1.087
which
Token which
Feature activation+1.890
looked
Token looked
Feature activation+2.016
had
Token had
Feature activation+1.666
been
Token been
Feature activation+1.665
transported
Token transported
Feature activation+1.056
straight
Token straight
Feature activation+0.000
out
Token out
Feature activation+0.000
Arms
Token Arms
Feature activation+0.000
F
Token F
Feature activation+0.000
iring
Tokeniring
Feature activation+0.000
School
Token School
Feature activation+0.075
âĢĶ
TokenâĢĶ
Feature activation+0.944
where
Tokenwhere
Feature activation+2.004
specially
Token specially
Feature activation+0.214
selected
Token selected
Feature activation+0.000
soldiers
Token soldiers
Feature activation+0.000
would
Token would
Feature activation+0.000
learn
Token learn
Feature activation+0.000
centre
Token centre
Feature activation+0.000
in
Token in
Feature activation+1.570
Tap
Token Tap
Feature activation+0.000
ach
Tokenach
Feature activation+0.000
ula
Tokenula
Feature activation+0.000
registered
Token registered
Feature activation+1.990
4
Token 4
Feature activation+0.000
42
Token42
Feature activation+0.000
African
Token African
Feature activation+0.000
migrants
Token migrants
Feature activation+0.000
in
Token in
Feature activation+0.000
,
Token,
Feature activation+0.000
Michigan
Token Michigan
Feature activation+0.000
,
Token,
Feature activation+0.000
assembly
Token assembly
Feature activation+0.000
plant
Token plant
Feature activation+0.000
located
Token located
Feature activation+1.921
12
Token 12
Feature activation+0.404
miles
Token miles
Feature activation+0.001
southwest
Token southwest
Feature activation+0.000
of
Token of
Feature activation+0.000
downtown
Token downtown
Feature activation+0.000
house
Tokenhouse
Feature activation+0.000
where
Token where
Feature activation+1.286
they
Token they
Feature activation+0.000
lived
Token lived
Feature activation+0.000
has
Token has
Feature activation+1.648
stood
Token stood
Feature activation+1.921
empty
Token empty
Feature activation+2.049
and
Token and
Feature activation+0.804
crumbling
Token crumbling
Feature activation+1.158
,
Token,
Feature activation+0.756
with
Token with
Feature activation+1.585

Top DFA by src position
MAX = 2.923

Ċ
TokenĊ
Feature activation-0.001
Top resid features:
The
TokenThe
Feature activation-0.012
Top resid features:
B
Token B
Feature activation-0.001
Top resid features:
oks
Tokenoks
Feature activation-0.008
Top resid features:
'
Token'
Feature activation+0.038
Top resid features:
hotel
Token hotel
Feature activation+2.803
Top resid features:
,
Token,
Feature activation+0.004
Top resid features:
which
Token which
Feature activation+0.028
Top resid features:
looked
Token looked
Feature activation+0.015
Top resid features:
had
Token had
Feature activation-0.004
Top resid features:
been
Token been
Feature activation+0.006
Top resid features:
the
Token the
Feature activation+0.001
Top resid features:
news
Token news
Feature activation+0.007
Top resid features:
that
Token that
Feature activation+0.003
Top resid features:
the
Token the
Feature activation+0.009
Top resid features:
GM
Token GM
Feature activation+0.048
Top resid features:
plant
Token plant
Feature activation+2.825
Top resid features:
was
Token was
Feature activation-0.009
Top resid features:
firing
Token firing
Feature activation-0.189
Top resid features:
750
Token 750
Feature activation+0.000
Top resid features:
people
Token people
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
The
TokenThe
Feature activation-0.004
Top resid features:
B
Token B
Feature activation+0.002
Top resid features:
oks
Tokenoks
Feature activation-0.011
Top resid features:
'
Token'
Feature activation+0.028
Top resid features:
hotel
Token hotel
Feature activation+2.695
Top resid features:
,
Token,
Feature activation+0.006
Top resid features:
which
Token which
Feature activation+0.045
Top resid features:
looked
Token looked
Feature activation+0.008
Top resid features:
had
Token had
Feature activation+0.009
Top resid features:
been
Token been
Feature activation+0.005
Top resid features:
One
TokenOne
Feature activation+0.007
Top resid features:
of
Token of
Feature activation-0.007
Top resid features:
the
Token the
Feature activation-0.005
Top resid features:
reasons
Token reasons
Feature activation+0.004
Top resid features:
that
Token that
Feature activation+0.003
Top resid features:
plant
Token plant
Feature activation+2.898
Top resid features:
got
Token got
Feature activation-0.024
Top resid features:
shut
Token shut
Feature activation-0.117
Top resid features:
down
Token down
Feature activation+0.000
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
$
Token $
Feature activation+0.000
Top resid features:
One
TokenOne
Feature activation+0.009
Top resid features:
of
Token of
Feature activation-0.004
Top resid features:
the
Token the
Feature activation-0.002
Top resid features:
reasons
Token reasons
Feature activation+0.004
Top resid features:
that
Token that
Feature activation+0.004
Top resid features:
plant
Token plant
Feature activation+2.923
Top resid features:
got
Token got
Feature activation-0.009
Top resid features:
shut
Token shut
Feature activation-0.111
Top resid features:
down
Token down
Feature activation-0.114
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
$
Token $
Feature activation+0.000
Top resid features:
own
Token own
Feature activation+0.007
Top resid features:
troops
Token troops
Feature activation-0.052
Top resid features:
for
Token for
Feature activation-0.076
Top resid features:
the
Token the
Feature activation+0.001
Top resid features:
field
Token field
Feature activation+0.107
Top resid features:
hospital
Token hospital
Feature activation+2.748
Top resid features:
,
Token,
Feature activation-0.088
Top resid features:
which
Token which
Feature activation-0.079
Top resid features:
G
Token G
Feature activation+0.000
Top resid features:
ij
Tokenij
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.000
Top resid features:
leaks
Token leaks
Feature activation+0.002
Top resid features:
reveal
Token reveal
Feature activation-0.054
Top resid features:
a
Token a
Feature activation-0.047
Top resid features:
hacking
Token hacking
Feature activation-0.015
Top resid features:
operations
Token operations
Feature activation+0.036
Top resid features:
center
Token center
Feature activation+2.839
Top resid features:
called
Token called
Feature activation-0.179
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
Center
Token Center
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
Cyber
Token Cyber
Feature activation+0.000
Top resid features:
to
Token to
Feature activation-0.006
Top resid features:
breath
Token breath
Feature activation-0.021
Top resid features:
after
Token after
Feature activation+0.028
Top resid features:
the
Token the
Feature activation-0.003
Top resid features:
medical
Token medical
Feature activation+0.060
Top resid features:
facility
Token facility
Feature activation+2.678
Top resid features:
was
Token was
Feature activation-0.024
Top resid features:
hit
Token hit
Feature activation-0.182
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
bombs
Token bombs
Feature activation+0.000
Top resid features:
containing
Token containing
Feature activation+0.000
Top resid features:
)
Token)
Feature activation-0.037
Top resid features:
Ċ
TokenĊ
Feature activation-0.041
Top resid features:
Ċ
TokenĊ
Feature activation-0.025
Top resid features:
The
TokenThe
Feature activation+0.003
Top resid features:
immigration
Token immigration
Feature activation-0.052
Top resid features:
centre
Token centre
Feature activation+2.699
Top resid features:
in
Token in
Feature activation-0.002
Top resid features:
Tap
Token Tap
Feature activation+0.081
Top resid features:
ach
Tokenach
Feature activation+0.034
Top resid features:
ula
Tokenula
Feature activation-0.019
Top resid features:
is
Token is
Feature activation-0.134
Top resid features:
,
Token,
Feature activation+0.002
Top resid features:
four
Token four
Feature activation-0.039
Top resid features:
out
Token out
Feature activation-0.002
Top resid features:
of
Token of
Feature activation-0.024
Top resid features:
seven
Token seven
Feature activation+0.002
Top resid features:
hospitals
Token hospitals
Feature activation+2.683
Top resid features:
are
Token are
Feature activation+0.037
Top resid features:
still
Token still
Feature activation+0.040
Top resid features:
in
Token in
Feature activation-0.117
Top resid features:
operation
Token operation
Feature activation+0.000
Top resid features:
serving
Token serving
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.020
Top resid features:
Ļ
TokenĻ
Feature activation-0.004
Top resid features:
s
Tokens
Feature activation+0.016
Top resid features:
first
Token first
Feature activation+0.007
Top resid features:
outlet
Token outlet
Feature activation+0.127
Top resid features:
mall
Token mall
Feature activation+2.387
Top resid features:
,
Token,
Feature activation+0.047
Top resid features:
scheduled
Token scheduled
Feature activation+0.009
Top resid features:
to
Token to
Feature activation+0.021
Top resid features:
open
Token open
Feature activation-0.095
Top resid features:
next
Token next
Feature activation+0.000
Top resid features:
to
Token to
Feature activation-0.007
Top resid features:
breath
Token breath
Feature activation-0.018
Top resid features:
after
Token after
Feature activation+0.011
Top resid features:
the
Token the
Feature activation-0.010
Top resid features:
medical
Token medical
Feature activation+0.054
Top resid features:
facility
Token facility
Feature activation+2.600
Top resid features:
was
Token was
Feature activation-0.085
Top resid features:
hit
Token hit
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
bombs
Token bombs
Feature activation+0.000
Top resid features:
containing
Token containing
Feature activation+0.000
Top resid features:
some
Token some
Feature activation+0.001
Top resid features:
specific
Token specific
Feature activation+0.003
Top resid features:
muscles
Token muscles
Feature activation-0.023
Top resid features:
?
Token?
Feature activation-0.008
Top resid features:
Gy
Token Gy
Feature activation+0.056
Top resid features:
ms
Tokenms
Feature activation+2.569
Top resid features:
are
Token are
Feature activation+0.033
Top resid features:
wonderful
Token wonderful
Feature activation-0.008
Top resid features:
for
Token for
Feature activation-0.090
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
family
Token family
Feature activation-0.001
Top resid features:
,
Token,
Feature activation+0.003
Top resid features:
the
Token the
Feature activation+0.011
Top resid features:
simple
Token simple
Feature activation-0.005
Top resid features:
stone
Token stone
Feature activation+0.005
Top resid features:
farm
Token farm
Feature activation+1.296
Top resid features:
house
Tokenhouse
Feature activation+1.242
Top resid features:
where
Token where
Feature activation+0.026
Top resid features:
they
Token they
Feature activation-0.001
Top resid features:
lived
Token lived
Feature activation+0.003
Top resid features:
has
Token has
Feature activation-0.004
Top resid features:
saying
Token saying
Feature activation-0.036
Top resid features:
he
Token he
Feature activation-0.010
Top resid features:
'll
Token'll
Feature activation+0.004
Top resid features:
keep
Token keep
Feature activation-0.052
Top resid features:
that
Token that
Feature activation-0.085
Top resid features:
plant
Token plant
Feature activation+2.733
Top resid features:
open
Token open
Feature activation-0.065
Top resid features:
,"
Token,"
Feature activation+0.000
Top resid features:
Ryan
Token Ryan
Feature activation+0.000
Top resid features:
said
Token said
Feature activation+0.000
Top resid features:
during
Token during
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
The
TokenThe
Feature activation-0.008
Top resid features:
B
Token B
Feature activation+0.003
Top resid features:
oks
Tokenoks
Feature activation-0.013
Top resid features:
'
Token'
Feature activation+0.009
Top resid features:
hotel
Token hotel
Feature activation+2.586
Top resid features:
,
Token,
Feature activation+0.021
Top resid features:
which
Token which
Feature activation-0.098
Top resid features:
looked
Token looked
Feature activation-0.064
Top resid features:
had
Token had
Feature activation+0.000
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
the
Tokenthe
Feature activation+0.022
Top resid features:
Small
Token Small
Feature activation-0.004
Top resid features:
Arms
Token Arms
Feature activation+0.005
Top resid features:
F
Token F
Feature activation+0.008
Top resid features:
iring
Tokeniring
Feature activation+0.008
Top resid features:
School
Token School
Feature activation+2.529
Top resid features:
âĢĶ
TokenâĢĶ
Feature activation-0.056
Top resid features:
where
Tokenwhere
Feature activation-0.123
Top resid features:
specially
Token specially
Feature activation+0.000
Top resid features:
selected
Token selected
Feature activation+0.000
Top resid features:
soldiers
Token soldiers
Feature activation+0.000
Top resid features:
Last
TokenLast
Feature activation+0.002
Top resid features:
year
Token year
Feature activation+0.003
Top resid features:
,
Token,
Feature activation-0.013
Top resid features:
the
Token the
Feature activation-0.003
Top resid features:
immigration
Token immigration
Feature activation-0.039
Top resid features:
centre
Token centre
Feature activation+2.567
Top resid features:
in
Token in
Feature activation+0.053
Top resid features:
Tap
Token Tap
Feature activation+0.007
Top resid features:
ach
Tokenach
Feature activation+0.010
Top resid features:
ula
Tokenula
Feature activation-0.061
Top resid features:
registered
Token registered
Feature activation-0.134
Top resid features:
born
Tokenborn
Feature activation-0.001
Top resid features:
,
Token,
Feature activation-0.006
Top resid features:
Michigan
Token Michigan
Feature activation-0.019
Top resid features:
,
Token,
Feature activation-0.004
Top resid features:
assembly
Token assembly
Feature activation+0.076
Top resid features:
plant
Token plant
Feature activation+2.284
Top resid features:
located
Token located
Feature activation-0.126
Top resid features:
12
Token 12
Feature activation+0.000
Top resid features:
miles
Token miles
Feature activation+0.000
Top resid features:
southwest
Token southwest
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
family
Token family
Feature activation-0.001
Top resid features:
,
Token,
Feature activation-0.002
Top resid features:
the
Token the
Feature activation+0.015
Top resid features:
simple
Token simple
Feature activation-0.025
Top resid features:
stone
Token stone
Feature activation-0.003
Top resid features:
farm
Token farm
Feature activation+1.114
Top resid features:
house
Tokenhouse
Feature activation+1.066
Top resid features:
where
Token where
Feature activation+0.126
Top resid features:
they
Token they
Feature activation+0.001
Top resid features:
lived
Token lived
Feature activation-0.018
Top resid features:
has
Token has
Feature activation+0.057
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.03

Head 2: 0.02

Head 3: 0.48

Head 4: 0.05

Head 5: 0.02

Head 6: 0.07

Head 7: 0.11

Head 8: 0.03

Head 9: 0.04

Head 10: 0.02

Head 11: 0.09

Positive logits

staffed1.97

reopened1.80

Located1.78

operated1.62

sprung1.61

opened1.53

Located1.53

accommodating1.51

reopen1.48

catering1.48

opened1.43

situated1.43

renovated1.42

closed1.41

ysc1.40

emptied1.40

entrances1.40

accommodate1.39

closed1.37

sprang1.36

Negative logits

MEP-1.33

mell-1.29

obin-1.22

ears-1.21

isms-1.20

pair-1.20

pardon-1.19

verbal-1.18

chrom-1.17

redo-1.17

speaking-1.17

phrases-1.17

vous-1.16

verse-1.13

lander-1.13

AUD-1.12

azeera-1.12

bitterness-1.11

obal-1.11

particles-1.11

INTERVAL 2.122 - 2.358
CONTAINS 0.001%

for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
field
Token field
Feature activation+0.000
hospital
Token hospital
Feature activation+0.000
,
Token,
Feature activation+0.845
which
Token which
Feature activation+2.150
G
Token G
Feature activation+0.000
ij
Tokenij
Feature activation+0.000
s
Tokens
Feature activation+0.000
expects
Token expects
Feature activation+0.000
to
Token to
Feature activation+0.208
the
Token the
Feature activation+0.000
reasons
Token reasons
Feature activation+0.000
that
Token that
Feature activation+0.000
plant
Token plant
Feature activation+0.323
got
Token got
Feature activation+1.837
shut
Token shut
Feature activation+2.276
down
Token down
Feature activation+2.178
was
Token was
Feature activation+0.000
$
Token $
Feature activation+0.671
4
Token4
Feature activation+0.000
gasoline
Token gasoline
Feature activation+0.000
reveal
Token reveal
Feature activation+0.000
a
Token a
Feature activation+0.000
hacking
Token hacking
Feature activation+0.000
operations
Token operations
Feature activation+0.000
center
Token center
Feature activation+0.000
called
Token called
Feature activation+2.142
the
Token the
Feature activation+0.652
Center
Token Center
Feature activation+0.000
for
Token for
Feature activation+0.045
Cyber
Token Cyber
Feature activation+0.000
Intelligence
Token Intelligence
Feature activation+0.011
Tus
Token Tus
Feature activation+0.000
can
Tokencan
Feature activation+0.000
setting
Token setting
Feature activation+0.000
,
Token,
Feature activation+0.000
was
Token was
Feature activation+2.358
surrounded
Token surrounded
Feature activation+2.285
by
Token by
Feature activation+1.030
manic
Token manic
Feature activation+0.000
ured
Tokenured
Feature activation+0.000
hed
Token hed
Feature activation+0.000
ges
Tokenges
Feature activation+0.460
centre
Token centre
Feature activation+0.000
in
Token in
Feature activation+1.393
Tap
Token Tap
Feature activation+0.000
ach
Tokenach
Feature activation+0.000
ula
Tokenula
Feature activation+0.000
is
Token is
Feature activation+2.128
the
Token the
Feature activation+1.879
largest
Token largest
Feature activation+1.715
in
Token in
Feature activation+1.681
Latin
Token Latin
Feature activation+0.000
America
Token America
Feature activation+0.000

INTERVAL 1.886 - 2.122
CONTAINS 0.002%

muscles
Token muscles
Feature activation+0.000
?
Token?
Feature activation+0.000
Gy
Token Gy
Feature activation+0.000
ms
Tokenms
Feature activation+0.589
are
Token are
Feature activation+1.341
wonderful
Token wonderful
Feature activation+1.920
for
Token for
Feature activation+2.090
that
Token that
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.131
unfortunately
Token unfortunately
Feature activation+0.000
breath
Token breath
Feature activation+0.000
after
Token after
Feature activation+0.000
the
Token the
Feature activation+0.000
medical
Token medical
Feature activation+0.000
facility
Token facility
Feature activation+0.000
was
Token was
Feature activation+2.096
hit
Token hit
Feature activation+2.135
by
Token by
Feature activation+1.607
bombs
Token bombs
Feature activation+0.025
containing
Token containing
Feature activation+0.000
toxic
Token toxic
Feature activation+0.000
,
Token,
Feature activation+0.000
Michigan
Token Michigan
Feature activation+0.000
,
Token,
Feature activation+0.000
assembly
Token assembly
Feature activation+0.000
plant
Token plant
Feature activation+0.000
located
Token located
Feature activation+1.921
12
Token 12
Feature activation+0.404
miles
Token miles
Feature activation+0.001
southwest
Token southwest
Feature activation+0.000
of
Token of
Feature activation+0.000
downtown
Token downtown
Feature activation+0.000
oks
Tokenoks
Feature activation+0.000
'
Token'
Feature activation+0.000
hotel
Token hotel
Feature activation+0.000
,
Token,
Feature activation+1.087
which
Token which
Feature activation+1.890
looked
Token looked
Feature activation+2.016
had
Token had
Feature activation+1.666
been
Token been
Feature activation+1.665
transported
Token transported
Feature activation+1.056
straight
Token straight
Feature activation+0.000
out
Token out
Feature activation+0.000
of
Token of
Feature activation+0.000
seven
Token seven
Feature activation+0.000
hospitals
Token hospitals
Feature activation+0.561
are
Token are
Feature activation+1.667
still
Token still
Feature activation+1.878
in
Token in
Feature activation+2.110
operation
Token operation
Feature activation+1.672
serving
Token serving
Feature activation+1.840
about
Token about
Feature activation+0.298
250
Token 250
Feature activation+0.388
,
Token,
Feature activation+0.000

