H1.5 top features

Top feature 0 in H1.5: (feature 5836)

TOP ACTIVATIONS
MAX = 3.843

women
Token women
Feature activation+0.000
 are
Token are
Feature activation+0.000
 more
Token more
Feature activation+0.000
 likely
Token likely
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.843
 to
Token to
Feature activation+0.000
 interrupt
Token interrupt
Feature activation+0.000
 their
Token their
Feature activation+0.000
 careers
Token careers
Feature activation+0.533
 for
Token for
Feature activation+0.000
An
Token An
Feature activation+0.000
 American
Token American
Feature activation+0.000
 tourist
Token tourist
Feature activation+0.000
 accused
Token accused
Feature activation+0.000
 two
Token two
Feature activation+0.000
 men
Token men
Feature activation+3.770
 of
Token of
Feature activation+0.000
 raping
Token raping
Feature activation+0.167
 her
Token her
Feature activation+0.242
 in
Token in
Feature activation+0.000
 D
Token D
Feature activation+0.000
compared
Token compared
Feature activation+0.000
 to
Token to
Feature activation+0.000
 almost
Token almost
Feature activation+0.000
 14
Token 14
Feature activation+0.000
 million
Token million
Feature activation+0.000
 men
Token men
Feature activation+3.678
 .
Token.
Feature activation+0.000
 In
Token In
Feature activation+0.000
 Seoul
Token Seoul
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 many
Token many
Feature activation+0.000
of
Token of
Feature activation+0.000
 seven
Token seven
Feature activation+0.000
 women
Token women
Feature activation+0.000
 and
Token and
Feature activation+0.000
 five
Token five
Feature activation+0.000
 men
Token men
Feature activation+3.659
 will
Token will
Feature activation+0.000
 have
Token have
Feature activation+0.000
 to
Token to
Feature activation+0.000
 come
Token come
Feature activation+0.000
 to
Token to
Feature activation+0.000
women
Token women
Feature activation+0.159
 might
Token might
Feature activation+0.000
 suffer
Token suffer
Feature activation+0.000
 less
Token less
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.576
 from
Token from
Feature activation+0.000
 declining
Token declining
Feature activation+0.000
 productivity
Token productivity
Feature activation+0.000
 as
Token as
Feature activation+0.000
 they
Token they
Feature activation+0.000
77
Token 77
Feature activation+0.000
 ¢
Token¢
Feature activation+0.000
 for
Token for
Feature activation+0.000
 every
Token every
Feature activation+0.000
 dollar
Token dollar
Feature activation+0.000
 men
Token men
Feature activation+3.371
 earn
Token earn
Feature activation+0.000
 ).
Token).
Feature activation+0.000
 This
Token This
Feature activation+0.000
 national
Token national
Feature activation+0.000
 trend
Token trend
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
 s
Tokens
Feature activation+0.000
 a
Token a
Feature activation+0.000
 problem
Token problem
Feature activation+0.000
 for
Token for
Feature activation+0.000
 men
Token men
Feature activation+3.346
 ,
Token,
Feature activation+0.000
 too
Token too
Feature activation+0.000
 .
Token.
Feature activation+0.000
 In
Token In
Feature activation+0.000
 just
Token just
Feature activation+0.000
more
Token more
Feature activation+0.000
 sensitive
Token sensitive
Feature activation+0.000
 to
Token to
Feature activation+0.000
 grades
Token grades
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.327
 ,
Token,
Feature activation+0.000
 and
Token and
Feature activation+0.000
 whether
Token whether
Feature activation+0.000
 the
Token the
Feature activation+0.000
 gender
Token gender
Feature activation+0.095
one
Token one
Feature activation+0.000
 out
Token out
Feature activation+0.000
 of
Token of
Feature activation+0.000
 every
Token every
Feature activation+0.000
 six
Token six
Feature activation+0.000
 men
Token men
Feature activation+3.302
 ,
Token,
Feature activation+0.000
 suffers
Token suffers
Feature activation+0.000
 such
Token such
Feature activation+0.000
 an
Token an
Feature activation+0.000
 assault
Token assault
Feature activation+0.000
are
Token are
Feature activation+0.000
 less
Token less
Feature activation+0.000
 likely
Token likely
Feature activation+0.000
 than
Token than
Feature activation+0.000
 American
Token American
Feature activation+0.000
 men
Token men
Feature activation+3.279
 to
Token to
Feature activation+0.000
 ride
Token ride
Feature activation+0.000
 bikes
Token bikes
Feature activation+0.000
 in
Token in
Feature activation+0.000
 cities
Token cities
Feature activation+0.000
Campbell
Token Campbell
Feature activation+0.000
 said
Token said
Feature activation+0.000
 domestic
Token domestic
Feature activation+0.000
 violence
Token violence
Feature activation+0.000
 against
Token against
Feature activation+0.000
 men
Token men
Feature activation+3.264
 was
Token was
Feature activation+0.000
 also
Token also
Feature activation+0.000
 a
Token a
Feature activation+0.000
 problem
Token problem
Feature activation+0.000
 but
Token but
Feature activation+0.000
-
Token-
Feature activation+0.000
 kil
Tokenkil
Feature activation+0.000
 ometer
Tokenometer
Feature activation+0.000
 hike
Token hike
Feature activation+0.000
 two
Token two
Feature activation+0.000
 men
Token men
Feature activation+3.248
 and
Token and
Feature activation+0.000
 two
Token two
Feature activation+0.000
 women
Token women
Feature activation+0.114
 <|endoftext|>
Token<|endoftext|>
Feature activation+0.000
 With
TokenWith
Feature activation+0.000
doing
Token doing
Feature activation+0.000
 the
Token the
Feature activation+0.000
 same
Token same
Feature activation+0.000
 jobs
Token jobs
Feature activation+0.709
 as
Token as
Feature activation+0.000
 men
Token men
Feature activation+3.217
 (
Token (
Feature activation+0.000
 the
Tokenthe
Feature activation+0.000
 Census
Token Census
Feature activation+0.000
 Bureau
Token Bureau
Feature activation+0.000
 concludes
Token concludes
Feature activation+0.000
ľ
Tokenľ
Feature activation+0.000
 we
Tokenwe
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
 men
Token men
Feature activation+3.209
 to
Token to
Feature activation+0.000
 not
Token not
Feature activation+0.000
 get
Token get
Feature activation+0.000
 blackout
Token blackout
Feature activation+0.000
 drunk
Token drunk
Feature activation+0.103
working
Token working
Feature activation+0.000
 women
Token women
Feature activation+0.000
 and
Token and
Feature activation+0.000
 all
Token all
Feature activation+0.000
 -
Token-
Feature activation+0.000
 male
Tokenmale
Feature activation+3.093
 establishments
Token establishments
Feature activation+0.000
 as
Token as
Feature activation+0.000
 heated
Token heated
Feature activation+0.000
 issues
Token issues
Feature activation+0.000
 in
Token in
Feature activation+0.000
r
Tokenr
Feature activation+0.000
 and
Token and
Feature activation+0.000
 less
Token less
Feature activation+0.000
 sexual
Token sexual
Feature activation+0.731
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.066
 ;
Token;
Feature activation+0.000
 and
Token and
Feature activation+0.000
 their
Token their
Feature activation+0.000
 role
Token role
Feature activation+0.000
 was
Token was
Feature activation+0.000
prising
Tokenprising
Feature activation+0.000
 is
Token is
Feature activation+0.000
 the
Token the
Feature activation+0.000
 revelation
Token revelation
Feature activation+0.000
 that
Token that
Feature activation+0.000
 men
Token men
Feature activation+3.032
 prefer
Token prefer
Feature activation+0.000
 higher
Token higher
Feature activation+0.000
 voices
Token voices
Feature activation+0.000
 in
Token in
Feature activation+0.000
 women
Token women
Feature activation+0.277
generations
Token generations
Feature activation+0.198
 of
Token of
Feature activation+0.000
 mothers
Token mothers
Feature activation+0.519
 without
Token without
Feature activation+0.000
 a
Token a
Feature activation+0.000
 man
Token man
Feature activation+2.959
 in
Token in
Feature activation+0.000
 the
Token the
Feature activation+0.000
 house
Token house
Feature activation+0.105
 âĢĶ
Token âĢĶ
Feature activation+0.000
 lone
Token lone
Feature activation+0.000
women
Token women
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 but
Token but
Feature activation+0.000
 not
Token not
Feature activation+0.000
 for
Token for
Feature activation+0.000
 men
Token men
Feature activation+2.871
 .
Token.
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 The
TokenThe
Feature activation+0.000
 researchers
Token researchers
Feature activation+0.000
on
Token on
Feature activation+0.000
 the
Token the
Feature activation+0.000
 part
Token part
Feature activation+0.000
 of
Token of
Feature activation+0.000
 all
Token all
Feature activation+0.000
 men
Token men
Feature activation+2.865
 to
Token to
Feature activation+0.000
 terror
Token terror
Feature activation+0.000
 ize
Tokenize
Feature activation+0.000
 all
Token all
Feature activation+0.000
 women
Token women
Feature activation+0.292

