H3.8 top features

Top feature 0 in H3.8: (feature 4319)

TOP ACTIVATIONS
MAX = 2.499

is
Token is
Feature activation+0.000
in
Token in
Feature activation+0.000
effect
Token effect
Feature activation+0.000
the
Token the
Feature activation+1.348
Bitcoin
Token Bitcoin
Feature activation+0.362
story
Token story
Feature activation+2.499
,
Token,
Feature activation+0.194
the
Token the
Feature activation+0.000
net
Token net
Feature activation+0.000
result
Token result
Feature activation+0.000
is
Token is
Feature activation+0.000
st
Token st
Feature activation+1.579
inging
Tokeninging
Feature activation+2.428
ev
Token ev
Feature activation+0.001
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+2.451
of
Token of
Feature activation+0.396
Dr
Token Dr
Feature activation+0.000
.
Token.
Feature activation+0.000
Mann
Token Mann
Feature activation+0.000
on
Token on
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
st
Token st
Feature activation+1.579
inging
Tokeninging
Feature activation+2.428
ev
Token ev
Feature activation+0.001
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+2.451
of
Token of
Feature activation+0.396
NATO
Token NATO
Feature activation+0.000
position
Token position
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
declared
Token declared
Feature activation+2.272
,
Token,
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
Roman
Token Roman
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
wore
Token wore
Feature activation+2.224
a
Token a
Feature activation+0.119
great
Token great
Feature activation+0.000
number
Token number
Feature activation+0.000
of
Token of
Feature activation+0.000
clothes
Token clothes
Feature activation+0.037
ached
Tokenached
Feature activation+0.000
.
Token.
Feature activation+0.000
In
Token In
Feature activation+0.000
other
Token other
Feature activation+0.000
words
Token words
Feature activation+0.000
lying
Token lying
Feature activation+2.181
to
Token to
Feature activation+0.000
Congress
Token Congress
Feature activation+0.000
is
Token is
Feature activation+0.000
not
Token not
Feature activation+0.000
only
Token only
Feature activation+0.000
the
Token the
Feature activation+0.000
media
Token media
Feature activation+0.000
at
Token at
Feature activation+0.000
all
Token all
Feature activation+0.000
this
Token this
Feature activation+0.000
season
Token season
Feature activation+2.096
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
C
TokenC
Feature activation+0.000
urt
Tokenurt
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.000
is
Token is
Feature activation+0.000
in
Token in
Feature activation+0.000
turn
Token turn
Feature activation+0.000
stretched
Token stretched
Feature activation+2.050
out
Token out
Feature activation+0.000
and
Token and
Feature activation+0.000
folded
Token folded
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
point
Token point
Feature activation+0.000
in
Token in
Feature activation+0.000
time
Token time
Feature activation+0.000
weak
Token weak
Feature activation+2.046
and
Token and
Feature activation+0.448
incapable
Token incapable
Feature activation+0.000
[
Token [
Feature activation+0.000
in
Tokenin
Feature activation+0.000
which
Token which
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
stays
Token stays
Feature activation+2.028
in
Token in
Feature activation+0.000
kind
Token kind
Feature activation+0.039
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
weird
Token weird
Feature activation+0.000
is
Token is
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
repeal
Token repeal
Feature activation+1.968
passes
Token passes
Feature activation+1.220
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Updated
TokenUpdated
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
weakening
Token weakening
Feature activation+1.936
the
Token the
Feature activation+0.000
foundations
Token foundations
Feature activation+0.019
of
Token of
Feature activation+0.000
Hogwarts
Token Hogwarts
Feature activation+0.000
,
Token,
Feature activation+0.000
of
Token of
Feature activation+0.000
complexity
Token complexity
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
particular
Token particular
Feature activation+0.000
some
Token some
Feature activation+1.928
history
Token history
Feature activation+0.782
of
Token of
Feature activation+0.133
mathematical
Token mathematical
Feature activation+0.000
approaches
Token approaches
Feature activation+0.000
and
Token and
Feature activation+0.000
But
Token But
Feature activation+0.000
,
Token,
Feature activation+0.000
they
Token they
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.000
do
Token do
Feature activation+1.922
exist
Token exist
Feature activation+0.011
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Not
TokenNot
Feature activation+0.000
media
Token media
Feature activation+0.000
sites
Token sites
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
particular
Token particular
Feature activation+0.000
Facebook
Token Facebook
Feature activation+1.892
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
report
Token report
Feature activation+0.000
,
Token,
Feature activation+0.000
at
Token at
Feature activation+0.000
which
Token which
Feature activation+0.000
point
Token point
Feature activation+0.164
the
Token the
Feature activation+1.308
custom
Token custom
Feature activation+1.809
was
Token was
Feature activation+0.240
put
Token put
Feature activation+0.051
to
Token to
Feature activation+0.000
rest
Token rest
Feature activation+0.000
,
Token,
Feature activation+0.000
while
Token while
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
cultivate
Token cultivate
Feature activation+1.752
the
Token the
Feature activation+0.001
honey
Token honey
Feature activation+0.512
that
Token that
Feature activation+0.000
bees
Token bees
Feature activation+0.000
produce
Token produce
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
very
Token very
Feature activation+0.000
time
Token time
Feature activation+0.000
they
Token they
Feature activation+0.767
made
Token made
Feature activation+1.727
that
Token that
Feature activation+0.000
invention
Token invention
Feature activation+0.000
.
Token.
Feature activation+0.000
He
Token He
Feature activation+0.000
worked
Token worked
Feature activation+0.000
they
Token they
Feature activation+0.000
are
Token are
Feature activation+0.000
without
Token without
Feature activation+0.000
a
Token a
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
part
Token part
Feature activation+1.679
of
Token of
Feature activation+0.312
my
Token my
Feature activation+0.000
chromosomes
Token chromosomes
Feature activation+0.000
,
Token,
Feature activation+0.000
DNA
Token DNA
Feature activation+0.000
.
Token.
Feature activation+0.000
At
Token At
Feature activation+0.000
first
Token first
Feature activation+0.000
glance
Token glance
Feature activation+0.000
the
Token the
Feature activation+0.899
car
Token car
Feature activation+1.630
appeared
Token appeared
Feature activation+0.000
as
Token as
Feature activation+0.000
described
Token described
Feature activation+0.000
but
Token but
Feature activation+0.000
when
Token when
Feature activation+0.000

Top DFA by src position
MAX = 3.193

Ċ
TokenĊ
Feature activation+0.026
Top resid features:
Ċ
TokenĊ
Feature activation+0.038
Top resid features:
This
TokenThis
Feature activation+0.076
Top resid features:
is
Token is
Feature activation+0.017
Top resid features:
in
Token in
Feature activation-0.063
Top resid features:
effect
Token effect
Feature activation+3.018
Top resid features:
the
Token the
Feature activation+0.124
Top resid features:
Bitcoin
Token Bitcoin
Feature activation+0.045
Top resid features:
story
Token story
Feature activation+0.022
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
funny
Token funny
Feature activation+0.013
Top resid features:
but
Token but
Feature activation+0.010
Top resid features:
at
Token at
Feature activation+0.068
Top resid features:
the
Token the
Feature activation+0.061
Top resid features:
same
Token same
Feature activation-0.040
Top resid features:
time
Token time
Feature activation+3.193
Top resid features:
st
Token st
Feature activation+0.014
Top resid features:
inging
Tokeninging
Feature activation+0.048
Top resid features:
ev
Token ev
Feature activation-0.035
Top resid features:
is
Tokenis
Feature activation-0.040
Top resid features:
cer
Tokencer
Feature activation+0.021
Top resid features:
funny
Token funny
Feature activation+0.005
Top resid features:
but
Token but
Feature activation-0.006
Top resid features:
at
Token at
Feature activation+0.019
Top resid features:
the
Token the
Feature activation+0.039
Top resid features:
same
Token same
Feature activation+0.066
Top resid features:
time
Token time
Feature activation+2.853
Top resid features:
st
Token st
Feature activation+0.042
Top resid features:
inging
Tokeninging
Feature activation+0.112
Top resid features:
ev
Token ev
Feature activation+0.000
Top resid features:
is
Tokenis
Feature activation+0.000
Top resid features:
cer
Tokencer
Feature activation+0.000
Top resid features:
high
Token high
Feature activation+0.001
Top resid features:
NATO
Token NATO
Feature activation-0.003
Top resid features:
position
Token position
Feature activation+0.039
Top resid features:
at
Token at
Feature activation+0.036
Top resid features:
that
Token that
Feature activation+0.087
Top resid features:
time
Token time
Feature activation+2.583
Top resid features:
declared
Token declared
Feature activation+0.086
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:
The
TokenThe
Feature activation+0.000
Top resid features:
deal
Token deal
Feature activation+0.005
Top resid features:
,
Token,
Feature activation+0.032
Top resid features:
and
Token and
Feature activation+0.030
Top resid features:
at
Token at
Feature activation+0.101
Top resid features:
that
Token that
Feature activation+0.059
Top resid features:
time
Token time
Feature activation+2.657
Top resid features:
wore
Token wore
Feature activation+0.096
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
great
Token great
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:
impe
Token impe
Feature activation-0.016
Top resid features:
ached
Tokenached
Feature activation+0.028
Top resid features:
.
Token.
Feature activation+0.051
Top resid features:
In
Token In
Feature activation+0.108
Top resid features:
other
Token other
Feature activation+0.015
Top resid features:
words
Token words
Feature activation+2.778
Top resid features:
lying
Token lying
Feature activation+0.038
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
Congress
Token Congress
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
not
Token not
Feature activation+0.000
Top resid features:
speak
Token speak
Feature activation+0.045
Top resid features:
with
Token with
Feature activation+0.023
Top resid features:
the
Token the
Feature activation+0.036
Top resid features:
media
Token media
Feature activation+0.017
Top resid features:
at
Token at
Feature activation-0.020
Top resid features:
all
Token all
Feature activation+2.578
Top resid features:
this
Token this
Feature activation+0.049
Top resid features:
season
Token season
Feature activation+0.012
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
iteration
Token iteration
Feature activation+0.001
Top resid features:
,
Token,
Feature activation+0.058
Top resid features:
which
Token which
Feature activation+0.083
Top resid features:
is
Token is
Feature activation+0.042
Top resid features:
in
Token in
Feature activation+0.105
Top resid features:
turn
Token turn
Feature activation+2.424
Top resid features:
stretched
Token stretched
Feature activation+0.032
Top resid features:
out
Token out
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
folded
Token folded
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
are
Token are
Feature activation+0.071
Top resid features:
at
Token at
Feature activation+0.022
Top resid features:
that
Token that
Feature activation+0.045
Top resid features:
point
Token point
Feature activation+0.158
Top resid features:
in
Token in
Feature activation+0.022
Top resid features:
time
Token time
Feature activation+2.396
Top resid features:
weak
Token weak
Feature activation+0.044
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
incapable
Token incapable
Feature activation+0.000
Top resid features:
[
Token [
Feature activation+0.000
Top resid features:
in
Tokenin
Feature activation+0.000
Top resid features:
Rome
Token Rome
Feature activation-0.007
Top resid features:
but
Token but
Feature activation+0.021
Top resid features:
at
Token at
Feature activation+0.043
Top resid features:
the
Token the
Feature activation+0.038
Top resid features:
same
Token same
Feature activation+0.081
Top resid features:
time
Token time
Feature activation+2.397
Top resid features:
stays
Token stays
Feature activation+0.133
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
kind
Token kind
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
replacement
Token replacement
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.027
Top resid features:
at
Token at
Feature activation+0.019
Top resid features:
the
Token the
Feature activation+0.077
Top resid features:
same
Token same
Feature activation+0.255
Top resid features:
time
Token time
Feature activation+2.105
Top resid features:
repeal
Token repeal
Feature activation+0.047
Top resid features:
passes
Token passes
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.023
Top resid features:
but
Token but
Feature activation-0.004
Top resid features:
at
Token at
Feature activation+0.083
Top resid features:
the
Token the
Feature activation+0.070
Top resid features:
same
Token same
Feature activation+0.153
Top resid features:
time
Token time
Feature activation+2.214
Top resid features:
weakening
Token weakening
Feature activation+0.042
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
foundations
Token foundations
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
Hogwarts
Token Hogwarts
Feature activation+0.000
Top resid features:
aspects
Token aspects
Feature activation-0.004
Top resid features:
of
Token of
Feature activation+0.012
Top resid features:
complexity
Token complexity
Feature activation-0.014
Top resid features:
,
Token,
Feature activation-0.008
Top resid features:
in
Token in
Feature activation+0.015
Top resid features:
particular
Token particular
Feature activation+2.680
Top resid features:
some
Token some
Feature activation+0.035
Top resid features:
history
Token history
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
mathematical
Token mathematical
Feature activation+0.000
Top resid features:
approaches
Token approaches
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.007
Top resid features:
But
Token But
Feature activation+0.039
Top resid features:
,
Token,
Feature activation+0.060
Top resid features:
they
Token they
Feature activation-0.062
Top resid features:
in
Token in
Feature activation-0.028
Top resid features:
fact
Token fact
Feature activation+2.689
Top resid features:
do
Token do
Feature activation+0.057
Top resid features:
exist
Token exist
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
social
Token social
Feature activation-0.014
Top resid features:
media
Token media
Feature activation+0.003
Top resid features:
sites
Token sites
Feature activation-0.003
Top resid features:
,
Token,
Feature activation+0.020
Top resid features:
in
Token in
Feature activation+0.084
Top resid features:
particular
Token particular
Feature activation+2.580
Top resid features:
Facebook
Token Facebook
Feature activation-0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
The
TokenThe
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.010
Top resid features:
death
Token death
Feature activation+0.007
Top resid features:
,
Token,
Feature activation+0.017
Top resid features:
at
Token at
Feature activation+0.034
Top resid features:
which
Token which
Feature activation+0.100
Top resid features:
point
Token point
Feature activation+1.948
Top resid features:
the
Token the
Feature activation+0.300
Top resid features:
custom
Token custom
Feature activation+0.114
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
put
Token put
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
population
Token population
Feature activation-0.002
Top resid features:
while
Token while
Feature activation+0.027
Top resid features:
at
Token at
Feature activation+0.084
Top resid features:
the
Token the
Feature activation+0.037
Top resid features:
same
Token same
Feature activation+0.120
Top resid features:
time
Token time
Feature activation+2.119
Top resid features:
cultivate
Token cultivate
Feature activation+0.161
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
honey
Token honey
Feature activation+0.000
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
bees
Token bees
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.010
Top resid features:
them
Token them
Feature activation+0.007
Top resid features:
at
Token at
Feature activation+0.041
Top resid features:
the
Token the
Feature activation+0.062
Top resid features:
very
Token very
Feature activation+0.041
Top resid features:
time
Token time
Feature activation+2.070
Top resid features:
they
Token they
Feature activation+0.093
Top resid features:
made
Token made
Feature activation+0.022
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
invention
Token invention
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.052
Top resid features:
they
Token they
Feature activation+0.002
Top resid features:
are
Token are
Feature activation+0.072
Top resid features:
without
Token without
Feature activation-0.006
Top resid features:
a
Token a
Feature activation+0.005
Top resid features:
doubt
Token doubt
Feature activation+2.328
Top resid features:
part
Token part
Feature activation-0.005
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
my
Token my
Feature activation+0.000
Top resid features:
chromosomes
Token chromosomes
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
after
Token after
Feature activation+0.010
Top resid features:
all
Token all
Feature activation+0.048
Top resid features:
.
Token.
Feature activation+0.028
Top resid features:
At
Token At
Feature activation-0.011
Top resid features:
first
Token first
Feature activation+0.043
Top resid features:
glance
Token glance
Feature activation+1.917
Top resid features:
the
Token the
Feature activation+0.310
Top resid features:
car
Token car
Feature activation-0.017
Top resid features:
appeared
Token appeared
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.000
Top resid features:
described
Token described
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.04

Head 2: 0.07

Head 3: 0.03

Head 4: 0.03

Head 5: 0.04

Head 6: 0.05

Head 7: 0.09

Head 8: 0.47

Head 9: 0.04

Head 10: 0.05

Head 11: 0.07

Positive logits

ippery1.49

idable1.48

risome1.47

irable1.42

aughtered1.40

ancial1.38

>>>>>>>>1.35

iosyn1.35

acknow1.33

rored1.32

irlf1.31

urally1.31

hematically1.30

apego1.30

toget1.27

INTO1.26

hetically1.23

avorite1.22

^^^^1.22

theless1.21

Negative logits

gam-1.51

Indies-1.38

locality-1.35

periphery-1.31

Platform-1.31

hindsight-1.31

Lum-1.29

amazon-1.27

Rum-1.24

Galaxy-1.23

persuasion-1.23

League-1.22

cloud-1.21

perspective-1.20

Highlands-1.20

Atl-1.20

peak-1.20

isphere-1.20

axy-1.18

finals-1.17

INTERVAL 2.249 - 2.499
CONTAINS 0.001%

at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
st
Token st
Feature activation+1.579
inging
Tokeninging
Feature activation+2.428
ev
Token ev
Feature activation+0.001
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+2.451
of
Token of
Feature activation+0.396
st
Token st
Feature activation+1.579
inging
Tokeninging
Feature activation+2.428
ev
Token ev
Feature activation+0.001
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+2.451
of
Token of
Feature activation+0.396
Dr
Token Dr
Feature activation+0.000
.
Token.
Feature activation+0.000
Mann
Token Mann
Feature activation+0.000
on
Token on
Feature activation+0.000
is
Token is
Feature activation+0.000
in
Token in
Feature activation+0.000
effect
Token effect
Feature activation+0.000
the
Token the
Feature activation+1.348
Bitcoin
Token Bitcoin
Feature activation+0.362
story
Token story
Feature activation+2.499
,
Token,
Feature activation+0.194
the
Token the
Feature activation+0.000
net
Token net
Feature activation+0.000
result
Token result
Feature activation+0.000
is
Token is
Feature activation+0.000
NATO
Token NATO
Feature activation+0.000
position
Token position
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
declared
Token declared
Feature activation+2.272
,
Token,
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
Roman
Token Roman
Feature activation+0.000

INTERVAL 2.000 - 2.249
CONTAINS 0.001%

ached
Tokenached
Feature activation+0.000
.
Token.
Feature activation+0.000
In
Token In
Feature activation+0.000
other
Token other
Feature activation+0.000
words
Token words
Feature activation+0.000
lying
Token lying
Feature activation+2.181
to
Token to
Feature activation+0.000
Congress
Token Congress
Feature activation+0.000
is
Token is
Feature activation+0.000
not
Token not
Feature activation+0.000
only
Token only
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
wore
Token wore
Feature activation+2.224
a
Token a
Feature activation+0.119
great
Token great
Feature activation+0.000
number
Token number
Feature activation+0.000
of
Token of
Feature activation+0.000
clothes
Token clothes
Feature activation+0.037
the
Token the
Feature activation+0.000
media
Token media
Feature activation+0.000
at
Token at
Feature activation+0.000
all
Token all
Feature activation+0.000
this
Token this
Feature activation+0.000
season
Token season
Feature activation+2.096
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
C
TokenC
Feature activation+0.000
urt
Tokenurt
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
stays
Token stays
Feature activation+2.028
in
Token in
Feature activation+0.000
kind
Token kind
Feature activation+0.039
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
weird
Token weird
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
point
Token point
Feature activation+0.000
in
Token in
Feature activation+0.000
time
Token time
Feature activation+0.000
weak
Token weak
Feature activation+2.046
and
Token and
Feature activation+0.448
incapable
Token incapable
Feature activation+0.000
[
Token [
Feature activation+0.000
in
Tokenin
Feature activation+0.000
which
Token which
Feature activation+0.000

INTERVAL 1.750 - 2.000
CONTAINS 0.001%

But
Token But
Feature activation+0.000
,
Token,
Feature activation+0.000
they
Token they
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.000
do
Token do
Feature activation+1.922
exist
Token exist
Feature activation+0.011
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Not
TokenNot
Feature activation+0.000
is
Token is
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
repeal
Token repeal
Feature activation+1.968
passes
Token passes
Feature activation+1.220
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Updated
TokenUpdated
Feature activation+0.000
,
Token,
Feature activation+0.000
at
Token at
Feature activation+0.000
which
Token which
Feature activation+0.000
point
Token point
Feature activation+0.164
the
Token the
Feature activation+1.308
custom
Token custom
Feature activation+1.809
was
Token was
Feature activation+0.240
put
Token put
Feature activation+0.051
to
Token to
Feature activation+0.000
rest
Token rest
Feature activation+0.000
,
Token,
Feature activation+0.000
media
Token media
Feature activation+0.000
sites
Token sites
Feature activation+0.000
,
Token,
Feature activation+0.000
in
Token in
Feature activation+0.000
particular
Token particular
Feature activation+0.000
Facebook
Token Facebook
Feature activation+1.892
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
report
Token report
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
weakening
Token weakening
Feature activation+1.936
the
Token the
Feature activation+0.000
foundations
Token foundations
Feature activation+0.019
of
Token of
Feature activation+0.000
Hogwarts
Token Hogwarts
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 1.500 - 1.750
CONTAINS 0.001%

suggest
Token suggest
Feature activation+0.000
that
Token that
Feature activation+0.000
under
Token under
Feature activation+0.000
certain
Token certain
Feature activation+0.000
circumstances
Token circumstances
Feature activation+0.000
scars
Token scars
Feature activation+1.595
may
Token may
Feature activation+0.000
advertise
Token advertise
Feature activation+0.000
valued
Token valued
Feature activation+0.000
information
Token information
Feature activation+0.000
about
Token about
Feature activation+0.000
naturally
Token naturally
Feature activation+0.000
âĢĶ
TokenâĢĶ
Feature activation+0.000
are
Tokenare
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.000
drugs
Token drugs
Feature activation+1.515
and
Token and
Feature activation+0.183
subject
Token subject
Feature activation+0.129
to
Token to
Feature activation+0.000
its
Token its
Feature activation+0.000
regulatory
Token regulatory
Feature activation+0.000
.
Token.
Feature activation+0.000
Wolves
Token Wolves
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
were
Token were
Feature activation+1.629
especially
Token especially
Feature activation+0.083
numerous
Token numerous
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
districts
Token districts
Feature activation+0.000
toilets
Token toilets
Feature activation+0.000
.
Token.
Feature activation+0.000
In
Token In
Feature activation+0.000
total
Token total
Feature activation+0.000
the
Token the
Feature activation+0.510
firm
Token firm
Feature activation+1.563
treats
Token treats
Feature activation+0.000
2
Token 2
Feature activation+0.000
.
Token.
Feature activation+0.000
8
Token8
Feature activation+0.000
billion
Token billion
Feature activation+0.000
in
Tokenin
Feature activation+0.000
.
Token .
Feature activation+0.000
In
Token In
Feature activation+0.000
that
Token that
Feature activation+0.000
case
Token case
Feature activation+0.033
use
Token use
Feature activation+1.567
the
Token the
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
help
Token help
Feature activation+0.000
<
Token <
Feature activation+0.000

INTERVAL 1.250 - 1.500
CONTAINS 0.003%

we
Token we
Feature activation+0.000
have
Token have
Feature activation+0.000
in
Token in
Feature activation+0.000
my
Token my
Feature activation+0.000
view
Token view
Feature activation+0.000
strong
Token strong
Feature activation+1.400
evidence
Token evidence
Feature activation+0.154
of
Token of
Feature activation+0.000
corpor
Token corpor
Feature activation+0.000
ates
Tokenates
Feature activation+0.000
having
Token having
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
forensic
Token forensic
Feature activation+0.000
scientific
Token scientific
Feature activation+0.000
sense
Token sense
Feature activation+0.370
and
Token and
Feature activation+1.261
barely
Token barely
Feature activation+0.000
the
Token the
Feature activation+0.000
simple
Token simple
Feature activation+0.000
image
Token image
Feature activation+0.000
it
Token it
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
And
TokenAnd
Feature activation+0.000
in
Token in
Feature activation+0.000
fact
Token fact
Feature activation+0.000
a
Token a
Feature activation+0.604
key
Token key
Feature activation+1.476
point
Token point
Feature activation+0.965
in
Token in
Feature activation+0.008
the
Token the
Feature activation+0.000
rhetorical
Token rhetorical
Feature activation+0.000
escalation
Token escalation
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
and
Token and
Feature activation+0.446
science
Token science
Feature activation+0.201
data
Token data
Feature activation+1.420
is
Token is
Feature activation+0.152
recorded
Token recorded
Feature activation+0.263
and
Token and
Feature activation+0.000
down
Token down
Feature activation+0.000
linked
Tokenlinked
Feature activation+0.000
in
Tokenin
Feature activation+0.000
other
Token other
Feature activation+0.000
words
Token words
Feature activation+0.000
,
Token,
Feature activation+0.000
overly
Token overly
Feature activation+0.770
focused
Token focused
Feature activation+1.491
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
sculpt
Token sculpt
Feature activation+0.000
ural
Tokenural
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000

INTERVAL 1.000 - 1.250
CONTAINS 0.007%

correlation
Token correlation
Feature activation+0.000
and
Token and
Feature activation+0.000
causation
Token causation
Feature activation+0.000
in
Token in
Feature activation+0.000
general
Token general
Feature activation+0.000
remains
Token remains
Feature activation+1.091
an
Token an
Feature activation+0.037
active
Token active
Feature activation+0.000
area
Token area
Feature activation+0.000
of
Token of
Feature activation+0.000
debate
Token debate
Feature activation+0.000
elections
Token elections
Feature activation+0.000
,
Token,
Feature activation+0.000
at
Token at
Feature activation+0.000
which
Token which
Feature activation+0.000
point
Token point
Feature activation+0.000
the
Token the
Feature activation+1.048
European
Token European
Feature activation+0.731
envoy
Token envoy
Feature activation+1.588
asserted
Token asserted
Feature activation+0.256
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
that
Token that
Feature activation+0.000
I
Token I
Feature activation+0.000
was
Token was
Feature activation+0.000
at
Token at
Feature activation+0.000
first
Token first
Feature activation+0.000
utterly
Token utterly
Feature activation+1.071
myst
Token myst
Feature activation+0.000
ified
Tokenified
Feature activation+0.221
as
Token as
Feature activation+0.000
to
Token to
Feature activation+0.000
how
Token how
Feature activation+0.000
to
Token to
Feature activation+0.000
firmware
Token firmware
Feature activation+0.000
in
Token in
Feature activation+0.000
no
Token no
Feature activation+0.000
way
Token way
Feature activation+0.000
provides
Token provides
Feature activation+1.230
access
Token access
Feature activation+0.000
to
Token to
Feature activation+0.000
data
Token data
Feature activation+0.000
protection
Token protection
Feature activation+0.000
class
Token class
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
repeal
Token repeal
Feature activation+1.968
passes
Token passes
Feature activation+1.220
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Updated
TokenUpdated
Feature activation+0.000
at
Token at
Feature activation+0.000

INTERVAL 0.750 - 1.000
CONTAINS 0.011%

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
general
Token general
Feature activation+0.000
becomes
Token becomes
Feature activation+0.837
aware
Token aware
Feature activation+0.000
of
Token of
Feature activation+0.000
this
Token this
Feature activation+0.000
movement
Token movement
Feature activation+0.000
,
Token,
Feature activation+0.000
by
Token by
Feature activation+0.000
no
Token no
Feature activation+0.000
means
Token means
Feature activation+0.000
an
Token an
Feature activation+1.236
altru
Token altru
Feature activation+0.235
ist
Tokenist
Feature activation+0.894
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
whole
Token whole
Feature activation+0.000
arrangement
Token arrangement
Feature activation+0.000
is
Token is
Feature activation+0.000
(
Token (
Feature activation+0.000
in
Tokenin
Feature activation+0.000
this
Token this
Feature activation+0.000
case
Token case
Feature activation+0.000
a
Token a
Feature activation+1.153
list
Token list
Feature activation+0.766
of
Token of
Feature activation+0.000
symbols
Token symbols
Feature activation+0.000
and
Token and
Feature activation+0.000
/
Token/
Feature activation+0.000
or
Tokenor
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
some
Token some
Feature activation+0.000
point
Token point
Feature activation+0.000
the
Token the
Feature activation+0.856
un
Token un
Feature activation+0.564
met
Tokenmet
Feature activation+0.000
demand
Token demand
Feature activation+0.354
for
Token for
Feature activation+0.000
housing
Token housing
Feature activation+0.000
-
Token-
Feature activation+0.000
depth
Tokendepth
Feature activation+0.000
behind
Token behind
Feature activation+0.000
the
Token the
Feature activation+0.000
scenes
Token scenes
Feature activation+0.000
entertainment
Token entertainment
Feature activation+0.920
and
Token and
Feature activation+0.000
grow
Token grow
Feature activation+0.000
the
Token the
Feature activation+0.000
prize
Token prize
Feature activation+0.000
pool
Token pool
Feature activation+0.000

INTERVAL 0.500 - 0.750
CONTAINS 0.023%

any
Token any
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
other
Token other
Feature activation+0.000
reasons
Token reasons
Feature activation+0.000
set
Token set
Feature activation+0.638
forth
Token forth
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
post
Token post
Feature activation+0.000
).
Token).
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.000
by
Token by
Feature activation+0.000
implication
Token implication
Feature activation+0.000
means
Token means
Feature activation+0.646
that
Token that
Feature activation+0.000
you
Token you
Feature activation+0.000
one
Token one
Feature activation+0.000
day
Token day
Feature activation+0.000
will
Token will
Feature activation+0.000
be
Token be
Feature activation+0.000
gone
Token gone
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
flash
Token flash
Feature activation+0.000
so
Token so
Feature activation+0.508
grab
Token grab
Feature activation+0.000
a
Token a
Feature activation+0.000
tort
Token tort
Feature activation+0.000
illa
Tokenilla
Feature activation+0.000
and
Token and
Feature activation+0.000
spend
Token spend
Feature activation+0.000
every
Token every
Feature activation+0.000
day
Token day
Feature activation+0.000
in
Token in
Feature activation+0.000
prison
Token prison
Feature activation+0.000
thinking
Token thinking
Feature activation+0.554
about
Token about
Feature activation+0.427
the
Token the
Feature activation+0.000
life
Token life
Feature activation+0.000
you
Token you
Feature activation+0.000
took
Token took
Feature activation+0.000
of
Token of
Feature activation+0.000
beer
Token beer
Feature activation+0.000
for
Token for
Feature activation+0.000
that
Token that
Feature activation+0.000
matter
Token matter
Feature activation+0.000
is
Token is
Feature activation+0.604
a
Token a
Feature activation+0.270
lot
Token lot
Feature activation+0.000
of
Token of
Feature activation+0.000
money
Token money
Feature activation+0.000
,"
Token,"
Feature activation+0.000

INTERVAL 0.250 - 0.500
CONTAINS 0.055%

economies
Token economies
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
its
Token its
Feature activation+0.511
cent
Token cent
Feature activation+0.297
enary
Tokenenary
Feature activation+0.000
comes
Token comes
Feature activation+0.000
around
Token around
Feature activation+0.000
in
Token in
Feature activation+0.000
20
Token 20
Feature activation+0.000
means
Token means
Feature activation+0.000
(
Token (
Feature activation+0.000
and
Tokenand
Feature activation+0.000
to
Token to
Feature activation+0.000
some
Token some
Feature activation+0.000
extent
Token extent
Feature activation+0.347
ends
Token ends
Feature activation+1.545
)
Token)
Feature activation+0.000
exist
Token exist
Feature activation+0.000
,
Token,
Feature activation+0.000
no
Token no
Feature activation+0.000
for
Token for
Feature activation+0.000
that
Token that
Feature activation+0.000
matter
Token matter
Feature activation+0.000
so
Token so
Feature activation+0.252
emotionally
Token emotionally
Feature activation+0.442
involved
Token involved
Feature activation+0.361
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
when
Token when
Feature activation+0.000
it
Token it
Feature activation+0.000
comes
Token comes
Feature activation+0.000
o
Tokeno
Feature activation+0.000
itte
Tokenitte
Feature activation+0.000
report
Token report
Feature activation+0.000
for
Token for
Feature activation+0.000
instance
Token instance
Feature activation+0.000
concluded
Token concluded
Feature activation+0.460
that
Token that
Feature activation+0.000
technology
Token technology
Feature activation+0.000
potentially
Token potentially
Feature activation+0.000
played
Token played
Feature activation+0.000
a
Token a
Feature activation+0.000
this
Token this
Feature activation+0.000
process
Token process
Feature activation+0.000
in
Token in
Feature activation+0.000
his
Token his
Feature activation+0.000
book
Token book
Feature activation+0.000
Germany
Token Germany
Feature activation+0.256
18
Token 18
Feature activation+0.000
66
Token66
Feature activation+0.000
-
Token-
Feature activation+0.000
1945
Token1945
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 0.000 - 0.250
CONTAINS 99.898%

since
Token since
Feature activation+0.000
.
Token.
Feature activation+0.000
It
Token It
Feature activation+0.000
was
Token was
Feature activation+0.000
not
Token not
Feature activation+0.000
clear
Token clear
Feature activation+0.000
on
Token on
Feature activation+0.000
which
Token which
Feature activation+0.000
website
Token website
Feature activation+0.000
the
Token the
Feature activation+0.000
two
Token two
Feature activation+0.000
.
Token.
Feature activation+0.000
3
Token 3
Feature activation+0.000
D
TokenD
Feature activation+0.000
printing
Token printing
Feature activation+0.000
itself
Token itself
Feature activation+0.000
took
Token took
Feature activation+0.000
place
Token place
Feature activation+0.000
on
Token on
Feature activation+0.000
a
Token a
Feature activation+0.000
Da
Token Da
Feature activation+0.000
Vin
Token Vin
Feature activation+0.000
me
Token me
Feature activation+0.000
anymore
Token anymore
Feature activation+0.000
.
Token.
Feature activation+0.000
Sure
Token Sure
Feature activation+0.000
a
Token a
Feature activation+0.000
woman
Token woman
Feature activation+0.000
looks
Token looks
Feature activation+0.000
hot
Token hot
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
tight
Token tight
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
F
TokenF
Feature activation+0.000
am
Tokenam
Feature activation+0.000
ilies
Tokenilies
Feature activation+0.000
started
Token started
Feature activation+0.000
to
Token to
Feature activation+0.000
collect
Token collect
Feature activation+0.000
wood
Token wood
Feature activation+0.000
for
Token for
Feature activation+0.000
their
Token their
Feature activation+0.000
bon
Token bon
Feature activation+0.000
affecting
Token affecting
Feature activation+0.000
the
Token the
Feature activation+0.000
appearance
Token appearance
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
image
Token image
Feature activation+0.000
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
prototype
Token prototype
Feature activation+0.000
version
Token version
Feature activation+0.000
is
Token is
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 H3.8: (feature 11889)

TOP ACTIVATIONS
MAX = 1.740

albums
Token albums
Feature activation+0.000
,
Token,
Feature activation+0.000
with
Token with
Feature activation+0.000
selected
Token selected
Feature activation+0.000
information
Token information
Feature activation+0.000
Title
Token Title
Feature activation+1.740
Album
Token Album
Feature activation+0.966
details
Token details
Feature activation+0.018
A
Token A
Feature activation+0.620
Picture
Token Picture
Feature activation+0.000
Perfect
Token Perfect
Feature activation+0.000
have
Token have
Feature activation+0.000
to
Token to
Feature activation+0.000
file
Token file
Feature activation+0.000
similar
Token similar
Feature activation+0.000
reports
Token reports
Feature activation+0.000
election
Token election
Feature activation+1.525
-
Token-
Feature activation+1.060
cycle
Tokencycle
Feature activation+0.767
after
Token after
Feature activation+0.000
election
Token election
Feature activation+0.000
-
Token-
Feature activation+0.000
short
Token short
Feature activation+0.000
-
Token-
Feature activation+0.000
term
Tokenterm
Feature activation+0.081
emissions
Token emissions
Feature activation+0.000
cuts
Token cuts
Feature activation+0.000
European
Token European
Feature activation+1.473
leaders
Token leaders
Feature activation+0.800
had
Token had
Feature activation+0.000
demanded
Token demanded
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
to
Token to
Feature activation+0.000
answer
Token answer
Feature activation+0.000
dozens
Token dozens
Feature activation+0.000
of
Token of
Feature activation+0.000
questions
Token questions
Feature activation+0.000
cert
Token cert
Feature activation+1.465
ifying
Tokenifying
Feature activation+0.437
her
Token her
Feature activation+0.000
background
Token background
Feature activation+0.000
information
Token information
Feature activation+0.052
.
Token.
Feature activation+0.000
the
Token the
Feature activation+0.000
extra
Token extra
Feature activation+0.000
52
Token 52
Feature activation+0.000
%
Token%
Feature activation+0.000
increase
Token increase
Feature activation+0.000
AoE
Token AoE
Feature activation+1.389
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
cycl
Token cycl
Feature activation+0.000
one
Tokenone
Feature activation+0.000
,
Token,
Feature activation+0.000
E
Token E
Feature activation+0.000
PL
TokenPL
Feature activation+0.000
rights
Token rights
Feature activation+0.000
supports
Token supports
Feature activation+0.089
claims
Token claims
Feature activation+0.384
Fox
Token Fox
Feature activation+1.386
Sports
Token Sports
Feature activation+1.069
had
Token had
Feature activation+0.387
been
Token been
Feature activation+0.218
compl
Token compl
Feature activation+0.000
acent
Tokenacent
Feature activation+0.000
issued
Token issued
Feature activation+0.000
a
Token a
Feature activation+0.000
brief
Token brief
Feature activation+0.000
statement
Token statement
Feature activation+0.000
commemor
Token commemor
Feature activation+1.209
ating
Tokenating
Feature activation+1.327
Dro
Token Dro
Feature activation+0.000
z
Tokenz
Feature activation+0.000
d
Tokend
Feature activation+0.000
ov
Tokenov
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
pricing
Token pricing
Feature activation+0.000
refers
Token refers
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
prices
Token prices
Feature activation+0.000
multinational
Token multinational
Feature activation+1.279
s
Tokens
Feature activation+0.927
set
Token set
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
charge
Token charge
Feature activation+0.000
Asked
TokenAsked
Feature activation+0.000
about
Token about
Feature activation+0.000
the
Token the
Feature activation+0.000
legal
Token legal
Feature activation+0.000
risks
Token risks
Feature activation+0.000
contractors
Token contractors
Feature activation+1.275
could
Token could
Feature activation+0.210
face
Token face
Feature activation+0.000
from
Token from
Feature activation+0.000
Canada
Token Canada
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
100
Token 100
Feature activation+0.000
-
Token-
Feature activation+0.000
plus
Tokenplus
Feature activation+0.000
question
Token question
Feature activation+0.000
survey
Token survey
Feature activation+1.260
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
if
Token if
Feature activation+0.000
poor
Token poor
Feature activation+0.000
results
Token results
Feature activation+0.000
are
Token are
Feature activation+0.000
just
Token just
Feature activation+0.000
having
Token having
Feature activation+0.000
a
Token a
Feature activation+0.000
blast
Token blast
Feature activation+0.000
playing
Token playing
Feature activation+1.254
music
Token music
Feature activation+0.000
together
Token together
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
of
Token of
Feature activation+0.000
.
Token.
Feature activation+0.000
But
Token But
Feature activation+0.000
with
Token with
Feature activation+0.000
plans
Token plans
Feature activation+0.000
starting
Token starting
Feature activation+1.244
at
Token at
Feature activation+0.000
low
Token low
Feature activation+0.000
rates
Token rates
Feature activation+0.000
,
Token,
Feature activation+0.000
from
Token from
Feature activation+0.000
're
Token're
Feature activation+0.000
facing
Token facing
Feature activation+0.000
big
Token big
Feature activation+0.000
problems
Token problems
Feature activation+0.000
reload
Token reload
Feature activation+1.086
ing
Tokening
Feature activation+1.214
their
Token their
Feature activation+0.000
cards
Token cards
Feature activation+0.000
at
Token at
Feature activation+0.000
Prest
Token Prest
Feature activation+0.000
o
Tokeno
Feature activation+0.000
,
Token,
Feature activation+0.000
issued
Token issued
Feature activation+0.000
a
Token a
Feature activation+0.000
brief
Token brief
Feature activation+0.000
statement
Token statement
Feature activation+0.000
commemor
Token commemor
Feature activation+1.209
ating
Tokenating
Feature activation+1.327
Dro
Token Dro
Feature activation+0.000
z
Tokenz
Feature activation+0.000
d
Tokend
Feature activation+0.000
ov
Tokenov
Feature activation+0.000
city
Token city
Feature activation+0.000
to
Token to
Feature activation+0.000
reduce
Token reduce
Feature activation+0.000
the
Token the
Feature activation+0.000
rates
Token rates
Feature activation+0.000
35
Token 35
Feature activation+1.204
%,
Token%,
Feature activation+0.009
though
Token though
Feature activation+0.000
Washington
Token Washington
Feature activation+0.000
disputes
Token disputes
Feature activation+0.000
how
Token how
Feature activation+0.115
absolute
Token absolute
Feature activation+0.000
--
Token--
Feature activation+0.000
a
Tokena
Feature activation+0.000
key
Token key
Feature activation+0.000
distinction
Token distinction
Feature activation+0.014
Kay
Token Kay
Feature activation+1.201
e
Tokene
Feature activation+0.816
identified
Token identified
Feature activation+0.000
as
Token as
Feature activation+0.000
an
Token an
Feature activation+0.000
independent
Token independent
Feature activation+0.000
sentences
Token sentences
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
palpable
Token palpable
Feature activation+0.000
excitement
Token excitement
Feature activation+0.000
crack
Token crack
Feature activation+1.196
ling
Tokenling
Feature activation+0.461
over
Token over
Feature activation+0.000
the
Token the
Feature activation+0.000
phone
Token phone
Feature activation+0.000
.
Token.
Feature activation+0.000
rador
Tokenrador
Feature activation+0.000
added
Token added
Feature activation+0.000
that
Token that
Feature activation+0.000
the
Token the
Feature activation+0.000
frustration
Token frustration
Feature activation+0.000
building
Token building
Feature activation+1.194
up
Token up
Feature activation+0.000
is
Token is
Feature activation+0.000
because
Token because
Feature activation+0.000
the
Token the
Feature activation+0.000
administration
Token administration
Feature activation+0.000
;
Token;
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
more
Token more
Feature activation+0.000
confidence
Token confidence
Feature activation+0.000
shooting
Token shooting
Feature activation+1.188
the
Token the
Feature activation+0.000
ball
Token ball
Feature activation+0.000
in
Token in
Feature activation+0.000
general
Token general
Feature activation+0.000
.
Token.
Feature activation+0.000
l
Tokenl
Feature activation+0.000
ovic
Tokenovic
Feature activation+0.000
made
Token made
Feature activation+0.000
wholesale
Token wholesale
Feature activation+0.000
changes
Token changes
Feature activation+0.000
for
Token for
Feature activation+1.183
his
Token his
Feature activation+0.000
first
Token first
Feature activation+0.000
game
Token game
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000