INTERVAL 1.650 - 1.886
CONTAINS 0.005%

Donetsk
Token Donetsk
Feature activation+0.000
airport
Token airport
Feature activation+0.000
,
Token,
Feature activation+0.902
which
Token which
Feature activation+1.539
has
Token has
Feature activation+1.540
been
Token been
Feature activation+1.875
under
Token under
Feature activation+1.717
fire
Token fire
Feature activation+0.829
from
Token from
Feature activation+0.000
Ukrainian
Token Ukrainian
Feature activation+0.000
forces
Token forces
Feature activation+0.000
aims
Token aims
Feature activation+1.669
to
Token to
Feature activation+1.579
inst
Token inst
Feature activation+0.602
ill
Tokenill
Feature activation+0.128
in
Token in
Feature activation+0.349
its
Token its
Feature activation+1.778
3
Token 3
Feature activation+0.000
-
Token-
Feature activation+0.000
to
Token to
Feature activation+0.000
5
Token 5
Feature activation+0.000
-
Token-
Feature activation+0.000
continue
Token continue
Feature activation+0.000
to
Token to
Feature activation+0.000
fill
Token fill
Feature activation+1.212
it
Token it
Feature activation+1.221
,
Token,
Feature activation+0.017
offers
Token offers
Feature activation+1.826
secure
Token secure
Feature activation+0.000
long
Token long
Feature activation+0.000
-
Token-
Feature activation+0.000
term
Tokenterm
Feature activation+0.000
revenues
Token revenues
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
The
Token The
Feature activation+0.000
Moscow
Token Moscow
Feature activation+0.000
plant
Token plant
Feature activation+0.000
would
Token would
Feature activation+1.207
focus
Token focus
Feature activation+1.830
on
Token on
Feature activation+1.336
production
Token production
Feature activation+0.047
of
Token of
Feature activation+0.000
Pro
Token Pro
Feature activation+0.000
ton
Tokenton
Feature activation+0.000
his
Token his
Feature activation+0.000
own
Token own
Feature activation+0.000
laundry
Token laundry
Feature activation+0.000
facility
Token facility
Feature activation+0.000
,
Token,
Feature activation+0.956
which
Token which
Feature activation+1.697
for
Token for
Feature activation+0.889
several
Token several
Feature activation+0.000
years
Token years
Feature activation+0.000
cleaned
Token cleaned
Feature activation+0.000
dresses
Token dresses
Feature activation+0.000

INTERVAL 1.415 - 1.650
CONTAINS 0.007%

hospitals
Token hospitals
Feature activation+0.000
would
Token would
Feature activation+0.972
also
Token also
Feature activation+0.949
be
Token be
Feature activation+1.459
invited
Token invited
Feature activation+0.844
to
Token to
Feature activation+1.542
discuss
Token discuss
Feature activation+1.090
disease
Token disease
Feature activation+0.000
control
Token control
Feature activation+0.000
tomorrow
Token tomorrow
Feature activation+0.000
.
Token.
Feature activation+0.000
compost
Token compost
Feature activation+0.000
ing
Tokening
Feature activation+0.000
facility
Token facility
Feature activation+0.132
in
Token in
Feature activation+0.381
Edmonton
Token Edmonton
Feature activation+0.000
had
Token had
Feature activation+1.486
such
Token such
Feature activation+0.715
structural
Token structural
Feature activation+0.248
damage
Token damage
Feature activation+0.048
that
Token that
Feature activation+0.000
it
Token it
Feature activation+1.392
Chicago
Token Chicago
Feature activation+0.000
Film
Token Film
Feature activation+0.000
Society
Token Society
Feature activation+0.000
,
Token,
Feature activation+0.000
will
Token will
Feature activation+1.617
be
Token be
Feature activation+1.642
restored
Token restored
Feature activation+1.656
,
Token,
Feature activation+0.194
and
Token and
Feature activation+0.067
offices
Token offices
Feature activation+0.000
for
Token for
Feature activation+0.400
stage
Token stage
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
arena
Token arena
Feature activation+0.067
will
Token will
Feature activation+1.245
host
Token host
Feature activation+1.591
the
Token the
Feature activation+0.308
third
Token third
Feature activation+0.000
and
Token and
Feature activation+0.000
final
Token final
Feature activation+0.000
presidential
Token presidential
Feature activation+0.000
While
Token While
Feature activation+0.000
the
Token the
Feature activation+0.000
power
Token power
Feature activation+0.000
station
Token station
Feature activation+0.000
reportedly
Token reportedly
Feature activation+0.781
uses
Token uses
Feature activation+1.501
coal
Token coal
Feature activation+0.124
with
Token with
Feature activation+0.000
a
Token a
Feature activation+0.000
volatility
Token volatility
Feature activation+0.000
rate
Token rate
Feature activation+0.000

INTERVAL 1.179 - 1.415
CONTAINS 0.017%

a
Token a
Feature activation+0.000
second
Token second
Feature activation+0.000
arena
Token arena
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
that
Token that
Feature activation+1.206
would
Token would
Feature activation+1.223
act
Token act
Feature activation+1.704
as
Token as
Feature activation+0.660
an
Token an
Feature activation+0.101
indoor
Token indoor
Feature activation+0.000
s
Tokens
Feature activation+0.000
Odd
Token Odd
Feature activation+0.000
itor
Tokenitor
Feature activation+0.000
ium
Tokenium
Feature activation+0.000
,
Token,
Feature activation+0.087
is
Token is
Feature activation+1.214
a
Token a
Feature activation+0.932
despair
Token despair
Feature activation+0.425
ing
Tokening
Feature activation+0.000
mismatch
Token mismatch
Feature activation+0.000
of
Token of
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
terminal
Token terminal
Feature activation+0.000
,
Token,
Feature activation+1.200
operated
Token operated
Feature activation+1.570
by
Token by
Feature activation+1.306
Massachusetts
Token Massachusetts
Feature activation+0.000
-
Token-
Feature activation+0.000
based
Tokenbased
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Many
TokenMany
Feature activation+0.000
hospitals
Token hospitals
Feature activation+0.306
including
Token including
Feature activation+1.228
the
Token the
Feature activation+0.921
children
Token children
Feature activation+0.000
's
Token's
Feature activation+0.000
hospital
Token hospital
Feature activation+0.000
,
Token,
Feature activation+0.977
the
Token the
Feature activation+0.000
simple
Token simple
Feature activation+0.000
stone
Token stone
Feature activation+0.000
farm
Token farm
Feature activation+0.000
house
Tokenhouse
Feature activation+0.000
where
Token where
Feature activation+1.286
they
Token they
Feature activation+0.000
lived
Token lived
Feature activation+0.000
has
Token has
Feature activation+1.648
stood
Token stood
Feature activation+1.921
empty
Token empty
Feature activation+2.049

INTERVAL 0.943 - 1.179
CONTAINS 0.029%

discussed
Token discussed
Feature activation+0.000
with
Token with
Feature activation+0.000
several
Token several
Feature activation+0.000
petrol
Token petrol
Feature activation+0.000
stations
Token stations
Feature activation+0.000
working
Token working
Feature activation+1.058
across
Token across
Feature activation+0.471
the
Token the
Feature activation+0.000
country
Token country
Feature activation+0.000
cance
Token cance
Feature activation+0.613
lling
Tokenlling
Feature activation+0.780
-
Token-
Feature activation+0.000
½
Token½
Feature activation+0.000
hours
Token hours
Feature activation+0.000
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.774
re
Token re
Feature activation+0.982
-
Token-
Feature activation+0.000
opened
Tokenopened
Feature activation+1.623
around
Token around
Feature activation+0.602
9
Token 9
Feature activation+0.000
a
Token a
Feature activation+0.000
at
Token at
Feature activation+0.281
this
Token this
Feature activation+0.000
location
Token location
Feature activation+0.296
since
Token since
Feature activation+0.165
it
Token it
Feature activation+0.469
opened
Token opened
Feature activation+1.054
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
It
TokenIt
Feature activation+0.000
medical
Token medical
Feature activation+0.000
attention
Token attention
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
jail
Token jail
Feature activation+0.000
and
Token and
Feature activation+0.952
medical
Token medical
Feature activation+0.419
personnel
Token personnel
Feature activation+0.000
ridiculed
Token ridiculed
Feature activation+0.000
plaintiff
Token plaintiff
Feature activation+0.000
,
Token,
Feature activation+0.000
machine
Tokenmachine
Feature activation+0.000
gun
Token gun
Feature activation+0.000
nests
Token nests
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.099
Ŀ
TokenĿ
Feature activation+0.000
were
Token were
Feature activation+1.109
being
Token being
Feature activation+0.876
set
Token set
Feature activation+0.594
up
Token up
Feature activation+0.235
in
Token in
Feature activation+0.000
anticipation
Token anticipation
Feature activation+0.000

INTERVAL 0.707 - 0.943
CONTAINS 0.048%

school
Token school
Feature activation+0.000
in
Token in
Feature activation+0.453
the
Token the
Feature activation+0.506
United
Token United
Feature activation+0.000
States
Token States
Feature activation+0.000
.
Token.
Feature activation+0.840
The
Token The
Feature activation+0.029
school
Token school
Feature activation+0.000
was
Token was
Feature activation+0.729
chart
Token chart
Feature activation+0.414
ered
Tokenered
Feature activation+0.000
,
Token,
Feature activation+0.000
while
Token while
Feature activation+0.000
grocery
Token grocery
Feature activation+0.000
stores
Token stores
Feature activation+0.000
mis
Token mis
Feature activation+0.599
label
Tokenlabel
Feature activation+0.841
ed
Tokened
Feature activation+0.727
their
Token their
Feature activation+0.807
products
Token products
Feature activation+0.000
27
Token 27
Feature activation+0.000
percent
Token percent
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.072
Ċ
TokenĊ
Feature activation+0.121
It
TokenIt
Feature activation+0.228
has
Token has
Feature activation+0.475
been
Token been
Feature activation+0.138
named
Token named
Feature activation+0.741
Iceland
Token Iceland
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
national
Token national
Feature activation+0.000
The
TokenThe
Feature activation+0.000
Student
Token Student
Feature activation+0.000
Health
Token Health
Feature activation+0.000
Center
Token Center
Feature activation+0.000
's
Token's
Feature activation+0.505
practice
Token practice
Feature activation+0.725
prohibits
Token prohibits
Feature activation+0.199
prescribing
Token prescribing
Feature activation+0.000
hormone
Token hormone
Feature activation+0.000
therapy
Token therapy
Feature activation+0.000
for
Token for
Feature activation+0.000
300
Token 300
Feature activation+0.000
-
Token-
Feature activation+0.000
seat
Tokenseat
Feature activation+0.000
theater
Token theater
Feature activation+0.000
on
Token on
Feature activation+0.782
the
Token the
Feature activation+0.721
second
Token second
Feature activation+0.000
floor
Token floor
Feature activation+0.000
,
Token,
Feature activation+0.773
which
Token which
Feature activation+0.747
was
Token was
Feature activation+0.776

INTERVAL 0.472 - 0.707
CONTAINS 0.085%

in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
world
Token world
Feature activation+0.000
,
Token,
Feature activation+0.155
first
Token first
Feature activation+0.176
assembled
Token assembled
Feature activation+0.504
in
Token in
Feature activation+0.000
9
Token 9
Feature activation+0.000
30
Token30
Feature activation+0.000
AD
Token AD
Feature activation+0.000
and
Token and
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
national
Token national
Feature activation+0.000
shrine
Token shrine
Feature activation+0.000
and
Token and
Feature activation+0.170
a
Token a
Feature activation+0.666
place
Token place
Feature activation+0.099
of
Token of
Feature activation+0.000
both
Token both
Feature activation+0.000
historical
Token historical
Feature activation+0.000
and
Token and
Feature activation+0.000
have
Token have
Feature activation+0.000
added
Token added
Feature activation+0.000
two
Token two
Feature activation+0.000
settlements
Token settlements
Feature activation+0.000
to
Token to
Feature activation+0.161
the
Token the
Feature activation+0.492
existing
Token existing
Feature activation+0.000
thirty
Token thirty
Feature activation+0.057
-
Token-
Feature activation+0.000
seven
Tokenseven
Feature activation+0.108
.
Token.
Feature activation+0.538
-
Token-
Feature activation+0.000
use
Tokenuse
Feature activation+0.000
development
Token development
Feature activation+0.000
that
Token that
Feature activation+0.017
will
Token will
Feature activation+0.414
include
Token include
Feature activation+0.628
residential
Token residential
Feature activation+0.000
towers
Token towers
Feature activation+0.000
,
Token,
Feature activation+0.000
town
Token town
Feature activation+0.000
houses
Tokenhouses
Feature activation+0.000
and
Token and
Feature activation+0.000
enjoyed
Token enjoyed
Feature activation+0.000
in
Token in
Feature activation+0.042
recent
Token recent
Feature activation+0.000
months
Token months
Feature activation+0.000
.
Token.
Feature activation+0.635
Put
Token Put
Feature activation+0.000
them
Token them
Feature activation+0.627
in
Token in
Feature activation+0.486
your
Token your
Feature activation+0.000
regular
Token regular
Feature activation+0.000

INTERVAL 0.236 - 0.472
CONTAINS 0.171%

city
Token city
Feature activation+0.000
that
Token that
Feature activation+0.000
is
Token is
Feature activation+0.044
a
Token a
Feature activation+0.156
headquarters
Token headquarters
Feature activation+0.009
for
Token for
Feature activation+0.448
Al
Token Al
Feature activation+0.000
Qaeda
Token Qaeda
Feature activation+0.000
in
Token in
Feature activation+0.000
Iraq
Token Iraq
Feature activation+0.000
.
Token.
Feature activation+0.000
The
TokenThe
Feature activation+0.000
College
Token College
Feature activation+0.000
was
Token was
Feature activation+0.405
named
Token named
Feature activation+0.629
Center
Token Center
Feature activation+0.000
of
Token of
Feature activation+0.380
Excellence
Token Excellence
Feature activation+0.579
in
Token in
Feature activation+0.280
Fisheries
Token Fisheries
Feature activation+0.071
Education
Token Education
Feature activation+0.000
by
Token by
Feature activation+0.000
white
Token white
Feature activation+0.000
tropical
Token tropical
Feature activation+0.000
building
Token building
Feature activation+0.000
that
Token that
Feature activation+0.000
was
Token was
Feature activation+0.339
sometimes
Token sometimes
Feature activation+0.344
called
Token called
Feature activation+0.155
Pentagon
Token Pentagon
Feature activation+0.000
East
Token East
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
will
Token will
Feature activation+0.165
be
Token be
Feature activation+0.276
greatly
Token greatly
Feature activation+0.363
limited
Token limited
Feature activation+0.498
,
Token,
Feature activation+0.262
available
Token available
Feature activation+0.368
only
Token only
Feature activation+0.002
at
Token at
Feature activation+0.000
meters
Token meters
Feature activation+0.000
or
Token or
Feature activation+0.200
for
Token for
Feature activation+0.063
destroyed
Token destroyed
Feature activation+1.132
and
Token and
Feature activation+0.193
several
Token several
Feature activation+0.142
homes
Token homes
Feature activation+0.272
were
Token were
Feature activation+0.099
damaged
Token damaged
Feature activation+0.416
.
Token .
Feature activation+0.007
But
Token But
Feature activation+0.000
no
Token no
Feature activation+0.000
one
Token one
Feature activation+0.000
was
Token was
Feature activation+0.000

INTERVAL 0.000 - 0.236
CONTAINS 99.636%

cannot
Token cannot
Feature activation+0.000
have
Token have
Feature activation+0.000
any
Token any
Feature activation+0.000
content
Token content
Feature activation+0.000
;
Token;
Feature activation+0.000
they
Token they
Feature activation+0.000
also
Token also
Feature activation+0.000
cannot
Token cannot
Feature activation+0.000
be
Token be
Feature activation+0.000
"
Token "
Feature activation+0.000
self
Tokenself
Feature activation+0.000
a
Token a
Feature activation+0.000
rate
Token rate
Feature activation+0.000
increase
Token increase
Feature activation+0.000
,
Token,
Feature activation+0.000
E
Token E
Feature activation+0.000
by
Tokenby
Feature activation+0.000
said
Token said
Feature activation+0.000
the
Token the
Feature activation+0.000
government
Token government
Feature activation+0.000
will
Token will
Feature activation+0.000
increase
Token increase
Feature activation+0.000
Pa
Token Pa
Feature activation+0.000
ik
Tokenik
Feature activation+0.000
(
Token (
Feature activation+0.000
19
Token19
Feature activation+0.000
32
Token32
Feature activation+0.000
-
Token-
Feature activation+0.000
2006
Token2006
Feature activation+0.000
),
Token),
Feature activation+0.000
the
Token the
Feature activation+0.000
Korean
Token Korean
Feature activation+0.000
-
Token-
Feature activation+0.000
up
Token up
Feature activation+0.000
the
Token the
Feature activation+0.000
formula
Token formula
Feature activation+0.000
.
Token.
Feature activation+0.000
Their
Token Their
Feature activation+0.000
Nintendo
Token Nintendo
Feature activation+0.000
branded
Token branded
Feature activation+0.000
unit
Token unit
Feature activation+0.000
charges
Token charges
Feature activation+0.000
Wii
Token Wii
Feature activation+0.000
rem
Token rem
Feature activation+0.000
of
Token of
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
The
TokenThe
Feature activation+0.000
two
Token two
Feature activation+0.000
associations
Token associations
Feature activation+0.000
representing
Token representing
Feature activation+0.000
police
Token police
Feature activation+0.000
chiefs
Token chiefs
Feature activation+0.000
in
Token in
Feature activation+0.000
B
Token B
Feature activation+0.000
.
Token.
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 7 in H4.3: (feature 11553)