Top DFA by src position
MAX = 4.881

He
Token He
Feature activation+0.036
Top resid features:
also
Token also
Feature activation+0.027
Top resid features:
argues
Token argues
Feature activation+0.028
Top resid features:
that
Token that
Feature activation+0.041
Top resid features:
because
Token because
Feature activation+0.034
Top resid features:
women
Token women
Feature activation+4.881
Top resid features:
are
Token are
Feature activation+0.059
Top resid features:
more
Token more
Feature activation+0.002
Top resid features:
likely
Token likely
Feature activation+0.098
Top resid features:
than
Token than
Feature activation+0.088
Top resid features:
men
Token men
Feature activation-0.230
Top resid features:
of
Token of
Feature activation+0.016
Top resid features:
crimes
Token crimes
Feature activation+0.037
Top resid features:
reported
Token reported
Feature activation+0.033
Top resid features:
by
Token by
Feature activation+0.023
Top resid features:
foreign
Token foreign
Feature activation+0.042
Top resid features:
women
Token women
Feature activation+3.287
Top resid features:
visiting
Token visiting
Feature activation+0.040
Top resid features:
India
Token India
Feature activation+0.046
Top resid features:
.
Token.
Feature activation+0.002
Top resid features:
An
Token An
Feature activation+0.033
Top resid features:
American
Token American
Feature activation+0.069
Top resid features:
June
Token June
Feature activation+0.006
Top resid features:
,
Token,
Feature activation+0.007
Top resid features:
nearly
Token nearly
Feature activation+0.007
Top resid features:
10
Token 10
Feature activation+0.009
Top resid features:
million
Token million
Feature activation+0.014
Top resid features:
women
Token women
Feature activation+1.806
Top resid features:
were
Token were
Feature activation+0.025
Top resid features:
employed
Token employed
Feature activation+0.022
Top resid features:
nationwide
Token nationwide
Feature activation+0.016
Top resid features:
,
Token,
Feature activation+0.019
Top resid features:
according
Token according
Feature activation+0.036
Top resid features:
,
Token,
Feature activation+0.035
Top resid features:
the
Token the
Feature activation+0.025
Top resid features:
panel
Token panel
Feature activation+0.031
Top resid features:
of
Token of
Feature activation+0.029
Top resid features:
seven
Token seven
Feature activation+0.050
Top resid features:
women
Token women
Feature activation+4.584
Top resid features:
and
Token and
Feature activation-0.015
Top resid features:
five
Token five
Feature activation+0.037
Top resid features:
men
Token men
Feature activation-0.086
Top resid features:
will
Token will
Feature activation+0.000
Top resid features:
have
Token have
Feature activation+0.000
Top resid features:
durable
Token durable
Feature activation+0.040
Top resid features:
.
Token.
Feature activation+0.007
Top resid features:
If
Token If
Feature activation+0.032
Top resid features:
so
Token so
Feature activation+0.020
Top resid features:
,
Token,
Feature activation+0.046
Top resid features:
women
Token women
Feature activation+3.071
Top resid features:
might
Token might
Feature activation+0.068
Top resid features:
suffer
Token suffer
Feature activation+0.028
Top resid features:
less
Token less
Feature activation+0.041
Top resid features:
than
Token than
Feature activation+0.058
Top resid features:
men
Token men
Feature activation-0.152
Top resid features:
the
Tokenthe
Feature activation+0.021
Top resid features:
Census
Token Census
Feature activation+0.028
Top resid features:
Bureau
Token Bureau
Feature activation+0.023
Top resid features:
concludes
Token concludes
Feature activation+0.037
Top resid features:
that
Token that
Feature activation+0.043
Top resid features:
women
Token women
Feature activation+2.271
Top resid features:
earn
Token earn
Feature activation+0.030
Top resid features:
77
Token 77
Feature activation+0.067
Top resid features:
¢
Token¢
Feature activation+0.071
Top resid features:
for
Token for
Feature activation+0.032
Top resid features:
every
Token every
Feature activation+0.060
Top resid features:
not
Token not
Feature activation+0.025
Top resid features:
solely
Token solely
Feature activation+0.014
Top resid features:
a
Token a
Feature activation+0.018
Top resid features:
problem
Token problem
Feature activation+0.013
Top resid features:
for
Token for
Feature activation+0.014
Top resid features:
women
Token women
Feature activation+3.183
Top resid features:
.
Token.
Feature activation+0.012
Top resid features:
It
Token It
Feature activation+0.029
Top resid features:
âĢ
TokenâĢ
Feature activation+0.032
Top resid features:
Ļ
TokenĻ
Feature activation+0.038
Top resid features:
s
Tokens
Feature activation+0.045
Top resid features:
-
Token-
Feature activation+0.006
Top resid features:
2016
Token2016
Feature activation+0.026
Top resid features:
to
Token to
Feature activation+0.027
Top resid features:
test
Token test
Feature activation+0.040
Top resid features:
whether
Token whether
Feature activation+0.050
Top resid features:
women
Token women
Feature activation+4.727
Top resid features:
are
Token are
Feature activation+0.035
Top resid features:
more
Token more
Feature activation+0.002
Top resid features:
sensitive
Token sensitive
Feature activation+0.060
Top resid features:
to
Token to
Feature activation+0.032
Top resid features:
grades
Token grades
Feature activation+0.079
Top resid features:
one
Token one
Feature activation+0.017
Top resid features:
out
Token out
Feature activation+0.016
Top resid features:
of
Token of
Feature activation+0.018
Top resid features:
every
Token every
Feature activation+0.021
Top resid features:
four
Token four
Feature activation+0.028
Top resid features:
women
Token women
Feature activation+4.418
Top resid features:
,
Token,
Feature activation+0.030
Top resid features:
and
Token and
Feature activation+0.003
Top resid features:
one
Token one
Feature activation+0.033
Top resid features:
out
Token out
Feature activation+0.016
Top resid features:
of
Token of
Feature activation+0.020
Top resid features:
's
Token's
Feature activation+0.018
Top resid features:
no
Token no
Feature activation+0.034
Top resid features:
secret
Token secret
Feature activation+0.027
Top resid features:
that
Token that
Feature activation+0.027
Top resid features:
American
Token American
Feature activation+0.053
Top resid features:
women
Token women
Feature activation+4.031
Top resid features:
are
Token are
Feature activation+0.057
Top resid features:
less
Token less
Feature activation+0.037
Top resid features:
likely
Token likely
Feature activation+0.108
Top resid features:
than
Token than
Feature activation+0.088
Top resid features:
American
Token American
Feature activation+0.099