Top DFA by src position
MAX = 2.069

of
Token of
Feature activation-0.004
Top resid features:
albums
Token albums
Feature activation-0.015
Top resid features:
,
Token,
Feature activation-0.036
Top resid features:
with
Token with
Feature activation+0.003
Top resid features:
selected
Token selected
Feature activation-0.015
Top resid features:
information
Token information
Feature activation+2.069
Top resid features:
Title
Token Title
Feature activation-0.131
Top resid features:
Album
Token Album
Feature activation+0.000
Top resid features:
details
Token details
Feature activation+0.000
Top resid features:
A
Token A
Feature activation+0.000
Top resid features:
Picture
Token Picture
Feature activation+0.000
Top resid features:
we
Token we
Feature activation+0.009
Top resid features:
have
Token have
Feature activation+0.003
Top resid features:
to
Token to
Feature activation+0.003
Top resid features:
file
Token file
Feature activation-0.077
Top resid features:
similar
Token similar
Feature activation-0.038
Top resid features:
reports
Token reports
Feature activation+1.884
Top resid features:
election
Token election
Feature activation-0.037
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
cycle
Tokencycle
Feature activation+0.000
Top resid features:
after
Token after
Feature activation+0.000
Top resid features:
election
Token election
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.014
Top resid features:
short
Token short
Feature activation-0.004
Top resid features:
-
Token-
Feature activation-0.003
Top resid features:
term
Tokenterm
Feature activation+0.020
Top resid features:
emissions
Token emissions
Feature activation+0.239
Top resid features:
cuts
Token cuts
Feature activation+1.556
Top resid features:
European
Token European
Feature activation-0.131
Top resid features:
leaders
Token leaders
Feature activation+0.000
Top resid features:
had
Token had
Feature activation+0.000
Top resid features:
demanded
Token demanded
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
required
Token required
Feature activation-0.008
Top resid features:
to
Token to
Feature activation+0.003
Top resid features:
answer
Token answer
Feature activation-0.005
Top resid features:
dozens
Token dozens
Feature activation-0.018
Top resid features:
of
Token of
Feature activation-0.008
Top resid features:
questions
Token questions
Feature activation+1.712
Top resid features:
cert
Token cert
Feature activation-0.025
Top resid features:
ifying
Tokenifying
Feature activation+0.000
Top resid features:
her
Token her
Feature activation+0.000
Top resid features:
background
Token background
Feature activation+0.000
Top resid features:
information
Token information
Feature activation+0.000
Top resid features:
With
Token With
Feature activation+0.011
Top resid features:
the
Token the
Feature activation+0.012
Top resid features:
extra
Token extra
Feature activation-0.013
Top resid features:
52
Token 52
Feature activation+0.027
Top resid features:
%
Token%
Feature activation+0.019
Top resid features:
increase
Token increase
Feature activation+1.494
Top resid features:
AoE
Token AoE
Feature activation+0.021
Top resid features:
from
Token from
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
cycl
Token cycl
Feature activation+0.000
Top resid features:
one
Tokenone
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.000
Top resid features:
E
Token E
Feature activation-0.006
Top resid features:
PL
TokenPL
Feature activation+0.001
Top resid features:
rights
Token rights
Feature activation+0.103
Top resid features:
supports
Token supports
Feature activation+0.140
Top resid features:
claims
Token claims
Feature activation+1.454
Top resid features:
Fox
Token Fox
Feature activation-0.086
Top resid features:
Sports
Token Sports
Feature activation+0.000
Top resid features:
had
Token had
Feature activation+0.000
Top resid features:
been
Token been
Feature activation+0.000
Top resid features:
compl
Token compl
Feature activation+0.000
Top resid features:
KGB
Token KGB
Feature activation+0.000
Top resid features:
,
Token,
Feature activation-0.034
Top resid features:
issued
Token issued
Feature activation-0.041
Top resid features:
a
Token a
Feature activation-0.040
Top resid features:
brief
Token brief
Feature activation+0.031
Top resid features:
statement
Token statement
Feature activation+1.671
Top resid features:
commemor
Token commemor
Feature activation-0.039
Top resid features:
ating
Tokenating
Feature activation-0.068
Top resid features:
Dro
Token Dro
Feature activation+0.000
Top resid features:
z
Tokenz
Feature activation+0.000
Top resid features:
d
Tokend
Feature activation+0.000
Top resid features:
Transfer
TokenTransfer
Feature activation-0.003
Top resid features:
pricing
Token pricing
Feature activation+0.006
Top resid features:
refers
Token refers
Feature activation-0.013
Top resid features:
to
Token to
Feature activation-0.014
Top resid features:
the
Token the
Feature activation+0.044
Top resid features:
prices
Token prices
Feature activation+1.479
Top resid features:
multinational
Token multinational
Feature activation-0.069
Top resid features:
s
Tokens
Feature activation+0.000
Top resid features:
set
Token set
Feature activation+0.000
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
they
Token they
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.000
Top resid features:
Asked
TokenAsked
Feature activation+0.006
Top resid features:
about
Token about
Feature activation+0.001
Top resid features:
the
Token the
Feature activation+0.006
Top resid features:
legal
Token legal
Feature activation+0.009
Top resid features:
risks
Token risks
Feature activation+1.404
Top resid features:
contractors
Token contractors
Feature activation+0.059
Top resid features:
could
Token could
Feature activation+0.000
Top resid features:
face
Token face
Feature activation+0.000
Top resid features:
from
Token from
Feature activation+0.000
Top resid features:
Canada
Token Canada
Feature activation+0.000
Top resid features:
out
Token out
Feature activation+0.005
Top resid features:
a
Token a
Feature activation-0.033
Top resid features:
100
Token 100
Feature activation-0.033
Top resid features:
-
Token-
Feature activation-0.004
Top resid features:
plus
Tokenplus
Feature activation-0.021
Top resid features:
question
Token question
Feature activation+1.524
Top resid features:
survey
Token survey
Feature activation+0.080
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
if
Token if
Feature activation+0.000
Top resid features:
poor
Token poor
Feature activation+0.000
Top resid features:
who
Token who
Feature activation+0.013
Top resid features:
are
Token are
Feature activation-0.008
Top resid features:
just
Token just
Feature activation+0.014
Top resid features:
having
Token having
Feature activation-0.009
Top resid features:
a
Token a
Feature activation-0.017
Top resid features:
blast
Token blast
Feature activation+1.537
Top resid features:
playing
Token playing
Feature activation-0.101
Top resid features:
music
Token music
Feature activation+0.000
Top resid features:
together
Token together
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
fan
Token fan
Feature activation-0.011
Top resid features:
of
Token of
Feature activation-0.006
Top resid features:
.
Token.
Feature activation-0.018
Top resid features:
But
Token But
Feature activation-0.007
Top resid features:
with
Token with
Feature activation+0.016
Top resid features:
plans
Token plans
Feature activation+1.503
Top resid features:
starting
Token starting
Feature activation-0.068
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
low
Token low
Feature activation+0.000
Top resid features:
rates
Token rates
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
say
Token say
Feature activation-0.014
Top resid features:
they
Token they
Feature activation+0.001
Top resid features:
're
Token're
Feature activation-0.004
Top resid features:
facing
Token facing
Feature activation-0.015
Top resid features:
big
Token big
Feature activation-0.001
Top resid features:
problems
Token problems
Feature activation+1.686
Top resid features:
reload
Token reload
Feature activation-0.197
Top resid features:
ing
Tokening
Feature activation-0.098
Top resid features:
their
Token their
Feature activation+0.000
Top resid features:
cards
Token cards
Feature activation+0.000
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
KGB
Token KGB
Feature activation+0.006
Top resid features:
,
Token,
Feature activation-0.032
Top resid features:
issued
Token issued
Feature activation-0.041
Top resid features:
a
Token a
Feature activation-0.061
Top resid features:
brief
Token brief
Feature activation+0.021
Top resid features:
statement
Token statement
Feature activation+1.536
Top resid features:
commemor
Token commemor
Feature activation-0.054
Top resid features:
ating
Tokenating
Feature activation+0.000
Top resid features:
Dro
Token Dro
Feature activation+0.000
Top resid features:
z
Tokenz
Feature activation+0.000
Top resid features:
d
Tokend
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
city
Token city
Feature activation+0.006
Top resid features:
to
Token to
Feature activation+0.011
Top resid features:
reduce
Token reduce
Feature activation-0.048
Top resid features:
the
Token the
Feature activation-0.007
Top resid features:
rates
Token rates
Feature activation+1.561
Top resid features:
35
Token 35
Feature activation-0.106
Top resid features:
%,
Token%,
Feature activation+0.000
Top resid features:
though
Token though
Feature activation+0.000
Top resid features:
Washington
Token Washington
Feature activation+0.000
Top resid features:
disputes
Token disputes
Feature activation+0.000
Top resid features:
is
Token is
Feature activation-0.036
Top resid features:
absolute
Token absolute
Feature activation-0.006
Top resid features:
--
Token--
Feature activation-0.028
Top resid features:
a
Tokena
Feature activation+0.003
Top resid features:
key
Token key
Feature activation+0.047
Top resid features:
distinction
Token distinction
Feature activation+1.493
Top resid features:
Kay
Token Kay
Feature activation-0.042
Top resid features:
e
Tokene
Feature activation+0.000
Top resid features:
identified
Token identified
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.000
Top resid features:
an
Token an
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.003
Top resid features:
sentences
Token sentences
Feature activation+0.058
Top resid features:
,
Token,
Feature activation-0.089
Top resid features:
a
Token a
Feature activation+0.008
Top resid features:
palpable
Token palpable
Feature activation+0.025
Top resid features:
excitement
Token excitement
Feature activation+1.460
Top resid features:
crack
Token crack
Feature activation-0.019
Top resid features:
ling
Tokenling
Feature activation+0.000
Top resid features:
over
Token over
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
phone
Token phone
Feature activation+0.000
Top resid features:
Lab
TokenLab
Feature activation-0.006
Top resid features:
rador
Tokenrador
Feature activation-0.000
Top resid features:
added
Token added
Feature activation-0.024
Top resid features:
that
Token that
Feature activation-0.032
Top resid features:
the
Token the
Feature activation-0.012
Top resid features:
frustration
Token frustration
Feature activation+1.481
Top resid features:
building
Token building
Feature activation-0.052
Top resid features:
up
Token up
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
because
Token because
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
shot
Token shot
Feature activation-0.022
Top resid features:
;
Token;
Feature activation-0.023
Top resid features:
I
Token I
Feature activation+0.023
Top resid features:
have
Token have
Feature activation-0.039
Top resid features:
more
Token more
Feature activation-0.013
Top resid features:
confidence
Token confidence
Feature activation+1.586
Top resid features:
shooting
Token shooting
Feature activation-0.093
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
ball
Token ball
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
general
Token general
Feature activation+0.000
Top resid features:
aj
Tokenaj
Feature activation-0.000
Top resid features:
l
Tokenl
Feature activation-0.003
Top resid features:
ovic
Tokenovic
Feature activation-0.000
Top resid features:
made
Token made
Feature activation-0.037
Top resid features:
wholesale
Token wholesale
Feature activation-0.003
Top resid features:
changes
Token changes
Feature activation+1.531
Top resid features:
for
Token for
Feature activation-0.105
Top resid features:
his
Token his
Feature activation+0.000
Top resid features:
first
Token first
Feature activation+0.000
Top resid features:
game
Token game
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.03

Head 2: 0.06

Head 3: 0.04

Head 4: 0.02

Head 5: 0.03

Head 6: 0.07

Head 7: 0.14

Head 8: 0.45

Head 9: 0.04

Head 10: 0.05

Head 11: 0.06

Positive logits

tesy1.26

etheless1.26

gently1.20

rily1.15

ersed1.12

oss1.10

ogether1.08

<-1.05

rid1.05

Cameron1.04

welcomed1.04

Erit1.03

isively1.03

Stern1.03

Cue1.02

xxxxxxxx1.01

spor1.00

icularly1.00

Rouhani1.00

Stock0.99

Negative logits

ドラ-1.42

ometric-1.30

bip-1.21

ゴン-1.20

arial-1.18

Divinity-1.14

Winged-1.08

enium-1.07

uma-1.06

axy-1.04

-1.04

Amendments-1.03

advertising-1.01

grain-1.00

FTWARE-0.99

Substance-0.98

artisan-0.98

phi-0.97

umatic-0.97

Platform-0.96

INTERVAL 1.566 - 1.740
CONTAINS 0.000%

albums
Token albums
Feature activation+0.000
,
Token,
Feature activation+0.000
with
Token with
Feature activation+0.000
selected
Token selected
Feature activation+0.000
information
Token information
Feature activation+0.000
Title
Token Title
Feature activation+1.740
Album
Token Album
Feature activation+0.966
details
Token details
Feature activation+0.018
A
Token A
Feature activation+0.620
Picture
Token Picture
Feature activation+0.000
Perfect
Token Perfect
Feature activation+0.000

INTERVAL 1.392 - 1.566
CONTAINS 0.000%

to
Token to
Feature activation+0.000
answer
Token answer
Feature activation+0.000
dozens
Token dozens
Feature activation+0.000
of
Token of
Feature activation+0.000
questions
Token questions
Feature activation+0.000
cert
Token cert
Feature activation+1.465
ifying
Tokenifying
Feature activation+0.437
her
Token her
Feature activation+0.000
background
Token background
Feature activation+0.000
information
Token information
Feature activation+0.052
.
Token.
Feature activation+0.000
have
Token have
Feature activation+0.000
to
Token to
Feature activation+0.000
file
Token file
Feature activation+0.000
similar
Token similar
Feature activation+0.000
reports
Token reports
Feature activation+0.000
election
Token election
Feature activation+1.525
-
Token-
Feature activation+1.060
cycle
Tokencycle
Feature activation+0.767
after
Token after
Feature activation+0.000
election
Token election
Feature activation+0.000
-
Token-
Feature activation+0.000
short
Token short
Feature activation+0.000
-
Token-
Feature activation+0.000
term
Tokenterm
Feature activation+0.081
emissions
Token emissions
Feature activation+0.000
cuts
Token cuts
Feature activation+0.000
European
Token European
Feature activation+1.473
leaders
Token leaders
Feature activation+0.800
had
Token had
Feature activation+0.000
demanded
Token demanded
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000

INTERVAL 1.218 - 1.392
CONTAINS 0.001%

the
Token the
Feature activation+0.000
extra
Token extra
Feature activation+0.000
52
Token 52
Feature activation+0.000
%
Token%
Feature activation+0.000
increase
Token increase
Feature activation+0.000
AoE
Token AoE
Feature activation+1.389
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
cycl
Token cycl
Feature activation+0.000
one
Tokenone
Feature activation+0.000
,
Token,
Feature activation+0.000
pricing
Token pricing
Feature activation+0.000
refers
Token refers
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
prices
Token prices
Feature activation+0.000
multinational
Token multinational
Feature activation+1.279
s
Tokens
Feature activation+0.927
set
Token set
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
charge
Token charge
Feature activation+0.000
are
Token are
Feature activation+0.000
just
Token just
Feature activation+0.000
having
Token having
Feature activation+0.000
a
Token a
Feature activation+0.000
blast
Token blast
Feature activation+0.000
playing
Token playing
Feature activation+1.254
music
Token music
Feature activation+0.000
together
Token together
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
Asked
TokenAsked
Feature activation+0.000
about
Token about
Feature activation+0.000
the
Token the
Feature activation+0.000
legal
Token legal
Feature activation+0.000
risks
Token risks
Feature activation+0.000
contractors
Token contractors
Feature activation+1.275
could
Token could
Feature activation+0.210
face
Token face
Feature activation+0.000
from
Token from
Feature activation+0.000
Canada
Token Canada
Feature activation+0.000
,
Token,
Feature activation+0.000
issued
Token issued
Feature activation+0.000
a
Token a
Feature activation+0.000
brief
Token brief
Feature activation+0.000
statement
Token statement
Feature activation+0.000
commemor
Token commemor
Feature activation+1.209
ating
Tokenating
Feature activation+1.327
Dro
Token Dro
Feature activation+0.000
z
Tokenz
Feature activation+0.000
d
Tokend
Feature activation+0.000
ov
Tokenov
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000

INTERVAL 1.044 - 1.218
CONTAINS 0.004%

PL
TokenPL
Feature activation+0.000
rights
Token rights
Feature activation+0.000
supports
Token supports
Feature activation+0.089
claims
Token claims
Feature activation+0.384
Fox
Token Fox
Feature activation+1.386
Sports
Token Sports
Feature activation+1.069
had
Token had
Feature activation+0.387
been
Token been
Feature activation+0.218
compl
Token compl
Feature activation+0.000
acent
Tokenacent
Feature activation+0.000
and
Token and
Feature activation+0.000
marketplace
Token marketplace
Feature activation+0.000
and
Token and
Feature activation+0.000
not
Token not
Feature activation+0.000
making
Token making
Feature activation+0.000
work
Token work
Feature activation+0.000
anybody
Token anybody
Feature activation+1.108
would
Token would
Feature activation+0.000
call
Token call
Feature activation+0.000
fresh
Token fresh
Feature activation+0.000
.
Token.
Feature activation+0.000
(
Token (
Feature activation+0.000
any
Token any
Feature activation+0.000
dis
Token dis
Feature activation+0.000
com
Tokencom
Feature activation+0.434
forts
Tokenforts
Feature activation+0.000
she
Token she
Feature activation+0.668
went
Token went
Feature activation+1.120
through
Token through
Feature activation+0.216
.
Token.
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
You
TokenYou
Feature activation+0.000
,
Token,
Feature activation+0.000
issued
Token issued
Feature activation+0.000
a
Token a
Feature activation+0.000
brief
Token brief
Feature activation+0.000
statement
Token statement
Feature activation+0.000
commemor
Token commemor
Feature activation+1.209
ating
Tokenating
Feature activation+1.327
Dro
Token Dro
Feature activation+0.000
z
Tokenz
Feature activation+0.000
d
Tokend
Feature activation+0.000
ov
Tokenov
Feature activation+0.000
absolute
Token absolute
Feature activation+0.000
--
Token--
Feature activation+0.000
a
Tokena
Feature activation+0.000
key
Token key
Feature activation+0.000
distinction
Token distinction
Feature activation+0.014
Kay
Token Kay
Feature activation+1.201
e
Tokene
Feature activation+0.816
identified
Token identified
Feature activation+0.000
as
Token as
Feature activation+0.000
an
Token an
Feature activation+0.000
independent
Token independent
Feature activation+0.000

INTERVAL 0.870 - 1.044
CONTAINS 0.007%

if
Token if
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+0.000
zero
Token zero
Feature activation+0.000
shots
Token shots
Feature activation+0.000
remaining
Token remaining
Feature activation+0.979
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
ret
Token ret
Feature activation+0.000
icle
Tokenicle
Feature activation+0.000
has
Token has
Feature activation+0.000
those
Token those
Feature activation+0.000
cuts
Token cuts
Feature activation+0.000
include
Token include
Feature activation+0.000
25
Token 25
Feature activation+0.000
,
Token,
Feature activation+0.000
000
Token000
Feature activation+0.989
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
oil
Token oil
Feature activation+0.000
services
Token services
Feature activation+0.000
sector
Token sector
Feature activation+0.000
lock
Token lock
Feature activation+0.000
ers
Tokeners
Feature activation+0.000
with
Token with
Feature activation+0.000
no
Token no
Feature activation+0.000
assurances
Token assurances
Feature activation+0.000
it
Token it
Feature activation+0.929
would
Token would
Feature activation+0.483
remain
Token remain
Feature activation+0.000
there
Token there
Feature activation+0.000
throughout
Token throughout
Feature activation+0.000
the
Token the
Feature activation+0.000
refers
Token refers
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
prices
Token prices
Feature activation+0.000
multinational
Token multinational
Feature activation+1.279
s
Tokens
Feature activation+0.927
set
Token set
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
charge
Token charge
Feature activation+0.000
their
Token their
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
if
Token if
Feature activation+0.000
the
Token the
Feature activation+0.000
improvement
Token improvement
Feature activation+0.000
between
Token between
Feature activation+0.913
two
Token two
Feature activation+0.000
measurements
Token measurements
Feature activation+0.000
is
Token is
Feature activation+0.000
far
Token far
Feature activation+0.000
larger
Token larger
Feature activation+0.000

INTERVAL 0.696 - 0.870
CONTAINS 0.018%

confront
Token confront
Feature activation+0.000
any
Token any
Feature activation+0.000
great
Token great
Feature activation+0.000
power
Token power
Feature activation+0.000
competition
Token competition
Feature activation+0.566
more
Token more
Feature activation+0.861
realistically
Token realistically
Feature activation+0.563
.
Token.
Feature activation+0.000
B
Token B
Feature activation+0.000
asing
Tokenasing
Feature activation+0.000
this
Token this
Feature activation+0.000
As
Token As
Feature activation+0.000
younger
Token younger
Feature activation+0.000
teachers
Token teachers
Feature activation+0.000
gain
Token gain
Feature activation+0.000
experience
Token experience
Feature activation+0.000
their
Token their
Feature activation+0.737
classroom
Token classroom
Feature activation+0.358
performance
Token performance
Feature activation+0.296
should
Token should
Feature activation+0.000
improve
Token improve
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+0.000
ambitious
Token ambitious
Feature activation+0.000
plans
Token plans
Feature activation+0.000
pay
Token pay
Feature activation+0.723
off
Token off
Feature activation+0.340
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
Google
Token Google
Feature activation+0.000
of
Token of
Feature activation+0.000
ado
Tokenado
Feature activation+0.000
did
Token did
Feature activation+0.000
a
Token a
Feature activation+0.000
good
Token good
Feature activation+0.000
job
Token job
Feature activation+0.000
finishing
Token finishing
Feature activation+0.719
her
Token her
Feature activation+0.000
high
Token high
Feature activation+0.000
school
Token school
Feature activation+0.000
diploma
Token diploma
Feature activation+0.044
while
Token while
Feature activation+0.000
the
Token the
Feature activation+0.000
sort
Token sort
Feature activation+0.000
of
Token of
Feature activation+0.000
orderly
Token orderly
Feature activation+0.000
withdrawal
Token withdrawal
Feature activation+0.000
provided
Token provided
Feature activation+0.704
for
Token for
Feature activation+0.143
by
Token by
Feature activation+0.000
Article
Token Article
Feature activation+0.000
50
Token 50
Feature activation+0.000
;
Token;
Feature activation+0.000

INTERVAL 0.522 - 0.696
CONTAINS 0.039%

asked
Token asked
Feature activation+0.000
if
Token if
Feature activation+0.000
they
Token they
Feature activation+0.000
have
Token have
Feature activation+0.000
difficulty
Token difficulty
Feature activation+0.000
speaking
Token speaking
Feature activation+0.581
in
Token in
Feature activation+0.000
front
Token front
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
group
Token group
Feature activation+0.000
to
Token to
Feature activation+0.000
hospitals
Token hospitals
Feature activation+0.000
for
Token for
Feature activation+0.000
their
Token their
Feature activation+0.000
feedback
Token feedback
Feature activation+0.000
last
Token last
Feature activation+0.669
week
Token week
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Article
TokenArticle
Feature activation+0.000
the
Token the
Feature activation+0.000
church
Token church
Feature activation+0.000
will
Token will
Feature activation+0.000
allow
Token allow
Feature activation+0.000
training
Token training
Feature activation+0.000
500
Token 500
Feature activation+0.616
pastors
Token pastors
Feature activation+0.386
,
Token,
Feature activation+0.000
servants
Token servants
Feature activation+0.000
and
Token and
Feature activation+0.000
priests
Token priests
Feature activation+0.000
s
Tokens
Feature activation+0.000
plenty
Token plenty
Feature activation+0.000
of
Token of
Feature activation+0.000
reasoned
Token reasoned
Feature activation+0.000
debate
Token debate
Feature activation+0.000
out
Token out
Feature activation+0.630
there
Token there
Feature activation+0.000
,
Token,
Feature activation+0.000
of
Token of
Feature activation+0.000
course
Token course
Feature activation+0.000
,
Token,
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
supply
Token supply
Feature activation+0.000
issues
Token issues
Feature activation+0.053
competing
Token competing
Feature activation+0.661
for
Token for
Feature activation+0.000
fibre
Token fibre
Feature activation+0.000
as
Token as
Feature activation+0.000
saw
Token saw
Feature activation+0.000
m
Tokenm
Feature activation+0.000

INTERVAL 0.348 - 0.522
CONTAINS 0.103%

Ċ
TokenĊ
Feature activation+0.000
Water
TokenWater
Feature activation+0.000
as
Token as
Feature activation+0.000
needed
Token needed
Feature activation+0.000
Instructions
Token Instructions
Feature activation+0.000
Add
Token Add
Feature activation+0.521
the
Token the
Feature activation+0.000
spinach
Token spinach
Feature activation+0.000
to
Token to
Feature activation+0.000
a
Token a
Feature activation+0.000
large
Token large
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
p
Token p
Feature activation+0.000
inches
Tokeninches
Feature activation+0.000
of
Token of
Feature activation+0.000
sugar
Token sugar
Feature activation+0.000
1
Token1
Feature activation+0.370
/
Token/
Feature activation+0.000
4
Token4
Feature activation+0.000
teaspoon
Token teaspoon
Feature activation+0.000
red
Token red
Feature activation+0.000
pepper
Token pepper
Feature activation+0.000
spy
Token spy
Feature activation+0.165
ware
Tokenware
Feature activation+0.000
or
Token or
Feature activation+0.000
other
Token other
Feature activation+0.000
tools
Token tools
Feature activation+0.000
intended
Token intended
Feature activation+0.465
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
used
Token used
Feature activation+0.000
for
Token for
Feature activation+0.000
unlawful
Token unlawful
Feature activation+0.000
or
Token or
Feature activation+0.000
pledging
Token pledging
Feature activation+0.000
to
Token to
Feature activation+0.000
cut
Token cut
Feature activation+0.000
taxes
Token taxes
Feature activation+0.000
in
Token in
Feature activation+0.450
his
Token his
Feature activation+0.000
next
Token next
Feature activation+0.000
budget
Token budget
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
people
Token people
Feature activation+0.000
with
Token with
Feature activation+0.000
attention
Token attention
Feature activation+0.000
deficit
Token deficit
Feature activation+0.000
hyper
Token hyper
Feature activation+0.000
activity
Tokenactivity
Feature activation+0.423
disorder
Token disorder
Feature activation+0.071
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
both
Token both
Feature activation+0.000
have
Token have
Feature activation+0.000

INTERVAL 0.174 - 0.348
CONTAINS 0.283%

,
Token,
Feature activation+0.000
fought
Token fought
Feature activation+0.000
to
Token to
Feature activation+0.000
protect
Token protect
Feature activation+0.000
spending
Token spending
Feature activation+0.000
it
Token it
Feature activation+0.176
sees
Token sees
Feature activation+0.000
as
Token as
Feature activation+0.000
essential
Token essential
Feature activation+0.000
to
Token to
Feature activation+0.000
boost
Token boost
Feature activation+0.000
concluded
Token concluded
Feature activation+0.000
and
Token and
Feature activation+0.000
what
Token what
Feature activation+0.000
a
Token a
Feature activation+0.000
ride
Token ride
Feature activation+0.000
it
Token it
Feature activation+0.248
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
been
Token been
Feature activation+0.000
!
Token!
Feature activation+0.000
C
TokenC
Feature activation+0.000
plan
Token plan
Feature activation+0.000
,
Token,
Feature activation+0.000
one
Token one
Feature activation+0.000
proposal
Token proposal
Feature activation+0.000
stood
Token stood
Feature activation+0.223
out
Token out
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
transportation
Token transportation
Feature activation+0.000
world
Token world
Feature activation+0.000
Organ
TokenOrgan
Feature activation+0.000
izer
Tokenizer
Feature activation+0.000
Att
Token Att
Feature activation+0.000
endants
Tokenendants
Feature activation+0.000
Events
Token Events
Feature activation+0.000
Registration
Token Registration
Feature activation+0.300
Attend
Token Attend
Feature activation+0.000
ance
Tokenance
Feature activation+0.000
Settings
Token Settings
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
often
Token often
Feature activation+0.000
have
Token have
Feature activation+0.000
a
Token a
Feature activation+0.000
hard
Token hard
Feature activation+0.000
time
Token time
Feature activation+0.000
understanding
Token understanding
Feature activation+0.192
this
Token this
Feature activation+0.143
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
U
Token U
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 0.000 - 0.174
CONTAINS 99.544%

their
Token their
Feature activation+0.000
troops
Token troops
Feature activation+0.000
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
withdrawn
Token withdrawn
Feature activation+0.000
from
Token from
Feature activation+0.000
Afghanistan
Token Afghanistan
Feature activation+0.000
,
Token,
Feature activation+0.000
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
so
Token so
Feature activation+0.000
statement
Token statement
Feature activation+0.000
.
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
probably
Token probably
Feature activation+0.000
matters
Token matters
Feature activation+0.000
little
Token little
Feature activation+0.000
to
Token to
Feature activation+0.000
people
Token people
Feature activation+0.000
comes
Token comes
Feature activation+0.000
just
Token just
Feature activation+0.000
a
Token a
Feature activation+0.000
week
Token week
Feature activation+0.000
after
Token after
Feature activation+0.000
she
Token she
Feature activation+0.000
posted
Token posted
Feature activation+0.000
two
Token two
Feature activation+0.000
videos
Token videos
Feature activation+0.000
on
Token on
Feature activation+0.000
YouTube
Token YouTube
Feature activation+0.000
Building
Token Building
Feature activation+0.000
inspectors
Token inspectors
Feature activation+0.000
will
Token will
Feature activation+0.000
be
Token be
Feature activation+0.000
out
Token out
Feature activation+0.000
sourced
Token sourced
Feature activation+0.000
as
Token as
Feature activation+0.000
will
Token will
Feature activation+0.000
city
Token city
Feature activation+0.000
plan
Token plan
Feature activation+0.000
check
Token check
Feature activation+0.000
/
Token/
Feature activation+0.000
--
Token --
Feature activation+0.000
Cedar
Token Cedar
Feature activation+0.000
Grove
Token Grove
Feature activation+0.000
Comp
Token Comp
Feature activation+0.000
ost
Tokenost
Feature activation+0.000
ing
Tokening
Feature activation+0.000
announced
Token announced
Feature activation+0.000
today
Token today
Feature activation+0.000
that
Token that
Feature activation+0.000
it
Token it
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 H3.8: (feature 10698)

TOP ACTIVATIONS
MAX = 2.200

--
Token --
Feature activation+0.000
about
Token about
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.146
first
Token first
Feature activation+2.200
large
Token large
Feature activation+1.054
open
Token open
Feature activation+0.279
-
Token-
Feature activation+0.000
o
Tokeno
Feature activation+0.000
cean
Tokencean
Feature activation+0.118
also
Token also
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
time
Token time
Feature activation+0.000
a
Token a
Feature activation+1.600
majority
Token majority
Feature activation+2.188
of
Token of
Feature activation+0.773
Democrats
Token Democrats
Feature activation+0.000
sided
Token sided
Feature activation+0.454
against
Token against
Feature activation+0.000
President
Token President
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
moment
Token moment
Feature activation+0.000
Dr
Token Dr
Feature activation+2.056
.
Token.
Feature activation+0.569
Gun
Token Gun
Feature activation+0.566
by
Tokenby
Feature activation+1.063
first
Token first
Feature activation+1.176
held
Token held
Feature activation+0.000
a
Token a
Feature activation+0.000
boost
Token boost
Feature activation+0.000
-
Token -
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
from
Token from
Feature activation+1.961
a
Token a
Feature activation+0.000
top
Token top
Feature activation+0.000
NASA
Token NASA
Feature activation+0.000
bo
Token bo
Feature activation+0.239
ff
Tokenff
Feature activation+0.000
global
Token global
Feature activation+0.000
economies
Token economies
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
its
Token its
Feature activation+1.953
cent
Token cent
Feature activation+1.112
enary
Tokenenary
Feature activation+0.000
comes
Token comes
Feature activation+0.118
around
Token around
Feature activation+0.000
in
Token in
Feature activation+0.000
same
Token same
Feature activation+0.000
practice
Token practice
Feature activation+0.000
âĢ¦
Token âĢ¦
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
engaged
Token engaged
Feature activation+1.806
in
Token in
Feature activation+0.000
by
Token by
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
."
Token."
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
st
Token st
Feature activation+1.236
inging
Tokeninging
Feature activation+1.791
ev
Token ev
Feature activation+0.000
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+1.739
of
Token of
Feature activation+0.114
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
minute
Token minute
Feature activation+0.000
an
Token an
Feature activation+1.777
Apple
Token Apple
Feature activation+0.771
store
Token store
Feature activation+1.047
appears
Token appears
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
st
Token st
Feature activation+1.236
inging
Tokeninging
Feature activation+1.791
ev
Token ev
Feature activation+0.000
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+1.739
of
Token of
Feature activation+0.114
Dr
Token Dr
Feature activation+0.000
.
Token.
Feature activation+0.000
Mann
Token Mann
Feature activation+0.000
on
Token on
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
moment
Token moment
Feature activation+0.000
his
Token his
Feature activation+1.690
father
Token father
Feature activation+1.286
cheated
Token cheated
Feature activation+0.000
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
woman
Token woman
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
very
Token very
Feature activation+0.000
time
Token time
Feature activation+0.000
they
Token they
Feature activation+0.452
made
Token made
Feature activation+1.671
that
Token that
Feature activation+0.000
invention
Token invention
Feature activation+0.000
.
Token.
Feature activation+0.000
He
Token He
Feature activation+0.000
worked
Token worked
Feature activation+0.000
seventeen
Token seventeen
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
she
Token she
Feature activation+0.341
claims
Token claims
Feature activation+1.663
she
Token she
Feature activation+0.000
slept
Token slept
Feature activation+0.000
with
Token with
Feature activation+0.000
Andrew
Token Andrew
Feature activation+0.000
.
Token.
Feature activation+0.000
on
Token on
Feature activation+0.000
Sunday
Token Sunday
Feature activation+0.000
,
Token,
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
from
Token from
Feature activation+1.637
Virginia
Token Virginia
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
cold
Token cold
Feature activation+0.000
same
Token same
Feature activation+0.000
area
Token area
Feature activation+0.000
,
Token,
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
about
Token about
Feature activation+1.628
a
Token a
Feature activation+0.327
man
Token man
Feature activation+0.000
running
Token running
Feature activation+0.000
in
Token in
Feature activation+0.000
traffic
Token traffic
Feature activation+0.000
was
Token was
Feature activation+0.000
also
Token also
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
time
Token time
Feature activation+0.000
a
Token a
Feature activation+1.600
majority
Token majority
Feature activation+2.188
of
Token of
Feature activation+0.773
Democrats
Token Democrats
Feature activation+0.000
sided
Token sided
Feature activation+0.454
against
Token against
Feature activation+0.000
Around
Token Around
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
Esper
Token Esper
Feature activation+1.245
anto
Tokenanto
Feature activation+1.574
made
Token made
Feature activation+0.692
its
Token its
Feature activation+0.000
way
Token way
Feature activation+0.000
to
Token to
Feature activation+0.000
East
Token East
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+0.920
Argent
Token Argent
Feature activation+0.800
inian
Tokeninian
Feature activation+1.542
pont
Token pont
Feature activation+1.175
iff
Tokeniff
Feature activation+0.847
has
Token has
Feature activation+0.000
met
Token met
Feature activation+0.788
with
Token with
Feature activation+0.000
is
Token is
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
repeal
Token repeal
Feature activation+1.511
passes
Token passes
Feature activation+0.893
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Updated
TokenUpdated
Feature activation+0.000
ure
Tokenure
Feature activation+0.000
coverage
Token coverage
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.501
enrollment
Token enrollment
Feature activation+0.648
period
Token period
Feature activation+0.000
is
Token is
Feature activation+0.000
over
Token over
Feature activation+0.000
.
Token.
Feature activation+0.000
allegations
Token allegations
Feature activation+0.000
stem
Token stem
Feature activation+0.000
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
H
Token H
Feature activation+1.496
aster
Tokenaster
Feature activation+0.862
t
Tokent
Feature activation+1.104
worked
Token worked
Feature activation+0.000
as
Token as
Feature activation+0.000
a
Token a
Feature activation+0.000

Top DFA by src position
MAX = 3.062

years
Token years
Feature activation+0.001
Top resid features:
ago
Token ago
Feature activation-0.001
Top resid features:
--
Token --
Feature activation+0.038
Top resid features:
about
Token about
Feature activation-0.216
Top resid features:
the
Token the
Feature activation+0.005
Top resid features:
time
Token time
Feature activation+3.062
Top resid features:
the
Token the
Feature activation+0.134
Top resid features:
first
Token first
Feature activation+0.078
Top resid features:
large
Token large
Feature activation+0.000
Top resid features:
open
Token open
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
It
TokenIt
Feature activation+0.002
Top resid features:
was
Token was
Feature activation+0.032
Top resid features:
also
Token also
Feature activation+0.040
Top resid features:
the
Token the
Feature activation+0.057
Top resid features:
first
Token first
Feature activation+0.027
Top resid features:
time
Token time
Feature activation+2.718
Top resid features:
a
Token a
Feature activation+0.243
Top resid features:
majority
Token majority
Feature activation-0.045
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
Democrats
Token Democrats
Feature activation+0.000
Top resid features:
sided
Token sided
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.002
Top resid features:
Ŀ
TokenĿ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation+0.006
Top resid features:
Ċ
TokenĊ
Feature activation+0.027
Top resid features:
The
TokenThe
Feature activation+0.095
Top resid features:
moment
Token moment
Feature activation+2.623
Top resid features:
Dr
Token Dr
Feature activation+0.085
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Gun
Token Gun
Feature activation+0.000
Top resid features:
by
Tokenby
Feature activation+0.000
Top resid features:
first
Token first
Feature activation+0.000
Top resid features:
received
Token received
Feature activation+0.073
Top resid features:
a
Token a
Feature activation+0.013
Top resid features:
boost
Token boost
Feature activation-0.006
Top resid features:
-
Token -
Feature activation+0.007
Top resid features:
this
Token this
Feature activation+0.217
Top resid features:
time
Token time
Feature activation+2.433
Top resid features:
from
Token from
Feature activation+0.120
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
top
Token top
Feature activation+0.000
Top resid features:
NASA
Token NASA
Feature activation+0.000
Top resid features:
bo
Token bo
Feature activation+0.000
Top resid features:
10
Token 10
Feature activation-0.004
Top resid features:
global
Token global
Feature activation-0.017
Top resid features:
economies
Token economies
Feature activation-0.020
Top resid features:
by
Token by
Feature activation-0.041
Top resid features:
the
Token the
Feature activation+0.010
Top resid features:
time
Token time
Feature activation+2.766
Top resid features:
its
Token its
Feature activation+0.090
Top resid features:
cent
Token cent
Feature activation+0.000
Top resid features:
enary
Tokenenary
Feature activation+0.000
Top resid features:
comes
Token comes
Feature activation+0.000
Top resid features:
around
Token around
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.002
Top resid features:
same
Token same
Feature activation-0.002
Top resid features:
practice
Token practice
Feature activation+0.010
Top resid features:
âĢ¦
Token âĢ¦
Feature activation+0.016
Top resid features:
this
Token this
Feature activation+0.175
Top resid features:
time
Token time
Feature activation+2.289
Top resid features:
engaged
Token engaged
Feature activation+0.127
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
Americans
Token Americans
Feature activation+0.000
Top resid features:
."
Token."
Feature activation+0.000
Top resid features:
funny
Token funny
Feature activation+0.000
Top resid features:
but
Token but
Feature activation-0.000
Top resid features:
at
Token at
Feature activation-0.047
Top resid features:
the
Token the
Feature activation+0.002
Top resid features:
same
Token same
Feature activation-0.026
Top resid features:
time
Token time
Feature activation+2.714
Top resid features:
st
Token st
Feature activation+0.061
Top resid features:
inging
Tokeninging
Feature activation-0.099
Top resid features:
ev
Token ev
Feature activation+0.000
Top resid features:
is
Tokenis
Feature activation+0.000
Top resid features:
cer
Tokencer
Feature activation+0.000
Top resid features:
ox
Tokenox
Feature activation+0.008
Top resid features:
.
Token.
Feature activation+0.009
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation+0.017
Top resid features:
The
TokenThe
Feature activation+0.069
Top resid features:
minute
Token minute
Feature activation+2.337
Top resid features:
an
Token an
Feature activation+0.278
Top resid features:
Apple
Token Apple
Feature activation+0.000
Top resid features:
store
Token store
Feature activation+0.000
Top resid features:
appears
Token appears
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
funny
Token funny
Feature activation+0.003
Top resid features:
but
Token but
Feature activation+0.008
Top resid features:
at
Token at
Feature activation-0.009
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
same
Token same
Feature activation-0.111
Top resid features:
time
Token time
Feature activation+2.980
Top resid features:
st
Token st
Feature activation+0.005
Top resid features:
inging
Tokeninging
Feature activation-0.135
Top resid features:
ev
Token ev
Feature activation+0.010
Top resid features:
is
Tokenis
Feature activation+0.002
Top resid features:
cer
Tokencer
Feature activation-0.047
Top resid features:
said
Token said
Feature activation-0.005
Top resid features:
.
Token.
Feature activation+0.023
Top resid features:
Ċ
TokenĊ
Feature activation-0.002
Top resid features:
Ċ
TokenĊ
Feature activation+0.004
Top resid features:
The
TokenThe
Feature activation+0.044
Top resid features:
moment
Token moment
Feature activation+2.365
Top resid features:
his
Token his
Feature activation+0.086
Top resid features:
father
Token father
Feature activation+0.000
Top resid features:
cheated
Token cheated
Feature activation+0.000
Top resid features:
on
Token on
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.003
Top resid features:
them
Token them
Feature activation-0.019
Top resid features:
at
Token at
Feature activation+0.007
Top resid features:
the
Token the
Feature activation+0.040
Top resid features:
very
Token very
Feature activation-0.009
Top resid features:
time
Token time
Feature activation+2.440
Top resid features:
they
Token they
Feature activation+0.137
Top resid features:
made
Token made
Feature activation-0.065
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
invention
Token invention
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Roberts
Token Roberts
Feature activation-0.006
Top resid features:
was
Token was
Feature activation+0.024
Top resid features:
seventeen
Token seventeen
Feature activation-0.101
Top resid features:
at
Token at
Feature activation+0.003
Top resid features:
the
Token the
Feature activation+0.033
Top resid features:
time
Token time
Feature activation+2.270
Top resid features:
she
Token she
Feature activation+0.169
Top resid features:
claims
Token claims
Feature activation-0.025
Top resid features:
she
Token she
Feature activation+0.000
Top resid features:
slept
Token slept
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
Station
Token Station
Feature activation+0.003
Top resid features:
on
Token on
Feature activation+0.004
Top resid features:
Sunday
Token Sunday
Feature activation+0.008
Top resid features:
,
Token,
Feature activation+0.027
Top resid features:
this
Token this
Feature activation+0.039
Top resid features:
time
Token time
Feature activation+2.327
Top resid features:
from
Token from
Feature activation+0.107
Top resid features:
Virginia
Token Virginia
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ļ
TokenĻ
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.003
Top resid features:
same
Token same
Feature activation-0.005
Top resid features:
area
Token area
Feature activation+0.008
Top resid features:
,
Token,
Feature activation+0.080
Top resid features:
this
Token this
Feature activation+0.073
Top resid features:
time
Token time
Feature activation+2.303
Top resid features:
about
Token about
Feature activation+0.060
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
man
Token man
Feature activation+0.000
Top resid features:
running
Token running
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
It
TokenIt
Feature activation-0.019
Top resid features:
was
Token was
Feature activation+0.093
Top resid features:
also
Token also
Feature activation+0.107
Top resid features:
the
Token the
Feature activation+0.047
Top resid features:
first
Token first
Feature activation-0.061
Top resid features:
time
Token time
Feature activation+2.182
Top resid features:
a
Token a
Feature activation+0.179
Top resid features:
majority
Token majority
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
Democrats
Token Democrats
Feature activation+0.000
Top resid features:
sided
Token sided
Feature activation+0.000
Top resid features:
earlier
Token earlier
Feature activation+0.004
Top resid features:
.
Token.
Feature activation+0.100
Top resid features:
Around
Token Around
Feature activation-0.036
Top resid features:
the
Token the
Feature activation+0.071
Top resid features:
same
Token same
Feature activation+0.022
Top resid features:
time
Token time
Feature activation+2.198
Top resid features:
Esper
Token Esper
Feature activation-0.005
Top resid features:
anto
Tokenanto
Feature activation+0.086
Top resid features:
made
Token made
Feature activation+0.000
Top resid features:
its
Token its
Feature activation+0.000
Top resid features:
way
Token way
Feature activation+0.000
Top resid features:
because
Token because
Feature activation-0.002
Top resid features:
it
Token it
Feature activation+0.006
Top resid features:
represents
Token represents
Feature activation-0.018
Top resid features:
the
Token the
Feature activation+0.034
Top resid features:
first
Token first
Feature activation+0.028
Top resid features:
time
Token time
Feature activation+2.048
Top resid features:
the
Token the
Feature activation+0.308
Top resid features:
Argent
Token Argent
Feature activation+0.008
Top resid features:
inian
Tokeninian
Feature activation+0.005
Top resid features:
pont
Token pont
Feature activation+0.000
Top resid features:
iff
Tokeniff
Feature activation+0.000
Top resid features:
replacement
Token replacement
Feature activation+0.023
Top resid features:
is
Token is
Feature activation+0.054
Top resid features:
at
Token at
Feature activation-0.005
Top resid features:
the
Token the
Feature activation+0.018
Top resid features:
same
Token same
Feature activation+0.005
Top resid features:
time
Token time
Feature activation+2.003
Top resid features:
repeal
Token repeal
Feature activation+0.151
Top resid features:
passes
Token passes
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
NS
TokenNS
Feature activation-0.005
Top resid features:
ure
Tokenure
Feature activation-0.006
Top resid features:
coverage
Token coverage
Feature activation-0.013
Top resid features:
by
Token by
Feature activation-0.052
Top resid features:
the
Token the
Feature activation-0.004
Top resid features:
time
Token time
Feature activation+2.350
Top resid features:
the
Token the
Feature activation+0.075
Top resid features:
enrollment
Token enrollment
Feature activation+0.000
Top resid features:
period
Token period
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
over
Token over
Feature activation+0.000
Top resid features:
The
TokenThe
Feature activation+0.015
Top resid features:
allegations
Token allegations
Feature activation-0.028
Top resid features:
stem
Token stem
Feature activation-0.057
Top resid features:
from
Token from
Feature activation-0.054
Top resid features:
the
Token the
Feature activation+0.077
Top resid features:
time
Token time
Feature activation+2.159
Top resid features:
H
Token H
Feature activation-0.042
Top resid features:
aster
Tokenaster
Feature activation+0.000
Top resid features:
t
Tokent
Feature activation+0.000
Top resid features:
worked
Token worked
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.05

Head 2: 0.07

Head 3: 0.03

Head 4: 0.03

Head 5: 0.04

Head 6: 0.06

Head 7: 0.09

Head 8: 0.45

Head 9: 0.03

Head 10: 0.06

Head 11: 0.06

Positive logits

orously1.72

iating1.15

resumed1.13

arrived1.12

regained1.11

tained1.10

orous1.08

died1.07

Lung1.07

say1.06

azz1.06

Cele1.05

wen1.05

Hey1.04

Untitled1.04

fax1.04

inally1.04

went1.03

Greenwich1.03

gets1.03

Negative logits

Medium-1.35

temperament-1.25

-1.20

uties-1.20

iasm-1.16

catentry-1.15

enlarge-1.14

;;;;;;;;;;;;-1.14

wagon-1.13

────-1.10

──-1.10

largeDownload-1.10

esm-1.09

apprehens-1.09

-1.08

apprehension-1.07

competitiveness-1.06

├──-1.05

upbringing-1.05

rolet-1.03

INTERVAL 1.980 - 2.200
CONTAINS 0.000%

--
Token --
Feature activation+0.000
about
Token about
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.146
first
Token first
Feature activation+2.200
large
Token large
Feature activation+1.054
open
Token open
Feature activation+0.279
-
Token-
Feature activation+0.000
o
Tokeno
Feature activation+0.000
cean
Tokencean
Feature activation+0.118
also
Token also
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
time
Token time
Feature activation+0.000
a
Token a
Feature activation+1.600
majority
Token majority
Feature activation+2.188
of
Token of
Feature activation+0.773
Democrats
Token Democrats
Feature activation+0.000
sided
Token sided
Feature activation+0.454
against
Token against
Feature activation+0.000
President
Token President
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
moment
Token moment
Feature activation+0.000
Dr
Token Dr
Feature activation+2.056
.
Token.
Feature activation+0.569
Gun
Token Gun
Feature activation+0.566
by
Tokenby
Feature activation+1.063
first
Token first
Feature activation+1.176
held
Token held
Feature activation+0.000