TOP ACTIVATIONS
MAX = 3.143

that
Token that
Feature activation+0.957
was
Token was
Feature activation+0.324
wrong
Token wrong
Feature activation+0.748
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.106
said
Token said
Feature activation+3.143
travelling
Token travelling
Feature activation+0.000
time
Token time
Feature activation+0.000
counts
Token counts
Feature activation+0.000
as
Token as
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
lead
Token lead
Feature activation+0.000
after
Token after
Feature activation+0.000
its
Token its
Feature activation+0.000
highest
Token highest
Feature activation+0.000
court
Token court
Feature activation+0.373
upheld
Token upheld
Feature activation+3.117
the
Token the
Feature activation+0.297
world
Token world
Feature activation+0.000
's
Token's
Feature activation+0.000
toughest
Token toughest
Feature activation+0.500
tobacco
Token tobacco
Feature activation+0.000
example
Token example
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.262
declined
Token declined
Feature activation+1.724
to
Token to
Feature activation+3.034
review
Token review
Feature activation+2.628
the
Token the
Feature activation+0.000
Tenth
Token Tenth
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.000
was
Token was
Feature activation+2.452
hearing
Token hearing
Feature activation+3.013
an
Token an
Feature activation+0.000
appeal
Token appeal
Feature activation+0.000
filed
Token filed
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+1.644
goes
Token goes
Feature activation+1.937
along
Token along
Feature activation+0.784
,
Token,
Feature activation+1.530
it
Token it
Feature activation+3.000
could
Token could
Feature activation+2.113
have
Token have
Feature activation+2.473
a
Token a
Feature activation+0.820
sweeping
Token sweeping
Feature activation+0.000
impact
Token impact
Feature activation+1.158
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
courts
Token courts
Feature activation+1.168
have
Token have
Feature activation+2.647
now
Token now
Feature activation+2.554
renewed
Token renewed
Feature activation+2.927
the
Token the
Feature activation+0.686
collection
Token collection
Feature activation+0.000
of
Token of
Feature activation+0.000
telephone
Token telephone
Feature activation+0.000
metadata
Token metadata
Feature activation+0.000
They
Token They
Feature activation+0.000
want
Token want
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+0.278
to
Token to
Feature activation+2.445
abandon
Token abandon
Feature activation+2.924
the
Token the
Feature activation+0.777
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.033
und
Tokenund
Feature activation+0.000
ue
Tokenue
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.079
has
Token has
Feature activation+2.498
a
Token a
Feature activation+2.424
discretion
Token discretion
Feature activation+2.862
to
Token to
Feature activation+1.276
refund
Token refund
Feature activation+1.719
part
Token part
Feature activation+0.000
or
Token or
Feature activation+0.000
all
Token all
Feature activation+0.000
Court
Token Court
Feature activation+0.000
of
Token of
Feature activation+0.044
Appeal
Token Appeal
Feature activation+0.825
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+1.186
s
Tokens
Feature activation+2.861
ruling
Token ruling
Feature activation+2.538
yesterday
Token yesterday
Feature activation+1.044
(
Token (
Feature activation+0.000
Friday
TokenFriday
Feature activation+0.406
).
Token).
Feature activation+0.012
later
Token later
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+0.000
has
Token has
Feature activation+2.279
spelled
Token spelled
Feature activation+2.816
out
Token out
Feature activation+0.790
for
Token for
Feature activation+0.070
everyone
Token everyone
Feature activation+0.000
.
Token.
Feature activation+0.153
The
Token The
Feature activation+0.000
the
Token the
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.279
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.747
s
Tokens
Feature activation+2.783
decision
Token decision
Feature activation+2.373
on
Token on
Feature activation+0.000
Panama
Token Panama
Feature activation+0.000
Le
Token Le
Feature activation+0.000
aks
Tokenaks
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.717
's
Token's
Feature activation+2.745
decision
Token decision
Feature activation+2.275
once
Token once
Feature activation+0.041
again
Token again
Feature activation+0.770
confirms
Token confirms
Feature activation+0.579
that
Token that
Feature activation+1.246
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+0.000
,
Token,
Feature activation+0.027
which
Token which
Feature activation+1.869
rejected
Token rejected
Feature activation+2.720
efforts
Token efforts
Feature activation+0.000
by
Token by
Feature activation+0.000
Iran
Token Iran
Feature activation+0.000
's
Token's
Feature activation+0.000
central
Token central
Feature activation+0.000
Australia
Token Australia
Feature activation+0.000
's
Token's
Feature activation+0.000
high
Token high
Feature activation+0.000
court
Token court
Feature activation+1.224
has
Token has
Feature activation+2.659
struck
Token struck
Feature activation+2.718
down
Token down
Feature activation+0.429
a
Token a
Feature activation+0.341
same
Token same
Feature activation+0.000
-
Token-
Feature activation+0.000
sex
Tokensex
Feature activation+0.000
and
Token and
Feature activation+0.000
Australia
Token Australia
Feature activation+0.000
's
Token's
Feature activation+0.000
high
Token high
Feature activation+0.000
court
Token court
Feature activation+1.224
has
Token has
Feature activation+2.659
struck
Token struck
Feature activation+2.718
down
Token down
Feature activation+0.429
a
Token a
Feature activation+0.341
same
Token same
Feature activation+0.000
-
Token-
Feature activation+0.000
of
Token of
Feature activation+0.000
legislation
Token legislation
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
courts
Token courts
Feature activation+1.168
have
Token have
Feature activation+2.647
now
Token now
Feature activation+2.554
renewed
Token renewed
Feature activation+2.927
the
Token the
Feature activation+0.686
collection
Token collection
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.000
,
Token,
Feature activation+0.315
said
Token said
Feature activation+2.036
it
Token it
Feature activation+2.646
is
Token is
Feature activation+1.272
likely
Token likely
Feature activation+0.000
the
Token the
Feature activation+0.377
initiative
Token initiative
Feature activation+0.000
will
Token will
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.205
had
Token had
Feature activation+2.335
allowed
Token allowed
Feature activation+2.643
Mass
Token Mass
Feature activation+0.000
imil
Tokenimil
Feature activation+0.000
iano
Tokeniano
Feature activation+0.000
L
Token L
Feature activation+0.000
ator
Tokenator
Feature activation+0.000
that
Token that
Feature activation+0.000
will
Token will
Feature activation+0.000
make
Token make
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.488
's
Token's
Feature activation+2.629
landmark
Token landmark
Feature activation+1.816
campaign
Token campaign
Feature activation+0.261
finance
Token finance
Feature activation+0.000
ruling
Token ruling
Feature activation+0.000
,
Token,
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.262
declined
Token declined
Feature activation+1.724
to
Token to
Feature activation+3.034
review
Token review
Feature activation+2.628
the
Token the
Feature activation+0.000
Tenth
Token Tenth
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+1.455

Top DFA by src position
MAX = 4.191

one
Token one
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.069
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
The
TokenThe
Feature activation-0.014
Top resid features:
court
Token court
Feature activation+3.821
Top resid features:
ruled
Token ruled
Feature activation+0.019
Top resid features:
that
Token that
Feature activation-0.036
Top resid features:
was
Token was
Feature activation-0.038
Top resid features:
wrong
Token wrong
Feature activation+0.029
Top resid features:
,
Token,
Feature activation+0.054
Top resid features:
's
Token's
Feature activation-0.006
Top resid features:
lead
Token lead
Feature activation+0.003
Top resid features:
after
Token after
Feature activation+0.062
Top resid features:
its
Token its
Feature activation-0.012
Top resid features:
highest
Token highest
Feature activation+0.050
Top resid features:
court
Token court
Feature activation+3.963
Top resid features:
upheld
Token upheld
Feature activation-0.012
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
world
Token world
Feature activation+0.000
Top resid features:
's
Token's
Feature activation+0.000
Top resid features:
toughest
Token toughest
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.037
Top resid features:
For
Token For
Feature activation+0.043
Top resid features:
example
Token example
Feature activation-0.043
Top resid features:
,
Token,
Feature activation+0.005
Top resid features:
the
Token the
Feature activation-0.023
Top resid features:
Court
Token Court
Feature activation+3.763
Top resid features:
declined
Token declined
Feature activation+0.098
Top resid features:
to
Token to
Feature activation+0.039
Top resid features:
review
Token review
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
Tenth
Token Tenth
Feature activation+0.000
Top resid features:
said
Token said
Feature activation-0.011
Top resid features:
.
Token.
Feature activation+0.009
Top resid features:
Ċ
TokenĊ
Feature activation+0.005
Top resid features:
Ċ
TokenĊ
Feature activation+0.007
Top resid features:
The
TokenThe
Feature activation-0.036
Top resid features:
court
Token court
Feature activation+4.191
Top resid features:
was
Token was
Feature activation-0.098
Top resid features:
hearing
Token hearing
Feature activation-0.166
Top resid features:
an
Token an
Feature activation+0.000
Top resid features:
appeal
Token appeal
Feature activation+0.000
Top resid features:
filed
Token filed
Feature activation+0.000
Top resid features:
benefit
Token benefit
Feature activation+0.009
Top resid features:
women
Token women
Feature activation+0.002
Top resid features:
.
Token.
Feature activation+0.071
Top resid features:
If
Token If
Feature activation+0.011
Top resid features:
the
Token the
Feature activation-0.017
Top resid features:
court
Token court
Feature activation+3.540
Top resid features:
goes
Token goes
Feature activation+0.081
Top resid features:
along
Token along
Feature activation+0.041
Top resid features:
,
Token,
Feature activation+0.077
Top resid features:
it
Token it
Feature activation-0.054
Top resid features:
could
Token could
Feature activation+0.000
Top resid features:
absence
Token absence
Feature activation+0.009
Top resid features:
of
Token of
Feature activation-0.015
Top resid features:
legislation
Token legislation
Feature activation-0.009
Top resid features:
,
Token,
Feature activation+0.026
Top resid features:
the
Token the
Feature activation+0.040
Top resid features:
courts
Token courts
Feature activation+3.234
Top resid features:
have
Token have
Feature activation+0.118
Top resid features:
now
Token now
Feature activation+0.163
Top resid features:
renewed
Token renewed
Feature activation+0.100
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
collection
Token collection
Feature activation+0.000
Top resid features:
purpose
Token purpose
Feature activation+0.014
Top resid features:
.
Token.
Feature activation+0.057
Top resid features:
They
Token They
Feature activation-0.031
Top resid features:
want
Token want
Feature activation+0.070
Top resid features:
the
Token the
Feature activation-0.062
Top resid features:
court
Token court
Feature activation+3.690
Top resid features:
to
Token to
Feature activation-0.005
Top resid features:
abandon
Token abandon
Feature activation-0.031
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
Appeal
Token Appeal
Feature activation+0.012
Top resid features:
are
Token are
Feature activation+0.008
Top resid features:
successful
Token successful
Feature activation+0.004
Top resid features:
,
Token,
Feature activation-0.004
Top resid features:
the
Token the
Feature activation+0.026
Top resid features:
Court
Token Court
Feature activation+3.438
Top resid features:
has
Token has
Feature activation+0.089
Top resid features:
a
Token a
Feature activation+0.124
Top resid features:
discretion
Token discretion
Feature activation-0.103
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
refund
Token refund
Feature activation+0.000
Top resid features:
be
Token be
Feature activation-0.033
Top resid features:
shut
Token shut
Feature activation+0.005
Top resid features:
down
Token down
Feature activation+0.001
Top resid features:
following
Token following
Feature activation+0.076
Top resid features:
the
Token the
Feature activation+0.039
Top resid features:
Court
Token Court
Feature activation+3.281
Top resid features:
of
Token of
Feature activation+0.142
Top resid features:
Appeal
Token Appeal
Feature activation+0.353
Top resid features:
âĢ
TokenâĢ
Feature activation+0.037
Top resid features:
Ļ
TokenĻ
Feature activation-0.020
Top resid features:
s
Tokens
Feature activation-0.080
Top resid features:
Three
TokenThree
Feature activation-0.013
Top resid features:
years
Token years
Feature activation-0.025
Top resid features:
later
Token later
Feature activation+0.015
Top resid features:
,
Token,
Feature activation-0.006
Top resid features:
the
Token the
Feature activation-0.017
Top resid features:
court
Token court
Feature activation+3.900
Top resid features:
has
Token has
Feature activation-0.127
Top resid features:
spelled
Token spelled
Feature activation-0.088
Top resid features:
out
Token out
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
everyone
Token everyone
Feature activation+0.000
Top resid features:
held
Token held
Feature activation-0.027
Top resid features:
to
Token to
Feature activation-0.010
Top resid features:
review
Token review
Feature activation+0.092
Top resid features:
the
Token the
Feature activation-0.018
Top resid features:
Supreme
Token Supreme
Feature activation+0.287
Top resid features:
Court
Token Court
Feature activation+3.365
Top resid features:
âĢ
TokenâĢ
Feature activation+0.061
Top resid features:
Ļ
TokenĻ
Feature activation-0.094
Top resid features:
s
Tokens
Feature activation+0.029
Top resid features:
decision
Token decision
Feature activation+0.000
Top resid features:
on
Token on
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.022
Top resid features:
Ċ
TokenĊ
Feature activation+0.004
Top resid features:
Ċ
TokenĊ
Feature activation+0.009
Top resid features:
"
Token"
Feature activation+0.138
Top resid features:
The
TokenThe
Feature activation-0.007
Top resid features:
court
Token court
Feature activation+3.504
Top resid features:
's
Token's
Feature activation-0.103
Top resid features:
decision
Token decision
Feature activation+0.000
Top resid features:
once
Token once
Feature activation+0.000
Top resid features:
again
Token again
Feature activation+0.000
Top resid features:
confirms
Token confirms
Feature activation+0.000
Top resid features:
wrote
Token wrote
Feature activation-0.037
Top resid features:
the
Token the
Feature activation-0.009
Top resid features:
opinion
Token opinion
Feature activation-0.048
Top resid features:
for
Token for
Feature activation+0.054
Top resid features:
the
Token the
Feature activation-0.005
Top resid features:
court
Token court
Feature activation+3.698
Top resid features:
,
Token,
Feature activation+0.056
Top resid features:
which
Token which
Feature activation-0.148
Top resid features:
rejected
Token rejected
Feature activation+0.010
Top resid features:
efforts
Token efforts
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
sex
Token sex
Feature activation+0.001
Top resid features:
and
Token and
Feature activation-0.013
Top resid features:
Australia
Token Australia
Feature activation+0.014
Top resid features:
's
Token's
Feature activation-0.003
Top resid features:
high
Token high
Feature activation+0.018
Top resid features:
court
Token court
Feature activation+3.579
Top resid features:
has
Token has
Feature activation-0.111
Top resid features:
struck
Token struck
Feature activation+0.107
Top resid features:
down
Token down
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
same
Token same
Feature activation+0.000
Top resid features:
sex
Token sex
Feature activation+0.001
Top resid features:
and
Token and
Feature activation+0.009
Top resid features:
Australia
Token Australia
Feature activation+0.042
Top resid features:
's
Token's
Feature activation+0.017
Top resid features:
high
Token high
Feature activation+0.029
Top resid features:
court
Token court
Feature activation+3.435
Top resid features:
has
Token has
Feature activation-0.019
Top resid features:
struck
Token struck
Feature activation+0.000
Top resid features:
down
Token down
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
same
Token same
Feature activation+0.000
Top resid features:
absence
Token absence
Feature activation-0.028
Top resid features:
of
Token of
Feature activation-0.009
Top resid features:
legislation
Token legislation
Feature activation+0.006
Top resid features:
,
Token,
Feature activation+0.001
Top resid features:
the
Token the
Feature activation+0.002
Top resid features:
courts
Token courts
Feature activation+3.511
Top resid features:
have
Token have
Feature activation-0.043
Top resid features:
now
Token now
Feature activation+0.000
Top resid features:
renewed
Token renewed
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
collection
Token collection
Feature activation+0.000
Top resid features:
The
Token The
Feature activation-0.005
Top resid features:
Denver
Token Denver
Feature activation+0.006
Top resid features:
-
Token-
Feature activation+0.002
Top resid features:
based
Tokenbased
Feature activation+0.002
Top resid features:
appellate
Token appellate
Feature activation+0.133
Top resid features:
court
Token court
Feature activation+3.288
Top resid features:
,
Token,
Feature activation+0.055
Top resid features:
one
Token one
Feature activation-0.049
Top resid features:
run
Token run
Feature activation-0.008
Top resid features:
g
Tokeng
Feature activation+0.007
Top resid features:
below
Token below
Feature activation-0.000
Top resid features:
.
Token.
Feature activation+0.053
Top resid features:
Ċ
TokenĊ
Feature activation+0.012
Top resid features:
Ċ
TokenĊ
Feature activation+0.014
Top resid features:
The
TokenThe
Feature activation-0.032
Top resid features:
Supreme
Token Supreme
Feature activation+0.416
Top resid features:
Court
Token Court
Feature activation+2.985
Top resid features:
had
Token had
Feature activation-0.072
Top resid features:
allowed
Token allowed
Feature activation+0.039
Top resid features:
Mass
Token Mass
Feature activation+0.000
Top resid features:
imil
Tokenimil
Feature activation+0.000
Top resid features:
iano
Tokeniano
Feature activation+0.000
Top resid features:
corporations
Token corporations
Feature activation+0.008
Top resid features:
that
Token that
Feature activation+0.011
Top resid features:
will
Token will
Feature activation-0.002
Top resid features:
make
Token make
Feature activation-0.062
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
Court
Token Court
Feature activation+3.640
Top resid features:
's
Token's
Feature activation-0.111
Top resid features:
landmark
Token landmark
Feature activation+0.000
Top resid features:
campaign
Token campaign
Feature activation+0.000
Top resid features:
finance
Token finance
Feature activation+0.000
Top resid features:
ruling
Token ruling
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.055
Top resid features:
For
Token For
Feature activation+0.022
Top resid features:
example
Token example
Feature activation-0.033
Top resid features:
,
Token,
Feature activation+0.009
Top resid features:
the
Token the
Feature activation-0.003
Top resid features:
Court
Token Court
Feature activation+3.326
Top resid features:
declined
Token declined
Feature activation+0.167
Top resid features:
to
Token to
Feature activation-0.084
Top resid features:
review
Token review
Feature activation+0.038
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
Tenth
Token Tenth
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.03

Head 2: 0.03

Head 3: 0.48

Head 4: 0.04

Head 5: 0.03

Head 6: 0.06

Head 7: 0.10

Head 8: 0.03

Head 9: 0.06

Head 10: 0.03

Head 11: 0.09

Positive logits

ctor1.55

sided1.55

alus1.48

nces1.44

examining1.42

investigates1.41

convened1.40

appoint1.38

jurisdiction1.37

investigating1.37

nsic1.35

hardt1.33

Exam1.33

exoner1.31

overseeing1.30

examine1.28

Cosponsors1.27

presiding1.27

otom1.26

debating1.25

Negative logits

weap-1.36

teammate-1.34

neighbour-1.23

Sold-1.22

wagen-1.18

Dad-1.17

marketed-1.15

Manager-1.14

manager-1.12

girlfriend-1.12

shipped-1.12

neighbor-1.11

ogle-1.10

Sold-1.10

purchaser-1.09

purchases-1.09

purchase-1.09

Turbo-1.08

merch-1.08

sold-1.07

INTERVAL 2.829 - 3.143
CONTAINS 0.001%

,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.079
has
Token has
Feature activation+2.498
a
Token a
Feature activation+2.424
discretion
Token discretion
Feature activation+2.862
to
Token to
Feature activation+1.276
refund
Token refund
Feature activation+1.719
part
Token part
Feature activation+0.000
or
Token or
Feature activation+0.000
all
Token all
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+1.644
goes
Token goes
Feature activation+1.937
along
Token along
Feature activation+0.784
,
Token,
Feature activation+1.530
it
Token it
Feature activation+3.000
could
Token could
Feature activation+2.113
have
Token have
Feature activation+2.473
a
Token a
Feature activation+0.820
sweeping
Token sweeping
Feature activation+0.000
impact
Token impact
Feature activation+1.158
that
Token that
Feature activation+0.957
was
Token was
Feature activation+0.324
wrong
Token wrong
Feature activation+0.748
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.106
said
Token said
Feature activation+3.143
travelling
Token travelling
Feature activation+0.000
time
Token time
Feature activation+0.000
counts
Token counts
Feature activation+0.000
as
Token as
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
Court
Token Court
Feature activation+0.000
of
Token of
Feature activation+0.044
Appeal
Token Appeal
Feature activation+0.825
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+1.186
s
Tokens
Feature activation+2.861
ruling
Token ruling
Feature activation+2.538
yesterday
Token yesterday
Feature activation+1.044
(
Token (
Feature activation+0.000
Friday
TokenFriday
Feature activation+0.406
).
Token).
Feature activation+0.012
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
courts
Token courts
Feature activation+1.168
have
Token have
Feature activation+2.647
now
Token now
Feature activation+2.554
renewed
Token renewed
Feature activation+2.927
the
Token the
Feature activation+0.686
collection
Token collection
Feature activation+0.000
of
Token of
Feature activation+0.000
telephone
Token telephone
Feature activation+0.000
metadata
Token metadata
Feature activation+0.000