Top resid features:
estimates
Token estimates
Feature activation+0.006
Top resid features:
up
Token up
Feature activation+0.002
Top resid features:
to
Token to
Feature activation+0.001
Top resid features:
one
Token one
Feature activation-0.003
Top resid features:
million
Token million
Feature activation+0.001
Top resid features:
women
Token women
Feature activation+4.238
Top resid features:
experience
Token experience
Feature activation+0.012
Top resid features:
at
Token at
Feature activation-0.004
Top resid features:
least
Token least
Feature activation+0.013
Top resid features:
one
Token one
Feature activation-0.004
Top resid features:
incident
Token incident
Feature activation+0.037
Top resid features:
since
Token since
Feature activation+0.008
Top resid features:
it
Token it
Feature activation-0.004
Top resid features:
remains
Token remains
Feature activation+0.022
Top resid features:
closed
Token closed
Feature activation+0.010
Top resid features:
to
Token to
Feature activation+0.009
Top resid features:
women
Token women
Feature activation+4.448
Top resid features:
.
Token.
Feature activation+0.023
Top resid features:
Ċ
TokenĊ
Feature activation+0.024
Top resid features:
Ċ
TokenĊ
Feature activation+0.028
Top resid features:
On
TokenOn
Feature activation-0.003
Top resid features:
the
Token the
Feature activation+0.020
Top resid features:
studies
Token studies
Feature activation+0.019
Top resid features:
suggest
Token suggest
Feature activation+0.018
Top resid features:
that
Token that
Feature activation+0.035
Top resid features:
in
Token in
Feature activation+0.021
Top resid features:
general
Token general
Feature activation+0.038
Top resid features:
women
Token women
Feature activation+3.026
Top resid features:
make
Token make
Feature activation+0.031
Top resid features:
less
Token less
Feature activation+0.038
Top resid features:
doing
Token doing
Feature activation+0.028
Top resid features:
the
Token the
Feature activation+0.027
Top resid features:
same
Token same
Feature activation+0.038
Top resid features:
.
Token.
Feature activation+0.005
Top resid features:
Ċ
TokenĊ
Feature activation+0.003
Top resid features:
Ċ
TokenĊ
Feature activation+0.004
Top resid features:
Suggest
TokenSuggest
Feature activation+0.007
Top resid features:
ing
Tokening
Feature activation-0.003
Top resid features:
women
Token women
Feature activation+4.334
Top resid features:
don
Token don
Feature activation+0.000
Top resid features:
âĢ
TokenâĢ
Feature activation+0.002
Top resid features:
Ļ
TokenĻ
Feature activation+0.001
Top resid features:
t
Tokent
Feature activation+0.002
Top resid features:
drink
Token drink
Feature activation+0.008
Top resid features:
of
Token of
Feature activation+0.023
Top resid features:
contraceptive
Token contraceptive
Feature activation+0.011
Top resid features:
rights
Token rights
Feature activation+0.019
Top resid features:
,
Token,
Feature activation+0.031
Top resid features:
working
Token working
Feature activation+0.065
Top resid features:
women
Token women
Feature activation+1.351
Top resid features:
and
Token and
Feature activation+0.010
Top resid features:
all
Token all
Feature activation+0.028
Top resid features:
-
Token-
Feature activation-0.010
Top resid features:
male
Tokenmale
Feature activation-0.285
Top resid features:
establishments
Token establishments
Feature activation+0.000
Top resid features:
of
Token of
Feature activation+0.016
Top resid features:
the
Token the
Feature activation+0.007
Top resid features:
male
Token male
Feature activation-0.179
Top resid features:
world
Token world
Feature activation+0.013
Top resid features:
;
Token;
Feature activation+0.008
Top resid features:
women
Token women
Feature activation+2.168
Top resid features:
were
Token were
Feature activation+0.028
Top resid features:
to
Token to
Feature activation+0.027
Top resid features:
be
Token be
Feature activation+0.022
Top resid features:
higher
Token higher
Feature activation+0.026
Top resid features:
,
Token,
Feature activation+0.013
Top resid features:
guys
Token guys
Feature activation+0.038
Top resid features:
prefer
Token prefer
Feature activation+0.010
Top resid features:
higher
Token higher
Feature activation+0.024
Top resid features:
voices
Token voices
Feature activation+0.006
Top resid features:
in
Token in
Feature activation+0.021
Top resid features:
women
Token women
Feature activation+4.220
Top resid features:
Ċ
TokenĊ
Feature activation+0.021
Top resid features:
Ċ
TokenĊ
Feature activation+0.021
Top resid features:
Equ
TokenEqu
Feature activation+0.028
Top resid features:
ally
Tokenally
Feature activation+0.042
Top resid features:
unsur
Token unsur
Feature activation+0.027
Top resid features:
Ċ
TokenĊ
Feature activation+0.013
Top resid features:
As
TokenAs
Feature activation+0.010
Top resid features:
many
Token many
Feature activation+0.008
Top resid features:
of
Token of
Feature activation+0.012
Top resid features:
these
Token these
Feature activation-0.005
Top resid features:
women
Token women
Feature activation+2.504
Top resid features:
become
Token become
Feature activation+0.013
Top resid features:
grandmother
Token grandmother
Feature activation+0.106
Top resid features:
s
Tokens
Feature activation+0.022
Top resid features:
,
Token,
Feature activation+0.029
Top resid features:
a
Token a
Feature activation+0.020
Top resid features:
significantly
Token significantly
Feature activation+0.007
Top resid features:
associated
Token associated
Feature activation+0.024
Top resid features:
with
Token with
Feature activation+0.023
Top resid features:
riding
Token riding
Feature activation+0.035
Top resid features:
for
Token for
Feature activation+0.010
Top resid features:
women
Token women
Feature activation+3.101
Top resid features:
,
Token,
Feature activation+0.024
Top resid features:
but
Token but
Feature activation+0.066
Top resid features:
not
Token not
Feature activation+0.070
Top resid features:
for
Token for
Feature activation+0.009
Top resid features:
men
Token men
Feature activation-0.207
Top resid features:
a
Token a
Feature activation-0.005
Top resid features:
conspiracy
Token conspiracy
Feature activation-0.003
Top resid features:
to
Token to
Feature activation+0.005
Top resid features:
subjug
Token subjug
Feature activation+0.007
Top resid features:
ate
Tokenate
Feature activation+0.003
Top resid features:
women
Token women
Feature activation+3.813
Top resid features:
,
Token,
Feature activation+0.009
Top resid features:
and
Token and
Feature activation+0.013
Top resid features:
just
Token just
Feature activation+0.001
Top resid features:
as
Token as
Feature activation+0.003
Top resid features:
the
Token the
Feature activation-0.005
Top resid features:

Decoder Weights Distribution

Head 0: 0.04

Head 1: 0.04

Head 2: 0.02

Head 3: 0.01

Head 4: 0.04

Head 5: 0.51

Head 6: 0.12

Head 7: 0.04

Head 8: 0.02

Head 9: 0.02

Head 10: 0.02

Head 11: 0.12

Positive logits

ocracy1.98

د1.97

lishes1.90

behind1.80

aki1.77

ignt1.72

default1.68

Requirements1.67

ollow1.67

iser1.66

aternity1.66

medium1.63

ikk1.63

only1.62

ards1.62

isers1.61

ocus1.60

Shade1.59

ِ1.57

center1.56

Negative logits

leased-1.81

GUN-1.81

smuggled-1.74

��-1.73

kidney-1.72

intestinal-1.66

gross-1.66

airborne-1.65

transports-1.65

zynski-1.63

Seah-1.62

incom-1.61

authorized-1.59

emis-1.57

sailed-1.56

lease-1.56

tten-1.56

lyak-1.55

undown-1.55

entreprene-1.54

INTERVAL 3.459 - 3.843
CONTAINS 0.001%

An
Token An
Feature activation+0.000
 American
Token American
Feature activation+0.000
 tourist
Token tourist
Feature activation+0.000
 accused
Token accused
Feature activation+0.000
 two
Token two
Feature activation+0.000
 men
Token men
Feature activation+3.770
 of
Token of
Feature activation+0.000
 raping
Token raping
Feature activation+0.167
 her
Token her
Feature activation+0.242
 in
Token in
Feature activation+0.000
 D
Token D
Feature activation+0.000
compared
Token compared
Feature activation+0.000
 to
Token to
Feature activation+0.000
 almost
Token almost
Feature activation+0.000
 14
Token 14
Feature activation+0.000
 million
Token million
Feature activation+0.000
 men
Token men
Feature activation+3.678
 .
Token.
Feature activation+0.000
 In
Token In
Feature activation+0.000
 Seoul
Token Seoul
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 many
Token many
Feature activation+0.000
of
Token of
Feature activation+0.000
 seven
Token seven
Feature activation+0.000
 women
Token women
Feature activation+0.000
 and
Token and
Feature activation+0.000
 five
Token five
Feature activation+0.000
 men
Token men
Feature activation+3.659
 will
Token will
Feature activation+0.000
 have
Token have
Feature activation+0.000
 to
Token to
Feature activation+0.000
 come
Token come
Feature activation+0.000
 to
Token to
Feature activation+0.000
women
Token women
Feature activation+0.000
 are
Token are
Feature activation+0.000
 more
Token more
Feature activation+0.000
 likely
Token likely
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.843
 to
Token to
Feature activation+0.000
 interrupt
Token interrupt
Feature activation+0.000
 their
Token their
Feature activation+0.000
 careers
Token careers
Feature activation+0.533
 for
Token for
Feature activation+0.000
women
Token women
Feature activation+0.159
 might
Token might
Feature activation+0.000
 suffer
Token suffer
Feature activation+0.000
 less
Token less
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.576
 from
Token from
Feature activation+0.000
 declining
Token declining
Feature activation+0.000
 productivity
Token productivity
Feature activation+0.000
 as
Token as
Feature activation+0.000
 they
Token they
Feature activation+0.000