INTERVAL 1.760 - 1.980
CONTAINS 0.001%

at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
st
Token st
Feature activation+1.236
inging
Tokeninging
Feature activation+1.791
ev
Token ev
Feature activation+0.000
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+1.739
of
Token of
Feature activation+0.114
global
Token global
Feature activation+0.000
economies
Token economies
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
its
Token its
Feature activation+1.953
cent
Token cent
Feature activation+1.112
enary
Tokenenary
Feature activation+0.000
comes
Token comes
Feature activation+0.118
around
Token around
Feature activation+0.000
in
Token in
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
minute
Token minute
Feature activation+0.000
an
Token an
Feature activation+1.777
Apple
Token Apple
Feature activation+0.771
store
Token store
Feature activation+1.047
appears
Token appears
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
a
Token a
Feature activation+0.000
boost
Token boost
Feature activation+0.000
-
Token -
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
from
Token from
Feature activation+1.961
a
Token a
Feature activation+0.000
top
Token top
Feature activation+0.000
NASA
Token NASA
Feature activation+0.000
bo
Token bo
Feature activation+0.239
ff
Tokenff
Feature activation+0.000
same
Token same
Feature activation+0.000
practice
Token practice
Feature activation+0.000
âĢ¦
Token âĢ¦
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
engaged
Token engaged
Feature activation+1.806
in
Token in
Feature activation+0.000
by
Token by
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
."
Token."
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000

INTERVAL 1.540 - 1.760
CONTAINS 0.001%

.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
moment
Token moment
Feature activation+0.000
his
Token his
Feature activation+1.690
father
Token father
Feature activation+1.286
cheated
Token cheated
Feature activation+0.000
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
woman
Token woman
Feature activation+0.000
Around
Token Around
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
Esper
Token Esper
Feature activation+1.245
anto
Tokenanto
Feature activation+1.574
made
Token made
Feature activation+0.692
its
Token its
Feature activation+0.000
way
Token way
Feature activation+0.000
to
Token to
Feature activation+0.000
East
Token East
Feature activation+0.000
on
Token on
Feature activation+0.000
Sunday
Token Sunday
Feature activation+0.000
,
Token,
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
from
Token from
Feature activation+1.637
Virginia
Token Virginia
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
cold
Token cold
Feature activation+0.000
st
Token st
Feature activation+1.236
inging
Tokeninging
Feature activation+1.791
ev
Token ev
Feature activation+0.000
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+1.739
of
Token of
Feature activation+0.114
Dr
Token Dr
Feature activation+0.000
.
Token.
Feature activation+0.000
Mann
Token Mann
Feature activation+0.000
on
Token on
Feature activation+0.000
seventeen
Token seventeen
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
she
Token she
Feature activation+0.341
claims
Token claims
Feature activation+1.663
she
Token she
Feature activation+0.000
slept
Token slept
Feature activation+0.000
with
Token with
Feature activation+0.000
Andrew
Token Andrew
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 1.320 - 1.540
CONTAINS 0.002%

had
Token had
Feature activation+0.000
healed
Token healed
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
officers
Token officers
Feature activation+1.366
returned
Token returned
Feature activation+0.565
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
so
Token so
Feature activation+0.000
by
Token by
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
we
Token we
Feature activation+0.546
got
Token got
Feature activation+1.478
to
Token to
Feature activation+0.000
Premiere
Token Premiere
Feature activation+0.000
4
Token 4
Feature activation+0.000
,
Token,
Feature activation+0.000
there
Token there
Feature activation+0.000
allegations
Token allegations
Feature activation+0.000
stem
Token stem
Feature activation+0.000
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
H
Token H
Feature activation+1.496
aster
Tokenaster
Feature activation+0.862
t
Tokent
Feature activation+1.104
worked
Token worked
Feature activation+0.000
as
Token as
Feature activation+0.000
a
Token a
Feature activation+0.000
WA
TokenWA
Feature activation+0.000
school
Token school
Feature activation+0.000
âĢĶ
Token âĢĶ
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.346
An
Token An
Feature activation+0.388
as
Tokenas
Feature activation+0.000
Ibn
Token Ibn
Feature activation+0.000
Malik
Token Malik
Feature activation+0.000
Prep
Token Prep
Feature activation+0.000
However
Token However
Feature activation+0.000
,
Token,
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
Greene
Token Greene
Feature activation+1.447
was
Token was
Feature activation+0.040
re
Token re
Feature activation+0.000
reading
Tokenreading
Feature activation+0.000
his
Token his
Feature activation+0.000
favorite
Token favorite
Feature activation+0.000

INTERVAL 1.100 - 1.320
CONTAINS 0.002%

moment
Token moment
Feature activation+0.000
Dr
Token Dr
Feature activation+2.056
.
Token.
Feature activation+0.569
Gun
Token Gun
Feature activation+0.566
by
Tokenby
Feature activation+1.063
first
Token first
Feature activation+1.176
held
Token held
Feature activation+0.000
up
Token up
Feature activation+0.000
the
Token the
Feature activation+0.000
baby
Token baby
Feature activation+0.000
was
Token was
Feature activation+0.000
ago
Token ago
Feature activation+0.000
--
Token --
Feature activation+0.000
about
Token about
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.146
first
Token first
Feature activation+2.200
large
Token large
Feature activation+1.054
open
Token open
Feature activation+0.279
-
Token-
Feature activation+0.000
o
Tokeno
Feature activation+0.000
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
my
Token my
Feature activation+0.765
Aunt
Token Aunt
Feature activation+0.801
gently
Token gently
Feature activation+1.174
(
Token (
Feature activation+0.000
ag
Tokenag
Feature activation+0.000
gress
Tokengress
Feature activation+0.000
ively
Tokenively
Feature activation+0.000
)
Token)
Feature activation+0.000
but
Token but
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
time
Token time
Feature activation+0.000
st
Token st
Feature activation+1.236
inging
Tokeninging
Feature activation+1.791
ev
Token ev
Feature activation+0.000
is
Tokenis
Feature activation+0.000
cer
Tokencer
Feature activation+0.000
ation
Tokenation
Feature activation+1.739
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
From
TokenFrom
Feature activation+0.000
the
Token the
Feature activation+0.000
moment
Token moment
Feature activation+0.000
a
Token a
Feature activation+1.480
revert
Token revert
Feature activation+1.151
to
Token to
Feature activation+0.000
Islam
Token Islam
Feature activation+0.000
ut
Token ut
Feature activation+0.000
ters
Tokenters
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 0.880 - 1.100
CONTAINS 0.005%

n
Token n
Feature activation+0.000
appy
Tokenappy
Feature activation+0.000
change
Token change
Feature activation+0.000
the
Token the
Feature activation+0.000
moment
Token moment
Feature activation+0.000
you
Token you
Feature activation+0.993
are
Token are
Feature activation+0.725
ready
Token ready
Feature activation+0.336
to
Token to
Feature activation+0.000
leave
Token leave
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
By
TokenBy
Feature activation+0.000
this
Token this
Feature activation+0.000
time
Token time
Feature activation+0.000
next
Token next
Feature activation+0.886
week
Token week
Feature activation+0.382
Prime
Token Prime
Feature activation+0.161
Minister
Token Minister
Feature activation+0.506
-
Token-
Feature activation+0.000
elect
Tokenelect
Feature activation+0.000
The
TokenThe
Feature activation+0.000
moment
Token moment
Feature activation+0.000
Dr
Token Dr
Feature activation+2.056
.
Token.
Feature activation+0.569
Gun
Token Gun
Feature activation+0.566
by
Tokenby
Feature activation+1.063
first
Token first
Feature activation+1.176
held
Token held
Feature activation+0.000
up
Token up
Feature activation+0.000
the
Token the
Feature activation+0.000
baby
Token baby
Feature activation+0.000
directive
Token directive
Feature activation+0.000
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
moment
Token moment
Feature activation+0.000
they
Token they
Feature activation+0.346
arrived
Token arrived
Feature activation+0.881
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
client
Token client
Feature activation+0.000
to
Token to
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
you
Token you
Feature activation+0.000
broke
Token broke
Feature activation+0.935
your
Token your
Feature activation+0.000
toe
Token toe
Feature activation+0.000
riding
Token riding
Feature activation+0.000
your
Token your
Feature activation+0.000
best
Token best
Feature activation+0.000

INTERVAL 0.660 - 0.880
CONTAINS 0.006%

of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
,
Token,
Feature activation+0.000
these
Token these
Feature activation+0.000
rumors
Token rumors
Feature activation+0.753
are
Token are
Feature activation+0.000
just
Token just
Feature activation+0.000
that
Token that
Feature activation+0.000
,
Token,
Feature activation+0.000
rumors
Token rumors
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
day
Token day
Feature activation+0.000
the
Token the
Feature activation+0.717
report
Token report
Feature activation+0.860
was
Token was
Feature activation+0.000
launched
Token launched
Feature activation+0.000
in
Token in
Feature activation+0.000
Moscow
Token Moscow
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
at
Token at
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
wore
Token wore
Feature activation+0.724
a
Token a
Feature activation+0.000
great
Token great
Feature activation+0.000
number
Token number
Feature activation+0.000
of
Token of
Feature activation+0.000
clothes
Token clothes
Feature activation+0.000
Wednesday
Token Wednesday
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.000
the
Token the
Feature activation+0.000
same
Token same
Feature activation+0.000
day
Token day
Feature activation+0.000
the
Token the
Feature activation+0.717
report
Token report
Feature activation+0.860
was
Token was
Feature activation+0.000
launched
Token launched
Feature activation+0.000
in
Token in
Feature activation+0.000
Moscow
Token Moscow
Feature activation+0.000
bl
Tokenbl
Feature activation+0.000
ading
Tokenading
Feature activation+0.000
.
Token.
Feature activation+0.000
Each
Token Each
Feature activation+0.000
time
Token time
Feature activation+0.000
Met
Token Met
Feature activation+0.820
el
Tokenel
Feature activation+0.000
ka
Tokenka
Feature activation+0.000
suggested
Token suggested
Feature activation+0.000
a
Token a
Feature activation+0.000
new
Token new
Feature activation+0.000

INTERVAL 0.440 - 0.660
CONTAINS 0.007%

institutions
Token institutions
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
present
Token present
Feature activation+0.000
time
Token time
Feature activation+0.000
to
Token to
Feature activation+0.441
resolve
Token resolve
Feature activation+0.000
the
Token the
Feature activation+0.000
ongoing
Token ongoing
Feature activation+0.000
dispute
Token dispute
Feature activation+0.000
over
Token over
Feature activation+0.000
in
Token in
Feature activation+0.000
August
Token August
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
month
Token month
Feature activation+0.000
the
Token the
Feature activation+0.625
city
Token city
Feature activation+0.096
empt
Token empt
Feature activation+0.000
ies
Tokenies
Feature activation+0.000
out
Token out
Feature activation+0.000
as
Token as
Feature activation+0.000
This
Token This
Feature activation+0.000
time
Token time
Feature activation+0.000
the
Token the
Feature activation+1.047
design
Token design
Feature activation+1.313
was
Token was
Feature activation+0.191
accepted
Token accepted
Feature activation+0.589
and
Token and
Feature activation+0.000
in
Token in
Feature activation+0.000
1922
Token 1922
Feature activation+0.000
the
Token the
Feature activation+0.000
contract
Token contract
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
That
TokenThat
Feature activation+0.000
time
Token time
Feature activation+0.000
they
Token they
Feature activation+0.025
said
Token said
Feature activation+0.569
their
Token their
Feature activation+0.000
slogans
Token slogans
Feature activation+0.395
'
Token '
Feature activation+0.000
rang
Tokenrang
Feature activation+0.000
through
Token through
Feature activation+0.000
League
Token League
Feature activation+0.000
father
Token father
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
year
Token year
Feature activation+0.000
Craig
Token Craig
Feature activation+0.507
James
Token James
Feature activation+0.093
over
Token over
Feature activation+0.000
respected
Token respected
Feature activation+0.004
professional
Token professional
Feature activation+0.000
Bruce
Token Bruce
Feature activation+0.000

INTERVAL 0.220 - 0.440
CONTAINS 0.013%

Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
believe
Token believe
Feature activation+0.000
the
Token the
Feature activation+0.000
night
Token night
Feature activation+0.000
he
Token he
Feature activation+0.244
endured
Token endured
Feature activation+0.786
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
to
Token to
Feature activation+0.000
finding
Token finding
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
to
Token to
Feature activation+0.000
read
Token read
Feature activation+0.282
this
Token this
Feature activation+0.000
article
Token article
Feature activation+0.000
.
Token.
Feature activation+0.000
To
Token To
Feature activation+0.000
an
Token an
Feature activation+0.000
lot
Token lot
Feature activation+0.000
of
Token of
Feature activation+0.000
time
Token time
Feature activation+0.000
m
Token m
Feature activation+0.107
udd
Tokenudd
Feature activation+0.000
ying
Tokenying
Feature activation+0.289
the
Token the
Feature activation+0.000
waters
Token waters
Feature activation+0.000
over
Token over
Feature activation+0.000
a
Token a
Feature activation+0.000
series
Token series
Feature activation+0.000
was
Token was
Feature activation+0.000
seventeen
Token seventeen
Feature activation+0.000
at
Token at
Feature activation+0.000
the
Token the
Feature activation+0.000
time
Token time
Feature activation+0.000
she
Token she
Feature activation+0.341
claims
Token claims
Feature activation+1.663
she
Token she
Feature activation+0.000
slept
Token slept
Feature activation+0.000
with
Token with
Feature activation+0.000
Andrew
Token Andrew
Feature activation+0.000
may
Token may
Feature activation+0.000
spend
Token spend
Feature activation+0.000
time
Token time
Feature activation+0.000
submitting
Token submitting
Feature activation+0.892
your
Token your
Feature activation+0.000
website
Token website
Feature activation+0.242
and
Token and
Feature activation+0.000
content
Token content
Feature activation+0.000
to
Token to
Feature activation+0.000
other
Token other
Feature activation+0.000
websites
Token websites
Feature activation+0.000

INTERVAL 0.000 - 0.220
CONTAINS 99.964%

in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
bad
Token bad
Feature activation+0.000
way
Token way
Feature activation+0.000
.
Token.
Feature activation+0.000
As
Token As
Feature activation+0.000
we
Token we
Feature activation+0.000
have
Token have
Feature activation+0.000
more
Token more
Feature activation+0.000
drought
Token drought
Feature activation+0.000
,
Token,
Feature activation+0.000
crash
Token crash
Feature activation+0.000
involving
Token involving
Feature activation+0.000
Morgan
Token Morgan
Feature activation+0.000
will
Token will
Feature activation+0.000
persuade
Token persuade
Feature activation+0.000
Congress
Token Congress
Feature activation+0.000
to
Token to
Feature activation+0.000
reject
Token reject
Feature activation+0.000
an
Token an
Feature activation+0.000
amendment
Token amendment
Feature activation+0.000
that
Token that
Feature activation+0.000
countries
Token countries
Feature activation+0.000
are
Token are
Feature activation+0.000
emitting
Token emitting
Feature activation+0.000
into
Token into
Feature activation+0.000
the
Token the
Feature activation+0.000
earth
Token earth
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
atmosphere
Token atmosphere
Feature activation+0.000
.
Token.
Feature activation+0.000
alert
Token alert
Feature activation+0.000
is
Token is
Feature activation+0.000
issued
Token issued
Feature activation+0.000
,
Token,
Feature activation+0.000
locking
Token locking
Feature activation+0.000
up
Token up
Feature activation+0.000
their
Token their
Feature activation+0.000
homes
Token homes
Feature activation+0.000
,
Token,
Feature activation+0.000
leaving
Token leaving
Feature activation+0.000
cattle
Token cattle
Feature activation+0.000
ors
Tokenors
Feature activation+0.000
ville
Tokenville
Feature activation+0.000
,
Token,
Feature activation+0.000
says
Token says
Feature activation+0.000
the
Token the
Feature activation+0.000
move
Token move
Feature activation+0.000
shouldn
Token shouldn
Feature activation+0.000
't
Token't
Feature activation+0.000
be
Token be
Feature activation+0.000
viewed
Token viewed
Feature activation+0.000
as
Token as
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 H3.8: (feature 17150)

TOP ACTIVATIONS
MAX = 2.126

once
Token once
Feature activation+0.000
it
Token it
Feature activation+0.000
became
Token became
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+0.831
didn
Token didn
Feature activation+2.126
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
have
Token have
Feature activation+0.000
enough
Token enough
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
played
Token played
Feature activation+0.279
any
Token any
Feature activation+0.000
role
Token role
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
signs
Token signs
Feature activation+0.000
they
Token they
Feature activation+0.712
made
Token made
Feature activation+1.889
some
Token some
Feature activation+0.000
progress
Token progress
Feature activation+0.366
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
been
Token been
Feature activation+0.000
made
Token made
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+1.752
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
going
Token going
Feature activation+0.922
to
Token to
Feature activation+0.000
committee
Token committee
Feature activation+0.000
believes
Token believes
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Trump
Token Trump
Feature activation+1.704
requested
Token requested
Feature activation+0.000
putting
Token putting
Feature activation+0.000
ally
Token ally
Feature activation+0.000
in
Token in
Feature activation+0.000
charge
Token charge
Feature activation+0.000
it
Token it
Feature activation+0.000
quickly
Token quickly
Feature activation+0.000
became
Token became
Feature activation+0.000
obvious
Token obvious
Feature activation+0.000
the
Token the
Feature activation+1.418
poor
Token poor
Feature activation+1.666
little
Token little
Feature activation+0.000
thing
Token thing
Feature activation+0.000
spoke
Token spoke
Feature activation+1.054
no
Token no
Feature activation+0.000
English
Token English
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
It
TokenIt
Feature activation+0.000
makes
Token makes
Feature activation+0.000
clear
Token clear
Feature activation+0.000
the
Token the
Feature activation+1.510
location
Token location
Feature activation+0.000
and
Token and
Feature activation+0.000
accessibility
Token accessibility
Feature activation+0.000
of
Token of
Feature activation+0.000
photographs
Token photographs
Feature activation+0.000
and
Token and
Feature activation+0.000
it
Token it
Feature activation+0.000
quickly
Token quickly
Feature activation+0.000
became
Token became
Feature activation+0.000
obvious
Token obvious
Feature activation+0.000
the
Token the
Feature activation+1.418
poor
Token poor
Feature activation+1.666
little
Token little
Feature activation+0.000
thing
Token thing
Feature activation+0.000
spoke
Token spoke
Feature activation+1.054
no
Token no
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
"
Token"
Feature activation+0.000
They
TokenThey
Feature activation+0.000
had
Token had
Feature activation+0.000
no
Token no
Feature activation+0.000
idea
Token idea
Feature activation+0.000
there
Token there
Feature activation+1.244
was
Token was
Feature activation+0.757
a
Token a
Feature activation+0.322
place
Token place
Feature activation+0.071
called
Token called
Feature activation+0.000
Rocky
Token Rocky
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
suggesting
Token suggesting
Feature activation+1.182
that
Token that
Feature activation+0.000
Al
Token Al
Feature activation+0.000
-
Token-
Feature activation+0.000
Qaeda
TokenQaeda
Feature activation+0.000
or
Token or
Feature activation+0.000
,
Token,
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
no
Token no
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
we
Token we
Feature activation+1.157
will
Token will
Feature activation+0.246
not
Token not
Feature activation+0.000
end
Token end
Feature activation+0.000
as
Token as
Feature activation+0.000
equal
Token equal
Feature activation+0.000
obviously
Token obviously
Feature activation+0.000
have
Token have
Feature activation+0.000
no
Token no
Feature activation+0.000
qual
Token qual
Feature activation+0.000
ms
Tokenms
Feature activation+0.000
there
Token there
Feature activation+1.097
,
Token,
Feature activation+0.000
either
Token either
Feature activation+0.000
.
Token.
Feature activation+0.000
But
Token But
Feature activation+0.000
what
Token what
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
played
Token played
Feature activation+0.279
any
Token any
Feature activation+0.000
role
Token role
Feature activation+0.000
in
Token in
Feature activation+0.000
his
Token his
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
That
TokenThat
Feature activation+0.000
's
Token's
Feature activation+0.000
clearly
Token clearly
Feature activation+0.000
Trump
Token Trump
Feature activation+0.171
's
Token's
Feature activation+1.084
plan
Token plan
Feature activation+0.556
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Live
TokenLive
Feature activation+0.000
no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Bend
Token Bend
Feature activation+1.016
andi
Tokenandi
Feature activation+0.487
even
Token even
Feature activation+0.509
made
Token made
Feature activation+1.067
the
Token the
Feature activation+0.000
prediction
Token prediction
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
obvious
Token obvious
Feature activation+0.000
the
Token the
Feature activation+1.418
poor
Token poor
Feature activation+1.666
little
Token little
Feature activation+0.000
thing
Token thing
Feature activation+0.000
spoke
Token spoke
Feature activation+1.054
no
Token no
Feature activation+0.000
English
Token English
Feature activation+0.000
at
Token at
Feature activation+0.000
all
Token all
Feature activation+0.000
.
Token.
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
played
Token played
Feature activation+0.279
any
Token any
Feature activation+0.000
It
TokenIt
Feature activation+0.000
is
Token is
Feature activation+0.000
clear
Token clear
Feature activation+0.000
both
Token both
Feature activation+0.167
sides
Token sides
Feature activation+0.000
wish
Token wish
Feature activation+1.021
to
Token to
Feature activation+0.000
approach
Token approach
Feature activation+0.000
these
Token these
Feature activation+0.000
talks
Token talks
Feature activation+0.000
construct
Token construct
Feature activation+0.000
say
Token say
Feature activation+0.000
there
Token there
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Bend
Token Bend
Feature activation+1.016
andi
Tokenandi
Feature activation+0.487
even
Token even
Feature activation+0.509
made
Token made
Feature activation+1.067
the
Token the
Feature activation+0.000
prediction
Token prediction
Feature activation+0.000

Top DFA by src position
MAX = 2.676

the
Token the
Feature activation+0.006
Top resid features:
race
Token race
Feature activation+0.003
Top resid features:
once
Token once
Feature activation-0.015
Top resid features:
it
Token it
Feature activation+0.036
Top resid features:
became
Token became
Feature activation+0.028
Top resid features:
clear
Token clear
Feature activation+2.676
Top resid features:
she
Token she
Feature activation-0.077
Top resid features:
didn
Token didn
Feature activation-0.030
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ļ
TokenĻ
Feature activation+0.000
Top resid features:
t
Tokent
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.013
Top resid features:
Ċ
TokenĊ
Feature activation-0.019
Top resid features:
There
TokenThere
Feature activation+0.026
Top resid features:
is
Token is
Feature activation+0.018
Top resid features:
no
Token no
Feature activation+0.058
Top resid features:
indication
Token indication
Feature activation+2.523
Top resid features:
Mike
Token Mike
Feature activation+0.089
Top resid features:
G
Token G
Feature activation-0.003
Top resid features:
atto
Tokenatto
Feature activation-0.108
Top resid features:
's
Token's
Feature activation+0.104
Top resid features:
political
Token political
Feature activation+0.000
Top resid features:
deficit
Token deficit
Feature activation+0.026
Top resid features:
,
Token,
Feature activation+0.008
Top resid features:
and
Token and
Feature activation-0.077
Top resid features:
there
Token there
Feature activation+0.070
Top resid features:
are
Token are
Feature activation+0.032
Top resid features:
signs
Token signs
Feature activation+2.130
Top resid features:
they
Token they
Feature activation+0.210
Top resid features:
made
Token made
Feature activation+0.046
Top resid features:
some
Token some
Feature activation+0.000
Top resid features:
progress
Token progress
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.052
Top resid features:
Ļ
TokenĻ
Feature activation-0.027
Top resid features:
s
Tokens
Feature activation+0.002
Top resid features:
been
Token been
Feature activation+0.012
Top resid features:
made
Token made
Feature activation-0.082
Top resid features:
clear
Token clear
Feature activation+2.340
Top resid features:
she
Token she
Feature activation+0.054
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ļ
TokenĻ
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.000
Top resid features:
going
Token going
Feature activation+0.000
Top resid features:
House
TokenHouse
Feature activation-0.008
Top resid features:
committee
Token committee
Feature activation-0.017
Top resid features:
believes
Token believes
Feature activation-0.065
Top resid features:
it
Token it
Feature activation+0.063
Top resid features:
has
Token has
Feature activation-0.014
Top resid features:
evidence
Token evidence
Feature activation+2.237
Top resid features:
Trump
Token Trump
Feature activation+0.037
Top resid features:
requested
Token requested
Feature activation+0.000
Top resid features:
putting
Token putting
Feature activation+0.000
Top resid features:
ally
Token ally
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
row
Tokenrow
Feature activation-0.003
Top resid features:
and
Token and
Feature activation-0.029
Top resid features:
it
Token it
Feature activation+0.009
Top resid features:
quickly
Token quickly
Feature activation-0.000
Top resid features:
became
Token became
Feature activation+0.014
Top resid features:
obvious
Token obvious
Feature activation+2.426
Top resid features:
the
Token the
Feature activation-0.029
Top resid features:
poor
Token poor
Feature activation-0.086
Top resid features:
little
Token little
Feature activation+0.000
Top resid features:
thing
Token thing
Feature activation+0.000
Top resid features:
spoke
Token spoke
Feature activation+0.000
Top resid features:
.
Token.
Feature activation-0.038
Top resid features:
âĢ
Token âĢ
Feature activation-0.157
Top resid features:
ľ
Tokenľ
Feature activation-0.031
Top resid features:
It
TokenIt
Feature activation-0.015
Top resid features:
makes
Token makes
Feature activation-0.032
Top resid features:
clear
Token clear
Feature activation+2.436
Top resid features:
the
Token the
Feature activation+0.038
Top resid features:
location
Token location
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
accessibility
Token accessibility
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
row
Tokenrow
Feature activation-0.002
Top resid features:
and
Token and
Feature activation-0.023
Top resid features:
it
Token it
Feature activation-0.005
Top resid features:
quickly
Token quickly
Feature activation-0.026
Top resid features:
became
Token became
Feature activation-0.060
Top resid features:
obvious
Token obvious
Feature activation+2.267
Top resid features:
the
Token the
Feature activation-0.025
Top resid features:
poor
Token poor
Feature activation+0.000
Top resid features:
little
Token little
Feature activation+0.000
Top resid features:
thing
Token thing
Feature activation+0.000
Top resid features:
spoke
Token spoke
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.044
Top resid features:
Ċ
TokenĊ
Feature activation-0.063
Top resid features:
There
TokenThere
Feature activation+0.096
Top resid features:
is
Token is
Feature activation+0.023
Top resid features:
no
Token no
Feature activation+0.055
Top resid features:
indication
Token indication
Feature activation+1.865
Top resid features:
Mike
Token Mike
Feature activation+0.159
Top resid features:
G
Token G
Feature activation+0.000
Top resid features:
atto
Tokenatto
Feature activation+0.000
Top resid features:
's
Token's
Feature activation+0.000
Top resid features:
political
Token political
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.031
Top resid features:
"
Token"
Feature activation-0.092
Top resid features:
They
TokenThey
Feature activation+0.013
Top resid features:
had
Token had
Feature activation-0.046
Top resid features:
no
Token no
Feature activation+0.060
Top resid features:
idea
Token idea
Feature activation+2.067
Top resid features:
there
Token there
Feature activation-0.062
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
place
Token place
Feature activation+0.000
Top resid features:
called
Token called
Feature activation+0.000
Top resid features:
.
Token.
Feature activation-0.005
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation-0.052
Top resid features:
There
TokenThere
Feature activation-0.122
Top resid features:
is
Token is
Feature activation+0.049
Top resid features:
no
Token no
Feature activation+0.100
Top resid features:
evidence
Token evidence
Feature activation+1.894
Top resid features:
suggesting
Token suggesting
Feature activation+0.035
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
Al
Token Al
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
Qaeda
TokenQaeda
Feature activation+0.000
Top resid features:
now
Token now
Feature activation+0.013
Top resid features:
,
Token,
Feature activation-0.009
Top resid features:
I
Token I
Feature activation-0.024
Top resid features:
have
Token have
Feature activation-0.060
Top resid features:
no
Token no
Feature activation-0.006
Top resid features:
doubt
Token doubt
Feature activation+1.918
Top resid features:
we
Token we
Feature activation+0.074
Top resid features:
will
Token will
Feature activation+0.000
Top resid features:
not
Token not
Feature activation+0.000
Top resid features:
end
Token end
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.000
Top resid features:
so
Token so
Feature activation-0.006
Top resid features:
I
Token I
Feature activation-0.035
Top resid features:
obviously
Token obviously
Feature activation-0.030
Top resid features:
have
Token have
Feature activation-0.047
Top resid features:
no
Token no
Feature activation+0.043
Top resid features:
qual
Token qual
Feature activation+1.485
Top resid features:
ms
Tokenms
Feature activation+0.523
Top resid features:
there
Token there
Feature activation-0.028
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
either
Token either
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation-0.014
Top resid features:
There
TokenThere
Feature activation+0.014
Top resid features:
is
Token is
Feature activation+0.018
Top resid features:
no
Token no
Feature activation+0.028
Top resid features:
indication
Token indication
Feature activation+1.767
Top resid features:
Mike
Token Mike
Feature activation+0.088
Top resid features:
G
Token G
Feature activation-0.012
Top resid features:
atto
Tokenatto
Feature activation-0.016
Top resid features:
's
Token's
Feature activation+0.011
Top resid features:
political
Token political
Feature activation-0.047
Top resid features:
Ŀ
TokenĿ
Feature activation-0.006
Top resid features:
Ċ
TokenĊ
Feature activation-0.021
Top resid features:
Ċ
TokenĊ
Feature activation-0.023
Top resid features:
That
TokenThat
Feature activation-0.002
Top resid features:
's
Token's
Feature activation+0.057
Top resid features:
clearly
Token clearly
Feature activation+1.706
Top resid features:
Trump
Token Trump
Feature activation-0.154
Top resid features:
's
Token's
Feature activation+0.083
Top resid features:
plan
Token plan
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
also
Token also
Feature activation-0.010
Top resid features:
say
Token say
Feature activation-0.152
Top resid features:
there
Token there
Feature activation+0.007
Top resid features:
is
Token is
Feature activation+0.039
Top resid features:
no
Token no
Feature activation+0.033
Top resid features:
evidence
Token evidence
Feature activation+1.913
Top resid features:
Bend
Token Bend
Feature activation-0.035
Top resid features:
andi
Tokenandi
Feature activation+0.012
Top resid features:
even
Token even
Feature activation-0.072
Top resid features:
made
Token made
Feature activation-0.011
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
row
Tokenrow
Feature activation+0.001
Top resid features:
and
Token and
Feature activation-0.011
Top resid features:
it
Token it
Feature activation+0.014
Top resid features:
quickly
Token quickly
Feature activation-0.019
Top resid features:
became
Token became
Feature activation+0.040
Top resid features:
obvious
Token obvious
Feature activation+1.729
Top resid features:
the
Token the
Feature activation+0.008
Top resid features:
poor
Token poor
Feature activation-0.015
Top resid features:
little
Token little
Feature activation-0.034
Top resid features:
thing
Token thing
Feature activation+0.034
Top resid features:
spoke
Token spoke
Feature activation-0.117
Top resid features:
Ċ
TokenĊ
Feature activation-0.019
Top resid features:
Ċ
TokenĊ
Feature activation-0.031
Top resid features:
There
TokenThere
Feature activation+0.025
Top resid features:
is
Token is
Feature activation+0.024
Top resid features:
no
Token no
Feature activation+0.059
Top resid features:
indication
Token indication
Feature activation+1.474
Top resid features:
Mike
Token Mike
Feature activation+0.077
Top resid features:
G
Token G
Feature activation-0.030
Top resid features:
atto
Tokenatto
Feature activation+0.018
Top resid features:
's
Token's
Feature activation+0.000
Top resid features:
political
Token political
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.006
Top resid features:
âĢ
TokenâĢ
Feature activation-0.065
Top resid features:
ľ
Tokenľ
Feature activation-0.019
Top resid features:
It
TokenIt
Feature activation+0.043
Top resid features:
is
Token is
Feature activation+0.047
Top resid features:
clear
Token clear
Feature activation+1.633
Top resid features:
both
Token both
Feature activation+0.176
Top resid features:
sides
Token sides
Feature activation-0.102
Top resid features:
wish
Token wish
Feature activation-0.188
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
approach
Token approach
Feature activation+0.000
Top resid features:
also
Token also
Feature activation-0.024
Top resid features:
say
Token say
Feature activation-0.065
Top resid features:
there
Token there
Feature activation-0.029
Top resid features:
is
Token is
Feature activation+0.059
Top resid features:
no
Token no
Feature activation+0.062
Top resid features:
evidence
Token evidence
Feature activation+1.677
Top resid features:
Bend
Token Bend
Feature activation-0.035
Top resid features:
andi
Tokenandi
Feature activation+0.000
Top resid features:
even
Token even
Feature activation+0.000
Top resid features:
made
Token made
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.08

Head 2: 0.06

Head 3: 0.03

Head 4: 0.03

Head 5: 0.04

Head 6: 0.07

Head 7: 0.07

Head 8: 0.43

Head 9: 0.05

Head 10: 0.05

Head 11: 0.05

Positive logits

forgot1.43

hadn1.41

hasn1.34

wore1.33

nea1.31

Rue1.31

wouldn1.28

couldn1.27

Aren1.25

je1.25

ctors1.21

psc1.19

slipped1.19

didn1.18

ONSORED1.18

HAS1.17

died1.17

complied1.16

Snow1.15

signed1.15

Negative logits

corpus-1.77

ivable-1.52

igious-1.42

aesthetic-1.39

stellar-1.38

uds-1.27

ieval-1.26

esthetic-1.25

resultant-1.23

Lucifer-1.23

uce-1.21

obin-1.20

exposition-1.18

orbital-1.17

medi-1.17

ellect-1.16

opener-1.15

oder-1.15

abus-1.15

poly-1.14

INTERVAL 1.913 - 2.126
CONTAINS 0.000%

once
Token once
Feature activation+0.000
it
Token it
Feature activation+0.000
became
Token became
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+0.831
didn
Token didn
Feature activation+2.126
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
have
Token have
Feature activation+0.000
enough
Token enough
Feature activation+0.000

INTERVAL 1.701 - 1.913
CONTAINS 0.000%

committee
Token committee
Feature activation+0.000
believes
Token believes
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Trump
Token Trump
Feature activation+1.704
requested
Token requested
Feature activation+0.000
putting
Token putting
Feature activation+0.000
ally
Token ally
Feature activation+0.000
in
Token in
Feature activation+0.000
charge
Token charge
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
signs
Token signs
Feature activation+0.000
they
Token they
Feature activation+0.712
made
Token made
Feature activation+1.889
some
Token some
Feature activation+0.000
progress
Token progress
Feature activation+0.366
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
played
Token played
Feature activation+0.279
any
Token any
Feature activation+0.000
role
Token role
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
been
Token been
Feature activation+0.000
made
Token made
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+1.752
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
going
Token going
Feature activation+0.922
to
Token to
Feature activation+0.000

INTERVAL 1.488 - 1.701
CONTAINS 0.000%

âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
It
TokenIt
Feature activation+0.000
makes
Token makes
Feature activation+0.000
clear
Token clear
Feature activation+0.000
the
Token the
Feature activation+1.510
location
Token location
Feature activation+0.000
and
Token and
Feature activation+0.000
accessibility
Token accessibility
Feature activation+0.000
of
Token of
Feature activation+0.000
photographs
Token photographs
Feature activation+0.000
it
Token it
Feature activation+0.000
quickly
Token quickly
Feature activation+0.000
became
Token became
Feature activation+0.000
obvious
Token obvious
Feature activation+0.000
the
Token the
Feature activation+1.418
poor
Token poor
Feature activation+1.666
little
Token little
Feature activation+0.000
thing
Token thing
Feature activation+0.000
spoke
Token spoke
Feature activation+1.054
no
Token no
Feature activation+0.000
English
Token English
Feature activation+0.000

INTERVAL 1.275 - 1.488
CONTAINS 0.000%

Ċ
TokenĊ
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
and
Token and
Feature activation+0.000
it
Token it
Feature activation+0.000
quickly
Token quickly
Feature activation+0.000
became
Token became
Feature activation+0.000
obvious
Token obvious
Feature activation+0.000
the
Token the
Feature activation+1.418
poor
Token poor
Feature activation+1.666
little
Token little
Feature activation+0.000
thing
Token thing
Feature activation+0.000
spoke
Token spoke
Feature activation+1.054
no
Token no
Feature activation+0.000

INTERVAL 1.063 - 1.275
CONTAINS 0.001%

no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Bend
Token Bend
Feature activation+1.016
andi
Tokenandi
Feature activation+0.487
even
Token even
Feature activation+0.509
made
Token made
Feature activation+1.067
the
Token the
Feature activation+0.000
prediction
Token prediction
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
There
TokenThere
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
suggesting
Token suggesting
Feature activation+1.182
that
Token that
Feature activation+0.000
Al
Token Al
Feature activation+0.000
-
Token-
Feature activation+0.000
Qaeda
TokenQaeda
Feature activation+0.000
or
Token or
Feature activation+0.000
"
Token"
Feature activation+0.000
They
TokenThey
Feature activation+0.000
had
Token had
Feature activation+0.000
no
Token no
Feature activation+0.000
idea
Token idea
Feature activation+0.000
there
Token there
Feature activation+1.244
was
Token was
Feature activation+0.757
a
Token a
Feature activation+0.322
place
Token place
Feature activation+0.071
called
Token called
Feature activation+0.000
Rocky
Token Rocky
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
That
TokenThat
Feature activation+0.000
's
Token's
Feature activation+0.000
clearly
Token clearly
Feature activation+0.000
Trump
Token Trump
Feature activation+0.171
's
Token's
Feature activation+1.084
plan
Token plan
Feature activation+0.556
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Live
TokenLive
Feature activation+0.000
,
Token,
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
no
Token no
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
we
Token we
Feature activation+1.157
will
Token will
Feature activation+0.246
not
Token not
Feature activation+0.000
end
Token end
Feature activation+0.000
as
Token as
Feature activation+0.000
equal
Token equal
Feature activation+0.000

INTERVAL 0.850 - 1.063
CONTAINS 0.001%

is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
indication
Token indication
Feature activation+0.000
Mike
Token Mike
Feature activation+1.307
G
Token G
Feature activation+0.166
atto
Tokenatto
Feature activation+1.038
's
Token's
Feature activation+1.906
political
Token political
Feature activation+0.000
career
Token career
Feature activation+1.092
played
Token played
Feature activation+0.279
any
Token any
Feature activation+0.000
as
Token as
Feature activation+0.000
the
Token the
Feature activation+0.000
government
Token government
Feature activation+0.000
made
Token made
Feature activation+0.000
clear
Token clear
Feature activation+0.000
in
Token in
Feature activation+0.851
an
Token an
Feature activation+0.000
In
Token In
Feature activation+0.000
format
Tokenformat
Feature activation+0.000
ization
Tokenization
Feature activation+0.000
Strategy
Token Strategy
Feature activation+0.000
am
Token am
Feature activation+0.000
gone
Token gone
Feature activation+0.000
,
Token,
Feature activation+0.000
no
Token no
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
you
Token you
Feature activation+0.859
shall
Token shall
Feature activation+0.025
be
Token be
Feature activation+0.000
sought
Token sought
Feature activation+0.000
for
Token for
Feature activation+0.000
my
Token my
Feature activation+0.000
I
TokenI
Feature activation+0.000
just
Token just
Feature activation+0.000
found
Token found
Feature activation+0.000
out
Token out
Feature activation+0.000
the
Token the
Feature activation+0.570
boss
Token boss
Feature activation+0.916
blew
Token blew
Feature activation+0.241
the
Token the
Feature activation+0.000
fuse
Token fuse
Feature activation+0.000
after
Token after
Feature activation+0.000
seeing
Token seeing
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+1.752
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
going
Token going
Feature activation+0.922
to
Token to
Feature activation+0.000
have
Token have
Feature activation+0.000
a
Token a
Feature activation+0.000
hand
Token hand
Feature activation+0.000
in
Token in
Feature activation+0.000

INTERVAL 0.638 - 0.850
CONTAINS 0.002%

they
Token they
Feature activation+0.000
are
Token are
Feature activation+0.000
without
Token without
Feature activation+0.000
a
Token a
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
part
Token part
Feature activation+0.646
of
Token of
Feature activation+0.000
my
Token my
Feature activation+0.000
chromosomes
Token chromosomes
Feature activation+0.000
,
Token,
Feature activation+0.000
DNA
Token DNA
Feature activation+0.000
that
Token that
Feature activation+0.000
it
Token it
Feature activation+0.000
makes
Token makes
Feature activation+0.000
explicit
Token explicit
Feature activation+0.000
the
Token the
Feature activation+0.721
link
Token link
Feature activation+0.706
between
Token between
Feature activation+0.000
peace
Token peace
Feature activation+0.000
and
Token and
Feature activation+0.000
justice
Token justice
Feature activation+0.000
.
Token.
Feature activation+0.000
With
TokenWith
Feature activation+0.000
May
Token May
Feature activation+0.000
having
Token having
Feature activation+0.000
made
Token made
Feature activation+0.000
clear
Token clear
Feature activation+0.000
she
Token she
Feature activation+0.849
plans
Token plans
Feature activation+0.818
for
Token for
Feature activation+0.000
Britain
Token Britain
Feature activation+0.000
to
Token to
Feature activation+0.000
leave
Token leave
Feature activation+0.000
to
Token to
Feature activation+0.000
find
Token find
Feature activation+0.000
out
Token out
Feature activation+0.000
my
Token my
Feature activation+0.566
best
Token best
Feature activation+0.956
friend
Token friend
Feature activation+0.800
âĢ¦
TokenâĢ¦
Feature activation+0.000
he
Tokenhe
Feature activation+0.000
set
Token set
Feature activation+0.000
her
Token her
Feature activation+0.000
place
Token place
Feature activation+0.000
They
TokenThey
Feature activation+0.000
had
Token had
Feature activation+0.000
no
Token no
Feature activation+0.000
idea
Token idea
Feature activation+0.000
there
Token there
Feature activation+1.244
was
Token was
Feature activation+0.757
a
Token a
Feature activation+0.322
place
Token place
Feature activation+0.071
called
Token called
Feature activation+0.000
Rocky
Token Rocky
Feature activation+0.000
Top
Token Top
Feature activation+0.000

INTERVAL 0.425 - 0.638
CONTAINS 0.003%

Ċ
TokenĊ
Feature activation+0.000
No
TokenNo
Feature activation+0.000
doubt
Token doubt
Feature activation+0.000
the
Token the
Feature activation+1.005
ATF
Token ATF
Feature activation+0.306
is
Token is
Feature activation+0.471
taking
Token taking
Feature activation+0.130
notes
Token notes
Feature activation+0.000
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
FBI
Token FBI
Feature activation+0.000
is
Token is
Feature activation+0.000
no
Token no
Feature activation+0.000
evidence
Token evidence
Feature activation+0.000
Bend
Token Bend
Feature activation+1.016
andi
Tokenandi
Feature activation+0.487
even
Token even
Feature activation+0.509
made
Token made
Feature activation+1.067
the
Token the
Feature activation+0.000
prediction
Token prediction
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.000
not
Token not
Feature activation+0.000
surprised
Token surprised
Feature activation+0.000
it
Token it
Feature activation+0.008
ended
Token ended
Feature activation+0.450
in
Token in
Feature activation+0.000
murder
Token murder
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
this
Token this
Feature activation+0.000
decision
Token decision
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
fact
Token fact
Feature activation+0.000
his
Token his
Feature activation+0.558
contract
Token contract
Feature activation+0.326
contains
Token contains
Feature activation+0.000
a
Token a
Feature activation+0.000
modified
Token modified
Feature activation+0.000
no
Token no
Feature activation+0.000
It
TokenIt
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
clear
Token clear
Feature activation+0.000
it
Token it
Feature activation+0.507
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
been
Token been
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 0.213 - 0.425
CONTAINS 0.007%

might
Token might
Feature activation+0.000
be
Token be
Feature activation+0.000
a
Token a
Feature activation+0.000
good
Token good
Feature activation+0.000
idea
Token idea
Feature activation+0.000
to
Token to
Feature activation+0.306
print
Token print
Feature activation+0.000
them
Token them
Feature activation+0.000
at
Token at
Feature activation+0.000
100
Token 100
Feature activation+0.000
%
Token%
Feature activation+0.000
have
Token have
Feature activation+0.000
to
Token to
Feature activation+0.000
file
Token file
Feature activation+0.000
similar
Token similar
Feature activation+0.000
reports
Token reports
Feature activation+0.000
election
Token election
Feature activation+0.328
-
Token-
Feature activation+0.227
cycle
Tokencycle
Feature activation+0.000
after
Token after
Feature activation+0.000
election
Token election
Feature activation+0.000
-
Token-
Feature activation+0.000
revelations
Token revelations
Feature activation+0.000
spurred
Token spurred
Feature activation+0.000
fresh
Token fresh
Feature activation+0.000
questions
Token questions
Feature activation+0.000
Thursday
Token Thursday
Feature activation+0.000
from
Token from
Feature activation+0.238
lawmakers
Token lawmakers
Feature activation+0.000
about
Token about
Feature activation+0.000
whether
Token whether
Feature activation+0.000
the
Token the
Feature activation+0.000
newly
Token newly
Feature activation+0.000
there
Token there
Feature activation+0.000
were
Token were
Feature activation+0.000
concerns
Token concerns
Feature activation+0.000
multinational
Token multinational
Feature activation+0.000
s
Tokens
Feature activation+0.000
were
Token were
Feature activation+0.332
not
Token not
Feature activation+0.000
paying
Token paying
Feature activation+0.000
their
Token their
Feature activation+0.000
fair
Token fair
Feature activation+0.000
share
Token share
Feature activation+0.000
had
Token had
Feature activation+0.000
no
Token no
Feature activation+0.000
idea
Token idea
Feature activation+0.000
there
Token there
Feature activation+1.244
was
Token was
Feature activation+0.757
a
Token a
Feature activation+0.322
place
Token place
Feature activation+0.071
called
Token called
Feature activation+0.000
Rocky
Token Rocky
Feature activation+0.000
Top
Token Top
Feature activation+0.000
when
Token when
Feature activation+0.000