INTERVAL 2.514 - 2.829
CONTAINS 0.002%

the
Token the
Feature activation+0.379
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.269
has
Token has
Feature activation+1.879
repeatedly
Token repeatedly
Feature activation+2.296
declined
Token declined
Feature activation+2.542
to
Token to
Feature activation+2.269
review
Token review
Feature activation+2.244
those
Token those
Feature activation+0.000
decisions
Token decisions
Feature activation+0.000
.
Token.
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.262
declined
Token declined
Feature activation+1.724
to
Token to
Feature activation+3.034
review
Token review
Feature activation+2.628
the
Token the
Feature activation+0.000
Tenth
Token Tenth
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+1.455
of
Token of
Feature activation+0.332
Inquiry
Token Inquiry
Feature activation+0.849
was
Token was
Feature activation+0.662
established
Token established
Feature activation+0.639
to
Token to
Feature activation+1.790
look
Token look
Feature activation+2.539
into
Token into
Feature activation+2.241
these
Token these
Feature activation+0.000
disturbances
Token disturbances
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
that
Token that
Feature activation+0.000
will
Token will
Feature activation+0.000
make
Token make
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.488
's
Token's
Feature activation+2.629
landmark
Token landmark
Feature activation+1.816
campaign
Token campaign
Feature activation+0.261
finance
Token finance
Feature activation+0.000
ruling
Token ruling
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
Australia
Token Australia
Feature activation+0.000
's
Token's
Feature activation+0.000
high
Token high
Feature activation+0.000
court
Token court
Feature activation+1.224
has
Token has
Feature activation+2.659
struck
Token struck
Feature activation+2.718
down
Token down
Feature activation+0.429
a
Token a
Feature activation+0.341
same
Token same
Feature activation+0.000
-
Token-
Feature activation+0.000

INTERVAL 2.200 - 2.514
CONTAINS 0.004%

goes
Token goes
Feature activation+1.937
along
Token along
Feature activation+0.784
,
Token,
Feature activation+1.530
it
Token it
Feature activation+3.000
could
Token could
Feature activation+2.113
have
Token have
Feature activation+2.473
a
Token a
Feature activation+0.820
sweeping
Token sweeping
Feature activation+0.000
impact
Token impact
Feature activation+1.158
on
Token on
Feature activation+0.000
access
Token access
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.000
was
Token was
Feature activation+2.452
hearing
Token hearing
Feature activation+3.013
an
Token an
Feature activation+0.000
appeal
Token appeal
Feature activation+0.000
filed
Token filed
Feature activation+0.000
by
Token by
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.128
âĢ
TokenâĢ
Feature activation+0.258
Ļ
TokenĻ
Feature activation+1.385
s
Tokens
Feature activation+2.271
ruling
Token ruling
Feature activation+2.432
last
Token last
Feature activation+0.000
month
Token month
Feature activation+0.477
that
Token that
Feature activation+1.103
his
Token his
Feature activation+0.000
party
Token party
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.717
's
Token's
Feature activation+2.745
decision
Token decision
Feature activation+2.275
once
Token once
Feature activation+0.041
again
Token again
Feature activation+0.770
confirms
Token confirms
Feature activation+0.579
that
Token that
Feature activation+1.246
the
Token the
Feature activation+0.000
high
Token high
Feature activation+0.000
court
Token court
Feature activation+0.171
used
Token used
Feature activation+2.286
to
Token to
Feature activation+0.766
invalid
Token invalid
Feature activation+1.778
ate
Tokenate
Feature activation+2.494
Washington
Token Washington
Feature activation+0.000
,
Token,
Feature activation+0.000
D
Token D
Feature activation+0.000
.
Token.
Feature activation+0.000
C
TokenC
Feature activation+0.000

INTERVAL 1.886 - 2.200
CONTAINS 0.004%

what
Token what
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+0.203
is
Token is
Feature activation+1.523
likely
Token likely
Feature activation+1.005
to
Token to
Feature activation+1.935
decide
Token decide
Feature activation+2.209
tomorrow
Token tomorrow
Feature activation+0.000
:
Token:
Feature activation+0.000
It
Token It
Feature activation+0.775
will
Token will
Feature activation+0.186
the
Token the
Feature activation+0.000
9
Token 9
Feature activation+0.000
th
Tokenth
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.000
,
Token,
Feature activation+0.768
which
Token which
Feature activation+1.914
followed
Token followed
Feature activation+1.158
a
Token a
Feature activation+0.000
hearing
Token hearing
Feature activation+0.000
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
men
Token men
Feature activation+0.000
.
Token.
Feature activation+0.000
Under
Token Under
Feature activation+0.000
the
Token the
Feature activation+0.000
court
Token court
Feature activation+0.000
's
Token's
Feature activation+1.969
interpretation
Token interpretation
Feature activation+1.640
of
Token of
Feature activation+0.000
Islamic
Token Islamic
Feature activation+0.000
law
Token law
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
appeals
Token appeals
Feature activation+0.000
court
Token court
Feature activation+0.061
judges
Token judges
Feature activation+0.000
must
Token must
Feature activation+1.611
apply
Token apply
Feature activation+2.151
the
Token the
Feature activation+0.144
same
Token same
Feature activation+1.675
test
Token test
Feature activation+0.316
the
Token the
Feature activation+0.000
high
Token high
Feature activation+0.000
But
TokenBut
Feature activation+0.000
Quebec
Token Quebec
Feature activation+0.000
's
Token's
Feature activation+0.000
top
Token top
Feature activation+0.000
court
Token court
Feature activation+0.000
ruled
Token ruled
Feature activation+2.150
in
Token in
Feature activation+0.688
favour
Token favour
Feature activation+1.243
of
Token of
Feature activation+0.000
an
Token an
Feature activation+0.000
execution
Token execution
Feature activation+0.000

INTERVAL 1.572 - 1.886
CONTAINS 0.006%

Denver
Token Denver
Feature activation+0.000
-
Token-
Feature activation+0.000
based
Tokenbased
Feature activation+0.000
appellate
Token appellate
Feature activation+0.000
court
Token court
Feature activation+0.172
,
Token,
Feature activation+1.879
one
Token one
Feature activation+1.591
run
Token run
Feature activation+1.040
g
Tokeng
Feature activation+0.000
below
Token below
Feature activation+0.000
the
Token the
Feature activation+0.000
flawed
Token flawed
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
court
Token court
Feature activation+0.000
can
Token can
Feature activation+1.794
dismiss
Token dismiss
Feature activation+1.611
the
Token the
Feature activation+0.000
cases
Token cases
Feature activation+0.000
,
Token,
Feature activation+0.000
uphold
Token uphold
Feature activation+0.766
the
Token the
Feature activation+0.000
member
Tokenmember
Feature activation+0.000
bench
Token bench
Feature activation+0.000
in
Token in
Feature activation+1.180
October
Token October
Feature activation+0.909
2016
Token 2016
Feature activation+0.669
to
Token to
Feature activation+1.844
start
Token start
Feature activation+1.846
hearings
Token hearings
Feature activation+0.396
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
case
Token case
Feature activation+0.000
me
Token me
Feature activation+0.000
is
Token is
Feature activation+0.000
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
courts
Token courts
Feature activation+0.943
might
Token might
Feature activation+1.783
choose
Token choose
Feature activation+2.202
a
Token a
Feature activation+0.442
m
Token m
Feature activation+0.000
uddled
Tokenuddled
Feature activation+0.138
half
Token half
Feature activation+0.054
hoped
Token hoped
Feature activation+0.000
the
Token the
Feature activation+0.000
high
Token high
Feature activation+0.000
court
Token court
Feature activation+0.000
would
Token would
Feature activation+2.221
now
Token now
Feature activation+1.880
settle
Token settle
Feature activation+2.472
the
Token the
Feature activation+0.000
matter
Token matter
Feature activation+0.000
promptly
Token promptly
Feature activation+0.000
.
Token.
Feature activation+0.263

INTERVAL 1.257 - 1.572
CONTAINS 0.006%

juvenile
Token juvenile
Feature activation+0.000
courts
Token courts
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.502
are
Token are
Feature activation+1.403
based
Token based
Feature activation+1.315
on
Token on
Feature activation+0.072
the
Token the
Feature activation+0.000
premise
Token premise
Feature activation+0.000
that
Token that
Feature activation+0.000
children
Token children
Feature activation+0.000
P
Token P
Feature activation+0.000
anch
Tokenanch
Feature activation+0.000
k
Tokenk
Feature activation+0.000
ula
Tokenula
Feature activation+0.000
on
Token on
Feature activation+0.000
Friday
Token Friday
Feature activation+1.282
.
Token.
Feature activation+0.293
Several
Token Several
Feature activation+0.000
north
Token north
Feature activation+0.000
Indian
Token Indian
Feature activation+0.000
cities
Token cities
Feature activation+0.000
Court
Token Court
Feature activation+0.000
seems
Token seems
Feature activation+1.160
to
Token to
Feature activation+1.938
be
Token be
Feature activation+1.606
saying
Token saying
Feature activation+2.324
:
Token:
Feature activation+1.290
This
Token This
Feature activation+0.239
is
Token is
Feature activation+0.000
a
Token a
Feature activation+0.000
state
Token state
Feature activation+0.000
-
Token-
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
Pennsylvania
Token Pennsylvania
Feature activation+0.000
appeals
Token appeals
Feature activation+0.000
court
Token court
Feature activation+0.435
ruled
Token ruled
Feature activation+1.466
this
Token this
Feature activation+0.000
week
Token week
Feature activation+1.106
.
Token.
Feature activation+0.395
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Court
Token Court
Feature activation+0.000
Justice
Token Justice
Feature activation+0.000
that
Token that
Feature activation+0.616
"
Token "
Feature activation+1.082
l
Tokenl
Feature activation+0.411
acks
Tokenacks
Feature activation+1.311
good
Token good
Feature activation+0.000
behavior
Token behavior
Feature activation+0.000
"
Token"
Feature activation+0.000
can
Token can
Feature activation+0.000
be
Token be
Feature activation+0.000

INTERVAL 0.943 - 1.257
CONTAINS 0.009%

week
Token week
Feature activation+0.000
while
Token while
Feature activation+0.000
the
Token the
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.000
decides
Token decides
Feature activation+0.976
on
Token on
Feature activation+0.403
the
Token the
Feature activation+0.000
outcome
Token outcome
Feature activation+0.000
of
Token of
Feature activation+0.000
his
Token his
Feature activation+0.000
.
Token.
Feature activation+0.000
Leon
Token Leon
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.116
s
Tokens
Feature activation+0.876
final
Token final
Feature activation+1.134
ruling
Token ruling
Feature activation+0.768
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
lawsuit
Token lawsuit
Feature activation+0.000
Tuesday
Token Tuesday
Feature activation+0.000
the
Token the
Feature activation+0.000
Court
Token Court
Feature activation+0.000
of
Token of
Feature activation+0.044
Appeal
Token Appeal
Feature activation+0.825
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+1.186
s
Tokens
Feature activation+2.861
ruling
Token ruling
Feature activation+2.538
yesterday
Token yesterday
Feature activation+1.044
(
Token (
Feature activation+0.000
Friday
TokenFriday
Feature activation+0.406
a
Token a
Feature activation+0.000
Supreme
Token Supreme
Feature activation+0.000
Court
Token Court
Feature activation+0.000
Justice
Token Justice
Feature activation+0.000
that
Token that
Feature activation+0.616
"
Token "
Feature activation+1.082
l
Tokenl
Feature activation+0.411
acks
Tokenacks
Feature activation+1.311
good
Token good
Feature activation+0.000
behavior
Token behavior
Feature activation+0.000
"
Token"
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
Fourth
Token Fourth
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.751
majority
Token majority
Feature activation+0.117
ruled
Token ruled
Feature activation+0.991
that
Token that
Feature activation+0.371
the
Token the
Feature activation+0.000
district
Token district
Feature activation+0.000
court
Token court
Feature activation+0.000
wrongly
Token wrongly
Feature activation+1.720

INTERVAL 0.629 - 0.943
CONTAINS 0.016%

Appeals
Token Appeals
Feature activation+0.317
Judge
Token Judge
Feature activation+0.000
Frank
Token Frank
Feature activation+0.000
Easter
Token Easter
Feature activation+0.000
brook
Tokenbrook
Feature activation+0.000
held
Token held
Feature activation+0.634
that
Token that
Feature activation+0.351
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
[
Token[
Feature activation+0.000
a
Tokena
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.014
ruled
Token ruled
Feature activation+1.936
that
Token that
Feature activation+0.957
was
Token was
Feature activation+0.324
wrong
Token wrong
Feature activation+0.748
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.106
said
Token said
Feature activation+3.143
travelling
Token travelling
Feature activation+0.000
time
Token time
Feature activation+0.000
unrelated
Token unrelated
Feature activation+0.000
case
Token case
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
High
Token High
Feature activation+0.000
Court
Token Court
Feature activation+0.891
ruled
Token ruled
Feature activation+2.577
in
Token in
Feature activation+0.941
April
Token April
Feature activation+1.222
that
Token that
Feature activation+2.670
the
Token the
Feature activation+0.000
three
Token three
Feature activation+0.000
-
Token-
Feature activation+0.000
jud
Tokenjud
Feature activation+0.000
ge
Tokenge
Feature activation+0.000
panel
Token panel
Feature activation+0.050
examined
Token examined
Feature activation+0.658
the
Token the
Feature activation+0.000
ballot
Token ballot
Feature activation+0.000
and
Token and
Feature activation+0.000
agreed
Token agreed
Feature activation+0.800
to
Token to
Feature activation+0.726
No
TokenNo
Feature activation+0.000
matter
Token matter
Feature activation+0.000
what
Token what
Feature activation+0.000
the
Token the
Feature activation+0.000
justices
Token justices
Feature activation+0.064
say
Token say
Feature activation+0.919
Tuesday
Token Tuesday
Feature activation+1.031
,
Token,
Feature activation+0.009
and
Token and
Feature activation+0.000
no
Token no
Feature activation+0.000
matter
Token matter
Feature activation+0.000

INTERVAL 0.314 - 0.629
CONTAINS 0.031%

that
Token that
Feature activation+0.000
Congress
Token Congress
Feature activation+0.000
was
Token was
Feature activation+0.017
intr
Token intr
Feature activation+0.378
uding
Tokenuding
Feature activation+0.615
into
Token into
Feature activation+0.454
the
Token the
Feature activation+0.000
business
Token business
Feature activation+0.000
of
Token of
Feature activation+0.000
federal
Token federal
Feature activation+0.000
courts
Token courts
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
court
Token court
Feature activation+0.070
ruling
Token ruling
Feature activation+0.768
came
Token came
Feature activation+0.000
on
Token on
Feature activation+0.502
a
Token a
Feature activation+0.619
statement
Token statement
Feature activation+1.028
filed
Token filed
Feature activation+0.000
by
Token by
Feature activation+0.000
federal
Token federal
Feature activation+0.000
sufficient
Token sufficient
Feature activation+0.000
wrongdoing
Token wrongdoing
Feature activation+0.000
by
Token by
Feature activation+0.000
Target
Token Target
Feature activation+0.000
to
Token to
Feature activation+0.000
throw
Token throw
Feature activation+0.466
out
Token out
Feature activation+0.000
that
Token that
Feature activation+0.000
election
Token election
Feature activation+0.000
result
Token result
Feature activation+0.000
,
Token,
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Seventh
Token Seventh
Feature activation+0.000
Circuit
Token Circuit
Feature activation+0.220
Cout
Token Cout
Feature activation+0.000
of
Token of
Feature activation+0.000
Appeals
Token Appeals
Feature activation+0.317
Judge
Token Judge
Feature activation+0.000
Frank
Token Frank
Feature activation+0.000
Easter
Token Easter
Feature activation+0.000
brook
Tokenbrook
Feature activation+0.000
held
Token held
Feature activation+0.634
igroup
Tokenigroup
Feature activation+0.000
settlement
Token settlement
Feature activation+0.000
.
Token.
Feature activation+0.000
Judge
Token Judge
Feature activation+0.000
Rak
Token Rak
Feature activation+0.000
off
Tokenoff
Feature activation+0.390
has
Token has
Feature activation+0.012
asked
Token asked
Feature activation+0.652
the
Token the
Feature activation+0.000
agency
Token agency
Feature activation+0.000
what
Token what
Feature activation+0.000

INTERVAL 0.000 - 0.314
CONTAINS 99.923%

of
Token of
Feature activation+0.000
car
Token car
Feature activation+0.000
accidents
Token accidents
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
benefit
Token benefit
Feature activation+0.000
of
Token of
Feature activation+0.000
those
Token those
Feature activation+0.000
that
Token that
Feature activation+0.000
are
Token are
Feature activation+0.000
causing
Token causing
Feature activation+0.000
sensitivity
Token sensitivity
Feature activation+0.000
setting
Token setting
Feature activation+0.000
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
uncontroll
Token uncontroll
Feature activation+0.000
able
Tokenable
Feature activation+0.000
.
Token.
Feature activation+0.000
It
Token It
Feature activation+0.000
is
Token is
Feature activation+0.000
definitely
Token definitely
Feature activation+0.000
the
Token the
Feature activation+0.000
ig
Tokenig
Feature activation+0.000
Michael
Token Michael
Feature activation+0.000
,
Token,
Feature activation+0.000
Ireland
Token Ireland
Feature activation+0.000
as
Token as
Feature activation+0.000
Ah
Token Ah
Feature activation+0.000
ch
Tokench
Feature activation+0.000
-
Token-
Feature activation+0.000
To
TokenTo
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
is
Token is
Feature activation+0.000
slated
Token slated
Feature activation+0.000
to
Token to
Feature activation+0.000
receive
Token receive
Feature activation+0.000
a
Token a
Feature activation+0.000
total
Token total
Feature activation+0.000
of
Token of
Feature activation+0.000
$
Token $
Feature activation+0.000
210
Token210
Feature activation+0.000
,
Token,
Feature activation+0.000
000
Token000
Feature activation+0.000
,
Token,
Feature activation+0.000
at
Token at
Feature activation+0.000
least
Token least
Feature activation+0.000
until
Token until
Feature activation+0.000
they
Token they
Feature activation+0.000
google
Token google
Feature activation+0.000
you
Token you
Feature activation+0.000
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
The
TokenThe
Feature activation+0.000
goal
Token goal
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 8 in H4.3: (feature 11622)