INTERVAL 3.074 - 3.459
CONTAINS 0.001%

more
Token more
Feature activation+0.000
 sensitive
Token sensitive
Feature activation+0.000
 to
Token to
Feature activation+0.000
 grades
Token grades
Feature activation+0.000
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.327
 ,
Token,
Feature activation+0.000
 and
Token and
Feature activation+0.000
 whether
Token whether
Feature activation+0.000
 the
Token the
Feature activation+0.000
 gender
Token gender
Feature activation+0.095
-
Token-
Feature activation+0.000
 kil
Tokenkil
Feature activation+0.000
 ometer
Tokenometer
Feature activation+0.000
 hike
Token hike
Feature activation+0.000
 two
Token two
Feature activation+0.000
 men
Token men
Feature activation+3.248
 and
Token and
Feature activation+0.000
 two
Token two
Feature activation+0.000
 women
Token women
Feature activation+0.114
 <|endoftext|>
Token<|endoftext|>
Feature activation+0.000
 With
TokenWith
Feature activation+0.000
one
Token one
Feature activation+0.000
 out
Token out
Feature activation+0.000
 of
Token of
Feature activation+0.000
 every
Token every
Feature activation+0.000
 six
Token six
Feature activation+0.000
 men
Token men
Feature activation+3.302
 ,
Token,
Feature activation+0.000
 suffers
Token suffers
Feature activation+0.000
 such
Token such
Feature activation+0.000
 an
Token an
Feature activation+0.000
 assault
Token assault
Feature activation+0.000
doing
Token doing
Feature activation+0.000
 the
Token the
Feature activation+0.000
 same
Token same
Feature activation+0.000
 jobs
Token jobs
Feature activation+0.709
 as
Token as
Feature activation+0.000
 men
Token men
Feature activation+3.217
 (
Token (
Feature activation+0.000
 the
Tokenthe
Feature activation+0.000
 Census
Token Census
Feature activation+0.000
 Bureau
Token Bureau
Feature activation+0.000
 concludes
Token concludes
Feature activation+0.000
Ļ
TokenĻ
Feature activation+0.000
 s
Tokens
Feature activation+0.000
 a
Token a
Feature activation+0.000
 problem
Token problem
Feature activation+0.000
 for
Token for
Feature activation+0.000
 men
Token men
Feature activation+3.346
 ,
Token,
Feature activation+0.000
 too
Token too
Feature activation+0.000
 .
Token.
Feature activation+0.000
 In
Token In
Feature activation+0.000
 just
Token just
Feature activation+0.000

INTERVAL 2.690 - 3.074
CONTAINS 0.001%

on
Token on
Feature activation+0.000
 the
Token the
Feature activation+0.000
 part
Token part
Feature activation+0.000
 of
Token of
Feature activation+0.000
 all
Token all
Feature activation+0.000
 men
Token men
Feature activation+2.865
 to
Token to
Feature activation+0.000
 terror
Token terror
Feature activation+0.000
 ize
Tokenize
Feature activation+0.000
 all
Token all
Feature activation+0.000
 women
Token women
Feature activation+0.292
women
Token women
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 but
Token but
Feature activation+0.000
 not
Token not
Feature activation+0.000
 for
Token for
Feature activation+0.000
 men
Token men
Feature activation+2.871
 .
Token.
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 The
TokenThe
Feature activation+0.000
 researchers
Token researchers
Feature activation+0.000
aren
Token aren
Feature activation+0.000
 âĢ
TokenâĢ
Feature activation+0.000
 Ļ
TokenĻ
Feature activation+0.000
 t
Tokent
Feature activation+0.000
 telling
Token telling
Feature activation+0.000
 men
Token men
Feature activation+2.787
 not
Token not
Feature activation+0.000
 to
Token to
Feature activation+0.000
 get
Token get
Feature activation+0.000
 blackout
Token blackout
Feature activation+0.000
 drunk
Token drunk
Feature activation+0.070
r
Tokenr
Feature activation+0.000
 and
Token and
Feature activation+0.000
 less
Token less
Feature activation+0.000
 sexual
Token sexual
Feature activation+0.731
 than
Token than
Feature activation+0.000
 men
Token men
Feature activation+3.066
 ;
Token;
Feature activation+0.000
 and
Token and
Feature activation+0.000
 their
Token their
Feature activation+0.000
 role
Token role
Feature activation+0.000
 was
Token was
Feature activation+0.000
âĢĶ
TokenâĢĶ
Feature activation+0.000
 after
Tokenafter
Feature activation+0.000
 all
Token all
Feature activation+0.000
 âĢĶ
TokenâĢĶ
Feature activation+0.000
 are
Tokenare
Feature activation+0.000
 children
Token children
Feature activation+2.736
 ,
Token,
Feature activation+0.000
 and
Token and
Feature activation+0.000
 it
Token it
Feature activation+0.098
 âĢ
TokenâĢ
Feature activation+0.000
 Ļ
TokenĻ
Feature activation+0.000

INTERVAL 2.306 - 2.690
CONTAINS 0.001%

male
Token male
Feature activation+1.776
 .
Token.
Feature activation+0.000
 âĢ
Token âĢ
Feature activation+0.000
 ľ
Tokenľ
Feature activation+0.000
 When
TokenWhen
Feature activation+0.000
 men
Token men
Feature activation+2.379
 make
Token make
Feature activation+0.000
 forceful
Token forceful
Feature activation+0.000
 gestures
Token gestures
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 they
Token they
Feature activation+0.000
careers
Token careers
Feature activation+0.533
 for
Token for
Feature activation+0.000
 reasons
Token reasons
Feature activation+0.000
 such
Token such
Feature activation+0.000
 as
Token as
Feature activation+0.000
 child
Token child
Feature activation+2.641
 care
Token care
Feature activation+0.512
 ,
Token,
Feature activation+0.000
 they
Token they
Feature activation+0.000
 tend
Token tend
Feature activation+0.000
 to
Token to
Feature activation+0.000
of
Token of
Feature activation+0.000
 Yaz
Token Yaz
Feature activation+0.000
 idi
Tokenidi
Feature activation+0.000
 women
Token women
Feature activation+0.000
 and
Token and
Feature activation+0.000
 children
Token children
Feature activation+2.313
 as
Token as
Feature activation+0.000
 slaves
Token slaves
Feature activation+0.000
 .
Token.
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
full
Token full
Feature activation+0.000
 -
Token-
Feature activation+0.000
 term
Tokenterm
Feature activation+0.000
 bree
Token bree
Feature activation+0.000
 ch
Tokench
Feature activation+0.000
 babies
Token babies
Feature activation+2.446
 opt
Token opt
Feature activation+0.000
 for
Token for
Feature activation+0.000
 ca
Token ca
Feature activation+0.000
 es
Tokenes
Feature activation+0.000
 are
Tokenare
Feature activation+0.000
connected
Token connected
Feature activation+0.000
 .
Token.
Feature activation+0.000
 One
Token One
Feature activation+0.000
 survey
Token survey
Feature activation+0.000
 of
Token of
Feature activation+0.000
 men
Token men
Feature activation+2.498
 in
Token in
Feature activation+0.000
 New
Token New
Feature activation+0.000
 York
Token York
Feature activation+0.000
 found
Token found
Feature activation+0.000
 that
Token that
Feature activation+0.000