INTERVAL 0.000 - 0.213
CONTAINS 99.985%

and
Token and
Feature activation+0.000
exploding
Token exploding
Feature activation+0.000
colour
Token colour
Feature activation+0.000
.
Token.
Feature activation+0.000
(
Token (
Feature activation+0.000
S
TokenS
Feature activation+0.000
uda
Tokenuda
Feature activation+0.000
calls
Token calls
Feature activation+0.000
the
Token the
Feature activation+0.000
look
Token look
Feature activation+0.000
a
Token a
Feature activation+0.000
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
of
Token of
Feature activation+0.000
epic
Token epic
Feature activation+0.000
proportions
Token proportions
Feature activation+0.000
.
Token.
Feature activation+0.000
Now
Token Now
Feature activation+0.000
that
Token that
Feature activation+0.000
Captain
Token Captain
Feature activation+0.000
man
Tokenman
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
People
TokenPeople
Feature activation+0.000
who
Token who
Feature activation+0.000
may
Token may
Feature activation+0.000
have
Token have
Feature activation+0.000
lost
Token lost
Feature activation+0.000
their
Token their
Feature activation+0.000
jobs
Token jobs
Feature activation+0.000
are
Token are
Feature activation+0.000
Mr
TokenMr
Feature activation+0.000
Cos
Token Cos
Feature activation+0.000
grave
Tokengrave
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
brother
Token brother
Feature activation+0.000
C
Token C
Feature activation+0.000
iar
Tokeniar
Feature activation+0.000
án
Tokenán
Feature activation+0.000
had
Token had
Feature activation+0.000
have
Token have
Feature activation+0.000
some
Token some
Feature activation+0.000
prior
Token prior
Feature activation+0.000
relationship
Token relationship
Feature activation+0.000
to
Token to
Feature activation+0.000
each
Token each
Feature activation+0.000
other
Token other
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
What
TokenWhat
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 H3.8: (feature 6576)

TOP ACTIVATIONS
MAX = 3.643

's
Token's
Feature activation+0.000
Who
Token Who
Feature activation+0.000
Do
Token Do
Feature activation+0.000
You
Token You
Feature activation+0.000
Think
Token Think
Feature activation+0.000
You
Token You
Feature activation+3.643
Are
Token Are
Feature activation+3.136
?,
Token?,
Feature activation+0.000
the
Token the
Feature activation+0.000
former
Token former
Feature activation+0.000
Top
Token Top
Feature activation+0.000
Who
Token Who
Feature activation+0.000
Do
Token Do
Feature activation+0.000
You
Token You
Feature activation+0.000
Think
Token Think
Feature activation+0.000
You
Token You
Feature activation+3.643
Are
Token Are
Feature activation+3.136
?,
Token?,
Feature activation+0.000
the
Token the
Feature activation+0.000
former
Token former
Feature activation+0.000
Top
Token Top
Feature activation+0.000
Gear
Token Gear
Feature activation+0.000
don
Token don
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
think
Token think
Feature activation+0.000
calling
Token calling
Feature activation+3.049
potential
Token potential
Feature activation+0.037
customers
Token customers
Feature activation+0.465
crazy
Token crazy
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
Do
TokenDo
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.084
all
Token all
Feature activation+2.988
federal
Token federal
Feature activation+0.388
funding
Token funding
Feature activation+1.536
for
Token for
Feature activation+2.057
Planned
Token Planned
Feature activation+0.000
Parenthood
Token Parenthood
Feature activation+0.000
65
Token65
Feature activation+0.000
percent
Token percent
Feature activation+0.000
of
Token of
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
think
Token think
Feature activation+0.000
our
Token our
Feature activation+2.855
economic
Token economic
Feature activation+0.967
system
Token system
Feature activation+1.829
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
un
Tokenun
Feature activation+0.000
besides
Token besides
Feature activation+0.000
dedication
Token dedication
Feature activation+0.000
that
Token that
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
a
Token a
Feature activation+2.797
beard
Token beard
Feature activation+1.491
-
Token-
Feature activation+0.000
we
Tokenwe
Feature activation+0.000
arer
Tokenarer
Feature activation+0.691
inherently
Token inherently
Feature activation+0.279
,
Token,
Feature activation+0.000
who
Token who
Feature activation+0.000
would
Token would
Feature activation+0.000
have
Token have
Feature activation+0.000
thought
Token thought
Feature activation+0.000
U
Token U
Feature activation+2.633
bi
Tokenbi
Feature activation+0.199
could
Token could
Feature activation+1.277
be
Token be
Feature activation+1.886
the
Token the
Feature activation+0.000
ones
Token ones
Feature activation+0.000
to
Token to
Feature activation+0.000
what
Token what
Feature activation+0.000
degree
Token degree
Feature activation+0.000
people
Token people
Feature activation+0.000
thought
Token thought
Feature activation+0.000
their
Token their
Feature activation+2.625
behavior
Token behavior
Feature activation+1.776
was
Token was
Feature activation+2.333
noticed
Token noticed
Feature activation+1.102
by
Token by
Feature activation+0.000
others
Token others
Feature activation+0.000
definition
Token definition
Feature activation+0.000
what
Token what
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
you
Token you
Feature activation+2.581
are
Token are
Feature activation+1.777
when
Token when
Feature activation+0.000
you
Token you
Feature activation+0.000
don
Token don
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
have
Token have
Feature activation+0.000
any
Token any
Feature activation+0.000
problems
Token problems
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
we
Token we
Feature activation+2.537
might
Token might
Feature activation+1.777
enjoy
Token enjoy
Feature activation+1.233
,
Token,
Feature activation+0.000
please
Token please
Feature activation+0.000
send
Token send
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
One
TokenOne
Feature activation+0.000
would
Token would
Feature activation+0.000
think
Token think
Feature activation+0.000
her
Token her
Feature activation+2.518
fear
Token fear
Feature activation+1.898
of
Token of
Feature activation+0.000
white
Token white
Feature activation+0.000
men
Token men
Feature activation+0.000
would
Token would
Feature activation+0.000
her
Token her
Feature activation+0.000
legs
Token legs
Feature activation+0.000
because
Token because
Feature activation+0.000
she
Token she
Feature activation+0.000
thought
Token thought
Feature activation+0.000
people
Token people
Feature activation+2.450
in
Token in
Feature activation+0.676
the
Token the
Feature activation+0.000
front
Token front
Feature activation+0.000
row
Token row
Feature activation+0.000
at
Token at
Feature activation+0.000
is
Token is
Feature activation+0.000
.
Token.
Feature activation+0.000
Most
Token Most
Feature activation+0.000
people
Token people
Feature activation+0.000
think
Token think
Feature activation+0.000
a
Token a
Feature activation+2.414
collaborative
Token collaborative
Feature activation+1.791
robot
Token robot
Feature activation+1.630
is
Token is
Feature activation+1.613
a
Token a
Feature activation+0.237
gentle
Token gentle
Feature activation+0.000
as
Token as
Feature activation+0.000
whether
Token whether
Feature activation+0.000
they
Token they
Feature activation+0.000
thought
Token thought
Feature activation+0.000
such
Token such
Feature activation+1.339
comments
Token comments
Feature activation+2.349
were
Token were
Feature activation+1.619
personal
Token personal
Feature activation+0.070
attacks
Token attacks
Feature activation+0.419
.
Token.
Feature activation+0.000
J
Token J
Feature activation+0.000
do
Token do
Feature activation+0.000
Manit
Token Manit
Feature activation+0.000
ob
Tokenob
Feature activation+0.000
ans
Tokenans
Feature activation+0.000
think
Token think
Feature activation+0.000
would
Token would
Feature activation+2.349
make
Token make
Feature activation+1.808
the
Token the
Feature activation+0.000
best
Token best
Feature activation+0.000
Premier
Token Premier
Feature activation+0.000
?
Token?
Feature activation+0.000
degree
Token degree
Feature activation+0.000
people
Token people
Feature activation+0.000
thought
Token thought
Feature activation+0.000
their
Token their
Feature activation+2.625
behavior
Token behavior
Feature activation+1.776
was
Token was
Feature activation+2.333
noticed
Token noticed
Feature activation+1.102
by
Token by
Feature activation+0.000
others
Token others
Feature activation+0.000
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
bus
Token bus
Feature activation+0.000
but
Token but
Feature activation+0.000
they
Token they
Feature activation+0.000
do
Token do
Feature activation+0.000
think
Token think
Feature activation+0.000
all
Token all
Feature activation+2.328
parents
Token parents
Feature activation+0.566
need
Token need
Feature activation+2.304
to
Token to
Feature activation+0.127
be
Token be
Feature activation+0.164
careful
Token careful
Feature activation+0.000
percent
Token percent
Feature activation+0.000
of
Token of
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
think
Token think
Feature activation+0.000
wealthy
Token wealthy
Feature activation+0.846
people
Token people
Feature activation+2.301
have
Token have
Feature activation+0.923
too
Token too
Feature activation+0.000
much
Token much
Feature activation+0.000
power
Token power
Feature activation+0.000
and
Token and
Feature activation+0.000
You
TokenYou
Feature activation+0.000
'd
Token'd
Feature activation+0.000
think
Token think
Feature activation+0.000
it
Token it
Feature activation+2.256
'd
Token'd
Feature activation+0.663
be
Token be
Feature activation+2.291
really
Token really
Feature activation+0.483
hard
Token hard
Feature activation+0.000
to
Token to
Feature activation+0.000
relate
Token relate
Feature activation+0.000
to
Token to
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
You
TokenYou
Feature activation+0.000
'd
Token'd
Feature activation+0.000
think
Token think
Feature activation+0.000
it
Token it
Feature activation+2.256
'd
Token'd
Feature activation+0.663
be
Token be
Feature activation+2.291
really
Token really
Feature activation+0.483
hard
Token hard
Feature activation+0.000
to
Token to
Feature activation+0.000

Top DFA by src position
MAX = 4.425

BBC
Token BBC
Feature activation+0.007
Top resid features:
's
Token's
Feature activation+0.024
Top resid features:
Who
Token Who
Feature activation-0.022
Top resid features:
Do
Token Do
Feature activation-0.052
Top resid features:
You
Token You
Feature activation+0.028
Top resid features:
Think
Token Think
Feature activation+4.425
Top resid features:
You
Token You
Feature activation+0.209
Top resid features:
Are
Token Are
Feature activation+0.000
Top resid features:
?,
Token?,
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
former
Token former
Feature activation+0.000
Top resid features:
BBC
Token BBC
Feature activation+0.040
Top resid features:
's
Token's
Feature activation+0.012
Top resid features:
Who
Token Who
Feature activation-0.008
Top resid features:
Do
Token Do
Feature activation+0.000
Top resid features:
You
Token You
Feature activation+0.017
Top resid features:
Think
Token Think
Feature activation+3.765
Top resid features:
You
Token You
Feature activation+0.281
Top resid features:
Are
Token Are
Feature activation-0.079
Top resid features:
?,
Token?,
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
former
Token former
Feature activation+0.000
Top resid features:
we
Token we
Feature activation-0.006
Top resid features:
don
Token don
Feature activation+0.024
Top resid features:
âĢ
TokenâĢ
Feature activation+0.107
Top resid features:
Ļ
TokenĻ
Feature activation+0.026
Top resid features:
t
Tokent
Feature activation+0.010
Top resid features:
think
Token think
Feature activation+3.633
Top resid features:
calling
Token calling
Feature activation+0.029
Top resid features:
potential
Token potential
Feature activation+0.000
Top resid features:
customers
Token customers
Feature activation+0.000
Top resid features:
crazy
Token crazy
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.012
Top resid features:
âĢ
Token âĢ
Feature activation+0.037
Top resid features:
ľ
Tokenľ
Feature activation+0.006
Top resid features:
Do
TokenDo
Feature activation+0.072
Top resid features:
you
Token you
Feature activation+0.043
Top resid features:
think
Token think
Feature activation+3.476
Top resid features:
all
Token all
Feature activation+0.127
Top resid features:
federal
Token federal
Feature activation+0.000
Top resid features:
funding
Token funding
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
Planned
Token Planned
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.011
Top resid features:
65
Token65
Feature activation+0.028
Top resid features:
percent
Token percent
Feature activation-0.022
Top resid features:
of
Token of
Feature activation+0.001
Top resid features:
Americans
Token Americans
Feature activation-0.028
Top resid features:
think
Token think
Feature activation+3.473
Top resid features:
our
Token our
Feature activation+0.232
Top resid features:
economic
Token economic
Feature activation+0.000
Top resid features:
system
Token system
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:
qualities
Token qualities
Feature activation+0.004
Top resid features:
besides
Token besides
Feature activation+0.018
Top resid features:
dedication
Token dedication
Feature activation+0.036
Top resid features:
that
Token that
Feature activation-0.002
Top resid features:
you
Token you
Feature activation+0.002
Top resid features:
think
Token think
Feature activation+3.449
Top resid features:
a
Token a
Feature activation+0.160
Top resid features:
beard
Token beard
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
we
Tokenwe
Feature activation+0.000
Top resid features:
arer
Tokenarer
Feature activation+0.000
Top resid features:
people
Token people
Feature activation+0.011
Top resid features:
,
Token,
Feature activation+0.015
Top resid features:
who
Token who
Feature activation+0.016
Top resid features:
would
Token would
Feature activation-0.009
Top resid features:
have
Token have
Feature activation+0.054
Top resid features:
thought
Token thought
Feature activation+3.133
Top resid features:
U
Token U
Feature activation+0.225
Top resid features:
bi
Tokenbi
Feature activation+0.000
Top resid features:
could
Token could
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
determine
Token determine
Feature activation+0.009
Top resid features:
to
Token to
Feature activation+0.022
Top resid features:
what
Token what
Feature activation+0.005
Top resid features:
degree
Token degree
Feature activation-0.015
Top resid features:
people
Token people
Feature activation+0.110
Top resid features:
thought
Token thought
Feature activation+3.185
Top resid features:
their
Token their
Feature activation+0.200
Top resid features:
behavior
Token behavior
Feature activation+0.000
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
noticed
Token noticed
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
that
Token that
Feature activation-0.003
Top resid features:
definition
Token definition
Feature activation+0.006
Top resid features:
what
Token what
Feature activation-0.005
Top resid features:
do
Token do
Feature activation-0.023
Top resid features:
you
Token you
Feature activation+0.021
Top resid features:
think
Token think
Feature activation+3.198
Top resid features:
you
Token you
Feature activation+0.144
Top resid features:
are
Token are
Feature activation+0.000
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
you
Token you
Feature activation+0.000
Top resid features:
don
Token don
Feature activation+0.000
Top resid features:
you
Token you
Feature activation+0.040
Top resid features:
have
Token have
Feature activation-0.009
Top resid features:
any
Token any
Feature activation+0.005
Top resid features:
problems
Token problems
Feature activation+0.055
Top resid features:
you
Token you
Feature activation+0.081
Top resid features:
think
Token think
Feature activation+2.931
Top resid features:
we
Token we
Feature activation+0.186
Top resid features:
might
Token might
Feature activation+0.000
Top resid features:
enjoy
Token enjoy
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
please
Token please
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.021
Top resid features:
Ċ
TokenĊ
Feature activation+0.022
Top resid features:
Ċ
TokenĊ
Feature activation+0.023
Top resid features:
One
TokenOne
Feature activation-0.031
Top resid features:
would
Token would
Feature activation-0.036
Top resid features:
think
Token think
Feature activation+3.148
Top resid features:
her
Token her
Feature activation+0.235
Top resid features:
fear
Token fear
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
white
Token white
Feature activation+0.000
Top resid features:
men
Token men
Feature activation+0.000
Top resid features:
shaved
Token shaved
Feature activation+0.004
Top resid features:
her
Token her
Feature activation+0.007
Top resid features:
legs
Token legs
Feature activation-0.011
Top resid features:
because
Token because
Feature activation-0.049
Top resid features:
she
Token she
Feature activation+0.014
Top resid features:
thought
Token thought
Feature activation+3.188
Top resid features:
people
Token people
Feature activation+0.166
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
front
Token front
Feature activation+0.000
Top resid features:
row
Token row
Feature activation+0.000
Top resid features:
robot
Token robot
Feature activation+0.006
Top resid features:
is
Token is
Feature activation-0.003
Top resid features:
.
Token.
Feature activation+0.076
Top resid features:
Most
Token Most
Feature activation+0.027
Top resid features:
people
Token people
Feature activation-0.049
Top resid features:
think
Token think
Feature activation+2.914
Top resid features:
a
Token a
Feature activation+0.240
Top resid features:
collaborative
Token collaborative
Feature activation+0.000
Top resid features:
robot
Token robot
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.008
Top resid features:
well
Token well
Feature activation-0.088
Top resid features:
as
Token as
Feature activation+0.022
Top resid features:
whether
Token whether
Feature activation+0.051
Top resid features:
they
Token they
Feature activation+0.017
Top resid features:
thought
Token thought
Feature activation+2.831
Top resid features:
such
Token such
Feature activation+0.157
Top resid features:
comments
Token comments
Feature activation+0.009
Top resid features:
were
Token were
Feature activation+0.000
Top resid features:
personal
Token personal
Feature activation+0.000
Top resid features:
attacks
Token attacks
Feature activation+0.000
Top resid features:
Who
TokenWho
Feature activation+0.135
Top resid features:
do
Token do
Feature activation+0.051
Top resid features:
Manit
Token Manit
Feature activation-0.010
Top resid features:
ob
Tokenob
Feature activation-0.004
Top resid features:
ans
Tokenans
Feature activation+0.000
Top resid features:
think
Token think
Feature activation+2.802
Top resid features:
would
Token would
Feature activation+0.139
Top resid features:
make
Token make
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
best
Token best
Feature activation+0.000
Top resid features:
Premier
Token Premier
Feature activation+0.000
Top resid features:
determine
Token determine
Feature activation+0.010
Top resid features:
to
Token to
Feature activation+0.012
Top resid features:
what
Token what
Feature activation+0.028
Top resid features:
degree
Token degree
Feature activation-0.025
Top resid features:
people
Token people
Feature activation+0.007
Top resid features:
thought
Token thought
Feature activation+2.914
Top resid features:
their
Token their
Feature activation+0.358
Top resid features:
behavior
Token behavior
Feature activation-0.008
Top resid features:
was
Token was
Feature activation-0.054
Top resid features:
noticed
Token noticed
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.006
Top resid features:
bus
Token bus
Feature activation+0.006
Top resid features:
but
Token but
Feature activation-0.019
Top resid features:
they
Token they
Feature activation+0.031
Top resid features:
do
Token do
Feature activation-0.013
Top resid features:
think
Token think
Feature activation+3.045
Top resid features:
all
Token all
Feature activation+0.159
Top resid features:
parents
Token parents
Feature activation+0.000
Top resid features:
need
Token need
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-0.015
Top resid features:
82
Token82
Feature activation+0.011
Top resid features:
percent
Token percent
Feature activation-0.022
Top resid features:
of
Token of
Feature activation-0.008
Top resid features:
Americans
Token Americans
Feature activation-0.006
Top resid features:
think
Token think
Feature activation+2.876
Top resid features:
wealthy
Token wealthy
Feature activation+0.412
Top resid features:
people
Token people
Feature activation-0.135
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
too
Token too
Feature activation+0.000
Top resid features:
much
Token much
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.028
Top resid features:
Ċ
TokenĊ
Feature activation+0.033
Top resid features:
Ċ
TokenĊ
Feature activation+0.033
Top resid features:
You
TokenYou
Feature activation-0.033
Top resid features:
'd
Token'd
Feature activation-0.008
Top resid features:
think
Token think
Feature activation+3.045
Top resid features:
it
Token it
Feature activation+0.072
Top resid features:
'd
Token'd
Feature activation+0.016
Top resid features:
be
Token be
Feature activation-0.039
Top resid features:
really
Token really
Feature activation+0.000
Top resid features:
hard
Token hard
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.029
Top resid features:
Ċ
TokenĊ
Feature activation+0.049
Top resid features:
Ċ
TokenĊ
Feature activation+0.044
Top resid features:
You
TokenYou
Feature activation-0.036
Top resid features:
'd
Token'd
Feature activation+0.079
Top resid features:
think
Token think
Feature activation+2.857
Top resid features:
it
Token it
Feature activation+0.068
Top resid features:
'd
Token'd
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
really
Token really
Feature activation+0.000
Top resid features:
hard
Token hard
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.01

Head 1: 0.05

Head 2: 0.07

Head 3: 0.05

Head 4: 0.01

Head 5: 0.03

Head 6: 0.05

Head 7: 0.06

Head 8: 0.43

Head 9: 0.08

Head 10: 0.07

Head 11: 0.08

Positive logits

ould1.45

should1.36

shouldn1.27

ought1.23

Rack1.23

Box1.15

fits1.14

Nurs1.13

quickShipAvailable1.11

spoiler1.10

Lean1.10

Mania1.08

facebook1.08

should1.08

Spoiler1.07

oyer1.06

horse1.05

icable1.05

Snow1.05

could1.03

Negative logits

habi-1.38

azeera-1.26

evangel-1.20

olan-1.20

覚醒-1.18

Avg-1.15

ennett-1.15

adic-1.12

voc-1.11

eyewitness-1.11

denomination-1.10

fellowship-1.07

witness-1.07

experien-1.05

anamo-1.04

ancest-1.03

ilateral-1.03

inately-1.03

itiner-1.00

FTA-1.00

INTERVAL 3.279 - 3.643
CONTAINS 0.000%

's
Token's
Feature activation+0.000
Who
Token Who
Feature activation+0.000
Do
Token Do
Feature activation+0.000
You
Token You
Feature activation+0.000
Think
Token Think
Feature activation+0.000
You
Token You
Feature activation+3.643
Are
Token Are
Feature activation+3.136
?,
Token?,
Feature activation+0.000
the
Token the
Feature activation+0.000
former
Token former
Feature activation+0.000
Top
Token Top
Feature activation+0.000

INTERVAL 2.914 - 3.279
CONTAINS 0.000%

Who
Token Who
Feature activation+0.000
Do
Token Do
Feature activation+0.000
You
Token You
Feature activation+0.000
Think
Token Think
Feature activation+0.000
You
Token You
Feature activation+3.643
Are
Token Are
Feature activation+3.136
?,
Token?,
Feature activation+0.000
the
Token the
Feature activation+0.000
former
Token former
Feature activation+0.000
Top
Token Top
Feature activation+0.000
Gear
Token Gear
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
Do
TokenDo
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.084
all
Token all
Feature activation+2.988
federal
Token federal
Feature activation+0.388
funding
Token funding
Feature activation+1.536
for
Token for
Feature activation+2.057
Planned
Token Planned
Feature activation+0.000
Parenthood
Token Parenthood
Feature activation+0.000
don
Token don
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
think
Token think
Feature activation+0.000
calling
Token calling
Feature activation+3.049
potential
Token potential
Feature activation+0.037
customers
Token customers
Feature activation+0.465
crazy
Token crazy
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 2.550 - 2.914
CONTAINS 0.001%

,
Token,
Feature activation+0.000
who
Token who
Feature activation+0.000
would
Token would
Feature activation+0.000
have
Token have
Feature activation+0.000
thought
Token thought
Feature activation+0.000
U
Token U
Feature activation+2.633
bi
Tokenbi
Feature activation+0.199
could
Token could
Feature activation+1.277
be
Token be
Feature activation+1.886
the
Token the
Feature activation+0.000
ones
Token ones
Feature activation+0.000
definition
Token definition
Feature activation+0.000
what
Token what
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
you
Token you
Feature activation+2.581
are
Token are
Feature activation+1.777
when
Token when
Feature activation+0.000
you
Token you
Feature activation+0.000
don
Token don
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
besides
Token besides
Feature activation+0.000
dedication
Token dedication
Feature activation+0.000
that
Token that
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
a
Token a
Feature activation+2.797
beard
Token beard
Feature activation+1.491
-
Token-
Feature activation+0.000
we
Tokenwe
Feature activation+0.000
arer
Tokenarer
Feature activation+0.691
inherently
Token inherently
Feature activation+0.279
65
Token65
Feature activation+0.000
percent
Token percent
Feature activation+0.000
of
Token of
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
think
Token think
Feature activation+0.000
our
Token our
Feature activation+2.855
economic
Token economic
Feature activation+0.967
system
Token system
Feature activation+1.829
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
un
Tokenun
Feature activation+0.000
to
Token to
Feature activation+0.000
what
Token what
Feature activation+0.000
degree
Token degree
Feature activation+0.000
people
Token people
Feature activation+0.000
thought
Token thought
Feature activation+0.000
their
Token their
Feature activation+2.625
behavior
Token behavior
Feature activation+1.776
was
Token was
Feature activation+2.333
noticed
Token noticed
Feature activation+1.102
by
Token by
Feature activation+0.000
others
Token others
Feature activation+0.000

INTERVAL 2.186 - 2.550
CONTAINS 0.001%

Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
One
TokenOne
Feature activation+0.000
would
Token would
Feature activation+0.000
think
Token think
Feature activation+0.000
her
Token her
Feature activation+2.518
fear
Token fear
Feature activation+1.898
of
Token of
Feature activation+0.000
white
Token white
Feature activation+0.000
men
Token men
Feature activation+0.000
would
Token would
Feature activation+0.000
percent
Token percent
Feature activation+0.000
of
Token of
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
think
Token think
Feature activation+0.000
wealthy
Token wealthy
Feature activation+0.846
people
Token people
Feature activation+2.301
have
Token have
Feature activation+0.923
too
Token too
Feature activation+0.000
much
Token much
Feature activation+0.000
power
Token power
Feature activation+0.000
and
Token and
Feature activation+0.000
bus
Token bus
Feature activation+0.000
but
Token but
Feature activation+0.000
they
Token they
Feature activation+0.000
do
Token do
Feature activation+0.000
think
Token think
Feature activation+0.000
all
Token all
Feature activation+2.328
parents
Token parents
Feature activation+0.566
need
Token need
Feature activation+2.304
to
Token to
Feature activation+0.127
be
Token be
Feature activation+0.164
careful
Token careful
Feature activation+0.000
have
Token have
Feature activation+0.000
any
Token any
Feature activation+0.000
problems
Token problems
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
we
Token we
Feature activation+2.537
might
Token might
Feature activation+1.777
enjoy
Token enjoy
Feature activation+1.233
,
Token,
Feature activation+0.000
please
Token please
Feature activation+0.000
send
Token send
Feature activation+0.000
do
Token do
Feature activation+0.000
Manit
Token Manit
Feature activation+0.000
ob
Tokenob
Feature activation+0.000
ans
Tokenans
Feature activation+0.000
think
Token think
Feature activation+0.000
would
Token would
Feature activation+2.349
make
Token make
Feature activation+1.808
the
Token the
Feature activation+0.000
best
Token best
Feature activation+0.000
Premier
Token Premier
Feature activation+0.000
?
Token?
Feature activation+0.000

INTERVAL 1.821 - 2.186
CONTAINS 0.003%

code
Token code
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
you
Token you
Feature activation+1.909
've
Token've
Feature activation+0.694
found
Token found
Feature activation+0.942
an
Token an
Feature activation+0.000
error
Token error
Feature activation+0.176
,
Token,
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
observers
Token observers
Feature activation+0.000
they
Token they
Feature activation+0.000
thought
Token thought
Feature activation+0.069
had
Token had
Feature activation+2.000
noticed
Token noticed
Feature activation+1.447
it
Token it
Feature activation+0.000
,
Token,
Feature activation+0.000
as
Token as
Feature activation+0.000
well
Token well
Feature activation+0.000
have
Token have
Feature activation+0.000
thought
Token thought
Feature activation+0.000
U
Token U
Feature activation+2.633
bi
Tokenbi
Feature activation+0.199
could
Token could
Feature activation+1.277
be
Token be
Feature activation+1.886
the
Token the
Feature activation+0.000
ones
Token ones
Feature activation+0.000
to
Token to
Feature activation+0.000
suddenly
Token suddenly
Feature activation+0.000
delay
Token delay
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
Do
TokenDo
Feature activation+0.000
you
Token you
Feature activation+0.000
think
Token think
Feature activation+0.000
they
Token they
Feature activation+1.861
would
Token would
Feature activation+1.894
allow
Token allow
Feature activation+2.034
themselves
Token themselves
Feature activation+0.000
to
Token to
Feature activation+0.000
behave
Token behave
Feature activation+0.000
shock
Token shock
Feature activation+0.000
that
Token that
Feature activation+0.000
parents
Token parents
Feature activation+0.000
think
Token think
Feature activation+0.000
it
Token it
Feature activation+1.636
's
Token's
Feature activation+2.126
okay
Token okay
Feature activation+1.260
to
Token to
Feature activation+0.450
publicly
Token publicly
Feature activation+0.000
shame
Token shame
Feature activation+0.000
their
Token their
Feature activation+0.000

INTERVAL 1.457 - 1.821
CONTAINS 0.005%

,
Token,
Feature activation+0.000
we
Token we
Feature activation+0.000
thought
Token thought
Feature activation+0.000
it
Token it
Feature activation+1.051
might
Token might
Feature activation+0.611
be
Token be
Feature activation+1.558
interesting
Token interesting
Feature activation+0.000
to
Token to
Feature activation+0.109
ask
Token ask
Feature activation+0.000
what
Token what
Feature activation+0.000
your
Token your
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
him
Token him
Feature activation+1.312
not
Token not
Feature activation+0.714
being
Token being
Feature activation+1.699
at
Token at
Feature activation+0.000
it
Token it
Feature activation+0.000
cost
Token cost
Feature activation+0.000
us
Token us
Feature activation+0.000
that
Token that
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
laboratory
Token laboratory
Feature activation+0.000
.
Token.
Feature activation+0.000
Think
Token Think
Feature activation+0.000
the
Token the
Feature activation+1.476
Large
Token Large
Feature activation+0.000
Had
Token Had
Feature activation+0.000
ron
Tokenron
Feature activation+0.000
Collider
Token Collider
Feature activation+0.000
in
Token in
Feature activation+0.000
of
Token of
Feature activation+0.000
criticism
Token criticism
Feature activation+0.000
.
Token.
Feature activation+0.000
You
Token You
Feature activation+0.000
think
Token think
Feature activation+0.000
you
Token you
Feature activation+1.758
are
Token are
Feature activation+0.550
creating
Token creating
Feature activation+0.272
a
Token a
Feature activation+0.000
masterpiece
Token masterpiece
Feature activation+0.000
,
Token,
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
historian
Token historian
Feature activation+0.000
,
Token,
Feature activation+0.000
thinks
Token thinks
Feature activation+0.000
a
Token a
Feature activation+1.693
national
Token national
Feature activation+0.000
two
Token two
Feature activation+0.000
-
Token-
Feature activation+0.000
party
Tokenparty
Feature activation+0.000
system
Token system
Feature activation+0.510

INTERVAL 1.093 - 1.457
CONTAINS 0.008%

one
Token one
Feature activation+0.000
might
Token might
Feature activation+0.000
think
Token think
Feature activation+0.000
there
Token there
Feature activation+1.130
would
Token would
Feature activation+0.795
be
Token be
Feature activation+1.387
increased
Token increased
Feature activation+0.273
interest
Token interest
Feature activation+0.000
in
Token in
Feature activation+0.000
this
Token this
Feature activation+0.000
week
Token week
Feature activation+0.000
Whatever
Token Whatever
Feature activation+0.000
.
Token.
Feature activation+0.000
You
Token You
Feature activation+0.000
think
Token think
Feature activation+0.000
I
Token I
Feature activation+1.082
shouldn
Token shouldn
Feature activation+1.203
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
feed
Token feed
Feature activation+0.000
him
Token him
Feature activation+0.000
ised
Tokenised
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+1.413
encouraged
Token encouraged
Feature activation+1.106
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
disenfranch
Token disenfranch
Feature activation+0.000
ised
Tokenised
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.000
we
Token we
Feature activation+0.000
think
Token think
Feature activation+0.000
will
Token will
Feature activation+1.202
be
Token be
Feature activation+1.164
a
Token a
Feature activation+0.412
complement
Token complement
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
residential
Token residential
Feature activation+0.000
.
Token.
Feature activation+0.000
And
Token And
Feature activation+0.000
what
Token what
Feature activation+0.000
he
Token he
Feature activation+0.000
thought
Token thought
Feature activation+0.000
was
Token was
Feature activation+1.417
that
Token that
Feature activation+0.000
Q
Token Q
Feature activation+0.000
E
TokenE
Feature activation+0.000
could
Token could
Feature activation+0.000
cure
Token cure
Feature activation+0.000

INTERVAL 0.729 - 1.093
CONTAINS 0.012%

but
Token but
Feature activation+0.000
I
Token I
Feature activation+0.000
still
Token still
Feature activation+0.000
think
Token think
Feature activation+0.000
Jones
Token Jones
Feature activation+0.433
is
Token is
Feature activation+0.768
looking
Token looking
Feature activation+0.301
for
Token for
Feature activation+0.000
a
Token a
Feature activation+0.000
bigger
Token bigger
Feature activation+0.000
ball
Token ball
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
We
TokenWe
Feature activation+0.000
think
Token think
Feature activation+0.000
this
Token this
Feature activation+0.472
is
Token is
Feature activation+0.917
the
Token the
Feature activation+0.018
simplest
Token simplest
Feature activation+0.000
,
Token,
Feature activation+0.000
most
Token most
Feature activation+0.000
straightforward
Token straightforward
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
never
Token never
Feature activation+0.000
think
Token think
Feature activation+0.000
it
Token it
Feature activation+0.907
's
Token's
Feature activation+1.561
going
Token going
Feature activation+0.824
to
Token to
Feature activation+0.813
be
Token be
Feature activation+0.780
as
Token as
Feature activation+0.000
bad
Token bad
Feature activation+0.000
as
Token as
Feature activation+0.000
if
Token if
Feature activation+0.000
he
Token he
Feature activation+0.000
thought
Token thought
Feature activation+0.000
that
Token that
Feature activation+0.000
would
Token would
Feature activation+0.000
help
Token help
Feature activation+0.776
him
Token him
Feature activation+0.000
save
Token save
Feature activation+0.000
his
Token his
Feature activation+0.000
skin
Token skin
Feature activation+0.000
.
Token.
Feature activation+0.000
.
Token.
Feature activation+0.000
At
Token At
Feature activation+0.000
first
Token first
Feature activation+0.000
he
Token he
Feature activation+0.000
thought
Token thought
Feature activation+0.000
the
Token the
Feature activation+0.835
bulbs
Token bulbs
Feature activation+0.254
were
Token were
Feature activation+0.485
part
Token part
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 0.364 - 0.729
CONTAINS 0.021%

'd
Token'd
Feature activation+0.000
think
Token think
Feature activation+0.000
it
Token it
Feature activation+2.256
'd
Token'd
Feature activation+0.663
be
Token be
Feature activation+2.291
really
Token really
Feature activation+0.483
hard
Token hard
Feature activation+0.000
to
Token to
Feature activation+0.000
relate
Token relate
Feature activation+0.000
to
Token to
Feature activation+0.000
a
Token a
Feature activation+0.000
I
Token I
Feature activation+0.000
honestly
Token honestly
Feature activation+0.000
don
Token don
Feature activation+0.000
't
Token't
Feature activation+0.000
think
Token think
Feature activation+0.000
they
Token they
Feature activation+0.462
'll
Token'll
Feature activation+0.000
be
Token be
Feature activation+0.440
another
Token another
Feature activation+0.000
period
Token period
Feature activation+0.000
like
Token like
Feature activation+0.000
which
Token which
Feature activation+0.000
we
Token we
Feature activation+0.000
think
Token think
Feature activation+0.000
will
Token will
Feature activation+1.202
be
Token be
Feature activation+1.164
a
Token a
Feature activation+0.412
complement
Token complement
Feature activation+0.000
to
Token to
Feature activation+0.000
the
Token the
Feature activation+0.000
residential
Token residential
Feature activation+0.000
,"
Token,"
Feature activation+0.000
court
Token court
Feature activation+0.000
if
Token if
Feature activation+0.000
they
Token they
Feature activation+0.000
feel
Token feel
Feature activation+0.000
their
Token their
Feature activation+0.000
rights
Token rights
Feature activation+0.494
are
Token are
Feature activation+0.269
being
Token being
Feature activation+0.000
infring
Token infring
Feature activation+0.000
ed
Tokened
Feature activation+0.000
upon
Token upon
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
We
TokenWe
Feature activation+0.000
thought
Token thought
Feature activation+0.000
we
Token we
Feature activation+0.390
were
Token were
Feature activation+0.464
immune
Token immune
Feature activation+0.173
to
Token to
Feature activation+0.000
their
Token their
Feature activation+0.000
jurisdiction
Token jurisdiction
Feature activation+0.000

INTERVAL 0.000 - 0.364
CONTAINS 99.949%

of
Token of
Feature activation+0.000
gender
Token gender
Feature activation+0.000
.'"
Token.'"
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
And
TokenAnd
Feature activation+0.000
finally
Token finally
Feature activation+0.000
,
Token,
Feature activation+0.000
they
Token they
Feature activation+0.000
asked
Token asked
Feature activation+0.000
if
Token if
Feature activation+0.000
,
Token,
Feature activation+0.000
pepper
Token pepper
Feature activation+0.000
,
Token,
Feature activation+0.000
ginger
Token ginger
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.000
and
Token and
Feature activation+0.000
opium
Token opium
Feature activation+0.000
,
Token,
Feature activation+0.000
which
Token which
Feature activation+0.000
did
Token did
Feature activation+0.000
not
Token not
Feature activation+0.000
memories
Token memories
Feature activation+0.000
are
Token are
Feature activation+0.000
short
Token short
Feature activation+0.000
,"
Token,"
Feature activation+0.000
says
Token says
Feature activation+0.000
Fitz
Token Fitz
Feature activation+0.000
g
Tokeng
Feature activation+0.000
ib
Tokenib
Feature activation+0.000
bon
Tokenbon
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
the
Token the
Feature activation+0.000
civil
Token civil
Feature activation+0.000
war
Token war
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
IDF
Token IDF
Feature activation+0.000
has
Token has
Feature activation+0.000
set
Token set
Feature activation+0.000
up
Token up
Feature activation+0.000
a
Token a
Feature activation+0.000
to
Token to
Feature activation+0.000
claim
Token claim
Feature activation+0.000
back
Token back
Feature activation+0.000
real
Token real
Feature activation+0.000
money
Token money
Feature activation+0.000
that
Token that
Feature activation+0.000
they
Token they
Feature activation+0.000
have
Token have
Feature activation+0.000
spent
Token spent
Feature activation+0.000
on
Token on
Feature activation+0.000
buying
Token buying
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 H3.8: (feature 1678)

TOP ACTIVATIONS
MAX = 3.333

quietly
Token quietly
Feature activation+0.000
feel
Token feel
Feature activation+0.000
mis
Token mis
Feature activation+1.544
g
Tokeng
Feature activation+0.000
iv
Tokeniv
Feature activation+0.000
ings
Tokenings
Feature activation+3.333
and
Token and
Feature activation+0.000
tw
Token tw
Feature activation+0.000
inges
Tokeninges
Feature activation+0.000
of
Token of
Feature activation+0.000
conscience
Token conscience
Feature activation+0.000
what
Token what
Feature activation+0.000
brands
Token brands
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.681
doing
Token doing
Feature activation+1.997
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
interesting
Token interesting
Feature activation+0.574
and
Token and
Feature activation+0.000
court
Token court
Feature activation+0.000
if
Token if
Feature activation+0.000
they
Token they
Feature activation+0.000
feel
Token feel
Feature activation+0.000
their
Token their
Feature activation+1.319
rights
Token rights
Feature activation+2.657
are
Token are
Feature activation+2.680
being
Token being
Feature activation+1.599
infring
Token infring
Feature activation+0.232
ed
Tokened
Feature activation+0.597
upon
Token upon
Feature activation+0.000
imperfect
Token imperfect
Feature activation+0.000
ions
Tokenions
Feature activation+0.000
that
Token that
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.341
truly
Token truly
Feature activation+0.537
an
Token an
Feature activation+0.000
important
Token important
Feature activation+0.354
factor
Token factor
Feature activation+0.465
of
Token of
Feature activation+0.000
alive
Token alive
Feature activation+0.921
,
Token,
Feature activation+0.323
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
more
Token more
Feature activation+0.000
focused
Token focused
Feature activation+2.172
,
Token,
Feature activation+0.282
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
more
Token more
Feature activation+0.000
energetic
Token energetic
Feature activation+0.451
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
government
Token government
Feature activation+0.000
felt
Token felt
Feature activation+0.000
children
Token children
Feature activation+2.141
were
Token were
Feature activation+0.991
easier
Token easier
Feature activation+0.000
to
Token to
Feature activation+0.000
mould
Token mould
Feature activation+0.000
than
Token than
Feature activation+0.000
list
Token list
Feature activation+0.000
of
Token of
Feature activation+0.000
territories
Token territories
Feature activation+0.000
China
Token China
Feature activation+0.000
feels
Token feels
Feature activation+0.000
were
Token were
Feature activation+2.101
unfairly
Token unfairly
Feature activation+0.297
taken
Token taken
Feature activation+0.430
away
Token away
Feature activation+0.000
when
Token when
Feature activation+0.000
it
Token it
Feature activation+0.000
.
Token .
Feature activation+0.000
.
Token .
Feature activation+0.000
.
Token .
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
assured
Token assured
Feature activation+2.016
of
Token of
Feature activation+1.291
the
Token the
Feature activation+0.000
police
Token police
Feature activation+0.000
department
Token department
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
of
Token of
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
it
Token it
Feature activation+0.000
felt
Token felt
Feature activation+0.000
very
Token very
Feature activation+0.000
much
Token much
Feature activation+2.006
like
Token like
Feature activation+0.000
a
Token a
Feature activation+0.000
coming
Token coming
Feature activation+0.000
together
Token together
Feature activation+0.000
of
Token of
Feature activation+0.000
brands
Token brands
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.681
doing
Token doing
Feature activation+1.997
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
interesting
Token interesting
Feature activation+0.574
and
Token and
Feature activation+0.000
innovative
Token innovative
Feature activation+0.000
out
Token out
Feature activation+0.000
that
Token that
Feature activation+0.000
D
Token D
Feature activation+0.000
olan
Tokenolan
Feature activation+0.000
felt
Token felt
Feature activation+0.000
âĢ
Token âĢ
Feature activation+1.919
ľ
Tokenľ
Feature activation+0.000
bet
Tokenbet
Feature activation+0.541
ray
Tokenray
Feature activation+0.000
ed
Tokened
Feature activation+0.543
âĢ
TokenâĢ
Feature activation+0.000
outline
Token outline
Feature activation+0.000
what
Token what
Feature activation+0.000
they
Token they
Feature activation+0.000
felt
Token felt
Feature activation+0.000
ought
Token ought
Feature activation+0.746
to
Token to
Feature activation+1.865
be
Token be
Feature activation+1.407
the
Token the
Feature activation+0.000
priorities
Token priorities
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
people
Token people
Feature activation+0.000
who
Token who
Feature activation+0.000
feel
Token feel
Feature activation+0.000
like
Token like
Feature activation+0.000
outsiders
Token outsiders
Feature activation+0.010
trying
Token trying
Feature activation+1.853
to
Token to
Feature activation+0.000
become
Token become
Feature activation+0.000
Dan
Token Dan
Feature activation+0.000
Harmon
Token Harmon
Feature activation+0.000
or
Token or
Feature activation+0.000
felt
Token felt
Feature activation+0.000
this
Token this
Feature activation+0.000
was
Token was
Feature activation+1.449
a
Token a
Feature activation+0.000
valid
Token valid
Feature activation+0.000
choice
Token choice
Feature activation+1.847
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
A
TokenA
Feature activation+0.000
mother
Token mother
Feature activation+0.000
Suddenly
TokenSuddenly
Feature activation+0.000
,
Token,
Feature activation+0.000
Weiss
Token Weiss
Feature activation+0.000
felt
Token felt
Feature activation+0.000
a
Token a
Feature activation+0.000
pair
Token pair
Feature activation+1.802
of
Token of
Feature activation+0.519
hands
Token hands
Feature activation+0.302
start
Token start
Feature activation+0.000
to
Token to
Feature activation+0.000
crawl
Token crawl
Feature activation+0.000
post
Token post
Feature activation+0.000
if
Token if
Feature activation+0.000
he
Token he
Feature activation+0.000
felt
Token felt
Feature activation+0.000
he
Token he
Feature activation+0.884
were
Token were
Feature activation+1.781
being
Token being
Feature activation+0.000
asked
Token asked
Feature activation+0.445
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
something
Token something
Feature activation+0.000
weather
Token weather
Feature activation+0.000
but
Token but
Feature activation+0.000
I
Token I
Feature activation+0.000
somehow
Token somehow
Feature activation+0.000
feel
Token feel
Feature activation+0.000
responsible
Token responsible
Feature activation+1.685
for
Token for
Feature activation+0.000
that
Token that
Feature activation+0.000
,
Token,
Feature activation+0.000
as
Token as
Feature activation+0.000
well
Token well
Feature activation+0.000
t
Tokent
Feature activation+0.000
wear
Token wear
Feature activation+0.000
short
Token short
Feature activation+0.000
skirts
Token skirts
Feature activation+0.000
feels
Token feels
Feature activation+0.000
wrong
Token wrong
Feature activation+1.655
,
Token,
Feature activation+0.000
Say
Token Say
Feature activation+0.000
re
Tokenre
Feature activation+0.000
said
Token said
Feature activation+0.000
,
Token,
Feature activation+0.000
have
Token have
Feature activation+0.000
started
Token started
Feature activation+0.000
to
Token to
Feature activation+0.000
feel
Token feel
Feature activation+0.000
my
Token my
Feature activation+0.318
shot
Token shot
Feature activation+1.603
;
Token;
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
more
Token more
Feature activation+0.000
confidence
Token confidence
Feature activation+0.000
I
Token I
Feature activation+0.000
do
Token do
Feature activation+0.000
not
Token not
Feature activation+0.000
feel
Token feel
Feature activation+0.000
this
Token this
Feature activation+0.870
is
Token is
Feature activation+1.601
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.505
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000