TOP ACTIVATIONS
MAX = 3.972

Toronto
Token Toronto
Feature activation+0.000
family
Token family
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
starts
Token starts
Feature activation+3.540
with
Token with
Feature activation+3.972
an
Token an
Feature activation+2.524
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
F
TokenF
Feature activation+0.000
,
Token,
Feature activation+0.000
Their
Token Their
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+3.083
be
Token be
Feature activation+3.887
added
Token added
Feature activation+3.207
to
Token to
Feature activation+3.890
the
Token the
Feature activation+3.009
endless
Token endless
Feature activation+0.000
list
Token list
Feature activation+0.086
of
Token of
Feature activation+0.000
victims
Token victims
Feature activation+0.000
injured
Token injured
Feature activation+0.000
.
Token.
Feature activation+0.000
Their
Token Their
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+3.083
be
Token be
Feature activation+3.887
added
Token added
Feature activation+3.207
to
Token to
Feature activation+3.890
the
Token the
Feature activation+3.009
endless
Token endless
Feature activation+0.000
list
Token list
Feature activation+0.086
proud
Token proud
Feature activation+0.000
to
Token to
Feature activation+0.000
have
Token have
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.120
on
Token on
Feature activation+3.797
,
Token,
Feature activation+1.545
thanks
Token thanks
Feature activation+0.000
to
Token to
Feature activation+0.035
a
Token a
Feature activation+0.000
partner
Token partner
Feature activation+0.000
Belgian
Token Belgian
Feature activation+0.000
farmer
Token farmer
Feature activation+0.000
,
Token,
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
was
Token was
Feature activation+3.617
Henri
Token Henri
Feature activation+0.000
Le
Token Le
Feature activation+0.000
j
Tokenj
Feature activation+0.000
oly
Tokenoly
Feature activation+0.000
,
Token,
Feature activation+0.000
Russo
Token Russo
Feature activation+0.000
had
Token had
Feature activation+0.000
thrown
Token thrown
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.000
in
Token in
Feature activation+3.576
the
Token the
Feature activation+3.207
hat
Token hat
Feature activation+2.983
,
Token,
Feature activation+0.019
and
Token and
Feature activation+0.626
they
Token they
Feature activation+0.000
the
Token the
Feature activation+0.000
Toronto
Token Toronto
Feature activation+0.000
family
Token family
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
starts
Token starts
Feature activation+3.540
with
Token with
Feature activation+3.972
an
Token an
Feature activation+2.524
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
F
TokenF
Feature activation+0.000
compounds
Token compounds
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
name
Token name
Feature activation+0.000
that
Token that
Feature activation+2.626
doesn
Token doesn
Feature activation+3.509
't
Token't
Feature activation+3.409
necessarily
Token necessarily
Feature activation+1.389
mean
Token mean
Feature activation+2.180
the
Token the
Feature activation+0.209
compounds
Token compounds
Feature activation+0.000
would
Token would
Feature activation+0.000
,
Token,
Feature activation+0.000
as
Token as
Feature activation+0.000
its
Token its
Feature activation+0.000
name
Token name
Feature activation+0.000
suggests
Token suggests
Feature activation+3.500
,
Token,
Feature activation+0.421
repeal
Token repeal
Feature activation+0.000
the
Token the
Feature activation+0.000
Patriot
Token Patriot
Feature activation+0.000
Act
Token Act
Feature activation+0.000
Wal
Token Wal
Feature activation+0.000
id
Tokenid
Feature activation+0.000
Sho
Token Sho
Feature activation+0.000
eb
Tokeneb
Feature activation+0.000
at
Tokenat
Feature activation+0.000
sounds
Token sounds
Feature activation+3.464
familiar
Token familiar
Feature activation+2.757
,
Token,
Feature activation+0.675
it
Token it
Feature activation+1.233
's
Token's
Feature activation+0.273
because
Token because
Feature activation+0.371
ened
Tokenened
Feature activation+0.000
to
Token to
Feature activation+0.000
see
Token see
Feature activation+0.000
his
Token his
Feature activation+0.000
name
Token name
Feature activation+0.000
listed
Token listed
Feature activation+3.432
.
Token.
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
I
TokenI
Feature activation+0.000
knew
Token knew
Feature activation+0.000
name
Token name
Feature activation+0.231
Charlotte
Token Charlotte
Feature activation+0.000
Cord
Token Cord
Feature activation+0.000
ay
Tokenay
Feature activation+0.000
can
Token can
Feature activation+1.437
be
Token be
Feature activation+3.432
seen
Token seen
Feature activation+2.817
on
Token on
Feature activation+2.712
the
Token the
Feature activation+1.812
paper
Token paper
Feature activation+0.000
held
Token held
Feature activation+0.000
so
Token so
Feature activation+0.000
that
Token that
Feature activation+0.000
anyone
Token anyone
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
is
Token is
Feature activation+3.415
targeted
Token targeted
Feature activation+2.538
is
Token is
Feature activation+0.000
now
Token now
Feature activation+0.000
informed
Token informed
Feature activation+0.000
about
Token about
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
name
Token name
Feature activation+0.000
that
Token that
Feature activation+2.626
doesn
Token doesn
Feature activation+3.509
't
Token't
Feature activation+3.409
necessarily
Token necessarily
Feature activation+1.389
mean
Token mean
Feature activation+2.180
the
Token the
Feature activation+0.209
compounds
Token compounds
Feature activation+0.000
come
Token come
Feature activation+0.000
and
Token and
Feature activation+0.000
politicians
Token politicians
Feature activation+0.000
whose
Token whose
Feature activation+0.000
names
Token names
Feature activation+0.000
barely
Token barely
Feature activation+3.125
register
Token register
Feature activation+3.377
with
Token with
Feature activation+1.473
the
Token the
Feature activation+0.616
general
Token general
Feature activation+0.000
public
Token public
Feature activation+0.000
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
,
Token,
Feature activation+0.000
as
Token as
Feature activation+0.000
the
Token the
Feature activation+0.000
name
Token name
Feature activation+0.133
states
Token states
Feature activation+3.376
,
Token,
Feature activation+0.656
a
Token a
Feature activation+0.112
coalition
Token coalition
Feature activation+0.000
.
Token.
Feature activation+0.000
Along
Token Along
Feature activation+0.000
Feel
Token Feel
Feature activation+0.000
ies
Tokenies
Feature activation+0.000
took
Token took
Feature activation+0.000
its
Token its
Feature activation+0.000
name
Token name
Feature activation+0.000
from
Token from
Feature activation+3.314
Ald
Token Ald
Feature activation+0.000
ous
Tokenous
Feature activation+0.000
H
Token H
Feature activation+0.000
ux
Tokenux
Feature activation+0.000
ley
Tokenley
Feature activation+0.000
Her
Token Her
Feature activation+0.000
identity
Token identity
Feature activation+0.000
was
Token was
Feature activation+2.608
not
Token not
Feature activation+3.235
immediately
Token immediately
Feature activation+2.544
released
Token released
Feature activation+3.304
pending
Token pending
Feature activation+0.523
notification
Token notification
Feature activation+0.000
of
Token of
Feature activation+0.000
her
Token her
Feature activation+0.000
next
Token next
Feature activation+0.000
scene
Token scene
Feature activation+0.000
.
Token.
Feature activation+0.000
Her
Token Her
Feature activation+0.000
identity
Token identity
Feature activation+0.000
was
Token was
Feature activation+2.608
not
Token not
Feature activation+3.235
immediately
Token immediately
Feature activation+2.544
released
Token released
Feature activation+3.304
pending
Token pending
Feature activation+0.523
notification
Token notification
Feature activation+0.000
of
Token of
Feature activation+0.000
had
Token had
Feature activation+0.000
thrown
Token thrown
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.000
in
Token in
Feature activation+3.576
the
Token the
Feature activation+3.207
hat
Token hat
Feature activation+2.983
,
Token,
Feature activation+0.019
and
Token and
Feature activation+0.626
they
Token they
Feature activation+0.000
responded
Token responded
Feature activation+0.000

Top DFA by src position
MAX = 4.564

in
Token in
Feature activation+0.004
Top resid features:
the
Token the
Feature activation+0.017
Top resid features:
Toronto
Token Toronto
Feature activation-0.033
Top resid features:
family
Token family
Feature activation-0.011
Top resid features:
whose
Token whose
Feature activation-0.125
Top resid features:
name
Token name
Feature activation+4.382
Top resid features:
starts
Token starts
Feature activation+0.234
Top resid features:
with
Token with
Feature activation+0.143
Top resid features:
an
Token an
Feature activation+0.000
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.004
Top resid features:
4
Token 4
Feature activation+0.002
Top resid features:
injured
Token injured
Feature activation-0.001
Top resid features:
.
Token.
Feature activation-0.019
Top resid features:
Their
Token Their
Feature activation-0.006
Top resid features:
names
Token names
Feature activation+4.286
Top resid features:
will
Token will
Feature activation-0.005
Top resid features:
be
Token be
Feature activation+0.104
Top resid features:
added
Token added
Feature activation+0.108
Top resid features:
to
Token to
Feature activation+0.084
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.008
Top resid features:
4
Token 4
Feature activation+0.003
Top resid features:
injured
Token injured
Feature activation-0.009
Top resid features:
.
Token.
Feature activation+0.009
Top resid features:
Their
Token Their
Feature activation-0.004
Top resid features:
names
Token names
Feature activation+4.564
Top resid features:
will
Token will
Feature activation+0.033
Top resid features:
be
Token be
Feature activation-0.047
Top resid features:
added
Token added
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
incredibly
Token incredibly
Feature activation-0.017
Top resid features:
proud
Token proud
Feature activation-0.030
Top resid features:
to
Token to
Feature activation-0.003
Top resid features:
have
Token have
Feature activation+0.042
Top resid features:
my
Token my
Feature activation+0.002
Top resid features:
name
Token name
Feature activation+4.374
Top resid features:
on
Token on
Feature activation+0.061
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
thanks
Token thanks
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
The
TokenThe
Feature activation-0.015
Top resid features:
Belgian
Token Belgian
Feature activation-0.074
Top resid features:
farmer
Token farmer
Feature activation-0.046
Top resid features:
,
Token,
Feature activation-0.049
Top resid features:
whose
Token whose
Feature activation-0.157
Top resid features:
name
Token name
Feature activation+4.301
Top resid features:
was
Token was
Feature activation+0.310
Top resid features:
Henri
Token Henri
Feature activation+0.000
Top resid features:
Le
Token Le
Feature activation+0.000
Top resid features:
j
Tokenj
Feature activation+0.000
Top resid features:
oly
Tokenoly
Feature activation+0.000
Top resid features:
Joe
TokenJoe
Feature activation+0.001
Top resid features:
Russo
Token Russo
Feature activation-0.006
Top resid features:
had
Token had
Feature activation-0.005
Top resid features:
thrown
Token thrown
Feature activation+0.089
Top resid features:
my
Token my
Feature activation+0.007
Top resid features:
name
Token name
Feature activation+4.187
Top resid features:
in
Token in
Feature activation-0.070
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
hat
Token hat
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.011
Top resid features:
the
Token the
Feature activation+0.018
Top resid features:
Toronto
Token Toronto
Feature activation-0.042
Top resid features:
family
Token family
Feature activation-0.011
Top resid features:
whose
Token whose
Feature activation-0.174
Top resid features:
name
Token name
Feature activation+4.310
Top resid features:
starts
Token starts
Feature activation+0.060
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
an
Token an
Feature activation+0.000
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
called
Tokencalled
Feature activation-0.010
Top resid features:
organic
Token organic
Feature activation-0.020
Top resid features:
compounds
Token compounds
Feature activation-0.022
Top resid features:
,
Token,
Feature activation-0.048
Top resid features:
a
Token a
Feature activation-0.057
Top resid features:
name
Token name
Feature activation+3.882
Top resid features:
that
Token that
Feature activation+0.229
Top resid features:
doesn
Token doesn
Feature activation+0.102
Top resid features:
't
Token't
Feature activation+0.000
Top resid features:
necessarily
Token necessarily
Feature activation+0.000
Top resid features:
mean
Token mean
Feature activation+0.000
Top resid features:
Act
Token Act
Feature activation-0.004
Top resid features:
would
Token would
Feature activation-0.006
Top resid features:
,
Token,
Feature activation-0.044
Top resid features:
as
Token as
Feature activation-0.052
Top resid features:
its
Token its
Feature activation-0.009
Top resid features:
name
Token name
Feature activation+4.123
Top resid features:
suggests
Token suggests
Feature activation+0.187
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
repeal
Token repeal
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
Patriot
Token Patriot
Feature activation+0.000
Top resid features:
cad
Token cad
Feature activation+0.000
Top resid features:
ets
Tokenets
Feature activation-0.002
Top resid features:
.
Token.
Feature activation-0.022
Top resid features:
If
Token If
Feature activation-0.017
Top resid features:
the
Token the
Feature activation-0.017
Top resid features:
name
Token name
Feature activation+4.097
Top resid features:
Wal
Token Wal
Feature activation-0.041
Top resid features:
id
Tokenid
Feature activation-0.009
Top resid features:
Sho
Token Sho
Feature activation+0.008
Top resid features:
eb
Tokeneb
Feature activation-0.060
Top resid features:
at
Tokenat
Feature activation+0.133
Top resid features:
sick
Token sick
Feature activation+0.011
Top resid features:
ened
Tokenened
Feature activation-0.048
Top resid features:
to
Token to
Feature activation+0.012
Top resid features:
see
Token see
Feature activation+0.190
Top resid features:
his
Token his
Feature activation-0.013
Top resid features:
name
Token name
Feature activation+3.962
Top resid features:
listed
Token listed
Feature activation-0.034
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
âĢ
Token âĢ
Feature activation+0.000
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
I
TokenI
Feature activation+0.000
Top resid features:
able
Tokenable
Feature activation-0.007
Top resid features:
hero
Token hero
Feature activation-0.018
Top resid features:
.
Token.
Feature activation-0.027
Top resid features:
Although
Token Although
Feature activation+0.062
Top resid features:
the
Token the
Feature activation+0.035
Top resid features:
name
Token name
Feature activation+3.810
Top resid features:
Charlotte
Token Charlotte
Feature activation-0.056
Top resid features:
Cord
Token Cord
Feature activation-0.016
Top resid features:
ay
Tokenay
Feature activation-0.042
Top resid features:
can
Token can
Feature activation+0.246
Top resid features:
be
Token be
Feature activation+0.001
Top resid features:
,
Token,
Feature activation-0.010
Top resid features:
so
Token so
Feature activation+0.007
Top resid features:
that
Token that
Feature activation+0.017
Top resid features:
anyone
Token anyone
Feature activation-0.042
Top resid features:
whose
Token whose
Feature activation-0.261
Top resid features:
name
Token name
Feature activation+4.272
Top resid features:
is
Token is
Feature activation+0.085
Top resid features:
targeted
Token targeted
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
now
Token now
Feature activation+0.000
Top resid features:
informed
Token informed
Feature activation+0.000
Top resid features:
called
Tokencalled
Feature activation-0.036
Top resid features:
organic
Token organic
Feature activation-0.012
Top resid features:
compounds
Token compounds
Feature activation-0.008
Top resid features:
,
Token,
Feature activation-0.028
Top resid features:
a
Token a
Feature activation-0.042
Top resid features:
name
Token name
Feature activation+3.746
Top resid features:
that
Token that
Feature activation+0.044
Top resid features:
doesn
Token doesn
Feature activation+0.283
Top resid features:
't
Token't
Feature activation+0.069
Top resid features:
necessarily
Token necessarily
Feature activation+0.000
Top resid features:
mean
Token mean
Feature activation+0.000
Top resid features:
primary
Token primary
Feature activation-0.002
Top resid features:
opponents
Token opponents
Feature activation-0.014
Top resid features:
and
Token and
Feature activation+0.009
Top resid features:
politicians
Token politicians
Feature activation+0.005
Top resid features:
whose
Token whose
Feature activation-0.118
Top resid features:
names
Token names
Feature activation+3.995
Top resid features:
barely
Token barely
Feature activation+0.168
Top resid features:
register
Token register
Feature activation-0.064
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
general
Token general
Feature activation+0.000
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation+0.079
Top resid features:
,
Token,
Feature activation-0.129
Top resid features:
as
Token as
Feature activation+0.139
Top resid features:
the
Token the
Feature activation-0.031
Top resid features:
name
Token name
Feature activation+3.954
Top resid features:
states
Token states
Feature activation+0.019
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
coalition
Token coalition
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.015
Top resid features:
Feel
Token Feel
Feature activation-0.001
Top resid features:
ies
Tokenies
Feature activation-0.028
Top resid features:
took
Token took
Feature activation+0.025
Top resid features:
its
Token its
Feature activation+0.003
Top resid features:
name
Token name
Feature activation+3.865
Top resid features:
from
Token from
Feature activation+0.115
Top resid features:
Ald
Token Ald
Feature activation+0.000
Top resid features:
ous
Tokenous
Feature activation+0.000
Top resid features:
H
Token H
Feature activation+0.000
Top resid features:
ux
Tokenux
Feature activation+0.000
Top resid features:
at
Token at
Feature activation+0.003
Top resid features:
the
Token the
Feature activation+0.001
Top resid features:
scene
Token scene
Feature activation-0.015
Top resid features:
.
Token.
Feature activation-0.050
Top resid features:
Her
Token Her
Feature activation+0.001
Top resid features:
identity
Token identity
Feature activation+4.002
Top resid features:
was
Token was
Feature activation+0.038
Top resid features:
not
Token not
Feature activation+0.052
Top resid features:
immediately
Token immediately
Feature activation-0.011
Top resid features:
released
Token released
Feature activation-0.036
Top resid features:
pending
Token pending
Feature activation+0.000
Top resid features:
at
Token at
Feature activation+0.005
Top resid features:
the
Token the
Feature activation+0.002
Top resid features:
scene
Token scene
Feature activation-0.010
Top resid features:
.
Token.
Feature activation-0.036
Top resid features:
Her
Token Her
Feature activation+0.006
Top resid features:
identity
Token identity
Feature activation+4.057
Top resid features:
was
Token was
Feature activation-0.007
Top resid features:
not
Token not
Feature activation+0.011
Top resid features:
immediately
Token immediately
Feature activation+0.000
Top resid features:
released
Token released
Feature activation+0.000
Top resid features:
pending
Token pending
Feature activation+0.000
Top resid features:
Joe
TokenJoe
Feature activation-0.002
Top resid features:
Russo
Token Russo
Feature activation-0.009
Top resid features:
had
Token had
Feature activation-0.010
Top resid features:
thrown
Token thrown
Feature activation+0.005
Top resid features:
my
Token my
Feature activation-0.094
Top resid features:
name
Token name
Feature activation+3.804
Top resid features:
in
Token in
Feature activation+0.152
Top resid features:
the
Token the
Feature activation+0.065
Top resid features:
hat
Token hat
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.04

Head 2: 0.02

Head 3: 0.48

Head 4: 0.03

Head 5: 0.03

Head 6: 0.05

Head 7: 0.12

Head 8: 0.02

Head 9: 0.06

Head 10: 0.01

Head 11: 0.10

Positive logits

trademarks1.50

initials1.50

1.49

likeness1.37

blurred1.35

Hatt1.35

amen1.31

trademark1.31

nationality1.30

ゴン1.27

tattoo1.23

OSH1.22

bearer1.19

pronounced1.19

ambig1.18

BUT1.18

Marks1.18

Nordic1.17

clarity1.14

tradem1.13

Negative logits

alore-1.43

rate-1.41

iless-1.40

issance-1.39

urious-1.34

succeeds-1.30

ocr-1.30

issions-1.29

successful-1.28

eat-1.28

akuya-1.25

trap-1.24

asters-1.23

workshop-1.22

meal-1.21

icago-1.20

loads-1.19

hooting-1.18

icer-1.18

ritis-1.18

INTERVAL 3.575 - 3.972
CONTAINS 0.001%

proud
Token proud
Feature activation+0.000
to
Token to
Feature activation+0.000
have
Token have
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.120
on
Token on
Feature activation+3.797
,
Token,
Feature activation+1.545
thanks
Token thanks
Feature activation+0.000
to
Token to
Feature activation+0.035
a
Token a
Feature activation+0.000
partner
Token partner
Feature activation+0.000
Russo
Token Russo
Feature activation+0.000
had
Token had
Feature activation+0.000
thrown
Token thrown
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.000
in
Token in
Feature activation+3.576
the
Token the
Feature activation+3.207
hat
Token hat
Feature activation+2.983
,
Token,
Feature activation+0.019
and
Token and
Feature activation+0.626
they
Token they
Feature activation+0.000
Their
Token Their
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+3.083
be
Token be
Feature activation+3.887
added
Token added
Feature activation+3.207
to
Token to
Feature activation+3.890
the
Token the
Feature activation+3.009
endless
Token endless
Feature activation+0.000
list
Token list
Feature activation+0.086
of
Token of
Feature activation+0.000
victims
Token victims
Feature activation+0.000
injured
Token injured
Feature activation+0.000
.
Token.
Feature activation+0.000
Their
Token Their
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+3.083
be
Token be
Feature activation+3.887
added
Token added
Feature activation+3.207
to
Token to
Feature activation+3.890
the
Token the
Feature activation+3.009
endless
Token endless
Feature activation+0.000
list
Token list
Feature activation+0.086
Belgian
Token Belgian
Feature activation+0.000
farmer
Token farmer
Feature activation+0.000
,
Token,
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
was
Token was
Feature activation+3.617
Henri
Token Henri
Feature activation+0.000
Le
Token Le
Feature activation+0.000
j
Tokenj
Feature activation+0.000
oly
Tokenoly
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 3.178 - 3.575
CONTAINS 0.002%

and
Token and
Feature activation+0.000
politicians
Token politicians
Feature activation+0.000
whose
Token whose
Feature activation+0.000
names
Token names
Feature activation+0.000
barely
Token barely
Feature activation+3.125
register
Token register
Feature activation+3.377
with
Token with
Feature activation+1.473
the
Token the
Feature activation+0.616
general
Token general
Feature activation+0.000
public
Token public
Feature activation+0.000
.
Token.
Feature activation+0.000
had
Token had
Feature activation+0.000
thrown
Token thrown
Feature activation+0.000
my
Token my
Feature activation+0.000
name
Token name
Feature activation+0.000
in
Token in
Feature activation+3.576
the
Token the
Feature activation+3.207
hat
Token hat
Feature activation+2.983
,
Token,
Feature activation+0.019
and
Token and
Feature activation+0.626
they
Token they
Feature activation+0.000
responded
Token responded
Feature activation+0.000
name
Token name
Feature activation+0.231
Charlotte
Token Charlotte
Feature activation+0.000
Cord
Token Cord
Feature activation+0.000
ay
Tokenay
Feature activation+0.000
can
Token can
Feature activation+1.437
be
Token be
Feature activation+3.432
seen
Token seen
Feature activation+2.817
on
Token on
Feature activation+2.712
the
Token the
Feature activation+1.812
paper
Token paper
Feature activation+0.000
held
Token held
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
name
Token name
Feature activation+0.000
that
Token that
Feature activation+2.626
doesn
Token doesn
Feature activation+3.509
't
Token't
Feature activation+3.409
necessarily
Token necessarily
Feature activation+1.389
mean
Token mean
Feature activation+2.180
the
Token the
Feature activation+0.209
compounds
Token compounds
Feature activation+0.000
come
Token come
Feature activation+0.000
the
Token the
Feature activation+0.000
Toronto
Token Toronto
Feature activation+0.000
family
Token family
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
starts
Token starts
Feature activation+3.540
with
Token with
Feature activation+3.972
an
Token an
Feature activation+2.524
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
F
TokenF
Feature activation+0.000