INTERVAL 1.921 - 2.306
CONTAINS 0.003%

the
Token the
Feature activation+0.000
 very
Token very
Feature activation+0.000
 high
Token high
Feature activation+0.000
 price
Token price
Feature activation+0.000
 of
Token of
Feature activation+0.000
 child
Token child
Feature activation+2.116
 -
Token-
Feature activation+0.000
 re
Tokenre
Feature activation+0.000
 aring
Tokenaring
Feature activation+0.000
 in
Token in
Feature activation+0.000
 Korea
Token Korea
Feature activation+0.000
lone
Token lone
Feature activation+0.000
 mum
Token mum
Feature activation+0.000
 and
Token and
Feature activation+0.000
 father
Token father
Feature activation+1.439
 less
Tokenless
Feature activation+0.000
 children
Token children
Feature activation+2.148
 ,
Token,
Feature activation+0.000
 all
Token all
Feature activation+0.000
 expecting
Token expecting
Feature activation+0.000
 the
Token the
Feature activation+0.000
 state
Token state
Feature activation+0.000
be
Token be
Feature activation+0.000
 traff
Token traff
Feature activation+0.268
 icked
Tokenicked
Feature activation+0.000
 in
Token in
Feature activation+0.000
 for
Token for
Feature activation+0.000
 men
Token men
Feature activation+1.976
 to
Token to
Feature activation+0.000
 marry
Token marry
Feature activation+0.286
 .
Token.
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
 Ċ
TokenĊ
Feature activation+0.000
state
Token state
Feature activation+0.000
 to
Token to
Feature activation+0.000
 stand
Token stand
Feature activation+0.000
 in
Token in
Feature activation+0.000
 for
Token for
Feature activation+0.000
 daddy
Token daddy
Feature activation+2.115
 ,
Token,
Feature activation+0.000
 as
Token as
Feature activation+0.000
 of
Token of
Feature activation+0.000
 right
Token right
Feature activation+0.000
 .
Token.
Feature activation+0.000
made
Token made
Feature activation+0.000
 up
Token up
Feature activation+0.000
 mostly
Token mostly
Feature activation+0.000
 of
Token of
Feature activation+0.000
 white
Token white
Feature activation+0.000
 men
Token men
Feature activation+1.937
 .
Token.
Feature activation+0.000
 K
Token K
Feature activation+0.000
 lo
Tokenlo
Feature activation+0.000
 os
Tokenos
Feature activation+0.000
 ,
Token,
Feature activation+0.000

INTERVAL 1.537 - 1.921
CONTAINS 0.002%

economical
Token economical
Feature activation+0.000
 to
Token to
Feature activation+0.000
 stay
Token stay
Feature activation+0.000
 behind
Token behind
Feature activation+0.000
 helping
Token helping
Feature activation+0.000
 children
Token children
Feature activation+1.571
 to
Token to
Feature activation+0.000
 do
Token do
Feature activation+0.000
 better
Token better
Feature activation+0.000
 in
Token in
Feature activation+0.000
 the
Token the
Feature activation+0.000
along
Token along
Feature activation+0.000
 with
Token with
Feature activation+0.000
 two
Token two
Feature activation+0.000
 Saudi
Token Saudi
Feature activation+0.000
 Arabian
Token Arabian
Feature activation+0.000
 men
Token men
Feature activation+1.719
 known
Token known
Feature activation+0.000
 only
Token only
Feature activation+0.000
 as
Token as
Feature activation+0.000
 Fah
Token Fah
Feature activation+0.000
 ad
Tokenad
Feature activation+0.000
d
Tokend
Feature activation+0.000
 be
Token be
Feature activation+0.000
 licking
Token licking
Feature activation+0.000
 up
Token up
Feature activation+0.000
 another
Token another
Feature activation+0.000
 man
Token man
Feature activation+1.691
 âĢ
TokenâĢ
Feature activation+0.000
 Ļ
TokenĻ
Feature activation+0.000
 s
Tokens
Feature activation+0.000
 sperm
Token sperm
Feature activation+0.000
 .
Token.
Feature activation+0.000
the
Token the
Feature activation+0.000
 physical
Token physical
Feature activation+0.000
 models
Token models
Feature activation+0.000
 have
Token have
Feature activation+0.000
 been
Token been
Feature activation+0.000
 male
Token male
Feature activation+1.776
 .
Token.
Feature activation+0.000
 âĢ
Token âĢ
Feature activation+0.000
 ľ
Tokenľ
Feature activation+0.000
 When
TokenWhen
Feature activation+0.000
 men
Token men
Feature activation+2.379
we
Token we
Feature activation+0.000
 have
Token have
Feature activation+0.000
 far
Token far
Feature activation+0.000
 more
Token more
Feature activation+0.000
 older
Token older
Feature activation+0.000
 guys
Token guys
Feature activation+1.585
 out
Token out
Feature activation+0.000
 there
Token there
Feature activation+0.000
 as
Token as
Feature activation+0.000
 a
Token a
Feature activation+0.000
 proportion
Token proportion
Feature activation+0.000