Top DFA by src position
MAX = 4.831

.
Token.
Feature activation+0.043
Top resid features:
Some
Token Some
Feature activation+0.074
Top resid features:
of
Token of
Feature activation+0.017
Top resid features:
you
Token you
Feature activation-0.028
Top resid features:
quietly
Token quietly
Feature activation-0.033
Top resid features:
feel
Token feel
Feature activation+4.831
Top resid features:
mis
Token mis
Feature activation+0.003
Top resid features:
g
Tokeng
Feature activation+0.010
Top resid features:
iv
Tokeniv
Feature activation+0.029
Top resid features:
ings
Tokenings
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.036
Top resid features:
what
Token what
Feature activation-0.124
Top resid features:
brands
Token brands
Feature activation+0.021
Top resid features:
do
Token do
Feature activation-0.045
Top resid features:
you
Token you
Feature activation-0.050
Top resid features:
feel
Token feel
Feature activation+4.339
Top resid features:
are
Token are
Feature activation+0.039
Top resid features:
doing
Token doing
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
most
Token most
Feature activation+0.000
Top resid features:
interesting
Token interesting
Feature activation+0.000
Top resid features:
to
Token to
Feature activation-0.003
Top resid features:
federal
Token federal
Feature activation-0.015
Top resid features:
court
Token court
Feature activation-0.040
Top resid features:
if
Token if
Feature activation-0.010
Top resid features:
they
Token they
Feature activation-0.028
Top resid features:
feel
Token feel
Feature activation+4.055
Top resid features:
their
Token their
Feature activation+0.310
Top resid features:
rights
Token rights
Feature activation-0.085
Top resid features:
are
Token are
Feature activation+0.000
Top resid features:
being
Token being
Feature activation+0.000
Top resid features:
infring
Token infring
Feature activation+0.000
Top resid features:
mild
Token mild
Feature activation-0.037
Top resid features:
imperfect
Token imperfect
Feature activation-0.053
Top resid features:
ions
Tokenions
Feature activation+0.010
Top resid features:
that
Token that
Feature activation+0.028
Top resid features:
I
Token I
Feature activation+0.050
Top resid features:
feel
Token feel
Feature activation+3.892
Top resid features:
are
Token are
Feature activation+0.112
Top resid features:
truly
Token truly
Feature activation+0.000
Top resid features:
an
Token an
Feature activation+0.000
Top resid features:
important
Token important
Feature activation+0.000
Top resid features:
factor
Token factor
Feature activation+0.000
Top resid features:
feel
Token feel
Feature activation+0.157
Top resid features:
more
Token more
Feature activation+0.009
Top resid features:
alive
Token alive
Feature activation+0.026
Top resid features:
,
Token,
Feature activation+0.110
Top resid features:
I
Token I
Feature activation+0.116
Top resid features:
feel
Token feel
Feature activation+2.835
Top resid features:
more
Token more
Feature activation+0.218
Top resid features:
focused
Token focused
Feature activation+0.151
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
I
Token I
Feature activation+0.000
Top resid features:
feel
Token feel
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.025
Top resid features:
Ċ
TokenĊ
Feature activation+0.014
Top resid features:
Ċ
TokenĊ
Feature activation+0.027
Top resid features:
The
TokenThe
Feature activation-0.058
Top resid features:
government
Token government
Feature activation+0.019
Top resid features:
felt
Token felt
Feature activation+3.330
Top resid features:
children
Token children
Feature activation+0.049
Top resid features:
were
Token were
Feature activation+0.000
Top resid features:
easier
Token easier
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
mould
Token mould
Feature activation+0.000
Top resid features:
a
Token a
Feature activation-0.012
Top resid features:
list
Token list
Feature activation-0.040
Top resid features:
of
Token of
Feature activation-0.012
Top resid features:
territories
Token territories
Feature activation-0.017
Top resid features:
China
Token China
Feature activation-0.006
Top resid features:
feels
Token feels
Feature activation+3.568
Top resid features:
were
Token were
Feature activation+0.151
Top resid features:
unfairly
Token unfairly
Feature activation+0.000
Top resid features:
taken
Token taken
Feature activation+0.000
Top resid features:
away
Token away
Feature activation+0.000
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.017
Top resid features:
.
Token .
Feature activation-0.006
Top resid features:
.
Token .
Feature activation-0.011
Top resid features:
.
Token .
Feature activation-0.027
Top resid features:
I
Token I
Feature activation+0.074
Top resid features:
feel
Token feel
Feature activation+3.346
Top resid features:
assured
Token assured
Feature activation+0.098
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
police
Token police
Feature activation+0.000
Top resid features:
department
Token department
Feature activation+0.000
Top resid features:
For
TokenFor
Feature activation-0.002
Top resid features:
much
Token much
Feature activation-0.009
Top resid features:
of
Token of
Feature activation+0.001
Top resid features:
Saturday
Token Saturday
Feature activation-0.005
Top resid features:
it
Token it
Feature activation-0.021
Top resid features:
felt
Token felt
Feature activation+3.222
Top resid features:
very
Token very
Feature activation+0.172
Top resid features:
much
Token much
Feature activation+0.147
Top resid features:
like
Token like
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
coming
Token coming
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.037
Top resid features:
what
Token what
Feature activation-0.088
Top resid features:
brands
Token brands
Feature activation+0.015
Top resid features:
do
Token do
Feature activation-0.011
Top resid features:
you
Token you
Feature activation-0.033
Top resid features:
feel
Token feel
Feature activation+3.483
Top resid features:
are
Token are
Feature activation+0.149
Top resid features:
doing
Token doing
Feature activation-0.037
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
most
Token most
Feature activation+0.000
Top resid features:
interesting
Token interesting
Feature activation+0.000
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation+0.247
Top resid features:
out
Token out
Feature activation-0.020
Top resid features:
that
Token that
Feature activation+0.059
Top resid features:
D
Token D
Feature activation+0.072
Top resid features:
olan
Tokenolan
Feature activation+0.010
Top resid features:
felt
Token felt
Feature activation+3.074
Top resid features:
âĢ
Token âĢ
Feature activation+0.110
Top resid features:
ľ
Tokenľ
Feature activation+0.000
Top resid features:
bet
Tokenbet
Feature activation+0.000
Top resid features:
ray
Tokenray
Feature activation+0.000
Top resid features:
ed
Tokened
Feature activation+0.000
Top resid features:
let
Token let
Feature activation-0.002
Top resid features:
them
Token them
Feature activation-0.019
Top resid features:
outline
Token outline
Feature activation-0.091
Top resid features:
what
Token what
Feature activation-0.069
Top resid features:
they
Token they
Feature activation-0.094
Top resid features:
felt
Token felt
Feature activation+3.610
Top resid features:
ought
Token ought
Feature activation+0.094
Top resid features:
to
Token to
Feature activation-0.009
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
priorities
Token priorities
Feature activation+0.000
Top resid features:
His
Token His
Feature activation-0.010
Top resid features:
audience
Token audience
Feature activation+0.025
Top resid features:
are
Token are
Feature activation-0.039
Top resid features:
people
Token people
Feature activation-0.013
Top resid features:
who
Token who
Feature activation+0.015
Top resid features:
feel
Token feel
Feature activation+2.612
Top resid features:
like
Token like
Feature activation+0.507
Top resid features:
outsiders
Token outsiders
Feature activation+0.250
Top resid features:
trying
Token trying
Feature activation-0.020
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
become
Token become
Feature activation+0.000
Top resid features:
asked
Token asked
Feature activation-0.018
Top resid features:
of
Token of
Feature activation+0.006
Top resid features:
a
Token a
Feature activation-0.007
Top resid features:
woman
Token woman
Feature activation+0.006
Top resid features:
who
Token who
Feature activation+0.035
Top resid features:
felt
Token felt
Feature activation+3.344
Top resid features:
this
Token this
Feature activation+0.063
Top resid features:
was
Token was
Feature activation+0.076
Top resid features:
a
Token a
Feature activation+0.045
Top resid features:
valid
Token valid
Feature activation-0.046
Top resid features:
choice
Token choice
Feature activation-0.026
Top resid features:
Ċ
TokenĊ
Feature activation+0.015
Top resid features:
Ċ
TokenĊ
Feature activation+0.016
Top resid features:
Suddenly
TokenSuddenly
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.024
Top resid features:
Weiss
Token Weiss
Feature activation+0.098
Top resid features:
felt
Token felt
Feature activation+2.833
Top resid features:
a
Token a
Feature activation+0.167
Top resid features:
pair
Token pair
Feature activation-0.013
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
hands
Token hands
Feature activation+0.000
Top resid features:
start
Token start
Feature activation+0.000
Top resid features:
leave
Token leave
Feature activation-0.007
Top resid features:
his
Token his
Feature activation+0.000
Top resid features:
post
Token post
Feature activation-0.017
Top resid features:
if
Token if
Feature activation+0.053
Top resid features:
he
Token he
Feature activation-0.023
Top resid features:
felt
Token felt
Feature activation+3.216
Top resid features:
he
Token he
Feature activation+0.173
Top resid features:
were
Token were
Feature activation-0.022
Top resid features:
being
Token being
Feature activation+0.000
Top resid features:
asked
Token asked
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.013
Top resid features:
weather
Token weather
Feature activation+0.009
Top resid features:
but
Token but
Feature activation+0.084
Top resid features:
I
Token I
Feature activation+0.075
Top resid features:
somehow
Token somehow
Feature activation-0.069
Top resid features:
feel
Token feel
Feature activation+3.041
Top resid features:
responsible
Token responsible
Feature activation+0.092
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
as
Token as
Feature activation+0.000
Top resid features:
Ļ
TokenĻ
Feature activation+0.034
Top resid features:
t
Tokent
Feature activation-0.005
Top resid features:
wear
Token wear
Feature activation-0.033
Top resid features:
short
Token short
Feature activation-0.011
Top resid features:
skirts
Token skirts
Feature activation+0.007
Top resid features:
feels
Token feels
Feature activation+3.116
Top resid features:
wrong
Token wrong
Feature activation+0.001
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
Say
Token Say
Feature activation+0.000
Top resid features:
re
Tokenre
Feature activation+0.000
Top resid features:
said
Token said
Feature activation+0.000
Top resid features:
âĢĵ
TokenâĢĵ
Feature activation-0.021
Top resid features:
I
Token I
Feature activation+0.019
Top resid features:
have
Token have
Feature activation-0.007
Top resid features:
started
Token started
Feature activation+0.217
Top resid features:
to
Token to
Feature activation+0.179
Top resid features:
feel
Token feel
Feature activation+2.131
Top resid features:
my
Token my
Feature activation+0.342
Top resid features:
shot
Token shot
Feature activation+0.085
Top resid features:
;
Token;
Feature activation+0.000
Top resid features:
I
Token I
Feature activation+0.000
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.038
Top resid features:
but
Token but
Feature activation-0.094
Top resid features:
I
Token I
Feature activation+0.011
Top resid features:
do
Token do
Feature activation+0.009
Top resid features:
not
Token not
Feature activation-0.097
Top resid features:
feel
Token feel
Feature activation+2.890
Top resid features:
this
Token this
Feature activation+0.446
Top resid features:
is
Token is
Feature activation-0.003
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
most
Token most
Feature activation+0.000
Top resid features:
important
Token important
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.08

Head 2: 0.08

Head 3: 0.03

Head 4: 0.02

Head 5: 0.03

Head 6: 0.06

Head 7: 0.09

Head 8: 0.42

Head 9: 0.07

Head 10: 0.06

Head 11: 0.05

Positive logits

fits1.53

ould1.36

should1.30

shouldn1.29

atched1.28

suited1.25

ija1.20

ropri1.20

Valiant1.18

orously1.17

wouldn1.13

lim1.13

Ground1.13

alert1.12

should1.12

Semitism1.11

tern1.10

Safe1.10

Normal1.09

fetched1.09

Negative logits

incorpor-1.30

Avg-1.22

GM-1.19

azeera-1.18

councillor-1.17

associate-1.15

adan-1.15

ocate-1.13

Gil-1.12

endment-1.11

��-1.09

welf-1.08

constitu-1.07

Calvin-1.06

ouston-1.05

incent-1.04

Expend-1.03

覚醒-1.03

EE-1.03

evangel-1.03

INTERVAL 3.000 - 3.333
CONTAINS 0.000%

quietly
Token quietly
Feature activation+0.000
feel
Token feel
Feature activation+0.000
mis
Token mis
Feature activation+1.544
g
Tokeng
Feature activation+0.000
iv
Tokeniv
Feature activation+0.000
ings
Tokenings
Feature activation+3.333
and
Token and
Feature activation+0.000
tw
Token tw
Feature activation+0.000
inges
Tokeninges
Feature activation+0.000
of
Token of
Feature activation+0.000
conscience
Token conscience
Feature activation+0.000

INTERVAL 2.667 - 3.000
CONTAINS 0.000%

what
Token what
Feature activation+0.000
brands
Token brands
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.681
doing
Token doing
Feature activation+1.997
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
interesting
Token interesting
Feature activation+0.574
and
Token and
Feature activation+0.000

INTERVAL 2.333 - 2.667
CONTAINS 0.000%

court
Token court
Feature activation+0.000
if
Token if
Feature activation+0.000
they
Token they
Feature activation+0.000
feel
Token feel
Feature activation+0.000
their
Token their
Feature activation+1.319
rights
Token rights
Feature activation+2.657
are
Token are
Feature activation+2.680
being
Token being
Feature activation+1.599
infring
Token infring
Feature activation+0.232
ed
Tokened
Feature activation+0.597
upon
Token upon
Feature activation+0.000
imperfect
Token imperfect
Feature activation+0.000
ions
Tokenions
Feature activation+0.000
that
Token that
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.341
truly
Token truly
Feature activation+0.537
an
Token an
Feature activation+0.000
important
Token important
Feature activation+0.354
factor
Token factor
Feature activation+0.465
of
Token of
Feature activation+0.000

INTERVAL 2.000 - 2.333
CONTAINS 0.001%

alive
Token alive
Feature activation+0.921
,
Token,
Feature activation+0.323
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
more
Token more
Feature activation+0.000
focused
Token focused
Feature activation+2.172
,
Token,
Feature activation+0.282
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
more
Token more
Feature activation+0.000
energetic
Token energetic
Feature activation+0.451
.
Token .
Feature activation+0.000
.
Token .
Feature activation+0.000
.
Token .
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
assured
Token assured
Feature activation+2.016
of
Token of
Feature activation+1.291
the
Token the
Feature activation+0.000
police
Token police
Feature activation+0.000
department
Token department
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
of
Token of
Feature activation+0.000
Saturday
Token Saturday
Feature activation+0.000
it
Token it
Feature activation+0.000
felt
Token felt
Feature activation+0.000
very
Token very
Feature activation+0.000
much
Token much
Feature activation+2.006
like
Token like
Feature activation+0.000
a
Token a
Feature activation+0.000
coming
Token coming
Feature activation+0.000
together
Token together
Feature activation+0.000
of
Token of
Feature activation+0.000
list
Token list
Feature activation+0.000
of
Token of
Feature activation+0.000
territories
Token territories
Feature activation+0.000
China
Token China
Feature activation+0.000
feels
Token feels
Feature activation+0.000
were
Token were
Feature activation+2.101
unfairly
Token unfairly
Feature activation+0.297
taken
Token taken
Feature activation+0.430
away
Token away
Feature activation+0.000
when
Token when
Feature activation+0.000
it
Token it
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
government
Token government
Feature activation+0.000
felt
Token felt
Feature activation+0.000
children
Token children
Feature activation+2.141
were
Token were
Feature activation+0.991
easier
Token easier
Feature activation+0.000
to
Token to
Feature activation+0.000
mould
Token mould
Feature activation+0.000
than
Token than
Feature activation+0.000

INTERVAL 1.667 - 2.000
CONTAINS 0.001%

out
Token out
Feature activation+0.000
that
Token that
Feature activation+0.000
D
Token D
Feature activation+0.000
olan
Tokenolan
Feature activation+0.000
felt
Token felt
Feature activation+0.000
âĢ
Token âĢ
Feature activation+1.919
ľ
Tokenľ
Feature activation+0.000
bet
Tokenbet
Feature activation+0.541
ray
Tokenray
Feature activation+0.000
ed
Tokened
Feature activation+0.543
âĢ
TokenâĢ
Feature activation+0.000
felt
Token felt
Feature activation+0.000
this
Token this
Feature activation+0.000
was
Token was
Feature activation+1.449
a
Token a
Feature activation+0.000
valid
Token valid
Feature activation+0.000
choice
Token choice
Feature activation+1.847
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
A
TokenA
Feature activation+0.000
mother
Token mother
Feature activation+0.000
brands
Token brands
Feature activation+0.000
do
Token do
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
are
Token are
Feature activation+2.681
doing
Token doing
Feature activation+1.997
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
interesting
Token interesting
Feature activation+0.574
and
Token and
Feature activation+0.000
innovative
Token innovative
Feature activation+0.000
people
Token people
Feature activation+0.000
who
Token who
Feature activation+0.000
feel
Token feel
Feature activation+0.000
like
Token like
Feature activation+0.000
outsiders
Token outsiders
Feature activation+0.010
trying
Token trying
Feature activation+1.853
to
Token to
Feature activation+0.000
become
Token become
Feature activation+0.000
Dan
Token Dan
Feature activation+0.000
Harmon
Token Harmon
Feature activation+0.000
or
Token or
Feature activation+0.000
outline
Token outline
Feature activation+0.000
what
Token what
Feature activation+0.000
they
Token they
Feature activation+0.000
felt
Token felt
Feature activation+0.000
ought
Token ought
Feature activation+0.746
to
Token to
Feature activation+1.865
be
Token be
Feature activation+1.407
the
Token the
Feature activation+0.000
priorities
Token priorities
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 1.333 - 1.667
CONTAINS 0.002%

contract
Token contract
Feature activation+0.000
.
Token.
Feature activation+0.000
So
Token So
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
I
Token I
Feature activation+1.414
can
Token can
Feature activation+0.572
definitely
Token definitely
Feature activation+0.000
handle
Token handle
Feature activation+0.000
30
Token 30
Feature activation+0.000
touches
Token touches
Feature activation+0.000
I
Token I
Feature activation+0.000
also
Token also
Feature activation+0.000
feel
Token feel
Feature activation+0.000
in
Token in
Feature activation+1.201
tune
Token tune
Feature activation+0.000
with
Token with
Feature activation+1.557
my
Token my
Feature activation+0.000
awareness
Token awareness
Feature activation+0.000
toward
Token toward
Feature activation+0.000
the
Token the
Feature activation+0.000
bre
Token bre
Feature activation+0.000
be
Token be
Feature activation+0.000
the
Token the
Feature activation+0.000
reason
Token reason
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
young
Token young
Feature activation+1.571
,
Token,
Feature activation+0.362
the
Token the
Feature activation+0.000
reason
Token reason
Feature activation+0.000
you
Token you
Feature activation+0.000
love
Token love
Feature activation+0.000
have
Token have
Feature activation+0.000
started
Token started
Feature activation+0.000
to
Token to
Feature activation+0.000
feel
Token feel
Feature activation+0.000
my
Token my
Feature activation+0.318
shot
Token shot
Feature activation+1.603
;
Token;
Feature activation+0.000
I
Token I
Feature activation+0.000
have
Token have
Feature activation+0.000
more
Token more
Feature activation+0.000
confidence
Token confidence
Feature activation+0.000
t
Tokent
Feature activation+0.000
wear
Token wear
Feature activation+0.000
short
Token short
Feature activation+0.000
skirts
Token skirts
Feature activation+0.000
feels
Token feels
Feature activation+0.000
wrong
Token wrong
Feature activation+1.655
,
Token,
Feature activation+0.000
Say
Token Say
Feature activation+0.000
re
Tokenre
Feature activation+0.000
said
Token said
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 1.000 - 1.333
CONTAINS 0.004%

.
Token.
Feature activation+0.000
"
Token "
Feature activation+0.000
I
TokenI
Feature activation+0.000
feel
Token feel
Feature activation+0.000
so
Token so
Feature activation+0.000
blessed
Token blessed
Feature activation+1.228
to
Token to
Feature activation+0.526
have
Token have
Feature activation+0.000
been
Token been
Feature activation+0.000
able
Token able
Feature activation+0.000
to
Token to
Feature activation+0.000
middle
Token middle
Feature activation+0.000
-
Token-
Feature activation+0.000
class
Tokenclass
Feature activation+0.000
Americans
Token Americans
Feature activation+0.000
felt
Token felt
Feature activation+0.000
themselves
Token themselves
Feature activation+1.295
to
Token to
Feature activation+0.594
be
Token be
Feature activation+1.345
on
Token on
Feature activation+0.000
relatively
Token relatively
Feature activation+0.000
secure
Token secure
Feature activation+0.000
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
difficult
Token difficult
Feature activation+0.000
or
Token or
Feature activation+0.000
feel
Token feel
Feature activation+0.000
important
Token important
Feature activation+1.128
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
because
Token because
Feature activation+0.000
we
Token we
Feature activation+0.000
know
Token know
Feature activation+0.000
because
Token because
Feature activation+0.000
I
Token I
Feature activation+0.000
feel
Token feel
Feature activation+0.000
this
Token this
Feature activation+0.308
film
Token film
Feature activation+0.038
is
Token is
Feature activation+1.152
difficult
Token difficult
Feature activation+0.219
to
Token to
Feature activation+0.000
dissect
Token dissect
Feature activation+0.000
without
Token without
Feature activation+0.000
going
Token going
Feature activation+0.000
âĢ
Token âĢ
Feature activation+1.128
ľ
Tokenľ
Feature activation+0.313
h
Tokenh
Feature activation+0.194
umbled
Tokenumbled
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+1.021
to
Token to
Feature activation+0.000
look
Token look
Feature activation+0.000
through
Token through
Feature activation+0.000
the
Token the
Feature activation+0.000
grav
Token grav
Feature activation+0.000

INTERVAL 0.667 - 1.000
CONTAINS 0.007%

.
Token.
Feature activation+0.000
Everyone
Token Everyone
Feature activation+0.000
will
Token will
Feature activation+0.000
feel
Token feel
Feature activation+0.000
good
Token good
Feature activation+0.000
about
Token about
Feature activation+0.775
what
Token what
Feature activation+0.000
they
Token they
Feature activation+0.000
do
Token do
Feature activation+0.000
for
Token for
Feature activation+0.000
a
Token a
Feature activation+0.000
officers
Token officers
Feature activation+0.000
"
Token "
Feature activation+0.000
felt
Tokenfelt
Feature activation+0.000
the
Token the
Feature activation+0.000
suspect
Token suspect
Feature activation+1.158
presented
Token presented
Feature activation+0.728
an
Token an
Feature activation+0.000
imminent
Token imminent
Feature activation+0.000
danger
Token danger
Feature activation+0.000
,"
Token,"
Feature activation+0.000
said
Token said
Feature activation+0.000
to
Token to
Feature activation+0.000
move
Token move
Feature activation+0.000
on
Token on
Feature activation+0.000
and
Token and
Feature activation+0.000
feel
Token feel
Feature activation+0.000
members
Token members
Feature activation+0.960
should
Token should
Feature activation+0.000
keep
Token keep
Feature activation+0.000
looking
Token looking
Feature activation+0.000
for
Token for
Feature activation+0.000
a
Token a
Feature activation+0.000
rankings
Token rankings
Feature activation+0.000
.
Token.
Feature activation+0.000
If
Token If
Feature activation+0.000
you
Token you
Feature activation+0.000
feel
Token feel
Feature activation+0.000
reckless
Token reckless
Feature activation+0.897
,
Token,
Feature activation+0.405
then
Token then
Feature activation+0.000
try
Token try
Feature activation+0.000
to
Token to
Feature activation+0.000
be
Token be
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
car
Token car
Feature activation+0.000
felt
Token felt
Feature activation+0.000
great
Token great
Feature activation+0.000
and
Token and
Feature activation+0.926
I
Token I
Feature activation+0.000
got
Token got
Feature activation+0.000
100
Token 100
Feature activation+0.000
%
Token%
Feature activation+0.000
out
Token out
Feature activation+0.000

INTERVAL 0.333 - 0.667
CONTAINS 0.011%

our
Token our
Feature activation+0.000
young
Token young
Feature activation+0.000
creative
Token creative
Feature activation+0.000
staff
Token staff
Feature activation+0.000
felt
Token felt
Feature activation+0.000
at
Token at
Feature activation+0.624
home
Token home
Feature activation+0.000
,"
Token,"
Feature activation+0.000
Prol
Token Prol
Feature activation+0.000
explains
Token explains
Feature activation+0.000
.
Token.
Feature activation+0.000
difficult
Token difficult
Feature activation+0.000
time
Token time
Feature activation+0.000
and
Token and
Feature activation+0.000
feel
Token feel
Feature activation+0.000
very
Token very
Feature activation+0.000
respectful
Token respectful
Feature activation+0.443
and
Token and
Feature activation+0.489
grateful
Token grateful
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
actor
Token actor
Feature activation+0.000
to
Token to
Feature activation+0.000
feel
Token feel
Feature activation+0.000
the
Token the
Feature activation+0.000
character
Token character
Feature activation+0.414
.
Token.
Feature activation+0.000
God
Token God
Feature activation+0.000
IS
Token IS
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
.
Token.
Feature activation+0.000
Sell
Token Sell
Feature activation+0.000
ers
Tokeners
Feature activation+0.000
felt
Token felt
Feature activation+0.000
he
Token he
Feature activation+0.596
was
Token was
Feature activation+0.370
already
Token already
Feature activation+0.000
stretched
Token stretched
Feature activation+0.000
with
Token with
Feature activation+0.000
3
Token 3
Feature activation+0.000
roles
Token roles
Feature activation+0.000
D
Token D
Feature activation+0.000
olan
Tokenolan
Feature activation+0.000
felt
Token felt
Feature activation+0.000
âĢ
Token âĢ
Feature activation+1.919
ľ
Tokenľ
Feature activation+0.000
bet
Tokenbet
Feature activation+0.541
ray
Tokenray
Feature activation+0.000
ed
Tokened
Feature activation+0.543
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.336
by
Token by
Feature activation+0.000

INTERVAL 0.000 - 0.333
CONTAINS 99.975%

that
Token that
Feature activation+0.000
now
Token now
Feature activation+0.000
I
Token I
Feature activation+0.000
not
Token not
Feature activation+0.000
only
Token only
Feature activation+0.000
have
Token have
Feature activation+0.000
to
Token to
Feature activation+0.000
deal
Token deal
Feature activation+0.000
with
Token with
Feature activation+0.000
the
Token the
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
Comb
TokenComb
Feature activation+0.000
ine
Tokenine
Feature activation+0.000
gin
Token gin
Feature activation+0.000
&
Token &
Feature activation+0.000
l
Token l
Feature activation+0.000
ique
Tokenique
Feature activation+0.000
ur
Tokenur
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
cocktail
Token cocktail
Feature activation+0.000
sh
Token sh
Feature activation+0.000
.
Token.
Feature activation+0.000
Apart
Token Apart
Feature activation+0.000
from
Token from
Feature activation+0.000
ex
Token ex
Feature activation+0.000
uber
Tokenuber
Feature activation+0.000
ant
Tokenant
Feature activation+0.000
dancing
Token dancing
Feature activation+0.000
,
Token,
Feature activation+0.000
u
Token u
Feature activation+0.000
k
Tokenk
Feature activation+0.000
ule
Tokenule
Feature activation+0.000
ulic
Tokenulic
Feature activation+0.000
said
Token said
Feature activation+0.000
he
Token he
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
first
Token first
Feature activation+0.000
player
Token player
Feature activation+0.000
born
Token born
Feature activation+0.000
and
Token and
Feature activation+0.000
raised
Token raised
Feature activation+0.000
in
Token in
Feature activation+0.000
expert
Token expert
Feature activation+0.000
groups
Token groups
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
"
Token "
Feature activation+0.000
personal
Tokenpersonal
Feature activation+0.000
capacity
Token capacity
Feature activation+0.000
."
Token."
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Other
TokenOther
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 H3.8: (feature 4662)

TOP ACTIVATIONS
MAX = 3.433

<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
only
Token only
Feature activation+0.000
make
Token make
Feature activation+0.000
employers
Token employers
Feature activation+2.181
'
Token'
Feature activation+1.860
jobs
Token jobs
Feature activation+3.433
harder
Token harder
Feature activation+0.162
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
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+0.000
make
Token make
Feature activation+0.000
getting
Token getting
Feature activation+3.432
at
Token at
Feature activation+0.689
customer
Token customer
Feature activation+0.000
data
Token data
Feature activation+0.274
possible
Token possible
Feature activation+0.000
."
Token."
Feature activation+0.000
,
Token,
Feature activation+0.000
cities
Token cities
Feature activation+0.000
)
Token)
Feature activation+0.000
can
Token can
Feature activation+0.000
make
Token make
Feature activation+0.000
riding
Token riding
Feature activation+3.329
more
Token more
Feature activation+0.000
appealing
Token appealing
Feature activation+0.022
to
Token to
Feature activation+0.000
women
Token women
Feature activation+0.000
by
Token by
Feature activation+0.000
status
Token status
Feature activation+0.000
of
Token of
Feature activation+0.000
software
Token software
Feature activation+0.000
makes
Token makes
Feature activation+0.000
tink
Token tink
Feature activation+2.147
ering
Tokenering
Feature activation+3.312
a
Token a
Feature activation+0.000
su
Token su
Feature activation+0.000
able
Tokenable
Feature activation+0.000
offense
Token offense
Feature activation+0.254
.
Token.
Feature activation+0.000
which
Token which
Feature activation+0.000
does
Token does
Feature activation+0.000
not
Token not
Feature activation+0.000
make
Token make
Feature activation+0.000
capital
Token capital
Feature activation+0.594
punishment
Token punishment
Feature activation+3.266
obligatory
Token obligatory
Feature activation+0.000
for
Token for
Feature activation+0.000
drug
Token drug
Feature activation+0.000
crimes
Token crimes
Feature activation+0.000
.
Token.
Feature activation+0.000
economic
Token economic
Feature activation+0.000
growth
Token growth
Feature activation+0.000
,
Token,
Feature activation+0.000
making
Token making
Feature activation+0.000
domestic
Token domestic
Feature activation+2.165
affairs
Token affairs
Feature activation+3.264
and
Token and
Feature activation+1.005
social
Token social
Feature activation+0.000
policies
Token policies
Feature activation+0.838
easier
Token easier
Feature activation+0.000
to
Token to
Feature activation+0.000
add
Token add
Feature activation+0.000
visual
Token visual
Feature activation+0.000
appeal
Token appeal
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
sharing
Token sharing
Feature activation+3.250
attractive
Token attractive
Feature activation+0.000
.
Token.
Feature activation+0.000
For
Token For
Feature activation+0.000
example
Token example
Feature activation+0.000
,
Token,
Feature activation+0.000
rent
Token rent
Feature activation+0.000
enough
Token enough
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
city
Token city
Feature activation+1.190
living
Token living
Feature activation+3.155
more
Token more
Feature activation+0.000
popular
Token popular
Feature activation+0.000
(
Token (
Feature activation+0.000
in
Tokenin
Feature activation+0.000
which
Token which
Feature activation+0.000
petition
Token petition
Feature activation+0.000
demanding
Token demanding
Feature activation+0.000
that
Token that
Feature activation+0.000
legislators
Token legislators
Feature activation+0.000
make
Token make
Feature activation+0.000
passing
Token passing
Feature activation+3.126
a
Token a
Feature activation+0.000
bill
Token bill
Feature activation+0.434
banning
Token banning
Feature activation+1.314
F
Token F
Feature activation+0.000
GM
TokenGM
Feature activation+0.000
from
Token from
Feature activation+0.000
different
Token different
Feature activation+0.000
angles
Token angles
Feature activation+0.000
,
Token,
Feature activation+0.000
making
Token making
Feature activation+0.000
holding
Token holding
Feature activation+3.015
the
Token the
Feature activation+0.000
sites
Token sites
Feature activation+0.000
a
Token a
Feature activation+0.000
challenge
Token challenge
Feature activation+0.000
.
Token.
Feature activation+0.000
s
Tokens
Feature activation+0.000
on
Token on
Feature activation+0.000
and
Token and
Feature activation+0.000
they
Token they
Feature activation+0.000
make
Token make
Feature activation+0.000
getting
Token getting
Feature activation+3.015
smashed
Token smashed
Feature activation+0.000
more
Token more
Feature activation+0.000
fun
Token fun
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Trump
TokenTrump
Feature activation+0.000
made
Token made
Feature activation+0.000
repeal
Token repeal
Feature activation+2.987
and
Token and
Feature activation+0.481
replace
Token replace
Feature activation+0.000
of
Token of
Feature activation+0.000
his
Token his
Feature activation+0.000
predecessor
Token predecessor
Feature activation+0.000
literature
Token literature
Feature activation+0.000
's
Token's
Feature activation+0.000
failure
Token failure
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
illness
Token illness
Feature activation+2.956
one
Token one
Feature activation+0.000
of
Token of
Feature activation+0.520
its
Token its
Feature activation+0.000
"
Token "
Feature activation+0.000
prime
Tokenprime
Feature activation+0.000
de
Token de
Feature activation+0.000
formed
Tokenformed
Feature activation+0.000
foot
Token foot
Feature activation+0.000
made
Token made
Feature activation+0.000
physical
Token physical
Feature activation+0.713
activity
Token activity
Feature activation+2.929
difficult
Token difficult
Feature activation+0.741
,
Token,
Feature activation+0.321
and
Token and
Feature activation+0.000
his
Token his
Feature activation+0.000
weight
Token weight
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
way
Token way
Feature activation+0.000
that
Token that
Feature activation+0.000
makes
Token makes
Feature activation+0.000
China
Token China
Feature activation+2.914
look
Token look
Feature activation+1.985
very
Token very
Feature activation+0.000
good
Token good
Feature activation+0.000
to
Token to
Feature activation+0.000
invent
Token invent
Feature activation+0.000
GOP
Token GOP
Feature activation+0.000
nominees
Token nominees
Feature activation+0.000
made
Token made
Feature activation+0.000
eyebrow
Token eyebrow
Feature activation+0.020
-
Token-
Feature activation+1.360
raising
Tokenraising
Feature activation+2.905
remarks
Token remarks
Feature activation+0.000
about
Token about
Feature activation+0.000
rape
Token rape
Feature activation+0.000
.
Token.
Feature activation+0.000
Missouri
Token Missouri
Feature activation+0.000
o
Tokeno
Feature activation+0.611
-
Token-
Feature activation+0.005
sy
Tokensy
Feature activation+1.684
nd
Tokennd
Feature activation+0.039
ical
Tokenical
Feature activation+1.179
ism
Tokenism
Feature activation+2.881
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
force
Token force
Feature activation+0.313
in
Token in
Feature activation+0.000
and
Token and
Feature activation+0.000
they
Token they
Feature activation+0.000
're
Token're
Feature activation+0.000
making
Token making
Feature activation+0.000
the
Token the
Feature activation+0.000
situation
Token situation
Feature activation+2.865
right
Token right
Feature activation+0.000
,
Token,
Feature activation+0.000
immediately
Token immediately
Feature activation+0.000
confirming
Token confirming
Feature activation+0.000
Nathan
Token Nathan
Feature activation+0.000
other
Token other
Feature activation+0.000
things
Token things
Feature activation+0.000
designed
Token designed
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
buying
Token buying
Feature activation+2.848
electronics
Token electronics
Feature activation+0.000
more
Token more
Feature activation+0.000
fun
Token fun
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
ways
Token ways
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
the
Token the
Feature activation+0.000
user
Token user
Feature activation+1.386
experience
Token experience
Feature activation+2.843
even
Token even
Feature activation+0.470
better
Token better
Feature activation+0.097
,
Token,
Feature activation+0.000
including
Token including
Feature activation+0.000
improving
Token improving
Feature activation+0.000

Top DFA by src position
MAX = 4.275

<|endoftext|>
Token<|endoftext|>
Feature activation-0.061
Top resid features:
only
Token only
Feature activation-0.074
Top resid features:
make
Token make
Feature activation+3.932
Top resid features:
employers
Token employers
Feature activation+0.105
Top resid features:
'
Token'
Feature activation+0.221
Top resid features:
jobs
Token jobs
Feature activation-0.026
Top resid features:
harder
Token harder
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
leap
Token leap
Feature activation-0.002
Top resid features:
to
Token to
Feature activation+0.010
Top resid features:
say
Token say
Feature activation-0.020
Top resid features:
it
Token it
Feature activation-0.073
Top resid features:
would
Token would
Feature activation-0.012
Top resid features:
make
Token make
Feature activation+4.275
Top resid features:
getting
Token getting
Feature activation-0.015
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
customer
Token customer
Feature activation+0.000
Top resid features:
data
Token data
Feature activation+0.000
Top resid features:
possible
Token possible
Feature activation+0.000
Top resid features:
broadly
Token broadly
Feature activation+0.000
Top resid features:
,
Token,
Feature activation-0.001
Top resid features:
cities
Token cities
Feature activation+0.001
Top resid features:
)
Token)
Feature activation-0.014
Top resid features:
can
Token can
Feature activation-0.040
Top resid features:
make
Token make
Feature activation+4.149
Top resid features:
riding
Token riding
Feature activation+0.043
Top resid features:
more
Token more
Feature activation+0.000
Top resid features:
appealing
Token appealing
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
women
Token women
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.004
Top resid features:
legal
Token legal
Feature activation-0.005
Top resid features:
status
Token status
Feature activation-0.001
Top resid features:
of
Token of
Feature activation+0.014
Top resid features:
software
Token software
Feature activation+0.019
Top resid features:
makes
Token makes
Feature activation+3.937
Top resid features:
tink
Token tink
Feature activation-0.024
Top resid features:
ering
Tokenering
Feature activation+0.094
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
su
Token su
Feature activation+0.000
Top resid features:
able
Tokenable
Feature activation+0.000
Top resid features:
law
Token law
Feature activation+0.013
Top resid features:
,
Token,
Feature activation+0.032
Top resid features:
which
Token which
Feature activation+0.010
Top resid features:
does
Token does
Feature activation-0.033
Top resid features:
not
Token not
Feature activation-0.032
Top resid features:
make
Token make
Feature activation+3.681
Top resid features:
capital
Token capital
Feature activation+0.192
Top resid features:
punishment
Token punishment
Feature activation+0.116
Top resid features:
obligatory
Token obligatory
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
drug
Token drug
Feature activation+0.000
Top resid features:
fortune
Token fortune
Feature activation-0.006
Top resid features:
of
Token of
Feature activation+0.008
Top resid features:
economic
Token economic
Feature activation-0.013
Top resid features:
growth
Token growth
Feature activation+0.004
Top resid features:
,
Token,
Feature activation+0.058
Top resid features:
making
Token making
Feature activation+3.905
Top resid features:
domestic
Token domestic
Feature activation+0.007
Top resid features:
affairs
Token affairs
Feature activation+0.063
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
social
Token social
Feature activation+0.000
Top resid features:
policies
Token policies
Feature activation+0.000
Top resid features:
we
Token we
Feature activation+0.007
Top resid features:
add
Token add
Feature activation-0.001
Top resid features:
visual
Token visual
Feature activation+0.062
Top resid features:
appeal
Token appeal
Feature activation+0.033
Top resid features:
to
Token to
Feature activation+0.012
Top resid features:
make
Token make
Feature activation+3.819
Top resid features:
sharing
Token sharing
Feature activation+0.030
Top resid features:
attractive
Token attractive
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
For
Token For
Feature activation+0.000
Top resid features:
example
Token example
Feature activation+0.000
Top resid features:
buildings
Token buildings
Feature activation-0.017
Top resid features:
'
Token'
Feature activation+0.025
Top resid features:
rent
Token rent
Feature activation-0.017
Top resid features:
enough
Token enough
Feature activation+0.039
Top resid features:
to
Token to
Feature activation+0.052
Top resid features:
make
Token make
Feature activation+3.699
Top resid features:
city
Token city
Feature activation+0.242
Top resid features:
living
Token living
Feature activation-0.033
Top resid features:
more
Token more
Feature activation+0.000
Top resid features:
popular
Token popular
Feature activation+0.000
Top resid features:
(
Token (
Feature activation+0.000
Top resid features:
our
Token our
Feature activation-0.007
Top resid features:
petition
Token petition
Feature activation-0.018
Top resid features:
demanding
Token demanding
Feature activation+0.010
Top resid features:
that
Token that
Feature activation+0.044
Top resid features:
legislators
Token legislators
Feature activation-0.049
Top resid features:
make
Token make
Feature activation+3.879
Top resid features:
passing
Token passing
Feature activation-0.004
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
bill
Token bill
Feature activation+0.000
Top resid features:
banning
Token banning
Feature activation+0.000
Top resid features:
F
Token F
Feature activation+0.000
Top resid features:
points
Token points
Feature activation+0.003
Top resid features:
from
Token from
Feature activation-0.003
Top resid features:
different
Token different
Feature activation-0.002
Top resid features:
angles
Token angles
Feature activation+0.006
Top resid features:
,
Token,
Feature activation-0.001
Top resid features:
making
Token making
Feature activation+3.872
Top resid features:
holding
Token holding
Feature activation-0.058
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
sites
Token sites
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
challenge
Token challenge
Feature activation+0.000
Top resid features:
off
Token off
Feature activation-0.036
Top resid features:
s
Tokens
Feature activation-0.001
Top resid features:
on
Token on
Feature activation-0.028
Top resid features:
and
Token and
Feature activation+0.028
Top resid features:
they
Token they
Feature activation-0.050
Top resid features:
make
Token make
Feature activation+3.949
Top resid features:
getting
Token getting
Feature activation-0.051
Top resid features:
smashed
Token smashed
Feature activation+0.000
Top resid features:
more
Token more
Feature activation+0.000
Top resid features:
fun
Token fun
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Act
Token Act
Feature activation-0.003
Top resid features:
.
Token.
Feature activation+0.015
Top resid features:
Ċ
TokenĊ
Feature activation+0.013
Top resid features:
Ċ
TokenĊ
Feature activation+0.010
Top resid features:
Trump
TokenTrump
Feature activation+0.023
Top resid features:
made
Token made
Feature activation+3.500
Top resid features:
repeal
Token repeal
Feature activation+0.169
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
replace
Token replace
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
his
Token his
Feature activation+0.000
Top resid features:
ans
Tokenans
Feature activation+0.005
Top resid features:
literature
Token literature
Feature activation+0.013
Top resid features:
's
Token's
Feature activation-0.006
Top resid features:
failure
Token failure
Feature activation+0.017
Top resid features:
to
Token to
Feature activation+0.077
Top resid features:
make
Token make
Feature activation+3.409
Top resid features:
illness
Token illness
Feature activation+0.095
Top resid features:
one
Token one
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
its
Token its
Feature activation+0.000
Top resid features:
"
Token "
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.003
Top resid features:
his
Token his
Feature activation-0.013
Top resid features:
de
Token de
Feature activation-0.007
Top resid features:
formed
Tokenformed
Feature activation+0.014
Top resid features:
foot
Token foot
Feature activation-0.009
Top resid features:
made
Token made
Feature activation+3.495
Top resid features:
physical
Token physical
Feature activation+0.104
Top resid features:
activity
Token activity
Feature activation+0.068
Top resid features:
difficult
Token difficult
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:
system
Token system
Feature activation-0.001
Top resid features:
in
Token in
Feature activation+0.004
Top resid features:
a
Token a
Feature activation+0.016
Top resid features:
way
Token way
Feature activation+0.017
Top resid features:
that
Token that
Feature activation+0.019
Top resid features:
makes
Token makes
Feature activation+3.460
Top resid features:
China
Token China
Feature activation+0.110
Top resid features:
look
Token look
Feature activation+0.000
Top resid features:
very
Token very
Feature activation+0.000
Top resid features:
good
Token good
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
rocked
Token rocked
Feature activation+0.009
Top resid features:
when
Token when
Feature activation+0.002
Top resid features:
the
Token the
Feature activation-0.001
Top resid features:
GOP
Token GOP
Feature activation+0.001
Top resid features:
nominees
Token nominees
Feature activation+0.034
Top resid features:
made
Token made
Feature activation+3.433
Top resid features:
eyebrow
Token eyebrow
Feature activation+0.135
Top resid features:
-
Token-
Feature activation+0.034
Top resid features:
raising
Tokenraising
Feature activation-0.027
Top resid features:
remarks
Token remarks
Feature activation+0.000
Top resid features:
about
Token about
Feature activation+0.000
Top resid features:
this
Token this
Feature activation+0.043
Top resid features:
unique
Token unique
Feature activation-0.015
Top resid features:
workers
Token workers
Feature activation-0.001
Top resid features:
movement
Token movement
Feature activation+0.009
Top resid features:
which
Token which
Feature activation+0.015
Top resid features:
made
Token made
Feature activation+3.336
Top resid features:
anarch
Token anarch
Feature activation+0.058
Top resid features:
o
Tokeno
Feature activation+0.006
Top resid features:
-
Token-
Feature activation+0.034
Top resid features:
sy
Tokensy
Feature activation+0.025
Top resid features:
nd
Tokennd
Feature activation-0.006
Top resid features:
here
Token here
Feature activation-0.011
Top resid features:
,
Token,
Feature activation+0.027
Top resid features:
and
Token and
Feature activation+0.055
Top resid features:
they
Token they
Feature activation-0.000
Top resid features:
're
Token're
Feature activation-0.023
Top resid features:
making
Token making
Feature activation+3.555
Top resid features:
the
Token the
Feature activation+0.018
Top resid features:
situation
Token situation
Feature activation-0.006
Top resid features:
right
Token right
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
immediately
Token immediately
Feature activation+0.000
Top resid features:
and
Token and
Feature activation-0.002
Top resid features:
other
Token other
Feature activation+0.002
Top resid features:
things
Token things
Feature activation+0.024
Top resid features:
designed
Token designed
Feature activation-0.022
Top resid features:
to
Token to
Feature activation+0.053
Top resid features:
make
Token make
Feature activation+3.632
Top resid features:
buying
Token buying
Feature activation-0.064
Top resid features:
electronics
Token electronics
Feature activation+0.000
Top resid features:
more
Token more
Feature activation+0.000
Top resid features:
fun
Token fun
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
is
Token is
Feature activation-0.000
Top resid features:
working
Token working
Feature activation-0.004
Top resid features:
on
Token on
Feature activation+0.002
Top resid features:
ways
Token ways
Feature activation+0.050
Top resid features:
to
Token to
Feature activation+0.024
Top resid features:
make
Token make
Feature activation+3.508
Top resid features:
the
Token the
Feature activation+0.063
Top resid features:
user
Token user
Feature activation-0.005
Top resid features:
experience
Token experience
Feature activation-0.033
Top resid features:
even
Token even
Feature activation+0.000
Top resid features:
better
Token better
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.04