INTERVAL 2.781 - 3.178
CONTAINS 0.002%

the
Token the
Feature activation+1.362
n
Token n
Feature activation+0.000
eng
Tokeneng
Feature activation+0.000
Åį
TokenÅį
Feature activation+0.000
was
Token was
Feature activation+1.824
changed
Token changed
Feature activation+2.943
to
Token to
Feature activation+2.958
Ky
Token Ky
Feature activation+0.000
Å«
Tokenū
Feature activation+0.000
an
Tokenan
Feature activation+0.000
.
Token .
Feature activation+0.000
of
Token of
Feature activation+1.161
the
Token the
Feature activation+1.500
country
Token country
Feature activation+0.000
was
Token was
Feature activation+2.191
not
Token not
Feature activation+2.469
included
Token included
Feature activation+2.860
within
Token within
Feature activation+2.218
the
Token the
Feature activation+0.662
design
Token design
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.282
from
Token from
Feature activation+0.000
a
Token a
Feature activation+0.000
group
Token group
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
contained
Token contained
Feature activation+2.990
the
Token the
Feature activation+1.493
word
Token word
Feature activation+0.000
As
Token As
Feature activation+0.000
at
Tokenat
Feature activation+0.000
ru
Tokenru
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+3.083
be
Token be
Feature activation+3.887
added
Token added
Feature activation+3.207
to
Token to
Feature activation+3.890
the
Token the
Feature activation+3.009
endless
Token endless
Feature activation+0.000
list
Token list
Feature activation+0.086
of
Token of
Feature activation+0.000
victims
Token victims
Feature activation+0.000
of
Token of
Feature activation+0.000
in
Token in
Feature activation+0.000
English
Token English
Feature activation+0.010
,
Token,
Feature activation+0.000
was
Token was
Feature activation+2.960
one
Token one
Feature activation+2.475
of
Token of
Feature activation+2.785
two
Token two
Feature activation+0.805
forwarded
Token forwarded
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
Spokane
Token Spokane
Feature activation+0.000

INTERVAL 2.383 - 2.781
CONTAINS 0.003%

name
Token name
Feature activation+0.177
of
Token of
Feature activation+1.161
the
Token the
Feature activation+1.500
country
Token country
Feature activation+0.000
was
Token was
Feature activation+2.191
not
Token not
Feature activation+2.469
included
Token included
Feature activation+2.860
within
Token within
Feature activation+2.218
the
Token the
Feature activation+0.662
design
Token design
Feature activation+0.000
,
Token,
Feature activation+0.000
family
Token family
Feature activation+0.000
whose
Token whose
Feature activation+0.000
name
Token name
Feature activation+0.000
starts
Token starts
Feature activation+3.540
with
Token with
Feature activation+3.972
an
Token an
Feature activation+2.524
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
F
TokenF
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
is
Token is
Feature activation+0.000
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
names
Token names
Feature activation+0.000
are
Token are
Feature activation+1.739
confusing
Token confusing
Feature activation+2.555
and
Token and
Feature activation+0.502
you
Token you
Feature activation+0.000
get
Token get
Feature activation+0.000
your
Token your
Feature activation+0.000
Sister
Token Sister
Feature activation+0.000
alien
Token alien
Feature activation+0.000
âĢĵ
TokenâĢĵ
Feature activation+0.000
whose
Tokenwhose
Feature activation+0.000
identity
Token identity
Feature activation+0.000
is
Token is
Feature activation+1.489
kept
Token kept
Feature activation+2.430
anonymous
Token anonymous
Feature activation+1.862
in
Token in
Feature activation+0.927
the
Token the
Feature activation+0.115
indictment
Token indictment
Feature activation+0.000
âĢĵ
TokenâĢĵ
Feature activation+0.000
.
Token.
Feature activation+0.000
Her
Token Her
Feature activation+0.000
identity
Token identity
Feature activation+0.000
was
Token was
Feature activation+2.608
not
Token not
Feature activation+3.235
immediately
Token immediately
Feature activation+2.544
released
Token released
Feature activation+3.304
pending
Token pending
Feature activation+0.523
notification
Token notification
Feature activation+0.000
of
Token of
Feature activation+0.000
her
Token her
Feature activation+0.000

INTERVAL 1.986 - 2.383
CONTAINS 0.004%

iness
Tokeniness
Feature activation+0.000
about
Token about
Feature activation+0.000
having
Token having
Feature activation+0.000
his
Token his
Feature activation+0.000
name
Token name
Feature activation+0.000
and
Token and
Feature activation+2.196
game
Token game
Feature activation+0.239
compared
Token compared
Feature activation+0.000
and
Token and
Feature activation+0.082
contrasted
Token contrasted
Feature activation+0.000
with
Token with
Feature activation+0.000
name
Token name
Feature activation+0.000
of
Token of
Feature activation+1.052
the
Token the
Feature activation+1.269
team
Token team
Feature activation+0.000
was
Token was
Feature activation+1.729
announced
Token announced
Feature activation+2.232
on
Token on
Feature activation+2.043
May
Token May
Feature activation+0.000
11
Token 11
Feature activation+0.000
,
Token,
Feature activation+0.000
2006
Token 2006
Feature activation+0.000
not
Token not
Feature activation+0.000
depend
Token depend
Feature activation+0.000
upon
Token upon
Feature activation+0.000
names
Token names
Feature activation+0.000
drawn
Token drawn
Feature activation+1.653
from
Token from
Feature activation+2.334
a
Token a
Feature activation+1.374
fish
Token fish
Feature activation+0.000
bowl
Tokenbowl
Feature activation+0.000
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
columns
Token columns
Feature activation+0.000
so
Token so
Feature activation+0.000
the
Token the
Feature activation+0.000
names
Token names
Feature activation+0.000
will
Token will
Feature activation+1.786
not
Token not
Feature activation+2.176
surprise
Token surprise
Feature activation+1.473
anyone
Token anyone
Feature activation+0.092
.
Token.
Feature activation+0.198
We
Token We
Feature activation+0.000
will
Token will
Feature activation+0.000
have
Token have
Feature activation+0.000
a
Token a
Feature activation+0.000
certain
Token certain
Feature activation+0.000
name
Token name
Feature activation+0.059
,
Token,
Feature activation+1.440
like
Token like
Feature activation+2.066
a
Token a
Feature activation+1.068
comic
Token comic
Feature activation+0.000
-
Token-
Feature activation+0.000
like
Tokenlike
Feature activation+0.180
look
Token look
Feature activation+0.169

INTERVAL 1.589 - 1.986
CONTAINS 0.003%

âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
âĢĶ
Token âĢĶ
Feature activation+0.000
another
Token another
Feature activation+0.000
name
Token name
Feature activation+0.424
for
Token for
Feature activation+1.771
marijuana
Token marijuana
Feature activation+0.000
âĢĶ
Token âĢĶ
Feature activation+0.000
to
Token to
Feature activation+0.000
promote
Token promote
Feature activation+0.000
their
Token their
Feature activation+0.000
Arizona
Token Arizona
Feature activation+0.000
.
Token.
Feature activation+0.000
His
Token His
Feature activation+0.000
nickname
Token nickname
Feature activation+0.000
came
Token came
Feature activation+1.241
from
Token from
Feature activation+1.790
his
Token his
Feature activation+0.233
tendency
Token tendency
Feature activation+0.000
to
Token to
Feature activation+0.000
"
Token "
Feature activation+0.000
buck
Tokenbuck
Feature activation+0.000
reckon
Token reckon
Feature activation+0.398
will
Token will
Feature activation+2.299
be
Token be
Feature activation+2.292
there
Token there
Feature activation+2.086
is
Token is
Feature activation+0.566
the
Token the
Feature activation+1.776
centre
Token centre
Feature activation+0.296
Ben
Token Ben
Feature activation+0.000
Te
Token Te
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
Franco
Token Franco
Feature activation+0.000
-
Token-
Feature activation+0.000
English
TokenEnglish
Feature activation+0.000
name
Token name
Feature activation+0.000
of
Token of
Feature activation+1.923
the
Token the
Feature activation+1.633
is
Token is
Feature activation+1.458
lets
Tokenlets
Feature activation+0.553
derives
Token derives
Feature activation+2.116
from
Token from
Feature activation+2.483
Le
Token Le
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Add
TokenAdd
Feature activation+0.000
your
Token your
Feature activation+0.000
name
Token name
Feature activation+0.000
to
Token to
Feature activation+1.824
millions
Token millions
Feature activation+0.258
demanding
Token demanding
Feature activation+0.000
that
Token that
Feature activation+0.000
Congress
Token Congress
Feature activation+0.000
take
Token take
Feature activation+0.000

INTERVAL 1.192 - 1.589
CONTAINS 0.006%

politicians
Token politicians
Feature activation+0.000
whose
Token whose
Feature activation+0.000
names
Token names
Feature activation+0.000
barely
Token barely
Feature activation+3.125
register
Token register
Feature activation+3.377
with
Token with
Feature activation+1.473
the
Token the
Feature activation+0.616
general
Token general
Feature activation+0.000
public
Token public
Feature activation+0.000
.
Token.
Feature activation+0.000
Many
Token Many
Feature activation+0.000
,
Token,
Feature activation+0.000
Mormons
Token Mormons
Feature activation+0.000
,
Token,
Feature activation+0.000
you
Token you
Feature activation+0.000
name
Token name
Feature activation+0.000
it
Token it
Feature activation+1.301
âĢĶ
Token âĢĶ
Feature activation+0.032
from
Token from
Feature activation+0.000
soccer
Token soccer
Feature activation+0.000
moms
Token moms
Feature activation+0.000
to
Token to
Feature activation+0.000
The
Token The
Feature activation+0.000
name
Token name
Feature activation+0.000
d
Token d
Feature activation+0.000
iox
Tokeniox
Feature activation+0.000
in
Tokenin
Feature activation+0.000
refers
Token refers
Feature activation+1.557
to
Token to
Feature activation+3.033
a
Token a
Feature activation+0.403
group
Token group
Feature activation+0.000
of
Token of
Feature activation+0.000
highly
Token highly
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Company
TokenCompany
Feature activation+0.000
names
Token names
Feature activation+0.000
[
Token [
Feature activation+0.416
edit
Token edit
Feature activation+0.080
]
Token ]
Feature activation+1.431
Ċ
TokenĊ
Feature activation+1.867
Ċ
TokenĊ
Feature activation+2.116
The
TokenThe
Feature activation+0.141
various
Token various
Feature activation+0.000
names
Token names
Feature activation+0.632
names
Token names
Feature activation+0.000
of
Token of
Feature activation+0.262
previous
Token previous
Feature activation+0.000
developers
Token developers
Feature activation+0.000
to
Token to
Feature activation+1.207
a
Token a
Feature activation+1.332
list
Token list
Feature activation+1.653
of
Token of
Feature activation+0.000
names
Token names
Feature activation+0.867
that
Token that
Feature activation+0.433
can
Token can
Feature activation+1.127

INTERVAL 0.794 - 1.192
CONTAINS 0.010%

wanted
Token wanted
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
a
Token a
Feature activation+0.000
name
Token name
Feature activation+0.000
for
Token for
Feature activation+0.982
himself
Token himself
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
blog
Token blog
Feature activation+0.000
,
Token,
Feature activation+0.000
our
Token our
Feature activation+0.000
favorite
Token favorite
Feature activation+0.000
people
Token people
Feature activation+0.000
on
Token on
Feature activation+1.044
the
Token the
Feature activation+0.686
top
Token top
Feature activation+0.825
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
What
TokenWhat
Feature activation+0.000
to
Token to
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
name
Token name
Feature activation+0.000
,
Token,
Feature activation+0.683
your
Token your
Feature activation+0.818
name
Token name
Feature activation+0.030
,
Token,
Feature activation+0.469
home
Token home
Feature activation+0.699
address
Token address
Feature activation+0.000
and
Token and
Feature activation+0.636
the
Token the
Feature activation+0.000
Common
Token Common
Feature activation+0.000
Name
Token Name
Feature activation+0.000
(
Token (
Feature activation+0.000
CN
TokenCN
Feature activation+0.000
)
Token)
Feature activation+1.012
field
Token field
Feature activation+0.125
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
TLS
Token TLS
Feature activation+0.000
certificate
Token certificate
Feature activation+0.000
also
Token also
Feature activation+0.000
took
Token took
Feature activation+0.000
its
Token its
Feature activation+0.000
name
Token name
Feature activation+0.000
from
Token from
Feature activation+3.181
the
Token the
Feature activation+1.106
work
Token work
Feature activation+0.000
of
Token of
Feature activation+0.000
writer
Token writer
Feature activation+0.000
Ald
Token Ald
Feature activation+0.000
ous
Tokenous
Feature activation+0.000

INTERVAL 0.397 - 0.794
CONTAINS 0.020%

and
Token and
Feature activation+1.223
my
Token my
Feature activation+0.281
band
Token band
Feature activation+0.000
name
Token name
Feature activation+0.092
can
Token can
Feature activation+0.000
give
Token give
Feature activation+0.620
it
Token it
Feature activation+2.243
up
Token up
Feature activation+0.493
.
Token.
Feature activation+0.449
The
Token The
Feature activation+0.000
joke
Token joke
Feature activation+0.000
Sho
Token Sho
Feature activation+0.000
eb
Tokeneb
Feature activation+0.000
at
Tokenat
Feature activation+0.000
sounds
Token sounds
Feature activation+3.464
familiar
Token familiar
Feature activation+2.757
,
Token,
Feature activation+0.675
it
Token it
Feature activation+1.233
's
Token's
Feature activation+0.273
because
Token because
Feature activation+0.371
CNN
Token CNN
Feature activation+0.000
just
Token just
Feature activation+0.000
of
Token of
Feature activation+0.033
the
Token the
Feature activation+0.035
type
Token type
Feature activation+0.000
,
Token,
Feature activation+0.129
or
Token or
Feature activation+0.735
a
Token a
Feature activation+0.681
goal
Token goal
Feature activation+0.000
of
Token of
Feature activation+0.000
that
Token that
Feature activation+0.000
type
Token type
Feature activation+0.000
.
Token.
Feature activation+0.000
.
Token.
Feature activation+0.000
Moreover
Token Moreover
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
name
Token name
Feature activation+0.170
sp
Token sp
Feature activation+0.428
anned
Tokenanned
Feature activation+1.915
the
Token the
Feature activation+0.765
history
Token history
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
s
Tokens
Feature activation+0.000
name
Token name
Feature activation+0.000
,
Token,
Feature activation+0.683
your
Token your
Feature activation+0.818
name
Token name
Feature activation+0.030
,
Token,
Feature activation+0.469
home
Token home
Feature activation+0.699
address
Token address
Feature activation+0.000
and
Token and
Feature activation+0.636
phone
Token phone
Feature activation+0.507
number
Token number
Feature activation+0.000

INTERVAL 0.000 - 0.397
CONTAINS 99.949%

restricted
Token restricted
Feature activation+0.000
to
Token to
Feature activation+0.000
kins
Token kins
Feature activation+0.000
hip
Tokenhip
Feature activation+0.000
terms
Token terms
Feature activation+0.000
.
Token.
Feature activation+0.000
In
Token In
Feature activation+0.000
terms
Token terms
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
sounds
Token sounds
Feature activation+0.000
growth
Token growth
Feature activation+0.000
.
Token.
Feature activation+0.000
It
Token It
Feature activation+0.000
's
Token's
Feature activation+0.000
a
Token a
Feature activation+0.000
process
Token process
Feature activation+0.000
known
Token known
Feature activation+0.000
as
Token as
Feature activation+0.000
d
Token d
Feature activation+0.000
end
Tokenend
Feature activation+0.000
ro
Tokenro
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
It
TokenIt
Feature activation+0.000
added
Token added
Feature activation+0.000
that
Token that
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
't
Token't
Feature activation+0.000
going
Token going
Feature activation+0.000
to
Token to
Feature activation+0.000
attend
Token attend
Feature activation+0.000
because
Token because
Feature activation+0.000
of
Token of
Feature activation+0.000
"
Token "
Feature activation+0.000
the
Tokenthe
Feature activation+0.000
president
Token president
Feature activation+0.000
's
Token's
Feature activation+0.000
continued
Token continued
Feature activation+0.000
doesn
Token doesn
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
matter
Token matter
Feature activation+0.000
that
Token that
Feature activation+0.000
a
Token a
Feature activation+0.000
significant
Token significant
Feature activation+0.000
portion
Token portion
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000

Top feature 9 in H4.3: (feature 15267)