INTERVAL 1.153 - 1.537
CONTAINS 0.004%

wind
Tokenwind
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 while
Token while
Feature activation+0.000
 women
Token women
Feature activation+0.216
 and
Token and
Feature activation+0.000
 people
Token people
Feature activation+1.462
 of
Token of
Feature activation+0.000
 color
Token color
Feature activation+0.000
 must
Token must
Feature activation+0.000
 push
Token push
Feature activation+0.000
 against
Token against
Feature activation+0.000
,
Token,
Feature activation+0.000
 the
Token the
Feature activation+0.000
 woman
Token woman
Feature activation+0.000
 who
Token who
Feature activation+0.000
 played
Token played
Feature activation+0.000
 his
Token his
Feature activation+1.207
 prec
Token prec
Feature activation+0.000
 ocious
Tokenocious
Feature activation+0.000
 daughter
Token daughter
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 Rudy
Token Rudy
Feature activation+0.000
"
Token"
Feature activation+0.000
 to
Token to
Feature activation+0.000
 be
Token be
Feature activation+0.000
 women
Token women
Feature activation+0.417
 as
Token as
Feature activation+0.000
 he
Token he
Feature activation+1.435
 and
Token and
Feature activation+0.000
 his
Token his
Feature activation+0.674
 guest
Token guest
Feature activation+0.000
 went
Token went
Feature activation+0.000
 back
Token back
Feature activation+0.000
izing
Tokenizing
Feature activation+0.000
 :
Token:
Feature activation+0.000
 Naked
Token Naked
Feature activation+0.000
 women
Token women
Feature activation+0.122
 on
Token on
Feature activation+0.000
 dog
Token dog
Feature activation+1.294
 le
Token le
Feature activation+0.185
 ashes
Tokenashes
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 crawling
Token crawling
Feature activation+0.000
 in
Token in
Feature activation+0.000
affordable
Token affordable
Feature activation+0.000
 and
Token and
Feature activation+0.000
 more
Token more
Feature activation+0.000
 available
Token available
Feature activation+0.000
 as
Token as
Feature activation+0.000
 children
Token children
Feature activation+1.209
 get
Token get
Feature activation+0.000
 older
Token older
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 with
Token with
Feature activation+0.000
 infant
Token infant
Feature activation+0.010

INTERVAL 0.769 - 1.153
CONTAINS 0.007%

the
Token the
Feature activation+0.000
 opposite
Token opposite
Feature activation+0.000
 of
Token of
Feature activation+0.000
 their
Token their
Feature activation+0.000
 actual
Token actual
Feature activation+0.000
 biological
Token biological
Feature activation+0.837
 sex
Token sex
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 even
Token even
Feature activation+0.000
 when
Token when
Feature activation+0.000
 people
Token people
Feature activation+0.593
'd
Token'd
Feature activation+0.000
 settle
Token settle
Feature activation+0.000
 down
Token down
Feature activation+0.000
 and
Token and
Feature activation+0.000
 get
Token get
Feature activation+0.000
 married
Token married
Feature activation+0.833
 ,
Token,
Feature activation+0.000
 since
Token since
Feature activation+0.000
 that
Token that
Feature activation+0.000
 keeps
Token keeps
Feature activation+0.000
 men
Token men
Feature activation+2.728
female
Token female
Feature activation+0.000
 chess
Token chess
Feature activation+0.000
 community
Token community
Feature activation+0.000
 after
Token after
Feature activation+0.000
 claiming
Token claiming
Feature activation+0.000
 men
Token men
Feature activation+0.915
 are
Token are
Feature activation+0.000
 âĢ
Token âĢ
Feature activation+0.000
 ľ
Tokenľ
Feature activation+0.000
 hard
Tokenhard
Feature activation+0.000
 wired
Tokenwired
Feature activation+0.000
across
Token across
Feature activation+0.000
 the
Token the
Feature activation+0.000
 globe
Token globe
Feature activation+0.000
 .
Token.
Feature activation+0.000
 <|endoftext|>
Token<|endoftext|>
Feature activation+0.000
 Children
TokenChildren
Feature activation+0.958
 in
Token in
Feature activation+0.000
 lower
Token lower
Feature activation+0.000
 -
Token-
Feature activation+0.000
 income
Tokenincome
Feature activation+0.000
 families
Token families
Feature activation+0.000
to
Token to
Feature activation+0.000
 women
Token women
Feature activation+0.000
 âĢ
TokenâĢ
Feature activation+0.000
 Ļ
TokenĻ
Feature activation+0.000
 s
Tokens
Feature activation+0.000
 hair
Token hair
Feature activation+0.929
 and
Token and
Feature activation+0.000
 noted
Token noted
Feature activation+0.000
 that
Token that
Feature activation+0.000
 âĢ
Token âĢ
Feature activation+0.000
 ľ
Tokenľ
Feature activation+0.000

INTERVAL 0.384 - 0.769
CONTAINS 0.013%

balance
Token balance
Feature activation+0.000
 between
Token between
Feature activation+0.000
 their
Token their
Feature activation+0.109
 jobs
Token jobs
Feature activation+0.249
 and
Token and
Feature activation+0.000
 families
Token families
Feature activation+0.765
 .
Token.
Feature activation+0.000
 "
Token "
Feature activation+0.000
 Because
TokenBecause
Feature activation+0.000
 of
Token of
Feature activation+0.000
 the
Token the
Feature activation+0.000
woman
Token woman
Feature activation+0.181
 who
Token who
Feature activation+0.000
 said
Token said
Feature activation+0.000
 she
Token she
Feature activation+0.000
 was
Token was
Feature activation+0.000
 raped
Token raped
Feature activation+0.618
 while
Token while
Feature activation+0.000
 a
Token a
Feature activation+0.000
 student
Token student
Feature activation+0.000
 at
Token at
Feature activation+0.000
 the
Token the
Feature activation+0.000
obese
Token obese
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 she
Token she
Feature activation+0.112
 'll
Token'll
Feature activation+0.000
 hand
Token hand
Feature activation+0.109
 them
Token them
Feature activation+0.611
 a
Token a
Feature activation+0.000
 letter
Token letter
Feature activation+0.000
 to
Token to
Feature activation+0.000
 give
Token give
Feature activation+0.000
 their
Token their
Feature activation+0.475
ľ
Tokenľ
Feature activation+0.000
 pot
Tokenpot
Feature activation+0.000
 ent
Tokenent
Feature activation+0.000
 benefits
Token benefits
Feature activation+0.000
 of
Token of
Feature activation+0.000
 human
Token human
Feature activation+0.736
 milk
Token milk
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 âĢ
TokenâĢ
Feature activation+0.000
 Ŀ
TokenĿ
Feature activation+0.000
 all
Token all
Feature activation+0.000
.
Token.
Feature activation+0.000
 I
Token I
Feature activation+0.000
 was
Token was
Feature activation+0.000
 aware
Token aware
Feature activation+0.000
 of
Token of
Feature activation+0.000
 people
Token people
Feature activation+0.441
 practicing
Token practicing
Feature activation+0.000
 it
Token it
Feature activation+0.000
 behind
Token behind
Feature activation+0.000
 closed
Token closed
Feature activation+0.000
 doors
Token doors
Feature activation+0.000