Head 2: 0.08

Head 3: 0.04

Head 4: 0.02

Head 5: 0.03

Head 6: 0.07

Head 7: 0.06

Head 8: 0.42

Head 9: 0.13

Head 10: 0.05

Head 11: 0.05

Positive logits

NPR1.16

faire1.15

pload1.14

isen1.14

olation1.12

Pry1.12

unb1.11

disappear1.10

essage1.07

Blizz1.05

noises1.04

shorth1.04

appear1.02

ibel1.02

undle1.01

recomm1.01

paste1.00

happen1.00

feature1.00

xf1.00

Negative logits

illard-1.28

ascript-1.23

assetsadobe-1.23

ATES-1.21

xual-1.18

ADE-1.17

Dealer-1.16

atan-1.14

ARI-1.13

illery-1.12

>:-1.11

arius-1.11

soType-1.10

quickShipAvailable-1.09

srfAttach-1.08

wayne-1.08

WARD-1.07

ares-1.06

ategory-1.06

iceps-1.05

INTERVAL 3.090 - 3.433
CONTAINS 0.001%

which
Token which
Feature activation+0.000
does
Token does
Feature activation+0.000
not
Token not
Feature activation+0.000
make
Token make
Feature activation+0.000
capital
Token capital
Feature activation+0.594
punishment
Token punishment
Feature activation+3.266
obligatory
Token obligatory
Feature activation+0.000
for
Token for
Feature activation+0.000
drug
Token drug
Feature activation+0.000
crimes
Token crimes
Feature activation+0.000
.
Token.
Feature activation+0.000
petition
Token petition
Feature activation+0.000
demanding
Token demanding
Feature activation+0.000
that
Token that
Feature activation+0.000
legislators
Token legislators
Feature activation+0.000
make
Token make
Feature activation+0.000
passing
Token passing
Feature activation+3.126
a
Token a
Feature activation+0.000
bill
Token bill
Feature activation+0.434
banning
Token banning
Feature activation+1.314
F
Token F
Feature activation+0.000
GM
TokenGM
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+0.000
make
Token make
Feature activation+0.000
getting
Token getting
Feature activation+3.432
at
Token at
Feature activation+0.689
customer
Token customer
Feature activation+0.000
data
Token data
Feature activation+0.274
possible
Token possible
Feature activation+0.000
."
Token."
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
only
Token only
Feature activation+0.000
make
Token make
Feature activation+0.000
employers
Token employers
Feature activation+2.181
'
Token'
Feature activation+1.860
jobs
Token jobs
Feature activation+3.433
harder
Token harder
Feature activation+0.162
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token"
Feature activation+0.000
rent
Token rent
Feature activation+0.000
enough
Token enough
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
city
Token city
Feature activation+1.190
living
Token living
Feature activation+3.155
more
Token more
Feature activation+0.000
popular
Token popular
Feature activation+0.000
(
Token (
Feature activation+0.000
in
Tokenin
Feature activation+0.000
which
Token which
Feature activation+0.000

INTERVAL 2.747 - 3.090
CONTAINS 0.002%

other
Token other
Feature activation+0.000
things
Token things
Feature activation+0.000
designed
Token designed
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
buying
Token buying
Feature activation+2.848
electronics
Token electronics
Feature activation+0.000
more
Token more
Feature activation+0.000
fun
Token fun
Feature activation+0.000
.
Token.
Feature activation+0.000
There
Token There
Feature activation+0.000
,
Token,
Feature activation+0.000
designed
Token designed
Feature activation+0.000
mostly
Token mostly
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
interaction
Token interaction
Feature activation+2.776
easier
Token easier
Feature activation+0.394
.
Token.
Feature activation+0.000
These
Token These
Feature activation+0.000
included
Token included
Feature activation+0.000
a
Token a
Feature activation+0.000
o
Tokeno
Feature activation+0.611
-
Token-
Feature activation+0.005
sy
Tokensy
Feature activation+1.684
nd
Tokennd
Feature activation+0.039
ical
Tokenical
Feature activation+1.179
ism
Tokenism
Feature activation+2.881
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
force
Token force
Feature activation+0.313
in
Token in
Feature activation+0.000
years
Token years
Feature activation+0.000
have
Token have
Feature activation+0.000
made
Token made
Feature activation+0.000
this
Token this
Feature activation+0.521
remarkable
Token remarkable
Feature activation+0.000
accomplishment
Token accomplishment
Feature activation+2.765
possible
Token possible
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
A
TokenA
Feature activation+0.000
GOP
Token GOP
Feature activation+0.000
nominees
Token nominees
Feature activation+0.000
made
Token made
Feature activation+0.000
eyebrow
Token eyebrow
Feature activation+0.020
-
Token-
Feature activation+1.360
raising
Tokenraising
Feature activation+2.905
remarks
Token remarks
Feature activation+0.000
about
Token about
Feature activation+0.000
rape
Token rape
Feature activation+0.000
.
Token.
Feature activation+0.000
Missouri
Token Missouri
Feature activation+0.000

INTERVAL 2.403 - 2.747
CONTAINS 0.002%

are
Token are
Feature activation+0.000
going
Token going
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
our
Token our
Feature activation+0.000
country
Token country
Feature activation+2.556
great
Token great
Feature activation+0.209
again
Token again
Feature activation+1.586
,"
Token,"
Feature activation+0.000
Trump
Token Trump
Feature activation+0.000
told
Token told
Feature activation+0.000
electr
Token electr
Feature activation+0.000
ifying
Tokenifying
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
makes
Token makes
Feature activation+0.000
playing
Token playing
Feature activation+2.408
every
Token every
Feature activation+0.000
class
Token class
Feature activation+0.000
an
Token an
Feature activation+0.000
absolute
Token absolute
Feature activation+0.000
blast
Token blast
Feature activation+0.000
on
Token on
Feature activation+0.000
what
Token what
Feature activation+0.000
would
Token would
Feature activation+0.000
make
Token make
Feature activation+0.000
good
Token good
Feature activation+0.000
research
Token research
Feature activation+2.423
assignments
Token assignments
Feature activation+1.528
for
Token for
Feature activation+0.000
his
Token his
Feature activation+0.000
graduate
Token graduate
Feature activation+0.000
students
Token students
Feature activation+0.000
tends
Token tends
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
people
Token people
Feature activation+0.966
slow
Token slow
Feature activation+1.479
down
Token down
Feature activation+2.737
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
And
TokenAnd
Feature activation+0.000
March
Token March
Feature activation+0.000
global
Token global
Feature activation+0.000
threats
Token threats
Feature activation+0.000
make
Token make
Feature activation+0.000
cases
Token cases
Feature activation+0.885
such
Token such
Feature activation+0.961
as
Token as
Feature activation+2.551
Moh
Token Moh
Feature activation+0.000
am
Tokenam
Feature activation+0.000
ud
Tokenud
Feature activation+0.000
's
Token's
Feature activation+0.000
more
Token more
Feature activation+0.000

INTERVAL 2.060 - 2.403
CONTAINS 0.004%

IONS
TokenIONS
Feature activation+0.000
:
Token:
Feature activation+0.000
--
Token --
Feature activation+0.000
inc
Tokeninc
Feature activation+0.000
make
Token make
Feature activation+0.000
reverse
Token reverse
Feature activation+2.173
incremental
Token incremental
Feature activation+0.000
backup
Token backup
Feature activation+1.040
--
Token --
Feature activation+0.000
dry
Tokendry
Feature activation+0.000
run
Token run
Feature activation+0.000
team
Token team
Feature activation+0.000
let
Token let
Feature activation+0.000
those
Token those
Feature activation+0.000
babies
Token babies
Feature activation+0.000
make
Token make
Feature activation+0.000
babies
Token babies
Feature activation+2.105
,
Token,
Feature activation+0.362
then
Token then
Feature activation+0.000
let
Token let
Feature activation+0.000
those
Token those
Feature activation+0.000
babies
Token babies
Feature activation+0.010
an
Token an
Feature activation+0.000
air
Token air
Feature activation+0.000
filter
Token filter
Feature activation+0.000
,
Token,
Feature activation+0.000
making
Token making
Feature activation+0.000
job
Token job
Feature activation+2.271
more
Token more
Feature activation+0.102
stressful
Token stressful
Feature activation+1.652
,
Token,
Feature activation+0.646
and
Token and
Feature activation+0.000
triggering
Token triggering
Feature activation+0.000
you
Token you
Feature activation+0.000
make
Token make
Feature activation+0.000
all
Token all
Feature activation+0.206
of
Token of
Feature activation+0.962
this
Token this
Feature activation+0.671
content
Token content
Feature activation+2.162
play
Token play
Feature activation+0.261
seamlessly
Token seamlessly
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
living
Token living
Feature activation+0.000
An
Token An
Feature activation+0.000
ac
Tokenac
Feature activation+0.000
onda
Tokenonda
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
Cobra
Token Cobra
Feature activation+2.264
Re
Token Re
Feature activation+0.938
ale
Tokenale
Feature activation+0.504
(
Token (
Feature activation+0.306
8
Token8
Feature activation+0.000
c
Tokenc
Feature activation+0.000

INTERVAL 1.717 - 2.060
CONTAINS 0.006%

made
Token made
Feature activation+0.000
Mu
Token Mu
Feature activation+2.086
q
Tokenq
Feature activation+0.210
rin
Tokenrin
Feature activation+1.148
crown
Token crown
Feature activation+1.198
prince
Token prince
Feature activation+1.916
,
Token,
Feature activation+0.464
as
Token as
Feature activation+0.000
was
Token was
Feature activation+0.000
expected
Token expected
Feature activation+0.000
,
Token,
Feature activation+0.000
the
Token the
Feature activation+0.000
popup
Token popup
Feature activation+0.000
and
Token and
Feature activation+0.000
make
Token make
Feature activation+0.000
it
Token it
Feature activation+0.000
live
Token live
Feature activation+1.843
for
Token for
Feature activation+0.000
3
Token 3
Feature activation+0.000
seconds
Token seconds
Feature activation+0.000
(
Token (
Feature activation+0.000
echo
Tokenecho
Feature activation+0.000
you
Token you
Feature activation+0.000
can
Token can
Feature activation+0.000
make
Token make
Feature activation+0.000
a
Token a
Feature activation+0.000
city
Token city
Feature activation+1.200
centre
Token centre
Feature activation+1.748
an
Token an
Feature activation+0.000
interesting
Token interesting
Feature activation+0.000
place
Token place
Feature activation+0.647
to
Token to
Feature activation+0.000
live
Token live
Feature activation+0.000
that
Token that
Feature activation+0.000
has
Token has
Feature activation+0.000
made
Token made
Feature activation+0.000
a
Token a
Feature activation+0.334
mockery
Token mockery
Feature activation+0.000
of
Token of
Feature activation+1.902
every
Token every
Feature activation+0.000
pret
Token pret
Feature activation+0.000
ense
Tokenense
Feature activation+0.000
of
Token of
Feature activation+0.000
international
Token international
Feature activation+0.000
the
Token the
Feature activation+0.000
system
Token system
Feature activation+0.000
,
Token,
Feature activation+0.000
making
Token making
Feature activation+0.000
the
Token the
Feature activation+0.640
technology
Token technology
Feature activation+1.870
user
Token user
Feature activation+1.687
friendly
Token friendly
Feature activation+0.172
,
Token,
Feature activation+0.617
overcoming
Token overcoming
Feature activation+0.000
maintenance
Token maintenance
Feature activation+0.000

INTERVAL 1.373 - 1.717
CONTAINS 0.011%

Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
usage
Token usage
Feature activation+0.000
pattern
Token pattern
Feature activation+0.000
made
Token made
Feature activation+0.000
this
Token this
Feature activation+1.386
data
Token data
Feature activation+1.704
transfer
Token transfer
Feature activation+0.402
a
Token a
Feature activation+0.000
perfect
Token perfect
Feature activation+0.000
client
Token client
Feature activation+0.000
an
Token an
Feature activation+0.000
ent
Token ent
Feature activation+0.131
ic
Tokenic
Feature activation+0.000
ingly
Tokeningly
Feature activation+0.000
different
Token different
Feature activation+0.233
experience
Token experience
Feature activation+1.567
.
Token.
Feature activation+0.000
I
Token I
Feature activation+0.000
will
Token will
Feature activation+0.000
be
Token be
Feature activation+0.000
playing
Token playing
Feature activation+0.000
feel
Token feel
Feature activation+0.000
better
Token better
Feature activation+0.000
by
Token by
Feature activation+0.000
making
Token making
Feature activation+0.000
other
Token other
Feature activation+0.268
people
Token people
Feature activation+1.401
feel
Token feel
Feature activation+0.219
worse
Token worse
Feature activation+0.000
totally
Token totally
Feature activation+0.000
tracks
Token tracks
Feature activation+0.000
for
Token for
Feature activation+0.000
well
Token well
Feature activation+0.000
and
Token and
Feature activation+0.000
can
Token can
Feature activation+0.000
make
Token make
Feature activation+0.000
morning
Token morning
Feature activation+0.289
hours
Token hours
Feature activation+1.473
food
Token food
Feature activation+0.000
more
Token more
Feature activation+0.000
filling
Token filling
Feature activation+0.367
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
which
Token which
Feature activation+0.000
made
Token made
Feature activation+0.000
anarch
Token anarch
Feature activation+2.591
o
Tokeno
Feature activation+0.611
-
Token-
Feature activation+0.005
sy
Tokensy
Feature activation+1.684
nd
Tokennd
Feature activation+0.039
ical
Tokenical
Feature activation+1.179
ism
Tokenism
Feature activation+2.881
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000

INTERVAL 1.030 - 1.373
CONTAINS 0.015%

acht
Tokenacht
Feature activation+0.000
also
Token also
Feature activation+0.000
regularly
Token regularly
Feature activation+0.000
made
Token made
Feature activation+0.000
use
Token use
Feature activation+0.718
of
Token of
Feature activation+1.219
captured
Token captured
Feature activation+0.000
Red
Token Red
Feature activation+0.000
Army
Token Army
Feature activation+0.000
winter
Token winter
Feature activation+0.000
uniforms
Token uniforms
Feature activation+0.000
port
Token port
Feature activation+0.000
ing
Tokening
Feature activation+0.000
an
Token an
Feature activation+0.000
app
Token app
Feature activation+0.000
made
Token made
Feature activation+0.000
for
Token for
Feature activation+1.042
a
Token a
Feature activation+0.000
5
Token 5
Feature activation+0.000
-
Token-
Feature activation+0.000
inch
Tokeninch
Feature activation+0.000
touchscreen
Token touchscreen
Feature activation+0.000
>
Token >
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
By
TokenBy
Feature activation+0.000
making
Token making
Feature activation+0.000
users
Token users
Feature activation+1.228
add
Token add
Feature activation+0.176
this
Token this
Feature activation+0.000
line
Token line
Feature activation+0.000
to
Token to
Feature activation+0.000
their
Token their
Feature activation+0.000
a
Token a
Feature activation+0.000
quite
Token quite
Feature activation+0.000
public
Token public
Feature activation+0.678
ult
Token ult
Feature activation+0.238
im
Tokenim
Feature activation+0.000
atum
Tokenatum
Feature activation+1.245
regarding
Token regarding
Feature activation+0.000
the
Token the
Feature activation+0.000
timeline
Token timeline
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
food
Tokenfood
Feature activation+0.000
chains
Token chains
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
any
Token any
Feature activation+0.120
meal
Token meal
Feature activation+1.342
with
Token with
Feature activation+0.229
a
Token a
Feature activation+0.000
toy
Token toy
Feature activation+0.000
more
Token more
Feature activation+0.000
nutritious
Token nutritious
Feature activation+0.000

INTERVAL 0.687 - 1.030
CONTAINS 0.031%

drawback
Token drawback
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
make
Token make
Feature activation+0.000
_
Token_
Feature activation+0.000
shared
Tokenshared
Feature activation+0.799
()
Token()
Feature activation+0.000
implementation
Token implementation
Feature activation+0.137
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
increase
Token increase
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
d
Tokend
Feature activation+0.000
make
Token make
Feature activation+0.000
all
Token all
Feature activation+0.003
commenters
Token commenters
Feature activation+0.830
log
Token log
Feature activation+0.835
in
Token in
Feature activation+0.000
with
Token with
Feature activation+0.000
Facebook
Token Facebook
Feature activation+0.000
or
Token or
Feature activation+0.000
some
Token some
Feature activation+0.000
made
Token made
Feature activation+0.000
231
Token 231
Feature activation+0.000
-
Token-
Feature activation+0.000
pound
Tokenpound
Feature activation+0.000
running
Token running
Feature activation+1.250
back
Token back
Feature activation+0.824
William
Token William
Feature activation+0.000
Stan
Token Stan
Feature activation+0.000
back
Tokenback
Feature activation+0.000
whiff
Token whiff
Feature activation+0.000
.
Token.
Feature activation+0.000
they
Token they
Feature activation+0.000
could
Token could
Feature activation+0.000
make
Token make
Feature activation+0.000
better
Token better
Feature activation+0.000
use
Token use
Feature activation+0.128
of
Token of
Feature activation+0.715
insurance
Token insurance
Feature activation+0.000
against
Token against
Feature activation+0.000
production
Token production
Feature activation+0.000
failures
Token failures
Feature activation+0.000
or
Token or
Feature activation+0.000
-
Token -
Feature activation+0.000
one
Token one
Feature activation+0.000
that
Token that
Feature activation+0.000
makes
Token makes
Feature activation+0.000
use
Token use
Feature activation+0.829
of
Token of
Feature activation+0.819
contin
Token contin
Feature activation+0.000
uations
Tokenuations
Feature activation+0.000
and
Token and
Feature activation+0.000
one
Token one
Feature activation+0.000
that
Token that
Feature activation+0.000

INTERVAL 0.343 - 0.687
CONTAINS 0.045%

was
Token was
Feature activation+0.000
making
Token making
Feature activation+0.000
his
Token his
Feature activation+0.000
own
Token own
Feature activation+0.000
legal
Token legal
Feature activation+0.053
counsel
Token counsel
Feature activation+0.507
,
Token,
Feature activation+0.234
Brian
Token Brian
Feature activation+0.000
H
Token H
Feature activation+0.000
aged
Tokenaged
Feature activation+0.000
orn
Tokenorn
Feature activation+0.000
not
Token not
Feature activation+0.000
directly
Token directly
Feature activation+0.000
made
Token made
Feature activation+0.000
the
Token the
Feature activation+0.000
tele
Token tele
Feature activation+0.051
market
Tokenmarket
Feature activation+0.547
ing
Tokening
Feature activation+1.453
calls
Token calls
Feature activation+0.000
in
Token in
Feature activation+0.000
question
Token question
Feature activation+0.000
:
Token:
Feature activation+0.000
's
Token's
Feature activation+0.000
Coco
Token Coco
Feature activation+0.000
continued
Token continued
Feature activation+0.000
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
magic
Token magic
Feature activation+0.416
in
Token in
Feature activation+0.000
its
Token its
Feature activation+0.000
second
Token second
Feature activation+0.000
weekend
Token weekend
Feature activation+0.000
,
Token,
Feature activation+0.000
night
Token night
Feature activation+0.000
last
Token last
Feature activation+0.000
week
Token week
Feature activation+0.000
making
Token making
Feature activation+0.000
fake
Token fake
Feature activation+0.379
milk
Token milk
Feature activation+0.353
cart
Token cart
Feature activation+0.000
ons
Tokenons
Feature activation+0.650
with
Token with
Feature activation+0.000
pictures
Token pictures
Feature activation+0.000
of
Token of
Feature activation+0.000
-
Token-
Feature activation+0.000
sale
Tokensale
Feature activation+0.000
,
Token,
Feature activation+0.000
has
Token has
Feature activation+0.000
made
Token made
Feature activation+0.000
a
Token a
Feature activation+0.523
quiet
Token quiet
Feature activation+1.523
debut
Token debut
Feature activation+0.056
on
Token on
Feature activation+0.000
the
Token the
Feature activation+0.000
Amazon
Token Amazon
Feature activation+0.000

INTERVAL 0.000 - 0.343
CONTAINS 99.885%

smack
Token smack
Feature activation+0.000
between
Token between
Feature activation+0.000
Cincinnati
Token Cincinnati
Feature activation+0.000
and
Token and
Feature activation+0.000
Dayton
Token Dayton
Feature activation+0.000
on
Token on
Feature activation+0.000
Interstate
Token Interstate
Feature activation+0.000
75
Token 75
Feature activation+0.000
,
Token,
Feature activation+0.000
where
Token where
Feature activation+0.000
his
Token his
Feature activation+0.000
are
Token are
Feature activation+0.000
Hispanic
Token Hispanic
Feature activation+0.000
,
Token,
Feature activation+0.000
including
Token including
Feature activation+0.000
management
Token management
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
Arizona
Token Arizona
Feature activation+0.000
Chamber
Token Chamber
Feature activation+0.000
school
Token school
Feature activation+0.000
board
Token board
Feature activation+0.000
is
Token is
Feature activation+0.000
left
Token left
Feature activation+0.000
guessing
Token guessing
Feature activation+0.000
with
Token with
Feature activation+0.000
what
Token what
Feature activation+0.000
may
Token may
Feature activation+0.000
happen
Token happen
Feature activation+0.000
and
Token and
Feature activation+0.000
how
Token how
Feature activation+0.000
best
Token best
Feature activation+0.000
âĢĵ
Token âĢĵ
Feature activation+0.000
pl
Token pl
Feature activation+0.000
am
Tokenam
Feature activation+0.000
asing
Tokenasing
Feature activation+0.000
everyone
Token everyone
Feature activation+0.000
in
Token in
Feature activation+0.000
sight
Token sight
Feature activation+0.000
,
Token,
Feature activation+0.000
looking
Token looking
Feature activation+0.000
like
Token like
Feature activation+0.000
.
Token.
Feature activation+0.000
Photo
Token Photo
Feature activation+0.000
courtesy
Token courtesy
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Damascus
Token Damascus
Feature activation+0.000
Police
Token Police
Feature activation+0.000
Force
Token Force
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
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 H3.8: (feature 19239)

TOP ACTIVATIONS
MAX = 2.626

lieutenant
Token lieutenant
Feature activation+0.000
who
Token who
Feature activation+0.000
survived
Token survived
Feature activation+0.000
a
Token a
Feature activation+1.376
mass
Token mass
Feature activation+1.533
shooting
Token shooting
Feature activation+2.626
at
Token at
Feature activation+0.771
a
Token a
Feature activation+0.000
Sikh
Token Sikh
Feature activation+0.000
temple
Token temple
Feature activation+0.000
in
Token in
Feature activation+0.000
or
Token or
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
escape
Token escape
Feature activation+0.000
certain
Token certain
Feature activation+1.654
destruction
Token destruction
Feature activation+2.338
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
stunts
Token stunts
Feature activation+0.000
hip
Tokenhip
Feature activation+0.000
used
Token used
Feature activation+0.000
illegal
Token illegal
Feature activation+0.000
aliens
Token aliens
Feature activation+0.000
who
Token who
Feature activation+0.000
are
Token are
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
enforcement
Token enforcement
Feature activation+2.215
of
Token of
Feature activation+0.312
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
At
Token At
Feature activation+0.000
Spurs
Token Spurs
Feature activation+0.000
he
Token he
Feature activation+0.000
survived
Token survived
Feature activation+0.000
being
Token being
Feature activation+2.185
quite
Token quite
Feature activation+0.204
clearly
Token clearly
Feature activation+0.717
targeted
Token targeted
Feature activation+0.000
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
.
Token.
Feature activation+0.000
Batman
Token Batman
Feature activation+0.000
narrowly
Token narrowly
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
the
Token the
Feature activation+1.038
destruction
Token destruction
Feature activation+1.988
ray
Token ray
Feature activation+0.073
by
Token by
Feature activation+0.039
j
Token j
Feature activation+0.000
ett
Tokenett
Feature activation+0.000
ison
Tokenison
Feature activation+0.000
survived
Token survived
Feature activation+0.000
the
Token the
Feature activation+1.488
no
Token no
Feature activation+0.508
-
Token-
Feature activation+0.000
confidence
Tokenconfidence
Feature activation+0.344
motion
Token motion
Feature activation+1.879
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
lower
Token lower
Feature activation+0.000
house
Token house
Feature activation+0.000
by
Token by
Feature activation+0.000
mud
Token mud
Feature activation+0.000
crab
Token crab
Feature activation+0.000
that
Token that
Feature activation+0.000
had
Token had
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
its
Token its
Feature activation+1.804
box
Token box
Feature activation+1.343
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.000
wandering
Token wandering
Feature activation+0.000
around
Token around
Feature activation+0.000
,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
escape
Token escape
Feature activation+0.000
certain
Token certain
Feature activation+1.654
destruction
Token destruction
Feature activation+2.338
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
stunts
Token stunts
Feature activation+0.000
hip
Tokenhip
Feature activation+0.000
police
Token police
Feature activation+0.000
lieutenant
Token lieutenant
Feature activation+0.000
who
Token who
Feature activation+0.000
survived
Token survived
Feature activation+0.000
a
Token a
Feature activation+1.376
mass
Token mass
Feature activation+1.533
shooting
Token shooting
Feature activation+2.626
at
Token at
Feature activation+0.771
a
Token a
Feature activation+0.000
Sikh
Token Sikh
Feature activation+0.000
temple
Token temple
Feature activation+0.000
believing
Token believing
Feature activation+0.000
that
Token that
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
misfortune
Token misfortune
Feature activation+1.499
,
Token,
Feature activation+0.000
he
Token he
Feature activation+0.000
found
Token found
Feature activation+0.000
himself
Token himself
Feature activation+0.000
in
Token in
Feature activation+0.000
,
Token,
Feature activation+0.000
Ber
Token Ber
Feature activation+0.000
lus
Tokenlus
Feature activation+0.000
coni
Tokenconi
Feature activation+0.000
survived
Token survived
Feature activation+0.000
the
Token the
Feature activation+1.488
no
Token no
Feature activation+0.508
-
Token-
Feature activation+0.000
confidence
Tokenconfidence
Feature activation+0.344
motion
Token motion
Feature activation+1.879
in
Token in
Feature activation+0.000
surface
Token surface
Feature activation+0.000
warships
Token warships
Feature activation+0.000
just
Token just
Feature activation+0.000
cannot
Token cannot
Feature activation+0.000
survive
Token survive
Feature activation+0.000
attacks
Token attacks
Feature activation+1.472
by
Token by
Feature activation+0.000
large
Token large
Feature activation+0.000
numbers
Token numbers
Feature activation+0.000
of
Token of
Feature activation+0.000
aircraft
Token aircraft
Feature activation+0.000
f
Token f
Feature activation+0.000
ates
Tokenates
Feature activation+0.000
of
Token of
Feature activation+0.000
families
Token families
Feature activation+0.000
fleeing
Token fleeing
Feature activation+0.000
persecution
Token persecution
Feature activation+1.453
from
Token from
Feature activation+0.000
its
Token its
Feature activation+0.000
northern
Token northern
Feature activation+0.000
neighbour
Token neighbour
Feature activation+0.000
but
Token but
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ver
TokenVer
Feature activation+0.000
lander
Tokenlander
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
third
Token third
Feature activation+1.403
-
Token-
Feature activation+0.990
inning
Tokeninning
Feature activation+0.673
jam
Token jam
Feature activation+0.851
:
Token:
Feature activation+0.000
Ver
Token Ver
Feature activation+0.000
a
Token a
Feature activation+0.000
police
Token police
Feature activation+0.000
lieutenant
Token lieutenant
Feature activation+0.000
who
Token who
Feature activation+0.000
survived
Token survived
Feature activation+0.000
a
Token a
Feature activation+1.376
mass
Token mass
Feature activation+1.533
shooting
Token shooting
Feature activation+2.626
at
Token at
Feature activation+0.771
a
Token a
Feature activation+0.000
Sikh
Token Sikh
Feature activation+0.000
David
TokenDavid
Feature activation+0.000
Davis
Token Davis
Feature activation+0.000
has
Token has
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
the
Token the
Feature activation+0.811
possibility
Token possibility
Feature activation+1.363
of
Token of
Feature activation+0.125
being
Token being
Feature activation+0.000
held
Token held
Feature activation+0.000
in
Token in
Feature activation+0.000
contempt
Token contempt
Feature activation+0.000
crab
Token crab
Feature activation+0.000
that
Token that
Feature activation+0.000
had
Token had
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
its
Token its
Feature activation+1.804
box
Token box
Feature activation+1.343
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.000
wandering
Token wandering
Feature activation+0.000
around
Token around
Feature activation+0.000
the
Token the
Feature activation+0.000
family
Token family
Feature activation+0.000
connections
Token connections
Feature activation+0.000
,
Token,
Feature activation+0.000
then
Token then
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
custody
Token custody
Feature activation+1.297
in
Token in
Feature activation+0.000
December
Token December
Feature activation+0.000
2007
Token 2007
Feature activation+0.000
while
Token while
Feature activation+0.000
on
Token on
Feature activation+0.000
key
Token key
Feature activation+0.000
to
Token to
Feature activation+0.000
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+1.223
in
Token in
Feature activation+0.000
demographics
Token demographics
Feature activation+0.000
and
Token and
Feature activation+0.000
changes
Token changes
Feature activation+0.000
in
Token in
Feature activation+0.000
insist
Token insist
Feature activation+0.000
upon
Token upon
Feature activation+0.000
entirely
Token entirely
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
the
Token the
Feature activation+0.170
heavy
Token heavy
Feature activation+1.207
hand
Token hand
Feature activation+1.183
of
Token of
Feature activation+0.000
government
Token government
Feature activation+0.000
are
Token are
Feature activation+0.000
dreaming
Token dreaming
Feature activation+0.000

Top DFA by src position
MAX = 3.612

,
Token,
Feature activation+0.010
Top resid features:
a
Token a
Feature activation+0.017
Top resid features:
police
Token police
Feature activation+0.007
Top resid features:
lieutenant
Token lieutenant
Feature activation-0.022
Top resid features:
who
Token who
Feature activation+0.016
Top resid features:
survived
Token survived
Feature activation+3.209
Top resid features:
a
Token a
Feature activation+0.387
Top resid features:
mass
Token mass
Feature activation+0.366
Top resid features:
shooting
Token shooting
Feature activation+0.089
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
Universe
Token Universe
Feature activation+0.025
Top resid features:
,
Token,
Feature activation+0.019
Top resid features:
or
Token or
Feature activation+0.049
Top resid features:
when
Token when
Feature activation+0.154
Top resid features:
they
Token they
Feature activation+0.020
Top resid features:
escape
Token escape
Feature activation+2.993
Top resid features:
certain
Token certain
Feature activation+0.554
Top resid features:
destruction
Token destruction
Feature activation+0.041
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
stunts
Token stunts
Feature activation+0.000
Top resid features:
only
Token only
Feature activation-0.007
Top resid features:
illegal
Token illegal
Feature activation-0.016
Top resid features:
aliens
Token aliens
Feature activation+0.004
Top resid features:
who
Token who
Feature activation+0.057
Top resid features:
are
Token are
Feature activation+0.033
Top resid features:
escaping
Token escaping
Feature activation+3.612
Top resid features:
enforcement
Token enforcement
Feature activation+0.061
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
nation
Token nation
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
impressed
Token impressed
Feature activation-0.013
Top resid features:
.
Token.
Feature activation+0.028
Top resid features:
At
Token At
Feature activation-0.002
Top resid features:
Spurs
Token Spurs
Feature activation-0.016
Top resid features:
he
Token he
Feature activation-0.008
Top resid features:
survived
Token survived
Feature activation+3.554
Top resid features:
being
Token being
Feature activation+0.176
Top resid features:
quite
Token quite
Feature activation+0.000
Top resid features:
clearly
Token clearly
Feature activation+0.000
Top resid features:
targeted
Token targeted
Feature activation+0.000
Top resid features:
for
Token for
Feature activation+0.000
Top resid features:
are
Token are
Feature activation-0.007
Top resid features:
killed
Token killed
Feature activation+0.005
Top resid features:
.
Token.
Feature activation+0.075
Top resid features:
Batman
Token Batman
Feature activation+0.028
Top resid features:
narrowly
Token narrowly
Feature activation+0.039
Top resid features:
escapes
Token escapes
Feature activation+2.794
Top resid features:
the
Token the
Feature activation+0.432
Top resid features:
destruction
Token destruction
Feature activation+0.094
Top resid features:
ray
Token ray
Feature activation+0.000
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
j
Token j
Feature activation+0.000
Top resid features:
tests
Token tests
Feature activation-0.011
Top resid features:
,
Token,
Feature activation+0.036
Top resid features:
Ber
Token Ber
Feature activation+0.017
Top resid features:
lus
Tokenlus
Feature activation+0.007
Top resid features:
coni
Tokenconi
Feature activation+0.031
Top resid features:
survived
Token survived
Feature activation+2.809
Top resid features:
the
Token the
Feature activation+0.362
Top resid features:
no
Token no
Feature activation+0.031
Top resid features:
-
Token-
Feature activation+0.068
Top resid features:
confidence
Tokenconfidence
Feature activation+0.090
Top resid features:
motion
Token motion
Feature activation-0.050
Top resid features:
large
Token large
Feature activation+0.008
Top resid features:
mud
Token mud
Feature activation+0.032
Top resid features:
crab
Token crab
Feature activation+0.032
Top resid features:
that
Token that
Feature activation+0.076
Top resid features:
had
Token had
Feature activation-0.028
Top resid features:
escaped
Token escaped
Feature activation+2.911
Top resid features:
its
Token its
Feature activation+0.215
Top resid features:
box
Token box
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
wandering
Token wandering
Feature activation+0.000
Top resid features:
Universe
Token Universe
Feature activation+0.023
Top resid features:
,
Token,
Feature activation+0.018
Top resid features:
or
Token or
Feature activation+0.045
Top resid features:
when
Token when
Feature activation+0.045
Top resid features:
they
Token they
Feature activation-0.037
Top resid features:
escape
Token escape
Feature activation+2.670
Top resid features:
certain
Token certain
Feature activation+0.242
Top resid features:
destruction
Token destruction
Feature activation+0.000
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
stunts
Token stunts
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.001
Top resid features:
a
Token a
Feature activation+0.012
Top resid features:
police
Token police
Feature activation+0.013
Top resid features:
lieutenant
Token lieutenant
Feature activation-0.048
Top resid features:
who
Token who
Feature activation+0.002
Top resid features:
survived
Token survived
Feature activation+2.451
Top resid features:
a
Token a
Feature activation+0.366
Top resid features:
mass
Token mass
Feature activation+0.136
Top resid features:
shooting
Token shooting
Feature activation+0.000
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.038
Top resid features:
believing
Token believing
Feature activation+0.038
Top resid features:
that
Token that
Feature activation+0.114
Top resid features:
it
Token it
Feature activation+0.109
Top resid features:
has
Token has
Feature activation+0.003
Top resid features:
escaped
Token escaped
Feature activation+2.730
Top resid features:
misfortune
Token misfortune
Feature activation+0.042
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
he
Token he
Feature activation+0.000
Top resid features:
found
Token found
Feature activation+0.000
Top resid features:
himself
Token himself
Feature activation+0.000
Top resid features:
tests
Token tests
Feature activation-0.026
Top resid features:
,
Token,
Feature activation+0.040
Top resid features:
Ber
Token Ber
Feature activation+0.005
Top resid features:
lus
Tokenlus
Feature activation+0.017
Top resid features:
coni
Tokenconi
Feature activation+0.034
Top resid features:
survived
Token survived
Feature activation+2.516
Top resid features:
the
Token the
Feature activation+0.282
Top resid features:
no
Token no
Feature activation+0.000
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
confidence
Tokenconfidence
Feature activation+0.000
Top resid features:
motion
Token motion
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.032
Top resid features:
surface
Token surface
Feature activation-0.008
Top resid features:
warships
Token warships
Feature activation+0.008
Top resid features:
just
Token just
Feature activation+0.048
Top resid features:
cannot
Token cannot
Feature activation+0.064
Top resid features:
survive
Token survive
Feature activation+2.636
Top resid features:
attacks
Token attacks
Feature activation+0.152
Top resid features:
by
Token by
Feature activation+0.000
Top resid features:
large
Token large
Feature activation+0.000
Top resid features:
numbers
Token numbers
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.015
Top resid features:
f
Token f
Feature activation-0.003
Top resid features:
ates
Tokenates
Feature activation+0.019
Top resid features:
of
Token of
Feature activation+0.012
Top resid features:
families
Token families
Feature activation+0.025
Top resid features:
fleeing
Token fleeing
Feature activation+2.795
Top resid features:
persecution
Token persecution
Feature activation+0.123
Top resid features:
from
Token from
Feature activation+0.000
Top resid features:
its
Token its
Feature activation+0.000
Top resid features:
northern
Token northern
Feature activation+0.000
Top resid features:
neighbour
Token neighbour
Feature activation+0.000
Top resid features:
ED
TokenED
Feature activation+0.005
Top resid features:
Ċ
TokenĊ
Feature activation+0.008
Top resid features:
Ċ
TokenĊ
Feature activation-0.001
Top resid features:
Ver
TokenVer
Feature activation+0.032
Top resid features:
lander
Tokenlander
Feature activation+0.009
Top resid features:
escapes
Token escapes
Feature activation+2.517
Top resid features:
third
Token third
Feature activation+0.235
Top resid features:
-
Token-
Feature activation+0.000
Top resid features:
inning
Tokeninning
Feature activation+0.000
Top resid features:
jam
Token jam
Feature activation+0.000
Top resid features:
:
Token:
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.004
Top resid features:
a
Token a
Feature activation+0.020
Top resid features:
police
Token police
Feature activation+0.001
Top resid features:
lieutenant
Token lieutenant
Feature activation+0.004
Top resid features:
who
Token who
Feature activation+0.014
Top resid features:
survived
Token survived
Feature activation+2.565
Top resid features:
a
Token a
Feature activation+0.204
Top resid features:
mass
Token mass
Feature activation+0.000
Top resid features:
shooting
Token shooting
Feature activation+0.000
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
a
Token a
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.028
Top resid features:
Ċ
TokenĊ
Feature activation+0.026
Top resid features:
David
TokenDavid
Feature activation+0.011
Top resid features:
Davis
Token Davis
Feature activation-0.002
Top resid features:
has
Token has
Feature activation-0.000
Top resid features:
escaped
Token escaped
Feature activation+2.500
Top resid features:
the
Token the
Feature activation+0.351
Top resid features:
possibility
Token possibility
Feature activation-0.044
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
being
Token being
Feature activation+0.000
Top resid features:
held
Token held
Feature activation+0.000
Top resid features:
large
Token large
Feature activation+0.005
Top resid features:
mud
Token mud
Feature activation+0.065
Top resid features:
crab
Token crab
Feature activation+0.091
Top resid features:
that
Token that
Feature activation+0.053
Top resid features:
had
Token had
Feature activation-0.065
Top resid features:
escaped
Token escaped
Feature activation+1.870
Top resid features:
its
Token its
Feature activation+0.703
Top resid features:
box
Token box
Feature activation+0.039
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
wandering
Token wandering
Feature activation+0.000
Top resid features:
his
Token his
Feature activation+0.002
Top resid features:
family
Token family
Feature activation+0.014
Top resid features:
connections
Token connections
Feature activation-0.003
Top resid features:
,
Token,
Feature activation+0.026
Top resid features:
then
Token then
Feature activation+0.101
Top resid features:
escaped
Token escaped
Feature activation+2.561
Top resid features:
custody
Token custody
Feature activation+0.086
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
December
Token December
Feature activation+0.000
Top resid features:
2007
Token 2007
Feature activation+0.000
Top resid features:
while
Token while
Feature activation+0.000
Top resid features:
is
Token is
Feature activation-0.008
Top resid features:
key
Token key
Feature activation-0.017
Top resid features:
to
Token to
Feature activation+0.027
Top resid features:
a
Token a
Feature activation+0.055
Top resid features:
society
Token society
Feature activation-0.008
Top resid features:
surviving
Token surviving
Feature activation+2.801
Top resid features:
changes
Token changes
Feature activation-0.010
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
demographics
Token demographics
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
changes
Token changes
Feature activation+0.000
Top resid features:
ists
Tokenists
Feature activation-0.001
Top resid features:
who
Token who
Feature activation+0.009
Top resid features:
insist
Token insist
Feature activation-0.013
Top resid features:
upon
Token upon
Feature activation+0.010
Top resid features:
entirely
Token entirely
Feature activation+0.052
Top resid features:
escaping
Token escaping
Feature activation+2.165
Top resid features:
the
Token the
Feature activation+0.425
Top resid features:
heavy
Token heavy
Feature activation+0.027
Top resid features:
hand
Token hand
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
government
Token government
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.04

Head 2: 0.09

Head 3: 0.04

Head 4: 0.02

Head 5: 0.04

Head 6: 0.06

Head 7: 0.05

Head 8: 0.41

Head 9: 0.08

Head 10: 0.05

Head 11: 0.08

Positive logits

cation1.79

imester1.58

ratch1.44

confines1.40

ossession1.35

assault1.33

rosso1.32

oho1.31

ptives1.30

orgetown1.29

cru1.29

alcohol1.28

fb1.27

mort1.26

death1.25

shack1.24

parts1.24

breeding1.22

violent1.22

frac1.21

Negative logits

glim-1.38

accordingly-1.35

Encyclopedia-1.31

behav-1.24

perspect-1.24

antioxid-1.23

empath-1.22

Filip-1.22

archaeologists-1.20

negotiator-1.20

encyclopedia-1.18

respons-1.18

councillors-1.17

emoji-1.16

Quote-1.16

Doodle-1.16

diagram-1.15

rug-1.15

interpol-1.15

envoy-1.13

INTERVAL 2.363 - 2.626
CONTAINS 0.000%

lieutenant
Token lieutenant
Feature activation+0.000
who
Token who
Feature activation+0.000
survived
Token survived
Feature activation+0.000
a
Token a
Feature activation+1.376
mass
Token mass
Feature activation+1.533
shooting
Token shooting
Feature activation+2.626
at
Token at
Feature activation+0.771
a
Token a
Feature activation+0.000
Sikh
Token Sikh
Feature activation+0.000
temple
Token temple
Feature activation+0.000
in
Token in
Feature activation+0.000

INTERVAL 2.101 - 2.363
CONTAINS 0.000%

.
Token.
Feature activation+0.000
At
Token At
Feature activation+0.000
Spurs
Token Spurs
Feature activation+0.000
he
Token he
Feature activation+0.000
survived
Token survived
Feature activation+0.000
being
Token being
Feature activation+2.185
quite
Token quite
Feature activation+0.204
clearly
Token clearly
Feature activation+0.717
targeted
Token targeted
Feature activation+0.000
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
or
Token or
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
escape
Token escape
Feature activation+0.000
certain
Token certain
Feature activation+1.654
destruction
Token destruction
Feature activation+2.338
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
stunts
Token stunts
Feature activation+0.000
hip
Tokenhip
Feature activation+0.000
used
Token used
Feature activation+0.000
illegal
Token illegal
Feature activation+0.000
aliens
Token aliens
Feature activation+0.000
who
Token who
Feature activation+0.000
are
Token are
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
enforcement
Token enforcement
Feature activation+2.215
of
Token of
Feature activation+0.312
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

INTERVAL 1.838 - 2.101
CONTAINS 0.000%

survived
Token survived
Feature activation+0.000
the
Token the
Feature activation+1.488
no
Token no
Feature activation+0.508
-
Token-
Feature activation+0.000
confidence
Tokenconfidence
Feature activation+0.344
motion
Token motion
Feature activation+1.879
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
lower
Token lower
Feature activation+0.000
house
Token house
Feature activation+0.000
by
Token by
Feature activation+0.000
.
Token.
Feature activation+0.000
Batman
Token Batman
Feature activation+0.000
narrowly
Token narrowly
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
the
Token the
Feature activation+1.038
destruction
Token destruction
Feature activation+1.988
ray
Token ray
Feature activation+0.073
by
Token by
Feature activation+0.039
j
Token j
Feature activation+0.000
ett
Tokenett
Feature activation+0.000
ison
Tokenison
Feature activation+0.000

INTERVAL 1.576 - 1.838
CONTAINS 0.000%

,
Token,
Feature activation+0.000
or
Token or
Feature activation+0.000
when
Token when
Feature activation+0.000
they
Token they
Feature activation+0.000
escape
Token escape
Feature activation+0.000
certain
Token certain
Feature activation+1.654
destruction
Token destruction
Feature activation+2.338
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
stunts
Token stunts
Feature activation+0.000
hip
Tokenhip
Feature activation+0.000
mud
Token mud
Feature activation+0.000
crab
Token crab
Feature activation+0.000
that
Token that
Feature activation+0.000
had
Token had
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
its
Token its
Feature activation+1.804
box
Token box
Feature activation+1.343
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.000
wandering
Token wandering
Feature activation+0.000
around
Token around
Feature activation+0.000

INTERVAL 1.313 - 1.576
CONTAINS 0.001%

David
TokenDavid
Feature activation+0.000
Davis
Token Davis
Feature activation+0.000
has
Token has
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
the
Token the
Feature activation+0.811
possibility
Token possibility
Feature activation+1.363
of
Token of
Feature activation+0.125
being
Token being
Feature activation+0.000
held
Token held
Feature activation+0.000
in
Token in
Feature activation+0.000
contempt
Token contempt
Feature activation+0.000
crab
Token crab
Feature activation+0.000
that
Token that
Feature activation+0.000
had
Token had
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
its
Token its
Feature activation+1.804
box
Token box
Feature activation+1.343
and
Token and
Feature activation+0.000
was
Token was
Feature activation+0.000
wandering
Token wandering
Feature activation+0.000
around
Token around
Feature activation+0.000
the
Token the
Feature activation+0.000
,
Token,
Feature activation+0.000
Ber
Token Ber
Feature activation+0.000
lus
Tokenlus
Feature activation+0.000
coni
Tokenconi
Feature activation+0.000
survived
Token survived
Feature activation+0.000
the
Token the
Feature activation+1.488
no
Token no
Feature activation+0.508
-
Token-
Feature activation+0.000
confidence
Tokenconfidence
Feature activation+0.344
motion
Token motion
Feature activation+1.879
in
Token in
Feature activation+0.000
believing
Token believing
Feature activation+0.000
that
Token that
Feature activation+0.000
it
Token it
Feature activation+0.000
has
Token has
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
misfortune
Token misfortune
Feature activation+1.499
,
Token,
Feature activation+0.000
he
Token he
Feature activation+0.000
found
Token found
Feature activation+0.000
himself
Token himself
Feature activation+0.000
in
Token in
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ver
TokenVer
Feature activation+0.000
lander
Tokenlander
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
third
Token third
Feature activation+1.403
-
Token-
Feature activation+0.990
inning
Tokeninning
Feature activation+0.673
jam
Token jam
Feature activation+0.851
:
Token:
Feature activation+0.000
Ver
Token Ver
Feature activation+0.000