TOP ACTIVATIONS
MAX = 3.191

smart
Token smart
Feature activation+0.146
and
Token and
Feature activation+0.000
cont
Token cont
Feature activation+0.000
oured
Tokenoured
Feature activation+0.000
face
Token face
Feature activation+0.000
can
Token can
Feature activation+3.191
enhance
Token enhance
Feature activation+2.111
your
Token your
Feature activation+0.000
beauty
Token beauty
Feature activation+0.000
.
Token.
Feature activation+0.000
Be
Token Be
Feature activation+0.000
in
Token in
Feature activation+0.000
next
Token next
Feature activation+0.000
year
Token year
Feature activation+0.000
's
Token's
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000
diverse
Token diverse
Feature activation+0.000
group
Token group
Feature activation+0.000
of
Token of
Feature activation+0.000
energy
Token energy
Feature activation+0.000
systems
Token systems
Feature activation+0.000
is
Token is
Feature activation+3.030
key
Token key
Feature activation+2.902
to
Token to
Feature activation+1.148
a
Token a
Feature activation+0.000
society
Token society
Feature activation+0.000
surviving
Token surviving
Feature activation+0.000
transl
Token transl
Feature activation+0.000
ators
Tokenators
Feature activation+0.000
here
Token here
Feature activation+0.000
has
Token has
Feature activation+2.894
been
Token been
Feature activation+2.275
a
Token a
Feature activation+2.934
really
Token really
Feature activation+1.777
big
Token big
Feature activation+2.333
plus
Token plus
Feature activation+0.251
,"
Token,"
Feature activation+0.000
Mes
Token Mes
Feature activation+0.000
group
Token group
Feature activation+0.000
of
Token of
Feature activation+0.000
energy
Token energy
Feature activation+0.000
systems
Token systems
Feature activation+0.000
is
Token is
Feature activation+3.030
key
Token key
Feature activation+2.902
to
Token to
Feature activation+1.148
a
Token a
Feature activation+0.000
society
Token society
Feature activation+0.000
surviving
Token surviving
Feature activation+0.000
changes
Token changes
Feature activation+0.000
having
Token having
Feature activation+0.000
the
Token the
Feature activation+0.000
transl
Token transl
Feature activation+0.000
ators
Tokenators
Feature activation+0.000
here
Token here
Feature activation+0.000
has
Token has
Feature activation+2.894
been
Token been
Feature activation+2.275
a
Token a
Feature activation+2.934
really
Token really
Feature activation+1.777
big
Token big
Feature activation+2.333
plus
Token plus
Feature activation+0.251
and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.000
truly
Token truly
Feature activation+0.000
good
Token good
Feature activation+0.000
teacher
Token teacher
Feature activation+0.512
is
Token is
Feature activation+2.867
much
Token much
Feature activation+2.497
more
Token more
Feature activation+1.875
rare
Token rare
Feature activation+2.272
than
Token than
Feature activation+0.000
you
Token you
Feature activation+0.000
ill
Tokenill
Feature activation+0.000
out
Token out
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
area
Token area
Feature activation+0.000
seems
Token seems
Feature activation+2.825
to
Token to
Feature activation+1.919
be
Token be
Feature activation+0.290
essential
Token essential
Feature activation+1.700
to
Token to
Feature activation+0.484
the
Token the
Feature activation+0.000
of
Token of
Feature activation+0.000
each
Token each
Feature activation+0.000
individual
Token individual
Feature activation+0.000
ride
Token ride
Feature activation+0.000
may
Token may
Feature activation+2.561
encourage
Token encourage
Feature activation+2.792
people
Token people
Feature activation+0.000
to
Token to
Feature activation+0.000
take
Token take
Feature activation+0.000
fewer
Token fewer
Feature activation+0.000
trips
Token trips
Feature activation+0.000
feeling
Token feeling
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.547
feels
Token feels
Feature activation+2.787
like
Token like
Feature activation+2.743
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.547
feels
Token feels
Feature activation+2.787
like
Token like
Feature activation+2.743
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000
,'
Token,'
Feature activation+0.000
she
Token she
Feature activation+0.000
maintains
Token maintains
Feature activation+0.000
,
Token,
Feature activation+0.000
"
Token "
Feature activation+0.000
t
Tokent
Feature activation+0.468
akes
Tokenakes
Feature activation+2.669
issues
Token issues
Feature activation+0.084
out
Token out
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
school
Token school
Feature activation+0.000
showed
Token showed
Feature activation+0.378
that
Token that
Feature activation+0.000
having
Token having
Feature activation+0.000
only
Token only
Feature activation+0.000
depression
Token depression
Feature activation+0.000
was
Token was
Feature activation+2.608
associated
Token associated
Feature activation+1.479
with
Token with
Feature activation+1.642
future
Token future
Feature activation+0.000
dementia
Token dementia
Feature activation+0.000
diagnosis
Token diagnosis
Feature activation+0.000
cost
Token cost
Feature activation+0.000
of
Token of
Feature activation+0.000
each
Token each
Feature activation+0.000
individual
Token individual
Feature activation+0.000
ride
Token ride
Feature activation+0.000
may
Token may
Feature activation+2.561
encourage
Token encourage
Feature activation+2.792
people
Token people
Feature activation+0.000
to
Token to
Feature activation+0.000
take
Token take
Feature activation+0.000
fewer
Token fewer
Feature activation+0.000
control
Token control
Feature activation+0.000
over
Token over
Feature activation+0.000
a
Token a
Feature activation+0.000
crowd
Token crowd
Feature activation+0.000
is
Token is
Feature activation+2.385
stimulating
Token stimulating
Feature activation+2.560
,
Token,
Feature activation+0.000
to
Token to
Feature activation+0.000
say
Token say
Feature activation+0.000
the
Token the
Feature activation+0.000
least
Token least
Feature activation+0.000
a
Token a
Feature activation+0.000
one
Token one
Feature activation+0.000
-
Token-
Feature activation+0.000
sided
Tokensided
Feature activation+0.000
hearing
Token hearing
Feature activation+0.000
undermines
Token undermines
Feature activation+2.515
the
Token the
Feature activation+0.000
credibility
Token credibility
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
committee
Token committee
Feature activation+0.000
their
Token their
Feature activation+0.000
licenses
Token licenses
Feature activation+0.000
suspended
Token suspended
Feature activation+0.000
or
Token or
Feature activation+0.000
revoked
Token revoked
Feature activation+0.000
doesn
Token doesn
Feature activation+2.512
't
Token't
Feature activation+2.260
seem
Token seem
Feature activation+2.129
to
Token to
Feature activation+1.407
be
Token be
Feature activation+1.218
much
Token much
Feature activation+0.731
a
Token a
Feature activation+0.000
truly
Token truly
Feature activation+0.000
good
Token good
Feature activation+0.000
teacher
Token teacher
Feature activation+0.512
is
Token is
Feature activation+2.867
much
Token much
Feature activation+2.497
more
Token more
Feature activation+1.875
rare
Token rare
Feature activation+2.272
than
Token than
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
next
Token next
Feature activation+0.000
year
Token year
Feature activation+0.000
's
Token's
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000
more
Token more
Feature activation+0.000
of
Token of
Feature activation+0.000
control
Token control
Feature activation+0.000
over
Token over
Feature activation+0.000
a
Token a
Feature activation+0.000
crowd
Token crowd
Feature activation+0.000
is
Token is
Feature activation+2.385
stimulating
Token stimulating
Feature activation+2.560
,
Token,
Feature activation+0.000
to
Token to
Feature activation+0.000
say
Token say
Feature activation+0.000
the
Token the
Feature activation+0.000

Top DFA by src position
MAX = 4.039

Energy
Token Energy
Feature activation-0.009
Top resid features:
last
Token last
Feature activation+0.000
Top resid features:
year
Token year
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.005
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.074
Top resid features:
Having
TokenHaving
Feature activation+4.039
Top resid features:
a
Token a
Feature activation+0.213
Top resid features:
smart
Token smart
Feature activation+0.153
Top resid features:
and
Token and
Feature activation-0.008
Top resid features:
cont
Token cont
Feature activation+0.016
Top resid features:
oured
Tokenoured
Feature activation+0.116
Top resid features:
are
Token are
Feature activation-0.005
Top resid features:
believed
Token believed
Feature activation-0.010
Top resid features:
possible
Token possible
Feature activation-0.009
Top resid features:
,
Token,
Feature activation+0.040
Top resid features:
and
Token and
Feature activation-0.150
Top resid features:
having
Token having
Feature activation+3.476
Top resid features:
two
Token two
Feature activation+0.165
Top resid features:
of
Token of
Feature activation+0.133
Top resid features:
the
Token the
Feature activation+0.191
Top resid features:
top
Token top
Feature activation+0.011
Top resid features:
11
Token 11
Feature activation+0.027
Top resid features:
true
Token true
Feature activation-0.008
Top resid features:
for
Token for
Feature activation-0.003
Top resid features:
electricity
Token electricity
Feature activation-0.037
Top resid features:
production
Token production
Feature activation-0.022
Top resid features:
.
Token.
Feature activation-0.034
Top resid features:
Having
Token Having
Feature activation+3.825
Top resid features:
a
Token a
Feature activation+0.153
Top resid features:
diverse
Token diverse
Feature activation+0.121
Top resid features:
group
Token group
Feature activation+0.066
Top resid features:
of
Token of
Feature activation+0.254
Top resid features:
energy
Token energy
Feature activation-0.034
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.029
Top resid features:
.
Token.
Feature activation+0.013
Top resid features:
So
Token So
Feature activation+0.029
Top resid features:
having
Token having
Feature activation+3.529
Top resid features:
the
Token the
Feature activation+0.102
Top resid features:
transl
Token transl
Feature activation+0.058
Top resid features:
ators
Tokenators
Feature activation+0.181
Top resid features:
here
Token here
Feature activation+0.097
Top resid features:
has
Token has
Feature activation+0.069
Top resid features:
true
Token true
Feature activation-0.016
Top resid features:
for
Token for
Feature activation-0.003
Top resid features:
electricity
Token electricity
Feature activation-0.017
Top resid features:
production
Token production
Feature activation-0.003
Top resid features:
.
Token.
Feature activation-0.094
Top resid features:
Having
Token Having
Feature activation+3.836
Top resid features:
a
Token a
Feature activation+0.075
Top resid features:
diverse
Token diverse
Feature activation+0.267
Top resid features:
group
Token group
Feature activation+0.096
Top resid features:
of
Token of
Feature activation+0.145
Top resid features:
energy
Token energy
Feature activation-0.012
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation+0.031
Top resid features:
.
Token.
Feature activation+0.035
Top resid features:
So
Token So
Feature activation+0.027
Top resid features:
having
Token having
Feature activation+3.727
Top resid features:
the
Token the
Feature activation+0.126
Top resid features:
transl
Token transl
Feature activation+0.049
Top resid features:
ators
Tokenators
Feature activation+0.261
Top resid features:
here
Token here
Feature activation+0.060
Top resid features:
has
Token has
Feature activation+0.012
Top resid features:
versus
Token versus
Feature activation+0.016
Top resid features:
a
Token a
Feature activation-0.010
Top resid features:
world
Token world
Feature activation-0.002
Top resid features:
champion
Token champion
Feature activation+0.005
Top resid features:
.
Token.
Feature activation+0.026
Top resid features:
Having
Token Having
Feature activation+3.630
Top resid features:
an
Token an
Feature activation+0.083
Top resid features:
instructor
Token instructor
Feature activation+0.053
Top resid features:
that
Token that
Feature activation+0.097
Top resid features:
is
Token is
Feature activation+0.010
Top resid features:
a
Token a
Feature activation-0.034
Top resid features:
seen
Token seen
Feature activation-0.031
Top resid features:
a
Token a
Feature activation-0.030
Top resid features:
few
Token few
Feature activation+0.003
Top resid features:
places
Token places
Feature activation-0.030
Top resid features:
.
Token.
Feature activation+0.017
Top resid features:
Having
Token Having
Feature activation+3.633
Top resid features:
a
Token a
Feature activation+0.130
Top resid features:
âĢ
Token âĢ
Feature activation+0.053
Top resid features:
ľ
Tokenľ
Feature activation+0.105
Top resid features:
ch
Tokench
Feature activation-0.003
Top resid features:
ill
Tokenill
Feature activation+0.084
Top resid features:
/
Token/
Feature activation-0.015
Top resid features:
Getty
TokenGetty
Feature activation-0.007
Top resid features:
Images
Token Images
Feature activation-0.007
Top resid features:
Ċ
TokenĊ
Feature activation-0.007
Top resid features:
Ċ
TokenĊ
Feature activation-0.003
Top resid features:
Having
TokenHaving
Feature activation+3.848
Top resid features:
to
Token to
Feature activation+0.048
Top resid features:
weigh
Token weigh
Feature activation+0.064
Top resid features:
the
Token the
Feature activation+0.005
Top resid features:
cost
Token cost
Feature activation+0.047
Top resid features:
of
Token of
Feature activation+0.026
Top resid features:
we
Token we
Feature activation+0.033
Top resid features:
're
Token're
Feature activation+0.013
Top resid features:
feeling
Token feeling
Feature activation+0.033
Top resid features:
what
Token what
Feature activation+0.093
Top resid features:
not
Token not
Feature activation+0.130
Top resid features:
having
Token having
Feature activation+3.601
Top resid features:
hope
Token hope
Feature activation+0.146
Top resid features:
feels
Token feels
Feature activation+0.166
Top resid features:
like
Token like
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Hope
Token Hope
Feature activation+0.000
Top resid features:
we
Token we
Feature activation+0.001
Top resid features:
're
Token're
Feature activation+0.006
Top resid features:
feeling
Token feeling
Feature activation+0.047
Top resid features:
what
Token what
Feature activation+0.034
Top resid features:
not
Token not
Feature activation+0.374
Top resid features:
having
Token having
Feature activation+3.557
Top resid features:
hope
Token hope
Feature activation-0.054
Top resid features:
feels
Token feels
Feature activation+0.278
Top resid features:
like
Token like
Feature activation-0.094
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Hope
Token Hope
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
principal
Token principal
Feature activation+0.003
Top resid features:
's
Token's
Feature activation+0.000
Top resid features:
office
Token office
Feature activation+0.006
Top resid features:
."
Token."
Feature activation+0.027
Top resid features:
Having
Token Having
Feature activation+3.189
Top resid features:
police
Token police
Feature activation+0.033
Top resid features:
in
Token in
Feature activation+0.014
Top resid features:
schools
Token schools
Feature activation-0.037
Top resid features:
,
Token,
Feature activation+0.121
Top resid features:
she
Token she
Feature activation+0.074
Top resid features:
depression
Token depression
Feature activation-0.009
Top resid features:
and
Token and
Feature activation-0.039
Top resid features:
anxiety
Token anxiety
Feature activation-0.007
Top resid features:
showed
Token showed
Feature activation+0.009
Top resid features:
that
Token that
Feature activation-0.080
Top resid features:
having
Token having
Feature activation+3.210
Top resid features:
only
Token only
Feature activation+0.409
Top resid features:
depression
Token depression
Feature activation+0.160
Top resid features:
was
Token was
Feature activation+0.269
Top resid features:
associated
Token associated
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
/
Token/
Feature activation-0.044
Top resid features:
Getty
TokenGetty
Feature activation-0.011
Top resid features:
Images
Token Images
Feature activation-0.011
Top resid features:
Ċ
TokenĊ
Feature activation-0.005
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Having
TokenHaving
Feature activation+3.733
Top resid features:
to
Token to
Feature activation+0.069
Top resid features:
weigh
Token weigh
Feature activation+0.051
Top resid features:
the
Token the
Feature activation+0.037
Top resid features:
cost
Token cost
Feature activation+0.042
Top resid features:
of
Token of
Feature activation+0.036
Top resid features:
dance
Token dance
Feature activation+0.002
Top resid features:
or
Token or
Feature activation-0.014
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
watch
Token watch
Feature activation-0.003
Top resid features:
.
Token.
Feature activation+0.023
Top resid features:
Having
Token Having
Feature activation+3.589
Top resid features:
that
Token that
Feature activation+0.225
Top resid features:
kind
Token kind
Feature activation+0.029
Top resid features:
of
Token of
Feature activation+0.035
Top resid features:
control
Token control
Feature activation+0.033
Top resid features:
over
Token over
Feature activation+0.029
Top resid features:
to
Token to
Feature activation+0.006
Top resid features:
understand
Token understand
Feature activation-0.012
Top resid features:
what
Token what
Feature activation-0.019
Top resid features:
happened
Token happened
Feature activation-0.004
Top resid features:
.
Token.
Feature activation-0.021
Top resid features:
Having
Token Having
Feature activation+3.482
Top resid features:
such
Token such
Feature activation+0.093
Top resid features:
a
Token a
Feature activation+0.027
Top resid features:
one
Token one
Feature activation+0.031
Top resid features:
-
Token-
Feature activation+0.026
Top resid features:
sided
Tokensided
Feature activation+0.090
Top resid features:
.
Token.
Feature activation+0.026
Top resid features:
For
TokenFor
Feature activation+0.009
Top resid features:
some
Token some
Feature activation-0.050
Top resid features:
drivers
Token drivers
Feature activation-0.070
Top resid features:
,
Token,
Feature activation+0.010
Top resid features:
having
Token having
Feature activation+3.485
Top resid features:
their
Token their
Feature activation+0.180
Top resid features:
licenses
Token licenses
Feature activation+0.063
Top resid features:
suspended
Token suspended
Feature activation+0.099
Top resid features:
or
Token or
Feature activation-0.120
Top resid features:
revoked
Token revoked
Feature activation+0.366
Top resid features:
versus
Token versus
Feature activation+0.013
Top resid features:
a
Token a
Feature activation-0.008
Top resid features:
world
Token world
Feature activation-0.002
Top resid features:
champion
Token champion
Feature activation+0.001
Top resid features:
.
Token.
Feature activation+0.010
Top resid features:
Having
Token Having
Feature activation+3.466
Top resid features:
an
Token an
Feature activation+0.049
Top resid features:
instructor
Token instructor
Feature activation+0.011
Top resid features:
that
Token that
Feature activation+0.081
Top resid features:
is
Token is
Feature activation+0.002
Top resid features:
a
Token a
Feature activation-0.012
Top resid features:
are
Token are
Feature activation+0.000
Top resid features:
believed
Token believed
Feature activation-0.017
Top resid features:
possible
Token possible
Feature activation-0.009
Top resid features:
,
Token,
Feature activation+0.036
Top resid features:
and
Token and
Feature activation-0.105
Top resid features:
having
Token having
Feature activation+3.428
Top resid features:
two
Token two
Feature activation+0.082
Top resid features:
of
Token of
Feature activation+0.082
Top resid features:
the
Token the
Feature activation+0.076
Top resid features:
top
Token top
Feature activation+0.002
Top resid features:
11
Token 11
Feature activation+0.016
Top resid features:
dance
Token dance
Feature activation+0.003
Top resid features:
or
Token or
Feature activation-0.018
Top resid features:
to
Token to
Feature activation+0.001
Top resid features:
watch
Token watch
Feature activation-0.003
Top resid features:
.
Token.
Feature activation+0.015
Top resid features:
Having
Token Having
Feature activation+3.514
Top resid features:
that
Token that
Feature activation+0.276
Top resid features:
kind
Token kind
Feature activation+0.049
Top resid features:
of
Token of
Feature activation+0.085
Top resid features:
control
Token control
Feature activation+0.013
Top resid features:
over
Token over
Feature activation+0.003
Top resid features:

Decoder Weights Distribution

Head 0: 0.04

Head 1: 0.04

Head 2: 0.03

Head 3: 0.48

Head 4: 0.04

Head 5: 0.04

Head 6: 0.09

Head 7: 0.08

Head 8: 0.04

Head 9: 0.04

Head 10: 0.02

Head 11: 0.06

Positive logits

quickShipAvailable1.61

hind1.41

affordability1.39

buyers1.39

newcom1.38

gain1.37

aback1.25

persu1.25

disadvantages1.25

downgrade1.25

renters1.24

oggle1.24

shoppers1.24

enormously1.23

greatly1.22

hog1.22

costing1.21

��1.20

hugely1.20

VERTISEMENT1.20

Negative logits

Archdemon-1.60

Tale-1.60

mol-1.55

chlor-1.54

lication-1.50

ques-1.43

-1.42

cies-1.41

](-1.39

jan-1.38

ologue-1.36

qt-1.35

Iter-1.35

ç-1.34

java-1.33

assi-1.30

ovi-1.30

Qiao-1.30

perse-1.29

atomic-1.28

INTERVAL 2.872 - 3.191
CONTAINS 0.001%

group
Token group
Feature activation+0.000
of
Token of
Feature activation+0.000
energy
Token energy
Feature activation+0.000
systems
Token systems
Feature activation+0.000
is
Token is
Feature activation+3.030
key
Token key
Feature activation+2.902
to
Token to
Feature activation+1.148
a
Token a
Feature activation+0.000
society
Token society
Feature activation+0.000
surviving
Token surviving
Feature activation+0.000
changes
Token changes
Feature activation+0.000
transl
Token transl
Feature activation+0.000
ators
Tokenators
Feature activation+0.000
here
Token here
Feature activation+0.000
has
Token has
Feature activation+2.894
been
Token been
Feature activation+2.275
a
Token a
Feature activation+2.934
really
Token really
Feature activation+1.777
big
Token big
Feature activation+2.333
plus
Token plus
Feature activation+0.251
,"
Token,"
Feature activation+0.000
Mes
Token Mes
Feature activation+0.000
smart
Token smart
Feature activation+0.146
and
Token and
Feature activation+0.000
cont
Token cont
Feature activation+0.000
oured
Tokenoured
Feature activation+0.000
face
Token face
Feature activation+0.000
can
Token can
Feature activation+3.191
enhance
Token enhance
Feature activation+2.111
your
Token your
Feature activation+0.000
beauty
Token beauty
Feature activation+0.000
.
Token.
Feature activation+0.000
Be
Token Be
Feature activation+0.000
diverse
Token diverse
Feature activation+0.000
group
Token group
Feature activation+0.000
of
Token of
Feature activation+0.000
energy
Token energy
Feature activation+0.000
systems
Token systems
Feature activation+0.000
is
Token is
Feature activation+3.030
key
Token key
Feature activation+2.902
to
Token to
Feature activation+1.148
a
Token a
Feature activation+0.000
society
Token society
Feature activation+0.000
surviving
Token surviving
Feature activation+0.000
in
Token in
Feature activation+0.000
next
Token next
Feature activation+0.000
year
Token year
Feature activation+0.000
's
Token's
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000