INTERVAL 0.000 - 0.384
CONTAINS 99.968%

harb
Token harb
Feature activation+0.000
 ored
Tokenored
Feature activation+0.000
 feature
Token feature
Feature activation+0.000
 dreams
Token dreams
Feature activation+0.000
 .
Token.
Feature activation+0.000
 With
Token With
Feature activation+0.000
 his
Token his
Feature activation+0.000
 FX
Token FX
Feature activation+0.000
 background
Token background
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 he
Token he
Feature activation+0.000
,
Token,
Feature activation+0.000
 260
Token260
Feature activation+0.000
 counterfeit
Token counterfeit
Feature activation+0.000
 G
Token G
Feature activation+0.000
 HD
TokenHD
Feature activation+0.000
 hair
Token hair
Feature activation+0.000
 -
Token-
Feature activation+0.000
 straight
Tokenstraight
Feature activation+0.000
 ening
Tokenening
Feature activation+0.000
 ir
Token ir
Feature activation+0.000
 ons
Tokenons
Feature activation+0.000
being
Token being
Feature activation+0.000
 taken
Token taken
Feature activation+0.000
 out
Token out
Feature activation+0.000
 context
Token context
Feature activation+0.000
 or
Token or
Feature activation+0.000
 something
Token something
Feature activation+0.000
 .
Token.
Feature activation+0.000
 https
Token https
Feature activation+0.000
 ://
Token://
Feature activation+0.000
 t
Tokent
Feature activation+0.000
 .
Token.
Feature activation+0.000
's
Token's
Feature activation+0.000
 moments
Token moments
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 even
Token even
Feature activation+0.000
 if
Token if
Feature activation+0.000
 England
Token England
Feature activation+0.000
 were
Token were
Feature activation+0.000
 behind
Token behind
Feature activation+0.000
 within
Token within
Feature activation+0.000
 30
Token 30
Feature activation+0.000
 seconds
Token seconds
Feature activation+0.000
just
Token just
Feature activation+0.000
 under
Token under
Feature activation+0.000
 1
Token 1
Feature activation+0.000
 gallon
Token gallon
Feature activation+0.000
 so
Token so
Feature activation+0.000
 I
Token I
Feature activation+0.000
 could
Token could
Feature activation+0.000
 fit
Token fit
Feature activation+0.000
 400
Token 400
Feature activation+0.000
 mL
Token mL
Feature activation+0.000
 of
Token of
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 H1.5: (feature 17102)

TOP ACTIVATIONS
MAX = 2.335

how
Token how
Feature activation+0.000
 long
Token long
Feature activation+0.000
 a
Token a
Feature activation+0.000
 dog
Token dog
Feature activation+0.000
 or
Token or
Feature activation+0.000
 cat
Token cat
Feature activation+2.335
 can
Token can
Feature activation+0.000
 spend
Token spend
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
:
Token:
Feature activation+0.000
 are
Token are
Feature activation+0.000
 dogs
Token dogs
Feature activation+1.682
 smarter
Token smarter
Feature activation+0.000
 than
Token than
Feature activation+0.000
 cats
Token cats
Feature activation+2.071
 because
Token because
Feature activation+0.000
 they
Token they
Feature activation+0.000
 understand
Token understand
Feature activation+0.000
 our
Token our
Feature activation+0.000
 communication
Token communication
Feature activation+0.000
them
Token them
Feature activation+0.000
 to
Token to
Feature activation+0.000
 ,
Token,
Feature activation+0.000
 or
Token or
Feature activation+0.000
 are
Token are
Feature activation+0.000
 cats
Token cats
Feature activation+1.842
 smarter
Token smarter
Feature activation+0.000
 because
Token because
Feature activation+0.000
 they
Token they
Feature activation+0.000
 don
Token don
Feature activation+0.000
 't
Token't
Feature activation+0.000
the
Token the
Feature activation+0.000
 dogs
Token dogs
Feature activation+0.000
 care
Token care
Feature activation+0.000
 and
Token and
Feature activation+0.000
 the
Token the
Feature activation+0.000
 cats
Token cats
Feature activation+1.756
 don
Token don
Feature activation+0.000
 't
Token't
Feature activation+0.000
 .
Token.
Feature activation+0.000
 The
Token The
Feature activation+0.000
 argument
Token argument
Feature activation+0.000
The
Token The
Feature activation+0.000
 argument
Token argument
Feature activation+0.000
 is
Token is
Feature activation+0.000
 :
Token:
Feature activation+0.000
 are
Token are
Feature activation+0.000
 dogs
Token dogs
Feature activation+1.682
 smarter
Token smarter
Feature activation+0.000
 than
Token than
Feature activation+0.000
 cats
Token cats
Feature activation+2.071
 because
Token because
Feature activation+0.000
 they
Token they
Feature activation+0.000
them
Token them
Feature activation+0.000
 meat
Token meat
Feature activation+0.000
 y
Tokeny
Feature activation+0.000
 cat
Token cat
Feature activation+0.088
 or
Token or
Feature activation+0.000
 dog
Token dog
Feature activation+1.456
 food
Token food
Feature activation+0.314
 and
Token and
Feature activation+0.000
 water
Token water
Feature activation+0.024
 âĢĵ
Token âĢĵ
Feature activation+0.000
 not
Token not
Feature activation+0.000
Try
TokenTry
Feature activation+0.006
 to
Token to
Feature activation+0.000
 guess
Token guess
Feature activation+0.107
 what
Token what
Feature activation+0.000
 the
Token the
Feature activation+0.000
 child
Token child
Feature activation+1.402
 has
Token has
Feature activation+0.000
 in
Token in
Feature activation+0.000
 his
Token his
Feature activation+0.086
 hand
Token hand
Feature activation+0.416
 ?
Token?
Feature activation+0.000
fell
Token fell
Feature activation+0.000
 off
Token off
Feature activation+0.000
 and
Token and
Feature activation+0.000
 the
Token the
Feature activation+0.000
 pit
Token pit
Feature activation+0.039
 bull
Token bull
Feature activation+1.398
 âĢ
TokenâĢ
Feature activation+0.000
 Ļ
TokenĻ
Feature activation+0.000
 s
Tokens
Feature activation+0.000
 jaws
Token jaws
Feature activation+0.540
 locked
Token locked
Feature activation+0.000
the
Token the
Feature activation+0.000
 average
Token average
Feature activation+0.000
 stay
Token stay
Feature activation+0.000
 for
Token for
Feature activation+0.000
 a
Token a
Feature activation+0.000
 dog
Token dog
Feature activation+1.341
 ,
Token,
Feature activation+0.000
 the
Token the
Feature activation+0.000
 charity
Token charity
Feature activation+0.000
 said