INTERVAL 1.050 - 1.313
CONTAINS 0.001%

insist
Token insist
Feature activation+0.000
upon
Token upon
Feature activation+0.000
entirely
Token entirely
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
the
Token the
Feature activation+0.170
heavy
Token heavy
Feature activation+1.207
hand
Token hand
Feature activation+1.183
of
Token of
Feature activation+0.000
government
Token government
Feature activation+0.000
are
Token are
Feature activation+0.000
dreaming
Token dreaming
Feature activation+0.000
you
Token you
Feature activation+0.000
are
Token are
Feature activation+0.000
trying
Token trying
Feature activation+0.000
to
Token to
Feature activation+0.000
escape
Token escape
Feature activation+0.000
writing
Token writing
Feature activation+1.117
it
Token it
Feature activation+0.000
.
Token.
Feature activation+0.000
By
Token By
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
family
Token family
Feature activation+0.000
connections
Token connections
Feature activation+0.000
,
Token,
Feature activation+0.000
then
Token then
Feature activation+0.000
escaped
Token escaped
Feature activation+0.000
custody
Token custody
Feature activation+1.297
in
Token in
Feature activation+0.000
December
Token December
Feature activation+0.000
2007
Token 2007
Feature activation+0.000
while
Token while
Feature activation+0.000
on
Token on
Feature activation+0.000
upon
Token upon
Feature activation+0.000
entirely
Token entirely
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
the
Token the
Feature activation+0.170
heavy
Token heavy
Feature activation+1.207
hand
Token hand
Feature activation+1.183
of
Token of
Feature activation+0.000
government
Token government
Feature activation+0.000
are
Token are
Feature activation+0.000
dreaming
Token dreaming
Feature activation+0.000
.
Token.
Feature activation+0.000
key
Token key
Feature activation+0.000
to
Token to
Feature activation+0.000
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+1.223
in
Token in
Feature activation+0.000
demographics
Token demographics
Feature activation+0.000
and
Token and
Feature activation+0.000
changes
Token changes
Feature activation+0.000
in
Token in
Feature activation+0.000

INTERVAL 0.788 - 1.050
CONTAINS 0.002%

lander
Tokenlander
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
third
Token third
Feature activation+1.403
-
Token-
Feature activation+0.990
inning
Tokeninning
Feature activation+0.673
jam
Token jam
Feature activation+0.851
:
Token:
Feature activation+0.000
Ver
Token Ver
Feature activation+0.000
lander
Tokenlander
Feature activation+0.000
had
Token had
Feature activation+0.000
runners
Token runners
Feature activation+0.000
he
Token he
Feature activation+0.000
fights
Token fights
Feature activation+0.000
to
Token to
Feature activation+0.000
overcome
Token overcome
Feature activation+0.000
a
Token a
Feature activation+0.000
calf
Token calf
Feature activation+0.792
problem
Token problem
Feature activation+0.254
.
Token.
Feature activation+0.000
It
Token It
Feature activation+0.000
is
Token is
Feature activation+0.000
hoped
Token hoped
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ver
TokenVer
Feature activation+0.000
lander
Tokenlander
Feature activation+0.000
escapes
Token escapes
Feature activation+0.000
third
Token third
Feature activation+1.403
-
Token-
Feature activation+0.990
inning
Tokeninning
Feature activation+0.673
jam
Token jam
Feature activation+0.851
:
Token:
Feature activation+0.000
Ver
Token Ver
Feature activation+0.000
lander
Tokenlander
Feature activation+0.000
promising
Token promising
Feature activation+0.000
to
Token to
Feature activation+0.000
help
Token help
Feature activation+0.000
them
Token them
Feature activation+0.000
overcome
Token overcome
Feature activation+0.000
work
Token work
Feature activation+0.812
or
Token or
Feature activation+0.000
love
Token love
Feature activation+0.000
difficulties
Token difficulties
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
survive
Token survive
Feature activation+0.000
?
Token?
Feature activation+0.000
Can
Token Can
Feature activation+0.000
Maggie
Token Maggie
Feature activation+0.000
survive
Token survive
Feature activation+0.000
this
Token this
Feature activation+0.973
ordeal
Token ordeal
Feature activation+0.803
?
Token?
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
I
TokenI
Feature activation+0.000

INTERVAL 0.525 - 0.788
CONTAINS 0.002%

Mind
Token Mind
Feature activation+0.000
Control
Token Control
Feature activation+0.000
will
Token will
Feature activation+0.000
never
Token never
Feature activation+0.000
overcome
Token overcome
Feature activation+0.000
D
Token D
Feature activation+0.710
EST
TokenEST
Feature activation+0.112
IN
TokenIN
Feature activation+0.000
Y
TokenY
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
"
Token "
Feature activation+0.000
won
Tokenwon
Feature activation+0.000
't
Token't
Feature activation+0.000
survive
Token survive
Feature activation+0.000
"
Token"
Feature activation+0.000
a
Token a
Feature activation+0.688
resurgence
Token resurgence
Feature activation+0.492
in
Token in
Feature activation+0.000
democracy
Token democracy
Feature activation+0.000
brought
Token brought
Feature activation+0.000
about
Token about
Feature activation+0.000
forced
Token forced
Feature activation+0.000
to
Token to
Feature activation+0.000
endure
Token endure
Feature activation+0.000
a
Token a
Feature activation+0.000
sustained
Token sustained
Feature activation+0.000
attack
Token attack
Feature activation+0.567
from
Token from
Feature activation+0.000
the
Token the
Feature activation+0.000
air
Token air
Feature activation+0.000
and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.000
apparent
Token apparent
Feature activation+0.000
bid
Token bid
Feature activation+0.000
to
Token to
Feature activation+0.000
evade
Token evade
Feature activation+0.000
federal
Token federal
Feature activation+0.000
record
Token record
Feature activation+0.680
-
Token-
Feature activation+0.000
keeping
Tokenkeeping
Feature activation+0.815
laws
Token laws
Feature activation+0.130
that
Token that
Feature activation+0.000
require
Token require
Feature activation+0.000
problems
Token problems
Feature activation+0.000
and
Token and
Feature activation+0.000
resolving
Token resolving
Feature activation+0.000
sl
Token sl
Feature activation+0.044
udge
Tokenudge
Feature activation+0.449
management
Token management
Feature activation+0.533
issues
Token issues
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Enjoy
TokenEnjoy
Feature activation+0.000

INTERVAL 0.263 - 0.525
CONTAINS 0.003%

anyone
Token anyone
Feature activation+0.000
who
Token who
Feature activation+0.000
survived
Token survived
Feature activation+0.000
the
Token the
Feature activation+1.106
bruising
Token bruising
Feature activation+1.038
campaigns
Token campaigns
Feature activation+0.324
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
1990
Token 1990
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
and
Token and
Feature activation+0.000
repair
Token repair
Feature activation+0.000
projects
Token projects
Feature activation+0.000
following
Token following
Feature activation+0.000
this
Token this
Feature activation+0.000
disaster
Token disaster
Feature activation+0.413
.
Token.
Feature activation+0.000
A
Token A
Feature activation+0.000
tool
Token tool
Feature activation+0.000
-
Token-
Feature activation+0.000
l
Tokenl
Feature activation+0.000
all
Token all
Feature activation+0.000
this
Token this
Feature activation+0.000
behind
Token behind
Feature activation+0.000
,
Token,
Feature activation+0.000
escaping
Token escaping
Feature activation+0.000
Syria
Token Syria
Feature activation+0.482
to
Token to
Feature activation+0.171
risk
Token risk
Feature activation+0.000
everything
Token everything
Feature activation+0.000
on
Token on
Feature activation+0.000
a
Token a
Feature activation+0.000
president
Token president
Feature activation+0.000
whose
Token whose
Feature activation+0.000
allies
Token allies
Feature activation+0.000
have
Token have
Feature activation+0.000
endured
Token endured
Feature activation+0.000
crises
Token crises
Feature activation+0.316
and
Token and
Feature activation+0.000
controversies
Token controversies
Feature activation+0.000
since
Token since
Feature activation+0.000
the
Token the
Feature activation+0.000
day
Token day
Feature activation+0.000
-
Token-
Feature activation+0.000
W
TokenW
Feature activation+0.000
ale
Tokenale
Feature activation+0.000
ed
Tokened
Feature activation+0.000
bin
Token bin
Feature activation+0.000
Tal
Token Tal
Feature activation+0.314
al
Tokenal
Feature activation+0.195
âĢĶ
TokenâĢĶ
Feature activation+0.000
is
Tokenis
Feature activation+0.000
timed
Token timed
Feature activation+0.000
to
Token to
Feature activation+0.000

INTERVAL 0.000 - 0.263
CONTAINS 99.990%

the
Token the
Feature activation+0.000
guise
Token guise
Feature activation+0.000
of
Token of
Feature activation+0.000
being
Token being
Feature activation+0.000
entertaining
Token entertaining
Feature activation+0.000
or
Token or
Feature activation+0.000
humorous
Token humorous
Feature activation+0.000
.
Token.
Feature activation+0.000
From
Token From
Feature activation+0.000
parents
Token parents
Feature activation+0.000
shaming
Token shaming
Feature activation+0.000
to
Token to
Feature activation+0.000
see
Token see
Feature activation+0.000
whether
Token whether
Feature activation+0.000
the
Token the
Feature activation+0.000
New
Token New
Feature activation+0.000
York
Token York
Feature activation+0.000
Post
Token Post
Feature activation+0.000
backs
Token backs
Feature activation+0.000
Trump
Token Trump
Feature activation+0.000
again
Token again
Feature activation+0.000
this
Token this
Feature activation+0.000
"
Token "
Feature activation+0.000
around
Tokenaround
Feature activation+0.000
the
Token the
Feature activation+0.000
clock
Token clock
Feature activation+0.000
"
Token"
Feature activation+0.000
to
Token to
Feature activation+0.000
keep
Token keep
Feature activation+0.000
the
Token the
Feature activation+0.000
tracks
Token tracks
Feature activation+0.000
clear
Token clear
Feature activation+0.000
.
Token.
Feature activation+0.000
on
Token on
Feature activation+0.000
to
Token to
Feature activation+0.000
earn
Token earn
Feature activation+0.000
a
Token a
Feature activation+0.000
degree
Token degree
Feature activation+0.000
in
Token in
Feature activation+0.000
philosophy
Token philosophy
Feature activation+0.000
from
Token from
Feature activation+0.000
Boston
Token Boston
Feature activation+0.000
University
Token University
Feature activation+0.000
,
Token,
Feature activation+0.000
from
Token from
Feature activation+0.000
it
Token it
Feature activation+0.000
and
Token and
Feature activation+0.000
we
Token we
Feature activation+0.000
have
Token have
Feature activation+0.000
taken
Token taken
Feature activation+0.000
the
Token the
Feature activation+0.000
steps
Token steps
Feature activation+0.000
necessary
Token necessary
Feature activation+0.000
to
Token to
Feature activation+0.000
bring
Token bring
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 H3.8: (feature 8249)

TOP ACTIVATIONS
MAX = 2.992

's
Token's
Feature activation+1.489
"
Token "
Feature activation+1.060
Science
TokenScience
Feature activation+0.000
Friday
Token Friday
Feature activation+0.000
"
Token"
Feature activation+1.202
host
Token host
Feature activation+2.992
I
Token I
Feature activation+0.000
ra
Tokenra
Feature activation+0.000
Flat
Token Flat
Feature activation+0.000
ow
Tokenow
Feature activation+0.000
on
Token on
Feature activation+0.000
he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
fellow
Token fellow
Feature activation+1.319
AM
Token AM
Feature activation+1.053
VA
TokenVA
Feature activation+0.000
members
Token members
Feature activation+2.897
.
Token.
Feature activation+0.265
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Despite
TokenDespite
Feature activation+0.000
the
Token the
Feature activation+0.000
PS
TokenPS
Feature activation+0.000
website
Token website
Feature activation+0.000
also
Token also
Feature activation+0.000
tells
Token tells
Feature activation+0.000
potential
Token potential
Feature activation+1.457
visitors
Token visitors
Feature activation+2.887
about
Token about
Feature activation+0.000
how
Token how
Feature activation+0.000
unpredictable
Token unpredictable
Feature activation+0.000
the
Token the
Feature activation+0.000
weather
Token weather
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Obama
Token Obama
Feature activation+0.000
told
Token told
Feature activation+0.089
reporters
Token reporters
Feature activation+1.408
at
Token at
Feature activation+2.705
the
Token the
Feature activation+0.000
end
Token end
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
private
Token private
Feature activation+0.000
of
Token of
Feature activation+0.000
Comey
Token Comey
Feature activation+0.000
telling
Token telling
Feature activation+0.000
a
Token a
Feature activation+1.464
congressional
Token congressional
Feature activation+1.634
committee
Token committee
Feature activation+2.688
that
Token that
Feature activation+0.693
she
Token she
Feature activation+0.000
had
Token had
Feature activation+0.000
been
Token been
Feature activation+0.000
unt
Token unt
Feature activation+0.000
."
Token."
Feature activation+0.000
He
Token He
Feature activation+0.000
tells
Token tells
Feature activation+0.000
a
Token a
Feature activation+0.826
police
Token police
Feature activation+1.400
sergeant
Token sergeant
Feature activation+2.611
that
Token that
Feature activation+0.974
his
Token his
Feature activation+0.000
step
Token step
Feature activation+0.000
-
Token-
Feature activation+0.000
s
Tokens
Feature activation+0.000
.
Token.
Feature activation+0.000
She
Token She
Feature activation+0.000
told
Token told
Feature activation+0.000
Boston
Token Boston
Feature activation+1.371
police
Token police
Feature activation+2.200
officers
Token officers
Feature activation+2.576
what
Token what
Feature activation+0.064
G
Token G
Feature activation+0.000
j
Tokenj
Feature activation+0.000
oni
Tokenoni
Feature activation+0.000
had
Token had
Feature activation+0.000
,
Token,
Feature activation+0.000
a
Token a
Feature activation+0.000
rep
Token rep
Feature activation+0.000
tells
Token tells
Feature activation+0.000
Cur
Token Cur
Feature activation+0.000
bed
Tokenbed
Feature activation+2.540
Atlanta
Token Atlanta
Feature activation+0.443
.
Token.
Feature activation+0.000
Else
Token Else
Feature activation+0.000
where
Tokenwhere
Feature activation+0.000
in
Token in
Feature activation+0.000
is
Token is
Feature activation+0.000
what
Token what
Feature activation+0.000
Sessions
Token Sessions
Feature activation+0.000
told
Token told
Feature activation+0.000
fellow
Token fellow
Feature activation+1.709
senators
Token senators
Feature activation+2.540
when
Token when
Feature activation+0.753
he
Token he
Feature activation+0.000
was
Token was
Feature activation+0.000
asked
Token asked
Feature activation+0.000
straightforward
Token straightforward
Feature activation+0.000
telling
Token telling
Feature activation+0.000
another
Token another
Feature activation+0.438
(
Token (
Feature activation+0.104
female
Tokenfemale
Feature activation+0.546
)
Token)
Feature activation+1.039
stream
Token stream
Feature activation+2.518
er
Tokener
Feature activation+2.069
âĢ
Token âĢ
Feature activation+1.488
ĺ
Tokenĺ
Feature activation+0.000
Eat
TokenEat
Feature activation+0.000
my
Token my
Feature activation+0.000
after
Token after
Feature activation+0.000
some
Token some
Feature activation+0.000
days
Token days
Feature activation+0.000
i
Token i
Feature activation+0.000
tell
Token tell
Feature activation+0.000
advance
Token advance
Feature activation+2.511
level
Token level
Feature activation+1.679
method
Token method
Feature activation+1.128
.
Token.
Feature activation+0.277
Th
TokenTh
Feature activation+0.000
ats
Tokenats
Feature activation+0.000
when
Token when
Feature activation+0.000
she
Token she
Feature activation+0.000
told
Token told
Feature activation+0.000
CNN
Token CNN
Feature activation+0.858
Gingrich
Token Gingrich
Feature activation+0.428
would
Token would
Feature activation+2.469
not
Token not
Feature activation+0.000
be
Token be
Feature activation+0.230
president
Token president
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.613
New
Token New
Feature activation+0.476
York
Token York
Feature activation+0.872
Police
Token Police
Feature activation+0.580
officer
Token officer
Feature activation+2.057
with
Token with
Feature activation+2.459
the
Token the
Feature activation+0.000
Manhattan
Token Manhattan
Feature activation+0.000
Special
Token Special
Feature activation+0.000
Victims
Token Victims
Feature activation+0.000
squad
Token squad
Feature activation+0.125
'm
Token'm
Feature activation+0.000
going
Token going
Feature activation+0.000
to
Token to
Feature activation+0.000
tell
Token tell
Feature activation+0.000
the
Token the
Feature activation+0.087
boss
Token boss
Feature activation+2.452
I
Token I
Feature activation+1.179
've
Token've
Feature activation+0.148
got
Token got
Feature activation+0.000
a
Token a
Feature activation+0.000
medical
Token medical
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Koh
Token Koh
Feature activation+0.000
told
Token told
Feature activation+0.000
a
Token a
Feature activation+0.332
Cleveland
Token Cleveland
Feature activation+0.245
audience
Token audience
Feature activation+2.414
.
Token.
Feature activation+0.181
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
K
TokenK
Feature activation+0.000
oh
Tokenoh
Feature activation+0.000
Bru
Token Bru
Feature activation+0.000
yne
Tokenyne
Feature activation+0.000
el
Tokenel
Feature activation+0.000
told
Token told
Feature activation+0.000
Belgian
Token Belgian
Feature activation+1.190
magazine
Token magazine
Feature activation+2.406
Hum
Token Hum
Feature activation+0.346
o
Tokeno
Feature activation+0.000
.
Token.
Feature activation+0.245
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
Sources
Token Sources
Feature activation+0.000
within
Token within
Feature activation+0.000
the
Token the
Feature activation+0.000
league
Token league
Feature activation+0.000
tell
Token tell
Feature activation+0.000
Goal
Token Goal
Feature activation+2.368
USA
Token USA
Feature activation+2.055
there
Token there
Feature activation+1.322
is
Token is
Feature activation+0.211
scarce
Token scarce
Feature activation+0.000
interest
Token interest
Feature activation+0.000
professor
Token professor
Feature activation+0.000
telling
Token telling
Feature activation+0.000
her
Token her
Feature activation+0.849
a
Token a
Feature activation+1.660
graduate
Token graduate
Feature activation+1.698
student
Token student
Feature activation+2.366
âĢ
TokenâĢ
Feature activation+1.696
Ļ
TokenĻ
Feature activation+0.113
s
Tokens
Feature activation+0.124
expected
Token expected
Feature activation+0.559
pay
Token pay
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
Gibson
Token Gibson
Feature activation+0.000
told
Token told
Feature activation+0.473
a
Token a
Feature activation+0.402
crowd
Token crowd
Feature activation+2.112
at
Token at
Feature activation+2.342
Terry
Token Terry
Feature activation+0.000
Schr
Token Schr
Feature activation+0.000
unk
Tokenunk
Feature activation+0.000
Plaza
Token Plaza
Feature activation+0.000
,
Token,
Feature activation+0.061
.
Token.
Feature activation+0.000
Whenever
Token Whenever
Feature activation+0.000
he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
friends
Token friends
Feature activation+1.739
and
Token and
Feature activation+2.321
colleagues
Token colleagues
Feature activation+0.302
about
Token about
Feature activation+0.000
his
Token his
Feature activation+0.000
gig
Token gig
Feature activation+0.000
,
Token,
Feature activation+0.000

Top DFA by src position
MAX = 3.508

joining
Token joining
Feature activation-0.042
Top resid features:
in
Token in
Feature activation-0.014
Top resid features:
,
Token,
Feature activation-0.003
Top resid features:
too
Token too
Feature activation-0.019
Top resid features:
,
Token,
Feature activation-0.027
Top resid features:
telling
Token telling
Feature activation+3.508
Top resid features:
NPR
Token NPR
Feature activation-0.064
Top resid features:
's
Token's
Feature activation+0.339
Top resid features:
"
Token "
Feature activation+0.098
Top resid features:
Science
TokenScience
Feature activation+0.048
Top resid features:
Friday
Token Friday
Feature activation-0.028
Top resid features:
animals
Token animals
Feature activation-0.002
Top resid features:
,
Token,
Feature activation-0.021
Top resid features:
âĢ
TokenâĢ
Feature activation-0.011
Top resid features:
Ŀ
TokenĿ
Feature activation-0.010
Top resid features:
he
Token he
Feature activation-0.036
Top resid features:
told
Token told
Feature activation+3.309
Top resid features:
fellow
Token fellow
Feature activation+0.298
Top resid features:
AM
Token AM
Feature activation+0.011
Top resid features:
VA
TokenVA
Feature activation+0.069
Top resid features:
members
Token members
Feature activation+0.023
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
The
TokenThe
Feature activation+0.005
Top resid features:
N
Token N
Feature activation-0.001
Top resid features:
PS
TokenPS
Feature activation+0.006
Top resid features:
website
Token website
Feature activation-0.044
Top resid features:
also
Token also
Feature activation-0.023
Top resid features:
tells
Token tells
Feature activation+3.157
Top resid features:
potential
Token potential
Feature activation+0.507
Top resid features:
visitors
Token visitors
Feature activation+0.031
Top resid features:
about
Token about
Feature activation+0.000
Top resid features:
how
Token how
Feature activation+0.000
Top resid features:
unpredictable
Token unpredictable
Feature activation+0.000
Top resid features:
serve
Token serve
Feature activation-0.005
Top resid features:
,
Token,
Feature activation-0.046
Top resid features:
âĢ
TokenâĢ
Feature activation-0.003
Top resid features:
Ŀ
TokenĿ
Feature activation-0.005
Top resid features:
Obama
Token Obama
Feature activation+0.089
Top resid features:
told
Token told
Feature activation+3.151
Top resid features:
reporters
Token reporters
Feature activation+0.368
Top resid features:
at
Token at
Feature activation-0.069
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
end
Token end
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.000
Top resid features:
Despite
Token Despite
Feature activation-0.025
Top resid features:
the
Token the
Feature activation-0.017
Top resid features:
humiliation
Token humiliation
Feature activation+0.053
Top resid features:
of
Token of
Feature activation-0.025
Top resid features:
Comey
Token Comey
Feature activation+0.014
Top resid features:
telling
Token telling
Feature activation+3.121
Top resid features:
a
Token a
Feature activation+0.266
Top resid features:
congressional
Token congressional
Feature activation+0.085
Top resid features:
committee
Token committee
Feature activation-0.038
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
she
Token she
Feature activation+0.000
Top resid features:
there
Token there
Feature activation-0.003
Top resid features:
right
Token right
Feature activation-0.001
Top resid features:
now
Token now
Feature activation-0.011
Top resid features:
."
Token."
Feature activation-0.012
Top resid features:
He
Token He
Feature activation-0.012
Top resid features:
tells
Token tells
Feature activation+2.912
Top resid features:
a
Token a
Feature activation+0.341
Top resid features:
police
Token police
Feature activation-0.000
Top resid features:
sergeant
Token sergeant
Feature activation+0.163
Top resid features:
that
Token that
Feature activation+0.000
Top resid features:
his
Token his
Feature activation+0.000
Top resid features:
âĢĶ
TokenâĢĶ
Feature activation-0.002
Top resid features:
numerous
Token numerous
Feature activation+0.005
Top resid features:
times
Token times
Feature activation+0.012
Top resid features:
.
Token.
Feature activation+0.014
Top resid features:
She
Token She
Feature activation-0.020
Top resid features:
told
Token told
Feature activation+2.857
Top resid features:
Boston
Token Boston
Feature activation+0.250
Top resid features:
police
Token police
Feature activation+0.039
Top resid features:
officers
Token officers
Feature activation+0.123
Top resid features:
what
Token what
Feature activation+0.000
Top resid features:
G
Token G
Feature activation+0.000
Top resid features:
000
Token000
Feature activation-0.003
Top resid features:
s
Tokens
Feature activation-0.000
Top resid features:
,
Token,
Feature activation-0.028
Top resid features:
a
Token a
Feature activation-0.104
Top resid features:
rep
Token rep
Feature activation+0.003
Top resid features:
tells
Token tells
Feature activation+2.809
Top resid features:
Cur
Token Cur
Feature activation+0.580
Top resid features:
bed
Tokenbed
Feature activation-0.039
Top resid features:
Atlanta
Token Atlanta
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Else
Token Else
Feature activation+0.000
Top resid features:
What
Token What
Feature activation-0.001
Top resid features:
matters
Token matters
Feature activation-0.030
Top resid features:
is
Token is
Feature activation+0.008
Top resid features:
what
Token what
Feature activation-0.040
Top resid features:
Sessions
Token Sessions
Feature activation+0.035
Top resid features:
told
Token told
Feature activation+2.836
Top resid features:
fellow
Token fellow
Feature activation+0.541
Top resid features:
senators
Token senators
Feature activation+0.032
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
he
Token he
Feature activation+0.000
Top resid features:
was
Token was
Feature activation+0.000
Top resid features:
that
Token that
Feature activation-0.017
Top resid features:
he
Token he
Feature activation-0.029
Top resid features:
was
Token was
Feature activation-0.005
Top resid features:
off
Token off
Feature activation+0.013
Top resid features:
,
Token,
Feature activation-0.070
Top resid features:
telling
Token telling
Feature activation+2.772
Top resid features:
another
Token another
Feature activation+0.618
Top resid features:
(
Token (
Feature activation+0.063
Top resid features:
female
Tokenfemale
Feature activation+0.044
Top resid features:
)
Token)
Feature activation-0.069
Top resid features:
stream
Token stream
Feature activation-0.104
Top resid features:
and
Token and
Feature activation+0.002
Top resid features:
after
Token after
Feature activation+0.052
Top resid features:
some
Token some
Feature activation-0.035
Top resid features:
days
Token days
Feature activation+0.123
Top resid features:
i
Token i
Feature activation-0.044
Top resid features:
tell
Token tell
Feature activation+2.936
Top resid features:
advance
Token advance
Feature activation+0.202
Top resid features:
level
Token level
Feature activation+0.000
Top resid features:
method
Token method
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Th
TokenTh
Feature activation+0.000
Top resid features:
headlines
Token headlines
Feature activation+0.036
Top resid features:
on
Token on
Feature activation-0.012
Top resid features:
Tuesday
Token Tuesday
Feature activation-0.006
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
she
Token she
Feature activation-0.014
Top resid features:
told
Token told
Feature activation+2.920
Top resid features:
CNN
Token CNN
Feature activation-0.070
Top resid features:
Gingrich
Token Gingrich
Feature activation+0.399
Top resid features:
would
Token would
Feature activation-0.141
Top resid features:
not
Token not
Feature activation+0.000
Top resid features:
be
Token be
Feature activation+0.000
Top resid features:
four
Token four
Feature activation-0.006
Top resid features:
hours
Token hours
Feature activation+0.028
Top resid features:
later
Token later
Feature activation+0.023
Top resid features:
,
Token,
Feature activation+0.002
Top resid features:
he
Token he
Feature activation-0.008
Top resid features:
told
Token told
Feature activation+2.925
Top resid features:
a
Token a
Feature activation+0.266
Top resid features:
New
Token New
Feature activation+0.012
Top resid features:
York
Token York
Feature activation+0.013
Top resid features:
Police
Token Police
Feature activation+0.018
Top resid features:
officer
Token officer
Feature activation+0.085
Top resid features:
"
Token"
Feature activation+0.010
Top resid features:
I
TokenI
Feature activation-0.025
Top resid features:
'm
Token'm
Feature activation-0.014
Top resid features:
going
Token going
Feature activation+0.003
Top resid features:
to
Token to
Feature activation-0.018
Top resid features:
tell
Token tell
Feature activation+2.849
Top resid features:
the
Token the
Feature activation+0.074
Top resid features:
boss
Token boss
Feature activation+0.295
Top resid features:
I
Token I
Feature activation+0.000
Top resid features:
've
Token've
Feature activation+0.000
Top resid features:
got
Token got
Feature activation+0.000
Top resid features:
rules
Token rules
Feature activation+0.002
Top resid features:
,
Token,
Feature activation-0.046
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
Ŀ
TokenĿ
Feature activation-0.012
Top resid features:
Koh
Token Koh
Feature activation+0.006
Top resid features:
told
Token told
Feature activation+2.763
Top resid features:
a
Token a
Feature activation+0.387
Top resid features:
Cleveland
Token Cleveland
Feature activation+0.111
Top resid features:
audience
Token audience
Feature activation-0.046
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation-0.040
Top resid features:
Ŀ
TokenĿ
Feature activation-0.012
Top resid features:
Bru
Token Bru
Feature activation+0.008
Top resid features:
yne
Tokenyne
Feature activation+0.009
Top resid features:
el
Tokenel
Feature activation+0.006
Top resid features:
told
Token told
Feature activation+2.728
Top resid features:
Belgian
Token Belgian
Feature activation+0.602
Top resid features:
magazine
Token magazine
Feature activation-0.171
Top resid features:
Hum
Token Hum
Feature activation+0.000
Top resid features:
o
Tokeno
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
?
Token?
Feature activation+0.002
Top resid features:
Sources
Token Sources
Feature activation-0.031
Top resid features:
within
Token within
Feature activation-0.057
Top resid features:
the
Token the
Feature activation-0.069
Top resid features:
league
Token league
Feature activation-0.048
Top resid features:
tell
Token tell
Feature activation+3.200
Top resid features:
Goal
Token Goal
Feature activation+0.153
Top resid features:
USA
Token USA
Feature activation+0.000
Top resid features:
there
Token there
Feature activation+0.000
Top resid features:
is
Token is
Feature activation+0.000
Top resid features:
scarce
Token scarce
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.008
Top resid features:
She
Token She
Feature activation+0.016
Top resid features:
recalls
Token recalls
Feature activation+0.050
Top resid features:
one
Token one
Feature activation+0.019
Top resid features:
professor
Token professor
Feature activation-0.076
Top resid features:
telling
Token telling
Feature activation+2.423
Top resid features:
her
Token her
Feature activation+0.123
Top resid features:
a
Token a
Feature activation+0.343
Top resid features:
graduate
Token graduate
Feature activation+0.158
Top resid features:
student
Token student
Feature activation+0.039
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
<|endoftext|>
Token<|endoftext|>
Feature activation+0.051
Top resid features:
Gibson
Token Gibson
Feature activation+0.009
Top resid features:
told
Token told
Feature activation+2.530
Top resid features:
a
Token a
Feature activation+0.289
Top resid features:
crowd
Token crowd
Feature activation+0.303
Top resid features:
at
Token at
Feature activation-0.080
Top resid features:
Terry
Token Terry
Feature activation+0.000
Top resid features:
Schr
Token Schr
Feature activation+0.000
Top resid features:
realization
Token realization
Feature activation+0.003
Top resid features:
hit
Token hit
Feature activation-0.006
Top resid features:
.
Token.
Feature activation+0.027
Top resid features:
Whenever
Token Whenever
Feature activation-0.011
Top resid features:
he
Token he
Feature activation-0.078
Top resid features:
told
Token told
Feature activation+2.492
Top resid features:
friends
Token friends
Feature activation+0.600
Top resid features:
and
Token and
Feature activation-0.017
Top resid features:
colleagues
Token colleagues
Feature activation+0.000
Top resid features:
about
Token about
Feature activation+0.000
Top resid features:
his
Token his
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.02

Head 1: 0.07

Head 2: 0.08

Head 3: 0.03

Head 4: 0.02

Head 5: 0.04

Head 6: 0.06

Head 7: 0.05

Head 8: 0.41

Head 9: 0.09

Head 10: 0.06

Head 11: 0.07

Positive logits

Leaks1.66

Offline1.58

Media1.44

NI1.38

HQ1.33

Ill1.26

Monitor1.24

BBC1.22

TV1.22

Line1.22

armac1.22

interviewer1.21

Hub1.19

JC1.17

anu1.17

Soc1.17

correspondent1.17

guiActiveUnfocused1.16

Interactive1.16

Courier1.16

Negative logits

largeDownload-1.50

BuyableInstoreAndOnline-1.49

thood-1.36

cannon-1.34

offensively-1.32

avorite-1.27

hunted-1.25

sugg-1.23

urated-1.20

redeem-1.20

cannons-1.17

oooooooo-1.17

��-1.17

subdu-1.16

reditary-1.16

defe-1.16

juven-1.16

morph-1.15

acknow-1.15

mell-1.14

INTERVAL 2.693 - 2.992
CONTAINS 0.000%

PS
TokenPS
Feature activation+0.000
website
Token website
Feature activation+0.000
also
Token also
Feature activation+0.000
tells
Token tells
Feature activation+0.000
potential
Token potential
Feature activation+1.457
visitors
Token visitors
Feature activation+2.887
about
Token about
Feature activation+0.000
how
Token how
Feature activation+0.000
unpredictable
Token unpredictable
Feature activation+0.000
the
Token the
Feature activation+0.000
weather
Token weather
Feature activation+0.000
he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
fellow
Token fellow
Feature activation+1.319
AM
Token AM
Feature activation+1.053
VA
TokenVA
Feature activation+0.000
members
Token members
Feature activation+2.897
.
Token.
Feature activation+0.265
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Despite
TokenDespite
Feature activation+0.000
the
Token the
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Obama
Token Obama
Feature activation+0.000
told
Token told
Feature activation+0.089
reporters
Token reporters
Feature activation+1.408
at
Token at
Feature activation+2.705
the
Token the
Feature activation+0.000
end
Token end
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
private
Token private
Feature activation+0.000
's
Token's
Feature activation+1.489
"
Token "
Feature activation+1.060
Science
TokenScience
Feature activation+0.000
Friday
Token Friday
Feature activation+0.000
"
Token"
Feature activation+1.202
host
Token host
Feature activation+2.992
I
Token I
Feature activation+0.000
ra
Tokenra
Feature activation+0.000
Flat
Token Flat
Feature activation+0.000
ow
Tokenow
Feature activation+0.000
on
Token on
Feature activation+0.000

INTERVAL 2.394 - 2.693
CONTAINS 0.001%

is
Token is
Feature activation+0.000
what
Token what
Feature activation+0.000
Sessions
Token Sessions
Feature activation+0.000
told
Token told
Feature activation+0.000
fellow
Token fellow
Feature activation+1.709
senators
Token senators
Feature activation+2.540
when
Token when
Feature activation+0.753
he
Token he
Feature activation+0.000
was
Token was
Feature activation+0.000
asked
Token asked
Feature activation+0.000
straightforward
Token straightforward
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Koh
Token Koh
Feature activation+0.000
told
Token told
Feature activation+0.000
a
Token a
Feature activation+0.332
Cleveland
Token Cleveland
Feature activation+0.245
audience
Token audience
Feature activation+2.414
.
Token.
Feature activation+0.181
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
K
TokenK
Feature activation+0.000
oh
Tokenoh
Feature activation+0.000
when
Token when
Feature activation+0.000
she
Token she
Feature activation+0.000
told
Token told
Feature activation+0.000
CNN
Token CNN
Feature activation+0.858
Gingrich
Token Gingrich
Feature activation+0.428
would
Token would
Feature activation+2.469
not
Token not
Feature activation+0.000
be
Token be
Feature activation+0.230
president
Token president
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.613
New
Token New
Feature activation+0.476
York
Token York
Feature activation+0.872
Police
Token Police
Feature activation+0.580
officer
Token officer
Feature activation+2.057
with
Token with
Feature activation+2.459
the
Token the
Feature activation+0.000
Manhattan
Token Manhattan
Feature activation+0.000
Special
Token Special
Feature activation+0.000
Victims
Token Victims
Feature activation+0.000
squad
Token squad
Feature activation+0.125
."
Token."
Feature activation+0.000
He
Token He
Feature activation+0.000
tells
Token tells
Feature activation+0.000
a
Token a
Feature activation+0.826
police
Token police
Feature activation+1.400
sergeant
Token sergeant
Feature activation+2.611
that
Token that
Feature activation+0.974
his
Token his
Feature activation+0.000
step
Token step
Feature activation+0.000
-
Token-
Feature activation+0.000
s
Tokens
Feature activation+0.000

INTERVAL 2.094 - 2.394
CONTAINS 0.003%

The
Token The
Feature activation+0.000
Obama
Token Obama
Feature activation+0.000
administration
Token administration
Feature activation+0.000
has
Token has
Feature activation+0.000
told
Token told
Feature activation+0.000
metropolitan
Token metropolitan
Feature activation+2.229
areas
Token areas
Feature activation+1.684
to
Token to
Feature activation+0.510
include
Token include
Feature activation+0.000
land
Token land
Feature activation+0.000
-
Token-
Feature activation+0.000
free
Token free
Feature activation+0.000
agent
Token agent
Feature activation+0.000
.
Token.
Feature activation+0.000
Sources
Token Sources
Feature activation+0.000
tell
Token tell
Feature activation+0.000
Goal
Token Goal
Feature activation+2.256
USA
Token USA
Feature activation+1.389
that
Token that
Feature activation+0.569
several
Token several
Feature activation+0.000
NASL
Token NASL
Feature activation+0.000
teams
Token teams
Feature activation+0.000
Duncan
Token Duncan
Feature activation+0.000
told
Token told
Feature activation+0.001
The
Token The
Feature activation+0.997
New
Token New
Feature activation+1.464
York
Token York
Feature activation+1.547
Times
Token Times
Feature activation+2.233
in
Token in
Feature activation+0.000
2012
Token 2012
Feature activation+0.000
,
Token,
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
same
Token same
Feature activation+0.000
thing
Token thing
Feature activation+0.000
I
Token I
Feature activation+0.000
tell
Token tell
Feature activation+0.000
anyone
Token anyone
Feature activation+1.381
who
Token who
Feature activation+2.180
invites
Token invites
Feature activation+0.000
me
Token me
Feature activation+0.000
to
Token to
Feature activation+0.000
discuss
Token discuss
Feature activation+0.000
the
Token the
Feature activation+0.000
had
Token had
Feature activation+0.000
been
Token been
Feature activation+0.000
telling
Token telling
Feature activation+0.000
myself
Token myself
Feature activation+1.086
stories
Token stories
Feature activation+0.000
like
Token like
Feature activation+2.306
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
Mar
TokenMar
Feature activation+0.000
riages
Tokenriages
Feature activation+0.000
never
Token never
Feature activation+0.000

INTERVAL 1.795 - 2.094
CONTAINS 0.004%

President
Token President
Feature activation+0.000
Vladimir
Token Vladimir
Feature activation+0.000
Putin
Token Putin
Feature activation+0.000
told
Token told
Feature activation+0.000
journalists
Token journalists
Feature activation+1.365
on
Token on
Feature activation+1.963
4
Token 4
Feature activation+0.000
March
Token March
Feature activation+0.000
:
Token:
Feature activation+0.000
"
Token "
Feature activation+0.000
We
TokenWe
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Mr
TokenMr
Feature activation+0.000
Trump
Token Trump
Feature activation+0.000
told
Token told
Feature activation+0.000
reporters
Token reporters
Feature activation+1.212
on
Token on
Feature activation+2.075
Tuesday
Token Tuesday
Feature activation+0.000
:
Token:
Feature activation+0.000
"
Token "
Feature activation+0.000
North
TokenNorth
Feature activation+0.000
Korea
Token Korea
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Police
TokenPolice
Feature activation+0.000
tell
Token tell
Feature activation+0.000
Alex
Token Alex
Feature activation+1.353
Jones
Token Jones
Feature activation+0.957
and
Token and
Feature activation+1.886
crew
Token crew
Feature activation+0.818
to
Token to
Feature activation+0.000
leave
Token leave
Feature activation+0.000
the
Token the
Feature activation+0.000
hotel
Token hotel
Feature activation+0.000
.
Token.
Feature activation+0.000
Bent
Token Bent
Feature activation+0.000
el
Tokenel
Feature activation+0.000
told
Token told
Feature activation+0.000
employees
Token employees
Feature activation+1.934
in
Token in
Feature activation+1.820
his
Token his
Feature activation+0.000
office
Token office
Feature activation+0.000
that
Token that
Feature activation+0.000
Secretary
Token Secretary
Feature activation+0.000
Clinton
Token Clinton
Feature activation+0.000
Sanford
Token Sanford
Feature activation+0.000
,
Token,
Feature activation+0.000
who
Token who
Feature activation+0.000
told
Token told
Feature activation+0.000
The
Token The
Feature activation+1.317
Post
Token Post
Feature activation+2.070
and
Token and
Feature activation+1.160
Courier
Token Courier
Feature activation+0.000
in
Token in
Feature activation+0.561
Charleston
Token Charleston
Feature activation+0.000
,
Token,
Feature activation+0.000

INTERVAL 1.496 - 1.795
CONTAINS 0.008%

humiliation
Token humiliation
Feature activation+0.000
of
Token of
Feature activation+0.000
Comey
Token Comey
Feature activation+0.000
telling
Token telling
Feature activation+0.000
a
Token a
Feature activation+1.464
congressional
Token congressional
Feature activation+1.634
committee
Token committee
Feature activation+2.688
that
Token that
Feature activation+0.693
she
Token she
Feature activation+0.000
had
Token had
Feature activation+0.000
been
Token been
Feature activation+0.000
a
Token a
Feature activation+0.000
family
Token family
Feature activation+0.000
friend
Token friend
Feature activation+0.000
told
Token told
Feature activation+0.000
the
Token the
Feature activation+0.032
press
Token press
Feature activation+1.644
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Those
TokenThose
Feature activation+0.000
who
Token who
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
the
Token the
Feature activation+0.520
newspaper
Token newspaper
Feature activation+1.763
A
Token A
Feature activation+0.217
ften
Tokenften
Feature activation+0.000
post
Tokenpost
Feature activation+1.036
en
Tokenen
Feature activation+0.000
.
Token.
Feature activation+0.588
ade
Tokenade
Feature activation+0.000
y
Tokeny
Feature activation+0.000
Jr
Token Jr
Feature activation+0.000
.
Token.
Feature activation+0.000
told
Token told
Feature activation+0.196
Ey
Token Ey
Feature activation+1.764
ewitness
Tokenewitness
Feature activation+1.277
News
Token News
Feature activation+0.639
on
Token on
Feature activation+1.185
Thursday
Token Thursday
Feature activation+0.000
.
Token.
Feature activation+0.000
ahn
Tokenahn
Feature activation+0.000
told
Token told
Feature activation+0.000
a
Token a
Feature activation+0.799
Port
Token Port
Feature activation+0.531
Authority
Token Authority
Feature activation+0.151
detective
Token detective
Feature activation+1.795
,
Token,
Feature activation+0.428
"
Token "
Feature activation+0.000
I
TokenI
Feature activation+0.000
have
Token have
Feature activation+0.000
diplomatic
Token diplomatic
Feature activation+0.000

INTERVAL 1.197 - 1.496
CONTAINS 0.010%

told
Token told
Feature activation+0.000
Mal
Token Mal
Feature activation+0.944
ay
Tokenay
Feature activation+0.000
Mail
Token Mail
Feature activation+1.370
Online
Token Online
Feature activation+1.713
when
Token when
Feature activation+1.242
contacted
Token contacted
Feature activation+0.000
,
Token,
Feature activation+0.000
using
Token using
Feature activation+0.000
the
Token the
Feature activation+0.000
Mal
Token Mal
Feature activation+0.000
(
Token (
Feature activation+0.104
female
Tokenfemale
Feature activation+0.546
)
Token)
Feature activation+1.039
stream
Token stream
Feature activation+2.518
er
Tokener
Feature activation+2.069
âĢ
Token âĢ
Feature activation+1.488
ĺ
Tokenĺ
Feature activation+0.000
Eat
TokenEat
Feature activation+0.000
my
Token my
Feature activation+0.000
bottom
Token bottom
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Media
Token Media
Feature activation+0.000
was
Token was
Feature activation+0.000
told
Token told
Feature activation+0.000
Opt
Token Opt
Feature activation+0.192
us
Tokenus
Feature activation+1.247
had
Token had
Feature activation+1.212
contacted
Token contacted
Feature activation+0.000
departing
Token departing
Feature activation+0.000
NRL
Token NRL
Feature activation+0.000
chief
Token chief
Feature activation+0.000
executive
Token executive
Feature activation+0.000
Police
Token Police
Feature activation+0.000
,
Token,
Feature activation+0.000
tells
Token tells
Feature activation+0.000
Syria
Token Syria
Feature activation+0.941
Direct
Token Direct
Feature activation+2.166
âĢ
TokenâĢ
Feature activation+1.290
Ļ
TokenĻ
Feature activation+0.081
s
Tokens
Feature activation+0.905
W
Token W
Feature activation+0.000
ale
Tokenale
Feature activation+0.000
ed
Tokened
Feature activation+0.000
senior
Token senior
Feature activation+0.000
official
Token official
Feature activation+0.000
told
Token told
Feature activation+0.906
Gulf
Token Gulf
Feature activation+1.530
News
Token News
Feature activation+0.000
on
Token on
Feature activation+1.234
Thursday
Token Thursday
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000