INTERVAL 2.553 - 2.872
CONTAINS 0.001%

showed
Token showed
Feature activation+0.378
that
Token that
Feature activation+0.000
having
Token having
Feature activation+0.000
only
Token only
Feature activation+0.000
depression
Token depression
Feature activation+0.000
was
Token was
Feature activation+2.608
associated
Token associated
Feature activation+1.479
with
Token with
Feature activation+1.642
future
Token future
Feature activation+0.000
dementia
Token dementia
Feature activation+0.000
diagnosis
Token diagnosis
Feature activation+0.000
and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.000
truly
Token truly
Feature activation+0.000
good
Token good
Feature activation+0.000
teacher
Token teacher
Feature activation+0.512
is
Token is
Feature activation+2.867
much
Token much
Feature activation+2.497
more
Token more
Feature activation+1.875
rare
Token rare
Feature activation+2.272
than
Token than
Feature activation+0.000
you
Token you
Feature activation+0.000
ill
Tokenill
Feature activation+0.000
out
Token out
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
area
Token area
Feature activation+0.000
seems
Token seems
Feature activation+2.825
to
Token to
Feature activation+1.919
be
Token be
Feature activation+0.290
essential
Token essential
Feature activation+1.700
to
Token to
Feature activation+0.484
the
Token the
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.547
feels
Token feels
Feature activation+2.787
like
Token like
Feature activation+2.743
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000
,'
Token,'
Feature activation+0.000
feeling
Token feeling
Feature activation+0.000
what
Token what
Feature activation+0.000
not
Token not
Feature activation+0.000
having
Token having
Feature activation+0.000
hope
Token hope
Feature activation+0.547
feels
Token feels
Feature activation+2.787
like
Token like
Feature activation+2.743
.
Token.
Feature activation+0.000
Hope
Token Hope
Feature activation+0.000
is
Token is
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000

INTERVAL 2.234 - 2.553
CONTAINS 0.001%

drivers
Token drivers
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
customer
Token customer
Feature activation+0.000
experience
Token experience
Feature activation+0.000
relies
Token relies
Feature activation+2.248
on
Token on
Feature activation+2.240
collecting
Token collecting
Feature activation+0.000
enormous
Token enormous
Feature activation+0.000
amounts
Token amounts
Feature activation+0.000
of
Token of
Feature activation+0.000
here
Token here
Feature activation+0.000
has
Token has
Feature activation+2.894
been
Token been
Feature activation+2.275
a
Token a
Feature activation+2.934
really
Token really
Feature activation+1.777
big
Token big
Feature activation+2.333
plus
Token plus
Feature activation+0.251
,"
Token,"
Feature activation+0.000
Mes
Token Mes
Feature activation+0.000
Am
Token Am
Feature activation+0.000
is
Tokenis
Feature activation+0.000
a
Token a
Feature activation+0.000
one
Token one
Feature activation+0.000
-
Token-
Feature activation+0.000
sided
Tokensided
Feature activation+0.000
hearing
Token hearing
Feature activation+0.000
undermines
Token undermines
Feature activation+2.515
the
Token the
Feature activation+0.000
credibility
Token credibility
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
committee
Token committee
Feature activation+0.000
their
Token their
Feature activation+0.000
licenses
Token licenses
Feature activation+0.000
suspended
Token suspended
Feature activation+0.000
or
Token or
Feature activation+0.000
revoked
Token revoked
Feature activation+0.000
doesn
Token doesn
Feature activation+2.512
't
Token't
Feature activation+2.260
seem
Token seem
Feature activation+2.129
to
Token to
Feature activation+1.407
be
Token be
Feature activation+1.218
much
Token much
Feature activation+0.731
next
Token next
Feature activation+0.000
year
Token year
Feature activation+0.000
's
Token's
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000
more
Token more
Feature activation+0.000

INTERVAL 1.915 - 2.234
CONTAINS 0.001%

and
Token and
Feature activation+0.000
cont
Token cont
Feature activation+0.000
oured
Tokenoured
Feature activation+0.000
face
Token face
Feature activation+0.000
can
Token can
Feature activation+3.191
enhance
Token enhance
Feature activation+2.111
your
Token your
Feature activation+0.000
beauty
Token beauty
Feature activation+0.000
.
Token.
Feature activation+0.000
Be
Token Be
Feature activation+0.000
it
Token it
Feature activation+0.000
out
Token out
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
area
Token area
Feature activation+0.000
seems
Token seems
Feature activation+2.825
to
Token to
Feature activation+1.919
be
Token be
Feature activation+0.290
essential
Token essential
Feature activation+1.700
to
Token to
Feature activation+0.484
the
Token the
Feature activation+0.000
community
Token community
Feature activation+0.000
I
Token I
Feature activation+0.000
can
Token can
Feature activation+0.000
Â
TokenÂ
Feature activation+0.000
Ĵ
TokenĴ
Feature activation+0.000
t
Tokent
Feature activation+0.118
help
Token help
Feature activation+2.067
but
Token but
Feature activation+0.268
think
Token think
Feature activation+0.000
this
Token this
Feature activation+0.000
was
Token was
Feature activation+0.000
politically
Token politically
Feature activation+0.000
suspended
Token suspended
Feature activation+0.000
or
Token or
Feature activation+0.000
revoked
Token revoked
Feature activation+0.000
doesn
Token doesn
Feature activation+2.512
't
Token't
Feature activation+2.260
seem
Token seem
Feature activation+2.129
to
Token to
Feature activation+1.407
be
Token be
Feature activation+1.218
much
Token much
Feature activation+0.731
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
films
Token films
Feature activation+0.000
and
Token and
Feature activation+0.000
television
Token television
Feature activation+0.000
would
Token would
Feature activation+1.869
be
Token be
Feature activation+1.668
a
Token a
Feature activation+2.190
true
Token true
Feature activation+0.547
comic
Token comic
Feature activation+0.000
book
Token book
Feature activation+0.000
movie
Token movie
Feature activation+0.000
event
Token event
Feature activation+0.000

INTERVAL 1.596 - 1.915
CONTAINS 0.001%

ators
Tokenators
Feature activation+0.000
here
Token here
Feature activation+0.000
has
Token has
Feature activation+2.894
been
Token been
Feature activation+2.275
a
Token a
Feature activation+2.934
really
Token really
Feature activation+1.777
big
Token big
Feature activation+2.333
plus
Token plus
Feature activation+0.251
,"
Token,"
Feature activation+0.000
Mes
Token Mes
Feature activation+0.000
Am
Token Am
Feature activation+0.000
the
Token the
Feature activation+0.000
films
Token films
Feature activation+0.000
and
Token and
Feature activation+0.000
television
Token television
Feature activation+0.000
would
Token would
Feature activation+1.869
be
Token be
Feature activation+1.668
a
Token a
Feature activation+2.190
true
Token true
Feature activation+0.547
comic
Token comic
Feature activation+0.000
book
Token book
Feature activation+0.000
movie
Token movie
Feature activation+0.000
truly
Token truly
Feature activation+0.000
good
Token good
Feature activation+0.000
teacher
Token teacher
Feature activation+0.512
is
Token is
Feature activation+2.867
much
Token much
Feature activation+2.497
more
Token more
Feature activation+1.875
rare
Token rare
Feature activation+2.272
than
Token than
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
.
Token.
Feature activation+0.000
having
Token having
Feature activation+0.000
only
Token only
Feature activation+0.000
depression
Token depression
Feature activation+0.000
was
Token was
Feature activation+2.608
associated
Token associated
Feature activation+1.479
with
Token with
Feature activation+1.642
future
Token future
Feature activation+0.000
dementia
Token dementia
Feature activation+0.000
diagnosis
Token diagnosis
Feature activation+0.000
but
Token but
Feature activation+0.000
a
Token a
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
area
Token area
Feature activation+0.000
seems
Token seems
Feature activation+2.825
to
Token to
Feature activation+1.919
be
Token be
Feature activation+0.290
essential
Token essential
Feature activation+1.700
to
Token to
Feature activation+0.484
the
Token the
Feature activation+0.000
community
Token community
Feature activation+0.000
building
Token building
Feature activation+0.000
that
Token that
Feature activation+0.000

INTERVAL 1.276 - 1.596
CONTAINS 0.001%

of
Token of
Feature activation+0.000
encouraging
Token encouraging
Feature activation+0.000
women
Token women
Feature activation+0.000
to
Token to
Feature activation+0.000
ride
Token ride
Feature activation+0.000
should
Token should
Feature activation+1.381
only
Token only
Feature activation+1.565
make
Token make
Feature activation+1.332
these
Token these
Feature activation+0.000
policies
Token policies
Feature activation+0.000
more
Token more
Feature activation+0.000
encouraging
Token encouraging
Feature activation+0.000
women
Token women
Feature activation+0.000
to
Token to
Feature activation+0.000
ride
Token ride
Feature activation+0.000
should
Token should
Feature activation+1.381
only
Token only
Feature activation+1.565
make
Token make
Feature activation+1.332
these
Token these
Feature activation+0.000
policies
Token policies
Feature activation+0.000
more
Token more
Feature activation+0.000
popular
Token popular
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000
more
Token more
Feature activation+0.000
talent
Token talent
Feature activation+0.000
.
Token.
Feature activation+0.000
However
Token However
Feature activation+0.000
own
Token own
Feature activation+0.000
children
Token children
Feature activation+0.000
for
Token for
Feature activation+0.000
six
Token six
Feature activation+0.000
months
Token months
Feature activation+0.000
is
Token is
Feature activation+1.304
a
Token a
Feature activation+0.663
choice
Token choice
Feature activation+0.000
I
Token I
Feature activation+0.000
would
Token would
Feature activation+0.000
personally
Token personally
Feature activation+0.000
women
Token women
Feature activation+0.000
to
Token to
Feature activation+0.000
ride
Token ride
Feature activation+0.000
should
Token should
Feature activation+1.381
only
Token only
Feature activation+1.565
make
Token make
Feature activation+1.332
these
Token these
Feature activation+0.000
policies
Token policies
Feature activation+0.000
more
Token more
Feature activation+0.000
popular
Token popular
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 0.957 - 1.276
CONTAINS 0.000%

year
Token year
Feature activation+0.000
's
Token's
Feature activation+0.000
draft
Token draft
Feature activation+0.000
is
Token is
Feature activation+3.033
also
Token also
Feature activation+2.492
going
Token going
Feature activation+0.984
to
Token to
Feature activation+2.364
provide
Token provide
Feature activation+1.385
even
Token even
Feature activation+0.000
more
Token more
Feature activation+0.000
talent
Token talent
Feature activation+0.000
revoked
Token revoked
Feature activation+0.000
doesn
Token doesn
Feature activation+2.512
't
Token't
Feature activation+2.260
seem
Token seem
Feature activation+2.129
to
Token to
Feature activation+1.407
be
Token be
Feature activation+1.218
much
Token much
Feature activation+0.731
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
deterrent
Token deterrent
Feature activation+0.301
,
Token,
Feature activation+0.000
preventing
Token preventing
Feature activation+0.000
another
Token another
Feature activation+0.000
loss
Token loss
Feature activation+0.000
of
Token of
Feature activation+0.000
life
Token life
Feature activation+0.000
is
Token is
Feature activation+1.124
the
Token the
Feature activation+0.620
reason
Token reason
Feature activation+0.062
I
Token I
Feature activation+0.000
get
Token get
Feature activation+0.000
up
Token up
Feature activation+0.000
of
Token of
Feature activation+0.000
energy
Token energy
Feature activation+0.000
systems
Token systems
Feature activation+0.000
is
Token is
Feature activation+3.030
key
Token key
Feature activation+2.902
to
Token to
Feature activation+1.148
a
Token a
Feature activation+0.000
society
Token society
Feature activation+0.000
surviving
Token surviving
Feature activation+0.000
changes
Token changes
Feature activation+0.000
in
Token in
Feature activation+0.000

INTERVAL 0.638 - 0.957
CONTAINS 0.001%

from
Token from
Feature activation+0.000
easy
Token easy
Feature activation+0.000
marks
Token marks
Feature activation+0.000
would
Token would
Feature activation+0.392
have
Token have
Feature activation+0.387
been
Token been
Feature activation+0.720
positively
Token positively
Feature activation+0.453
un
Token un
Feature activation+0.000
-
Token-
Feature activation+0.000
American
TokenAmerican
Feature activation+0.000
.
Token.
Feature activation+0.000
headset
Token headset
Feature activation+0.000
in
Token in
Feature activation+0.000
public
Token public
Feature activation+0.000
may
Token may
Feature activation+0.265
be
Token be
Feature activation+0.740
like
Token like
Feature activation+0.657
seeing
Token seeing
Feature activation+0.000
someone
Token someone
Feature activation+0.000
taking
Token taking
Feature activation+0.000
a
Token a
Feature activation+0.000
picture
Token picture
Feature activation+0.000
doesn
Token doesn
Feature activation+2.512
't
Token't
Feature activation+2.260
seem
Token seem
Feature activation+2.129
to
Token to
Feature activation+1.407
be
Token be
Feature activation+1.218
much
Token much
Feature activation+0.731
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
deterrent
Token deterrent
Feature activation+0.301
,
Token,
Feature activation+0.000
said
Token said
Feature activation+0.000
children
Token children
Feature activation+0.000
for
Token for
Feature activation+0.000
six
Token six
Feature activation+0.000
months
Token months
Feature activation+0.000
is
Token is
Feature activation+1.304
a
Token a
Feature activation+0.663
choice
Token choice
Feature activation+0.000
I
Token I
Feature activation+0.000
would
Token would
Feature activation+0.000
personally
Token personally
Feature activation+0.000
never
Token never
Feature activation+0.000
a
Token a
Feature activation+0.000
headset
Token headset
Feature activation+0.000
in
Token in
Feature activation+0.000
public
Token public
Feature activation+0.000
may
Token may
Feature activation+0.265
be
Token be
Feature activation+0.740
like
Token like
Feature activation+0.657
seeing
Token seeing
Feature activation+0.000
someone
Token someone
Feature activation+0.000
taking
Token taking
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 0.319 - 0.638
CONTAINS 0.004%

along
Token along
Feature activation+0.000
to
Token to
Feature activation+0.000
a
Token a
Feature activation+0.000
gun
Token gun
Feature activation+0.000
fight
Tokenfight
Feature activation+0.000
is
Token is
Feature activation+0.325
never
Token never
Feature activation+0.000
going
Token going
Feature activation+0.000
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
a
Token a
Feature activation+0.000
someone
Token someone
Feature activation+0.000
living
Token living
Feature activation+0.000
nearby
Token nearby
Feature activation+0.000
âĢĶ
Token âĢĶ
Feature activation+0.000
could
Token could
Feature activation+0.072
breed
Token breed
Feature activation+0.328
a
Token a
Feature activation+0.000
sense
Token sense
Feature activation+0.000
of
Token of
Feature activation+0.000
une
Token une
Feature activation+0.000
ase
Tokenase
Feature activation+0.000
,
Token,
Feature activation+0.000
she
Token she
Feature activation+0.000
maintains
Token maintains
Feature activation+0.000
,
Token,
Feature activation+0.000
"
Token "
Feature activation+0.000
t
Tokent
Feature activation+0.468
akes
Tokenakes
Feature activation+2.669
issues
Token issues
Feature activation+0.084
out
Token out
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
help
Token help
Feature activation+0.000
the
Token the
Feature activation+0.000
have
Token have
Feature activation+0.000
-
Token-
Feature activation+0.000
not
Tokennot
Feature activation+0.000
s
Tokens
Feature activation+0.368
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
De
Token De
Feature activation+0.000
aling
Tokenaling
Feature activation+0.000
with
Token with
Feature activation+0.000
her
Token her
Feature activation+0.000
will
Token will
Feature activation+0.000
involve
Token involve
Feature activation+0.592
keeping
Token keeping
Feature activation+0.000
your
Token your
Feature activation+0.000
fire
Token fire
Feature activation+0.000
team
Tokenteam
Feature activation+0.000
personnel
Token personnel
Feature activation+0.000

INTERVAL 0.000 - 0.319
CONTAINS 99.990%

ando
Tokenando
Feature activation+0.000
tut
Token tut
Feature activation+0.000
to
Tokento
Feature activation+0.000
.
Token.
Feature activation+0.000
E
Token E
Feature activation+0.000
inf
Token inf
Feature activation+0.000
ine
Tokenine
Feature activation+0.000
se
Token se
Feature activation+0.000
ne
Token ne
Feature activation+0.000
v
Token v
Feature activation+0.000
ann
Tokenann
Feature activation+0.000
Regardless
TokenRegardless
Feature activation+0.000
of
Token of
Feature activation+0.000
mis
Token mis
Feature activation+0.000
f
Tokenf
Feature activation+0.000
ortun
Tokenortun
Feature activation+0.000
es
Tokenes
Feature activation+0.000
and
Token and
Feature activation+0.000
temporary
Token temporary
Feature activation+0.000
setbacks
Token setbacks
Feature activation+0.000
that
Token that
Feature activation+0.000
are
Token are
Feature activation+0.000
to
Token to
Feature activation+0.000
us
Token us
Feature activation+0.000
about
Token about
Feature activation+0.000
this
Token this
Feature activation+0.000
request
Token request
Feature activation+0.000
,
Token,
Feature activation+0.000
we
Token we
Feature activation+0.000
have
Token have
Feature activation+0.000
decided
Token decided
Feature activation+0.000
to
Token to
Feature activation+0.000
deny
Token deny
Feature activation+0.000
ered
Tokenered
Feature activation+0.000
with
Token with
Feature activation+0.000
acet
Token acet
Feature activation+0.000
yl
Tokenyl
Feature activation+0.000
sal
Tokensal
Feature activation+0.000
icy
Tokenicy
Feature activation+0.000
lic
Tokenlic
Feature activation+0.000
acid
Token acid
Feature activation+0.000
on
Token on
Feature activation+0.000
lipid
Token lipid
Feature activation+0.000
per
Token per
Feature activation+0.000
,
Token,
Feature activation+0.000
fact
Token fact
Feature activation+0.000
-
Token-
Feature activation+0.000
checking
Tokenchecking
Feature activation+0.000
procedures
Token procedures
Feature activation+0.000
.
Token.
Feature activation+0.000
Death
Token Death
Feature activation+0.000
certificates
Token certificates
Feature activation+0.000
also
Token also
Feature activation+0.000
contain
Token contain
Feature activation+0.000
the
Token the
Feature activation+0.000

BOTTOM ACTIVATIONS
MIN = 0.000

atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
Development
Token Development
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
&
Token &
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Defence
Token Defence
Feature activation+0.000
Research
Token Research
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
rik
Tokenrik
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
ar
Tokenar
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
said
Token said
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
atre
Tokenatre
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
former
Token former
Feature activation+0.000
boss
Token boss
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
K
TokenK
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
defence
Token defence
Feature activation+0.000
ministry
Token ministry
Feature activation+0.000
committee
Token committee
Feature activation+0.000
headed
Token headed
Feature activation+0.000
by
Token by
Feature activation+0.000
V
Token V
Feature activation+0.000
.
Token.
Feature activation+0.000