INTERVAL 0.898 - 1.197
CONTAINS 0.014%

he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
MMA
Token MMA
Feature activation+0.666
j
Tokenj
Feature activation+0.000
unk
Tokenunk
Feature activation+0.000
ie
Tokenie
Feature activation+1.029
he
Token he
Feature activation+0.211
had
Token had
Feature activation+0.262
a
Token a
Feature activation+0.000
striking
Token striking
Feature activation+0.000
advantage
Token advantage
Feature activation+0.000
Ministry
Token Ministry
Feature activation+0.000
spokesman
Token spokesman
Feature activation+0.000
Lu
Token Lu
Feature activation+0.000
Kang
Token Kang
Feature activation+0.000
told
Token told
Feature activation+0.321
reporters
Token reporters
Feature activation+1.107
at
Token at
Feature activation+1.557
a
Token a
Feature activation+0.000
regular
Token regular
Feature activation+0.000
briefing
Token briefing
Feature activation+0.144
:
Token:
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Roberts
TokenRoberts
Feature activation+0.000
told
Token told
Feature activation+0.000
the
Token the
Feature activation+0.009
Senate
Token Senate
Feature activation+0.947
he
Token he
Feature activation+0.000
was
Token was
Feature activation+0.000
keen
Token keen
Feature activation+0.000
for
Token for
Feature activation+0.000
the
Token the
Feature activation+0.000
hhh
Tokenhhh
Feature activation+0.000
,
Token,
Feature activation+0.000
don
Token don
Feature activation+0.000
't
Token't
Feature activation+0.000
tell
Token tell
Feature activation+0.000
anybody
Token anybody
Feature activation+0.988
)
Token)
Feature activation+0.140
"
Token "
Feature activation+0.000
v
Tokenv
Feature activation+0.053
oter
Tokenoter
Feature activation+0.032
fraud
Token fraud
Feature activation+0.000
won
Token won
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
tell
Token tell
Feature activation+0.000
a
Token a
Feature activation+1.060
story
Token story
Feature activation+0.000
of
Token of
Feature activation+0.000
a
Token a
Feature activation+0.000
stronger
Token stronger
Feature activation+0.000
recovery
Token recovery
Feature activation+0.000

INTERVAL 0.598 - 0.898
CONTAINS 0.021%

MORE
Token MORE
Feature activation+0.000
for
Token for
Feature activation+0.000
a
Token a
Feature activation+0.000
tweet
Token tweet
Feature activation+0.000
warning
Token warning
Feature activation+0.000
former
Token former
Feature activation+0.758
FBI
Token FBI
Feature activation+0.000
Director
Token Director
Feature activation+0.032
James
Token James
Feature activation+0.000
Comey
Token Comey
Feature activation+0.009
about
Token about
Feature activation+0.000
head
Tokenhead
Feature activation+0.000
Dale
Token Dale
Feature activation+0.000
Ste
Token Ste
Feature activation+0.000
yn
Tokenyn
Feature activation+0.000
told
Token told
Feature activation+0.337
cricket
Token cricket
Feature activation+0.859
.
Token.
Feature activation+0.441
com
Tokencom
Feature activation+0.706
.
Token.
Feature activation+0.039
au
Tokenau
Feature activation+0.104
earlier
Token earlier
Feature activation+0.264
Police
Token Police
Feature activation+0.000
has
Token has
Feature activation+0.000
told
Token told
Feature activation+0.000
Human
Token Human
Feature activation+0.656
Services
Token Services
Feature activation+1.149
Minister
Token Minister
Feature activation+0.734
Alan
Token Alan
Feature activation+0.002
T
Token T
Feature activation+0.000
udge
Tokenudge
Feature activation+0.000
it
Token it
Feature activation+0.315
will
Token will
Feature activation+0.000
,
Token,
Feature activation+0.000
he
Token he
Feature activation+0.000
told
Token told
Feature activation+0.000
a
Token a
Feature activation+0.613
New
Token New
Feature activation+0.476
York
Token York
Feature activation+0.872
Police
Token Police
Feature activation+0.580
officer
Token officer
Feature activation+2.057
with
Token with
Feature activation+2.459
the
Token the
Feature activation+0.000
Manhattan
Token Manhattan
Feature activation+0.000
-
Token-
Feature activation+0.000
alcohol
Tokenalcohol
Feature activation+0.000
content
Token content
Feature activation+0.000
and
Token and
Feature activation+0.000
told
Token told
Feature activation+0.000
officers
Token officers
Feature activation+0.893
he
Token he
Feature activation+0.000
drank
Token drank
Feature activation+0.000
two
Token two
Feature activation+0.000
beers
Token beers
Feature activation+0.000
.
Token.
Feature activation+0.000

INTERVAL 0.299 - 0.598
CONTAINS 0.030%

cool
Token cool
Feature activation+0.000
.
Token.
Feature activation+0.000
I
Token I
Feature activation+0.000
told
Token told
Feature activation+0.000
fox
Token fox
Feature activation+1.410
y
Tokeny
Feature activation+0.438
I
Token I
Feature activation+0.539
lived
Token lived
Feature activation+0.595
not
Token not
Feature activation+0.000
far
Token far
Feature activation+0.000
and
Token and
Feature activation+0.000
Ben
Token Ben
Feature activation+0.000
io
Tokenio
Feature activation+0.000
ff
Tokenff
Feature activation+0.000
told
Token told
Feature activation+0.000
Entertainment
Token Entertainment
Feature activation+1.036
Weekly
Token Weekly
Feature activation+0.555
.
Token.
Feature activation+0.204
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
For
TokenFor
Feature activation+0.140
budgetary
Token budgetary
Feature activation+0.000
m
Tokenm
Feature activation+0.000
told
Token told
Feature activation+0.000
by
Token by
Feature activation+0.740
(
Token (
Feature activation+0.000
an
Tokenan
Feature activation+0.000
)
Token)
Feature activation+0.381
adviser
Token adviser
Feature activation+0.000
in
Token in
Feature activation+0.125
the
Token the
Feature activation+0.000
AS
Token AS
Feature activation+0.000
UC
TokenUC
Feature activation+0.000
Anthony
TokenAnthony
Feature activation+0.000
Hamilton
Token Hamilton
Feature activation+0.000
had
Token had
Feature activation+0.000
earlier
Token earlier
Feature activation+0.000
told
Token told
Feature activation+0.000
the
Token the
Feature activation+0.438
trial
Token trial
Feature activation+0.182
how
Token how
Feature activation+0.000
he
Token he
Feature activation+0.000
stopped
Token stopped
Feature activation+0.000
being
Token being
Feature activation+0.000
Comey
Token Comey
Feature activation+0.000
told
Token told
Feature activation+0.073
the
Token the
Feature activation+0.000
Senate
Token Senate
Feature activation+0.924
Intelligence
Token Intelligence
Feature activation+0.000
Committee
Token Committee
Feature activation+0.498
in
Token in
Feature activation+0.000
testimony
Token testimony
Feature activation+0.000
earlier
Token earlier
Feature activation+0.000
this
Token this
Feature activation+0.000
month
Token month
Feature activation+0.000

INTERVAL 0.000 - 0.299
CONTAINS 99.910%

rench
Tokenrench
Feature activation+0.000
ard
Tokenard
Feature activation+0.000
,
Token,
Feature activation+0.000
had
Token had
Feature activation+0.000
commanded
Token commanded
Feature activation+0.000
aircraft
Token aircraft
Feature activation+0.000
during
Token during
Feature activation+0.000
World
Token World
Feature activation+0.000
War
Token War
Feature activation+0.000
I
Token I
Feature activation+0.000
.
Token.
Feature activation+0.000
mayor
Token mayor
Feature activation+0.000
had
Token had
Feature activation+0.000
given
Token given
Feature activation+0.000
the
Token the
Feature activation+0.000
celebrations
Token celebrations
Feature activation+0.000
the
Token the
Feature activation+0.000
title
Token title
Feature activation+0.000
Paris
Token Paris
Feature activation+0.000
Se
Token Se
Feature activation+0.000
Lib
Token Lib
Feature activation+0.000
ere
Tokenere
Feature activation+0.000
32
Token 32
Feature activation+0.000
62
Token 62
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Which
TokenWhich
Feature activation+0.000
shows
Token shows
Feature activation+0.000
us
Token us
Feature activation+0.000
that
Token that
Feature activation+0.000
accum
Token accum
Feature activation+0.000
ulator
Tokenulator
Feature activation+0.000
value
Token value
Feature activation+0.000
participated
Token participated
Feature activation+0.000
in
Token in
Feature activation+0.000
talks
Token talks
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Reb
TokenReb
Feature activation+0.000
els
Tokenels
Feature activation+0.000
participating
Token participating
Feature activation+0.000
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
gear
Tokengear
Feature activation+0.000
get
Token get
Feature activation+0.000
up
Token up
Feature activation+0.000
,
Token,
Feature activation+0.000
and
Token and
Feature activation+0.000
the
Token the
Feature activation+0.000
two
Token two
Feature activation+0.000
exited
Token exited
Feature activation+0.000
the
Token the
Feature activation+0.000
now
Token now
Feature activation+0.000
barren
Token barren
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 H3.8: (feature 17598)

TOP ACTIVATIONS
MAX = 2.807

âĢ
Token âĢ
Feature activation+0.000
ĺ
Tokenĺ
Feature activation+0.000
Top
TokenTop
Feature activation+0.000
five
Token five
Feature activation+0.000
things
Token things
Feature activation+0.000
physicians
Token physicians
Feature activation+2.807
and
Token and
Feature activation+2.540
patients
Token patients
Feature activation+0.000
should
Token should
Feature activation+0.000
question
Token question
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
were
Token were
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
thing
Token thing
Feature activation+0.000
men
Token men
Feature activation+2.559
and
Token and
Feature activation+2.421
women
Token women
Feature activation+0.000
brought
Token brought
Feature activation+0.000
into
Token into
Feature activation+0.000
the
Token the
Feature activation+0.000
ĺ
Tokenĺ
Feature activation+0.000
Top
TokenTop
Feature activation+0.000
five
Token five
Feature activation+0.000
things
Token things
Feature activation+0.000
physicians
Token physicians
Feature activation+2.807
and
Token and
Feature activation+2.540
patients
Token patients
Feature activation+0.000
should
Token should
Feature activation+0.000
question
Token question
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
stupid
Token stupid
Feature activation+0.000
est
Tokenest
Feature activation+0.000
thing
Token thing
Feature activation+0.000
human
Token human
Feature activation+2.423
beings
Token beings
Feature activation+2.158
do
Token do
Feature activation+0.257
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
thing
Token thing
Feature activation+0.000
men
Token men
Feature activation+2.559
and
Token and
Feature activation+2.421
women
Token women
Feature activation+0.000
brought
Token brought
Feature activation+0.000
into
Token into
Feature activation+0.000
the
Token the
Feature activation+0.000
world
Token world
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
First
TokenFirst
Feature activation+0.000
things
Token things
Feature activation+0.000
first
Token first
Feature activation+2.413
,
Token,
Feature activation+0.000
we
Tokenwe
Feature activation+0.000
wa
Token wa
Feature activation+0.000
n
Token n
Feature activation+0.000
ted
Tokented
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
cliché
Token cliché
Feature activation+0.000
thing
Token thing
Feature activation+0.000
one
Token one
Feature activation+2.265
could
Token could
Feature activation+0.000
say
Token say
Feature activation+0.000
here
Token here
Feature activation+0.000
.
Token.
Feature activation+0.000
Everyone
Token Everyone
Feature activation+0.000
the
Token the
Feature activation+0.000
stupid
Token stupid
Feature activation+0.000
est
Tokenest
Feature activation+0.000
thing
Token thing
Feature activation+0.000
human
Token human
Feature activation+2.423
beings
Token beings
Feature activation+2.158
do
Token do
Feature activation+0.257
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Almost
Token Almost
Feature activation+0.000
maybe
Token maybe
Feature activation+0.000
some
Token some
Feature activation+0.000
function
Token function
Feature activation+0.000
.
Token.
Feature activation+0.000
Things
Token Things
Feature activation+0.000
Organ
Token Organ
Feature activation+2.032
ized
Tokenized
Feature activation+0.181
Ne
Token Ne
Feature activation+0.172
at
Tokenat
Feature activation+0.000
ly
Tokenly
Feature activation+0.000
(
Token (
Feature activation+0.000
Here
Token Here
Feature activation+0.000
are
Token are
Feature activation+0.000
four
Token four
Feature activation+0.000
things
Token things
Feature activation+0.000
NATO
Token NATO
Feature activation+1.553
needs
Token needs
Feature activation+1.936
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
when
Token when
Feature activation+0.000
its
Token its
Feature activation+0.000
leaders
Token leaders
Feature activation+0.000
kinda
Token kinda
Feature activation+0.000
fascinated
Token fascinated
Feature activation+0.000
with
Token with
Feature activation+0.000
all
Token all
Feature activation+0.000
things
Token things
Feature activation+0.000
British
Token British
Feature activation+1.713
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
only
Token only
Feature activation+0.000
things
Token things
Feature activation+0.000
we
Token we
Feature activation+0.000
love
Token love
Feature activation+0.000
with
Token with
Feature activation+0.000
all
Token all
Feature activation+0.000
things
Token things
Feature activation+0.000
Gall
Token Gall
Feature activation+0.541
ic
Tokenic
Feature activation+1.679
,
Token,
Feature activation+0.013
and
Token and
Feature activation+0.000
started
Token started
Feature activation+0.000
taking
Token taking
Feature activation+0.000
evening
Token evening
Feature activation+0.000
This
TokenThis
Feature activation+0.000
was
Token was
Feature activation+0.000
the
Token the
Feature activation+0.000
last
Token last
Feature activation+0.000
thing
Token thing
Feature activation+0.000
a
Token a
Feature activation+1.670
new
Token new
Feature activation+0.962
President
Token President
Feature activation+0.000
dealing
Token dealing
Feature activation+0.000
with
Token with
Feature activation+0.000
a
Token a
Feature activation+0.000
.
Token.
Feature activation+0.000
Here
Token Here
Feature activation+0.000
are
Token are
Feature activation+0.000
four
Token four
Feature activation+0.000
things
Token things
Feature activation+0.000
NATO
Token NATO
Feature activation+1.553
needs
Token needs
Feature activation+1.936
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
when
Token when
Feature activation+0.000
its
Token its
Feature activation+0.000
case
Token case
Feature activation+0.000
of
Token of
Feature activation+0.000
taking
Token taking
Feature activation+0.000
things
Token things
Feature activation+0.000
one
Token one
Feature activation+1.001
step
Token step
Feature activation+1.541
at
Token at
Feature activation+0.000
time
Token time
Feature activation+0.000
-
Token -
Feature activation+0.000
first
Token first
Feature activation+0.000
to
Token to
Feature activation+0.000
too
Token too
Feature activation+0.000
).
Token).
Feature activation+0.000
The
Token The
Feature activation+0.000
whole
Token whole
Feature activation+0.000
thing
Token thing
Feature activation+0.000
snow
Token snow
Feature activation+1.519
balls
Tokenballs
Feature activation+0.938
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Hist
TokenHist
Feature activation+0.000
,
Token,
Feature activation+0.000
hey
Token hey
Feature activation+0.000
,
Token,
Feature activation+0.000
one
Token one
Feature activation+0.000
thing
Token thing
Feature activation+0.000
the
Token the
Feature activation+1.473
Marvel
Token Marvel
Feature activation+0.402
Cinem
Token Cinem
Feature activation+0.103
atic
Tokenatic
Feature activation+0.197
Universe
Token Universe
Feature activation+0.291
does
Token does
Feature activation+0.000
certainly
Token certainly
Feature activation+0.000
not
Token not
Feature activation+0.000
the
Token the
Feature activation+0.000
worst
Token worst
Feature activation+0.000
thing
Token thing
Feature activation+0.000
Ron
Token Ron
Feature activation+1.465
Atkinson
Token Atkinson
Feature activation+0.006
ever
Token ever
Feature activation+0.000
said
Token said
Feature activation+0.000
,
Token,
Feature activation+0.000
but
Token but
Feature activation+0.000
?
Token?
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
One
TokenOne
Feature activation+0.000
thing
Token thing
Feature activation+0.000
is
Token is
Feature activation+1.461
clear
Token clear
Feature activation+0.782
:
Token:
Feature activation+0.000
they
Token they
Feature activation+0.000
won
Token won
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
admits
Token admits
Feature activation+0.000
graduates
Token graduates
Feature activation+0.000
could
Token could
Feature activation+0.000
find
Token find
Feature activation+0.000
things
Token things
Feature activation+0.000
difficult
Token difficult
Feature activation+1.436
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
current
Token current
Feature activation+0.000
financial
Token financial
Feature activation+0.000
climate
Token climate
Feature activation+0.000

Top DFA by src position
MAX = 4.917

the
Token the
Feature activation-0.032
Top resid features:
âĢ
Token âĢ
Feature activation-0.019
Top resid features:
ĺ
Tokenĺ
Feature activation-0.048
Top resid features:
Top
TokenTop
Feature activation-0.061
Top resid features:
five
Token five
Feature activation+0.000
Top resid features:
things
Token things
Feature activation+4.917
Top resid features:
physicians
Token physicians
Feature activation-0.144
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
patients
Token patients
Feature activation+0.000
Top resid features:
should
Token should
Feature activation+0.000
Top resid features:
question
Token question
Feature activation+0.000
Top resid features:
babies
Token babies
Feature activation+0.004
Top resid features:
were
Token were
Feature activation+0.001
Top resid features:
the
Token the
Feature activation+0.059
Top resid features:
most
Token most
Feature activation-0.057
Top resid features:
important
Token important
Feature activation+0.113
Top resid features:
thing
Token thing
Feature activation+3.588
Top resid features:
men
Token men
Feature activation+0.162
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
women
Token women
Feature activation+0.000
Top resid features:
brought
Token brought
Feature activation+0.000
Top resid features:
into
Token into
Feature activation+0.000
Top resid features:
the
Token the
Feature activation-0.031
Top resid features:
âĢ
Token âĢ
Feature activation-0.038
Top resid features:
ĺ
Tokenĺ
Feature activation-0.092
Top resid features:
Top
TokenTop
Feature activation-0.005
Top resid features:
five
Token five
Feature activation-0.030
Top resid features:
things
Token things
Feature activation+3.998
Top resid features:
physicians
Token physicians
Feature activation+0.164
Top resid features:
and
Token and
Feature activation+0.019
Top resid features:
patients
Token patients
Feature activation+0.000
Top resid features:
should
Token should
Feature activation+0.000
Top resid features:
question
Token question
Feature activation+0.000
Top resid features:
War
TokenWar
Feature activation+0.074
Top resid features:
is
Token is
Feature activation+0.036
Top resid features:
the
Token the
Feature activation+0.094
Top resid features:
stupid
Token stupid
Feature activation+0.028
Top resid features:
est
Tokenest
Feature activation-0.014
Top resid features:
thing
Token thing
Feature activation+3.543
Top resid features:
human
Token human
Feature activation+0.095
Top resid features:
beings
Token beings
Feature activation+0.000
Top resid features:
do
Token do
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
babies
Token babies
Feature activation-0.001
Top resid features:
were
Token were
Feature activation-0.003
Top resid features:
the
Token the
Feature activation+0.048
Top resid features:
most
Token most
Feature activation-0.053
Top resid features:
important
Token important
Feature activation+0.026
Top resid features:
thing
Token thing
Feature activation+3.458
Top resid features:
men
Token men
Feature activation+0.205
Top resid features:
and
Token and
Feature activation+0.095
Top resid features:
women
Token women
Feature activation+0.000
Top resid features:
brought
Token brought
Feature activation+0.000
Top resid features:
into
Token into
Feature activation+0.000
Top resid features:
him
Token him
Feature activation-0.007
Top resid features:
.
Token.
Feature activation+0.043
Top resid features:
Ċ
TokenĊ
Feature activation-0.005
Top resid features:
Ċ
TokenĊ
Feature activation-0.011
Top resid features:
First
TokenFirst
Feature activation+0.072
Top resid features:
things
Token things
Feature activation+3.590
Top resid features:
first
Token first
Feature activation+0.068
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
we
Tokenwe
Feature activation+0.000
Top resid features:
wa
Token wa
Feature activation+0.000
Top resid features:
n
Token n
Feature activation+0.000
Top resid features:
it
Token it
Feature activation-0.032
Top resid features:
is
Token is
Feature activation-0.017
Top resid features:
the
Token the
Feature activation+0.059
Top resid features:
most
Token most
Feature activation-0.050
Top resid features:
cliché
Token cliché
Feature activation+0.005
Top resid features:
thing
Token thing
Feature activation+3.573
Top resid features:
one
Token one
Feature activation+0.116
Top resid features:
could
Token could
Feature activation+0.000
Top resid features:
say
Token say
Feature activation+0.000
Top resid features:
here
Token here
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
War
TokenWar
Feature activation+0.038
Top resid features:
is
Token is
Feature activation+0.051
Top resid features:
the
Token the
Feature activation+0.055
Top resid features:
stupid
Token stupid
Feature activation+0.025
Top resid features:
est
Tokenest
Feature activation-0.064
Top resid features:
thing
Token thing
Feature activation+2.993
Top resid features:
human
Token human
Feature activation+0.222
Top resid features:
beings
Token beings
Feature activation+0.160
Top resid features:
do
Token do
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.000
Top resid features:
even
Token even
Feature activation-0.001
Top resid features:
maybe
Token maybe
Feature activation-0.008
Top resid features:
some
Token some
Feature activation+0.001
Top resid features:
function
Token function
Feature activation-0.023
Top resid features:
.
Token.
Feature activation+0.072
Top resid features:
Things
Token Things
Feature activation+3.328
Top resid features:
Organ
Token Organ
Feature activation+0.056
Top resid features:
ized
Tokenized
Feature activation+0.000
Top resid features:
Ne
Token Ne
Feature activation+0.000
Top resid features:
at
Tokenat
Feature activation+0.000
Top resid features:
ly
Tokenly
Feature activation+0.000
Top resid features:
response
Token response
Feature activation-0.014
Top resid features:
.
Token.
Feature activation+0.027
Top resid features:
Here
Token Here
Feature activation-0.072
Top resid features:
are
Token are
Feature activation-0.053
Top resid features:
four
Token four
Feature activation+0.178
Top resid features:
things
Token things
Feature activation+2.791
Top resid features:
NATO
Token NATO
Feature activation+0.318
Top resid features:
needs
Token needs
Feature activation+0.015
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
do
Token do
Feature activation+0.000
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
'm
Token'm
Feature activation-0.006
Top resid features:
kinda
Token kinda
Feature activation+0.004
Top resid features:
fascinated
Token fascinated
Feature activation-0.023
Top resid features:
with
Token with
Feature activation+0.033
Top resid features:
all
Token all
Feature activation-0.030
Top resid features:
things
Token things
Feature activation+2.660
Top resid features:
British
Token British
Feature activation+0.153
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
The
Token The
Feature activation+0.000
Top resid features:
only
Token only
Feature activation+0.000
Top resid features:
things
Token things
Feature activation+0.000
Top resid features:
fell
Token fell
Feature activation+0.003
Top resid features:
in
Token in
Feature activation+0.002
Top resid features:
love
Token love
Feature activation+0.014
Top resid features:
with
Token with
Feature activation+0.013
Top resid features:
all
Token all
Feature activation+0.010
Top resid features:
things
Token things
Feature activation+2.615
Top resid features:
Gall
Token Gall
Feature activation+0.145
Top resid features:
ic
Tokenic
Feature activation+0.138
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
and
Token and
Feature activation+0.000
Top resid features:
started
Token started
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation-0.006
Top resid features:
This
TokenThis
Feature activation-0.059
Top resid features:
was
Token was
Feature activation+0.017
Top resid features:
the
Token the
Feature activation+0.067
Top resid features:
last
Token last
Feature activation+0.084
Top resid features:
thing
Token thing
Feature activation+2.829
Top resid features:
a
Token a
Feature activation+0.046
Top resid features:
new
Token new
Feature activation+0.000
Top resid features:
President
Token President
Feature activation+0.000
Top resid features:
dealing
Token dealing
Feature activation+0.000
Top resid features:
with
Token with
Feature activation+0.000
Top resid features:
response
Token response
Feature activation-0.019
Top resid features:
.
Token.
Feature activation+0.036
Top resid features:
Here
Token Here
Feature activation-0.019
Top resid features:
are
Token are
Feature activation-0.096
Top resid features:
four
Token four
Feature activation+0.111
Top resid features:
things
Token things
Feature activation+2.612
Top resid features:
NATO
Token NATO
Feature activation+0.161
Top resid features:
needs
Token needs
Feature activation+0.000
Top resid features:
to
Token to
Feature activation+0.000
Top resid features:
do
Token do
Feature activation+0.000
Top resid features:
when
Token when
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.003
Top resid features:
a
Token a
Feature activation+0.011
Top resid features:
case
Token case
Feature activation-0.019
Top resid features:
of
Token of
Feature activation+0.057
Top resid features:
taking
Token taking
Feature activation+0.001
Top resid features:
things
Token things
Feature activation+2.532
Top resid features:
one
Token one
Feature activation+0.090
Top resid features:
step
Token step
Feature activation+0.132
Top resid features:
at
Token at
Feature activation+0.000
Top resid features:
time
Token time
Feature activation+0.000
Top resid features:
-
Token -
Feature activation+0.000
Top resid features:
,
Token,
Feature activation-0.004
Top resid features:
too
Token too
Feature activation-0.005
Top resid features:
).
Token).
Feature activation+0.070
Top resid features:
The
Token The
Feature activation+0.057
Top resid features:
whole
Token whole
Feature activation+0.112
Top resid features:
thing
Token thing
Feature activation+2.476
Top resid features:
snow
Token snow
Feature activation+0.144
Top resid features:
balls
Tokenballs
Feature activation+0.000
Top resid features:
.
Token.
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
Ċ
TokenĊ
Feature activation+0.000
Top resid features:
But
TokenBut
Feature activation+0.016
Top resid features:
,
Token,
Feature activation+0.008
Top resid features:
hey
Token hey
Feature activation-0.056
Top resid features:
,
Token,
Feature activation+0.006
Top resid features:
one
Token one
Feature activation-0.011
Top resid features:
thing
Token thing
Feature activation+2.544
Top resid features:
the
Token the
Feature activation+0.267
Top resid features:
Marvel
Token Marvel
Feature activation+0.000
Top resid features:
Cinem
Token Cinem
Feature activation+0.000
Top resid features:
atic
Tokenatic
Feature activation+0.000
Top resid features:
Universe
Token Universe
Feature activation+0.000
Top resid features:
s
Tokens
Feature activation+0.018
Top resid features:
certainly
Token certainly
Feature activation-0.004
Top resid features:
not
Token not
Feature activation+0.008
Top resid features:
the
Token the
Feature activation+0.082
Top resid features:
worst
Token worst
Feature activation+0.198
Top resid features:
thing
Token thing
Feature activation+2.057
Top resid features:
Ron
Token Ron
Feature activation+0.257
Top resid features:
Atkinson
Token Atkinson
Feature activation+0.000
Top resid features:
ever
Token ever
Feature activation+0.000
Top resid features:
said
Token said
Feature activation+0.000
Top resid features:
,
Token,
Feature activation+0.000
Top resid features:
case
Token case
Feature activation+0.000
Top resid features:
?
Token?
Feature activation+0.035
Top resid features:
Ċ
TokenĊ
Feature activation+0.032
Top resid features:
Ċ
TokenĊ
Feature activation+0.024
Top resid features:
One
TokenOne
Feature activation-0.058
Top resid features:
thing
Token thing
Feature activation+2.397
Top resid features:
is
Token is
Feature activation+0.351
Top resid features:
clear
Token clear
Feature activation+0.000
Top resid features:
:
Token:
Feature activation+0.000
Top resid features:
they
Token they
Feature activation+0.000
Top resid features:
won
Token won
Feature activation+0.000
Top resid features:
Skills
Token Skills
Feature activation-0.009
Top resid features:
admits
Token admits
Feature activation-0.022
Top resid features:
graduates
Token graduates
Feature activation+0.002
Top resid features:
could
Token could
Feature activation-0.087
Top resid features:
find
Token find
Feature activation-0.152
Top resid features:
things
Token things
Feature activation+2.882
Top resid features:
difficult
Token difficult
Feature activation+0.196
Top resid features:
in
Token in
Feature activation+0.000
Top resid features:
the
Token the
Feature activation+0.000
Top resid features:
current
Token current
Feature activation+0.000
Top resid features:
financial
Token financial
Feature activation+0.000
Top resid features:

Decoder Weights Distribution

Head 0: 0.03

Head 1: 0.07

Head 2: 0.07

Head 3: 0.03

Head 4: 0.03

Head 5: 0.03

Head 6: 0.06

Head 7: 0.14

Head 8: 0.41

Head 9: 0.04

Head 10: 0.05

Head 11: 0.05

Positive logits

orously1.28

does1.12

Courts1.12

Depend1.10

Fro1.09

blogspot1.09

married1.05

Reviewer1.05

Latter1.04

});1.02

bald1.01

mite1.01

aky0.99

strangers0.99

sulf0.97

...]0.97

hither0.97

gets0.96

TABLE0.96

fuss0.96

Negative logits

raints-1.11

auri-1.08

BILITIES-1.08

unctions-1.08

Atlantis-1.06

ourses-1.04

introduction-1.04

ーン-1.02

Arcane-1.02

Minotaur-1.01

eanor-1.01

loe-1.00

Armory-0.99

instance-0.98

ourse-0.97

Valkyrie-0.96

ttes-0.96

alion-0.96

ysical-0.95

uphem-0.95

INTERVAL 2.526 - 2.807
CONTAINS 0.000%

âĢ
Token âĢ
Feature activation+0.000
ĺ
Tokenĺ
Feature activation+0.000
Top
TokenTop
Feature activation+0.000
five
Token five
Feature activation+0.000
things
Token things
Feature activation+0.000
physicians
Token physicians
Feature activation+2.807
and
Token and
Feature activation+2.540
patients
Token patients
Feature activation+0.000
should
Token should
Feature activation+0.000
question
Token question
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
ĺ
Tokenĺ
Feature activation+0.000
Top
TokenTop
Feature activation+0.000
five
Token five
Feature activation+0.000
things
Token things
Feature activation+0.000
physicians
Token physicians
Feature activation+2.807
and
Token and
Feature activation+2.540
patients
Token patients
Feature activation+0.000
should
Token should
Feature activation+0.000
question
Token question
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
were
Token were
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
thing
Token thing
Feature activation+0.000
men
Token men
Feature activation+2.559
and
Token and
Feature activation+2.421
women
Token women
Feature activation+0.000
brought
Token brought
Feature activation+0.000
into
Token into
Feature activation+0.000
the
Token the
Feature activation+0.000

INTERVAL 2.245 - 2.526
CONTAINS 0.000%

is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
stupid
Token stupid
Feature activation+0.000
est
Tokenest
Feature activation+0.000
thing
Token thing
Feature activation+0.000
human
Token human
Feature activation+2.423
beings
Token beings
Feature activation+2.158
do
Token do
Feature activation+0.257
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
is
Token is
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
cliché
Token cliché
Feature activation+0.000
thing
Token thing
Feature activation+0.000
one
Token one
Feature activation+2.265
could
Token could
Feature activation+0.000
say
Token say
Feature activation+0.000
here
Token here
Feature activation+0.000
.
Token.
Feature activation+0.000
Everyone
Token Everyone
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
First
TokenFirst
Feature activation+0.000
things
Token things
Feature activation+0.000
first
Token first
Feature activation+2.413
,
Token,
Feature activation+0.000
we
Tokenwe
Feature activation+0.000
wa
Token wa
Feature activation+0.000
n
Token n
Feature activation+0.000
ted
Tokented
Feature activation+0.000
the
Token the
Feature activation+0.000
most
Token most
Feature activation+0.000
important
Token important
Feature activation+0.000
thing
Token thing
Feature activation+0.000
men
Token men
Feature activation+2.559
and
Token and
Feature activation+2.421
women
Token women
Feature activation+0.000
brought
Token brought
Feature activation+0.000
into
Token into
Feature activation+0.000
the
Token the
Feature activation+0.000
world
Token world
Feature activation+0.000

INTERVAL 1.965 - 2.245
CONTAINS 0.000%

the
Token the
Feature activation+0.000
stupid
Token stupid
Feature activation+0.000
est
Tokenest
Feature activation+0.000
thing
Token thing
Feature activation+0.000
human
Token human
Feature activation+2.423
beings
Token beings
Feature activation+2.158
do
Token do
Feature activation+0.257
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000
Almost
Token Almost
Feature activation+0.000
maybe
Token maybe
Feature activation+0.000
some
Token some
Feature activation+0.000
function
Token function
Feature activation+0.000
.
Token.
Feature activation+0.000
Things
Token Things
Feature activation+0.000
Organ
Token Organ
Feature activation+2.032
ized
Tokenized
Feature activation+0.181
Ne
Token Ne
Feature activation+0.172
at
Tokenat
Feature activation+0.000
ly
Tokenly
Feature activation+0.000
(
Token (
Feature activation+0.000

INTERVAL 1.684 - 1.965
CONTAINS 0.000%

Here
Token Here
Feature activation+0.000
are
Token are
Feature activation+0.000
four
Token four
Feature activation+0.000
things
Token things
Feature activation+0.000
NATO
Token NATO
Feature activation+1.553
needs
Token needs
Feature activation+1.936
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
when
Token when
Feature activation+0.000
its
Token its
Feature activation+0.000
leaders
Token leaders
Feature activation+0.000
kinda
Token kinda
Feature activation+0.000
fascinated
Token fascinated
Feature activation+0.000
with
Token with
Feature activation+0.000
all
Token all
Feature activation+0.000
things
Token things
Feature activation+0.000
British
Token British
Feature activation+1.713
.
Token.
Feature activation+0.000
The
Token The
Feature activation+0.000
only
Token only
Feature activation+0.000
things
Token things
Feature activation+0.000
we
Token we
Feature activation+0.000

INTERVAL 1.403 - 1.684
CONTAINS 0.001%

too
Token too
Feature activation+0.000
).
Token).
Feature activation+0.000
The
Token The
Feature activation+0.000
whole
Token whole
Feature activation+0.000
thing
Token thing
Feature activation+0.000
snow
Token snow
Feature activation+1.519
balls
Tokenballs
Feature activation+0.938
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Hist
TokenHist
Feature activation+0.000
admits
Token admits
Feature activation+0.000
graduates
Token graduates
Feature activation+0.000
could
Token could
Feature activation+0.000
find
Token find
Feature activation+0.000
things
Token things
Feature activation+0.000
difficult
Token difficult
Feature activation+1.436
in
Token in
Feature activation+0.000
the
Token the
Feature activation+0.000
current
Token current
Feature activation+0.000
financial
Token financial
Feature activation+0.000
climate
Token climate
Feature activation+0.000
?
Token?
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
One
TokenOne
Feature activation+0.000
thing
Token thing
Feature activation+0.000
is
Token is
Feature activation+1.461
clear
Token clear
Feature activation+0.782
:
Token:
Feature activation+0.000
they
Token they
Feature activation+0.000
won
Token won
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
case
Token case
Feature activation+0.000
of
Token of
Feature activation+0.000
taking
Token taking
Feature activation+0.000
things
Token things
Feature activation+0.000
one
Token one
Feature activation+1.001
step
Token step
Feature activation+1.541
at
Token at
Feature activation+0.000
time
Token time
Feature activation+0.000
-
Token -
Feature activation+0.000
first
Token first
Feature activation+0.000
to
Token to
Feature activation+0.000
This
TokenThis
Feature activation+0.000
was
Token was
Feature activation+0.000
the
Token the
Feature activation+0.000
last
Token last
Feature activation+0.000
thing
Token thing
Feature activation+0.000
a
Token a
Feature activation+1.670
new
Token new
Feature activation+0.962
President
Token President
Feature activation+0.000
dealing
Token dealing
Feature activation+0.000
with
Token with
Feature activation+0.000
a
Token a
Feature activation+0.000

INTERVAL 1.123 - 1.403
CONTAINS 0.001%

and
Token and
Feature activation+0.000
a
Token a
Feature activation+0.000
few
Token few
Feature activation+0.000
worthwhile
Token worthwhile
Feature activation+0.000
things
Token things
Feature activation+0.000
every
Token every
Feature activation+1.192
once
Token once
Feature activation+0.000
in
Token in
Feature activation+0.000
a
Token a
Feature activation+0.000
while
Token while
Feature activation+0.000
,
Token,
Feature activation+0.000
,
Token,
Feature activation+0.000
surprises
Token surprises
Feature activation+0.000
and
Token and
Feature activation+0.000
other
Token other
Feature activation+0.000
things
Token things
Feature activation+0.000
designed
Token designed
Feature activation+1.363
to
Token to
Feature activation+0.000
make
Token make
Feature activation+0.000
buying
Token buying
Feature activation+0.000
electronics
Token electronics
Feature activation+0.000
more
Token more
Feature activation+0.000
that
Token that
Feature activation+0.000
was
Token was
Feature activation+0.000
the
Token the
Feature activation+0.000
coolest
Token coolest
Feature activation+0.000
thing
Token thing
Feature activation+0.000
anyone
Token anyone
Feature activation+1.379
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
s
Tokens
Feature activation+0.000
ever
Token ever
Feature activation+0.000
done
Token done
Feature activation+0.000
One
TokenOne
Feature activation+0.000
of
Token of
Feature activation+0.000
my
Token my
Feature activation+0.000
favorite
Token favorite
Feature activation+0.000
things
Token things
Feature activation+0.000
is
Token is
Feature activation+1.177
seeing
Token seeing
Feature activation+0.000
a
Token a
Feature activation+0.000
project
Token project
Feature activation+0.000
where
Token where
Feature activation+0.000
all
Token all
Feature activation+0.000
âĢ
Token âĢ
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
10
Token10
Feature activation+0.000
Things
Token Things
Feature activation+0.000
I
Token I
Feature activation+0.000
Hate
Token Hate
Feature activation+1.372
About
Token About
Feature activation+0.000
You
Token You
Feature activation+0.003
.
Token.
Feature activation+0.000
âĢ
TokenâĢ
Feature activation+0.000
Ŀ
TokenĿ
Feature activation+0.000

INTERVAL 0.842 - 1.123
CONTAINS 0.002%

Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
The
TokenThe
Feature activation+0.000
only
Token only
Feature activation+0.000
thing
Token thing
Feature activation+0.000
show
Token show
Feature activation+1.115
runners
Tokenrunners
Feature activation+0.921
Lisa
Token Lisa
Feature activation+0.000
Joy
Token Joy
Feature activation+0.000
and
Token and
Feature activation+0.000
Jonah
Token Jonah
Feature activation+0.000
're
Token're
Feature activation+0.000
facing
Token facing
Feature activation+0.000
big
Token big
Feature activation+0.000
problems
Token problems
Feature activation+0.000
reload
Token reload
Feature activation+0.522
ing
Tokening
Feature activation+1.060
their
Token their
Feature activation+0.000
cards
Token cards
Feature activation+0.000
at
Token at
Feature activation+0.000
Prest
Token Prest
Feature activation+0.000
o
Tokeno
Feature activation+0.000
a
Token a
Feature activation+0.000
case
Token case
Feature activation+0.000
of
Token of
Feature activation+0.000
taking
Token taking
Feature activation+0.000
things
Token things
Feature activation+0.000
one
Token one
Feature activation+1.001
step
Token step
Feature activation+1.541
at
Token at
Feature activation+0.000
time
Token time
Feature activation+0.000
-
Token -
Feature activation+0.000
first
Token first
Feature activation+0.000
they
Token they
Feature activation+0.000
distract
Token distract
Feature activation+0.000
us
Token us
Feature activation+0.000
from
Token from
Feature activation+0.000
things
Token things
Feature activation+0.000
Trump
Token Trump
Feature activation+1.018
wants
Token wants
Feature activation+0.674
us
Token us
Feature activation+0.000
to
Token to
Feature activation+0.000
ignore
Token ignore
Feature activation+0.000
.
Token.
Feature activation+0.000
is
Token is
Feature activation+0.000
a
Token a
Feature activation+0.000
good
Token good
Feature activation+0.000
thing
Token thing
Feature activation+0.000
Wood
Token Wood
Feature activation+0.507
cock
Tokencock
Feature activation+0.933
mentioned
Token mentioned
Feature activation+0.000
unintended
Token unintended
Feature activation+0.000
consequences
Token consequences
Feature activation+0.000
.
Token.
Feature activation+0.000
Starting
Token Starting
Feature activation+0.000

INTERVAL 0.561 - 0.842
CONTAINS 0.003%

you
Token you
Feature activation+0.000
saw
Token saw
Feature activation+0.000
things
Token things
Feature activation+0.000
like
Token like
Feature activation+0.000
social
Token social
Feature activation+0.000
security
Token security
Feature activation+0.608
on
Token on
Feature activation+0.000
blockchain
Token blockchain
Feature activation+0.000
in
Token in
Feature activation+0.000
five
Token five
Feature activation+0.000
or
Token or
Feature activation+0.000
But
TokenBut
Feature activation+0.000
since
Token since
Feature activation+0.000
that
Token that
Feature activation+0.000
time
Token time
Feature activation+0.000
things
Token things
Feature activation+0.000
haven
Token haven
Feature activation+0.667
âĢ
TokenâĢ
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
t
Tokent
Feature activation+0.000
gone
Token gone
Feature activation+0.000
smoothly
Token smoothly
Feature activation+0.000
we
Token we
Feature activation+0.000
are
Token are
Feature activation+0.000
not
Token not
Feature activation+0.000
seeing
Token seeing
Feature activation+0.000
things
Token things
Feature activation+0.000
this
Token this
Feature activation+0.594
way
Token way
Feature activation+0.000
or
Token or
Feature activation+0.000
we
Token we
Feature activation+0.000
will
Token will
Feature activation+0.000
keep
Token keep
Feature activation+0.000
ing
Tokening
Feature activation+0.000
and
Token and
Feature activation+0.000
several
Token several
Feature activation+0.000
other
Token other
Feature activation+0.000
things
Token things
Feature activation+0.000
are
Token are
Feature activation+0.812
now
Token now
Feature activation+0.000
tracked
Token tracked
Feature activation+0.000
(
Token (
Feature activation+0.000
more
Tokenmore
Feature activation+0.000
info
Token info
Feature activation+0.000
.
Token.
Feature activation+0.000
"
Token "
Feature activation+0.000
The
TokenThe
Feature activation+0.000
main
Token main
Feature activation+0.000
thing
Token thing
Feature activation+0.000
for
Token for
Feature activation+0.563
the
Token the
Feature activation+0.000
president
Token president
Feature activation+0.000
to
Token to
Feature activation+0.000
do
Token do
Feature activation+0.000
is
Token is
Feature activation+0.000

INTERVAL 0.281 - 0.561
CONTAINS 0.006%

ch
Tokench
Feature activation+0.000
ia
Tokenia
Feature activation+0.000
rules
Token rules
Feature activation+0.000
everything
Token everything
Feature activation+0.000
east
Token east
Feature activation+0.000
and
Token and
Feature activation+0.349
north
Token north
Feature activation+0.000
of
Token of
Feature activation+0.000
the
Token the
Feature activation+0.000
Rh
Token Rh
Feature activation+0.000
ine
Tokenine
Feature activation+0.000
morning
Token morning
Feature activation+0.000
we
Token we
Feature activation+0.000
packed
Token packed
Feature activation+0.000
our
Token our
Feature activation+0.000
stuff
Token stuff
Feature activation+0.000
and
Token and
Feature activation+0.366
started
Token started
Feature activation+0.000
the
Token the
Feature activation+0.000
last
Token last
Feature activation+0.000
day
Token day
Feature activation+0.000
of
Token of
Feature activation+0.000
on
Token on
Feature activation+0.000
me
Token me
Feature activation+0.000
made
Token made
Feature activation+0.000
things
Token things
Feature activation+0.000
abundantly
Token abundantly
Feature activation+0.000
clear
Token clear
Feature activation+0.391
to
Token to
Feature activation+0.000
me
Token me
Feature activation+0.000
.
Token.
Feature activation+0.000
Brand
Token Brand
Feature activation+0.000
i
Tokeni
Feature activation+0.000
hiber
Token hiber
Feature activation+0.000
nation
Tokennation
Feature activation+0.000
.
Token.
Feature activation+0.000
Stranger
Token Stranger
Feature activation+0.000
things
Token things
Feature activation+0.000
have
Token have
Feature activation+0.394
happened
Token happened
Feature activation+0.000
.
Token.
Feature activation+0.000
<|endoftext|>
Token<|endoftext|>
Feature activation+0.000
President
TokenPresident
Feature activation+0.000
-
Token-
Feature activation+0.000
go
Token go
Feature activation+0.000
,
Token,
Feature activation+0.000
10
Token 10
Feature activation+0.000
things
Token things
Feature activation+0.000
you
Token you
Feature activation+0.000
can
Token can
Feature activation+0.307
do
Token do
Feature activation+0.000
.
Token.
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
#
Token#
Feature activation+0.000

INTERVAL 0.000 - 0.281
CONTAINS 99.986%

48
Token 48
Feature activation+0.000
143
Token143
Feature activation+0.000
Mü
Token Mü
Feature activation+0.000
n
Tokenn
Feature activation+0.000
ster
Tokenster
Feature activation+0.000
,
Token,
Feature activation+0.000
Germany
Token Germany
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
2
Token2
Feature activation+0.000
Unit
TokenUnit
Feature activation+0.000
nil
Token nil
Feature activation+0.000
1
Token 1
Feature activation+0.000
@
Token @
Feature activation+0.000
rec
Tokenrec
Feature activation+0.000
ieved
Tokenieved
Feature activation+0.000
_
Token_
Feature activation+0.000
message
Tokenmessage
Feature activation+0.000
#
Token #
Feature activation+0.000
=>
Token =>
Feature activation+0.000
nil
Token nil
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
unique
Token unique
Feature activation+0.000
in
Token in
Feature activation+0.000
that
Token that
Feature activation+0.000
a
Token a
Feature activation+0.000
publisher
Token publisher
Feature activation+0.000
has
Token has
Feature activation+0.000
never
Token never
Feature activation+0.000
assisted
Token assisted
Feature activation+0.000
a
Token a
Feature activation+0.000
large
Token large
Feature activation+0.000
Kickstarter
Token Kickstarter
Feature activation+0.000
attempting
Token attempting
Feature activation+0.000
to
Token to
Feature activation+0.000
show
Token show
Feature activation+0.000
how
Token how
Feature activation+0.000
conditioned
Token conditioned
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
Ċ
TokenĊ
Feature activation+0.000
we
Tokenwe
Feature activation+0.000
have
Token have
Feature activation+0.000
become
Token become
Feature activation+0.000
to
Token to
Feature activation+0.000
There
Token There
Feature activation+0.000
is
Token is
Feature activation+0.000
one
Token one
Feature activation+0.000
constant
Token constant
Feature activation+0.000
in
Token in
Feature activation+0.000
this
Token this
Feature activation+0.000
cruel
Token cruel
Feature activation+0.000
and
Token and
Feature activation+0.000
cap
Token cap
Feature activation+0.000
ric
Tokenric
Feature activation+0.000
ious
Tokenious
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