TOP ACTIVATIONS
MAX = 0.283

The girl eyed it hung rily . She thought it looked
both skipped and jumped mer rily as they made their way
S ki pping mer rily along , Hawk came across
. He pur red hung rily and decided it would be
Polly looked at it hung rily
happy and sailed away mer rily . <|endoftext|> Once there was
watched as she skipped mer rily away into the sunset .
He pointed his magn ifying glass at different classmates and
flashlight , and a magn ifying glass . As
borrowed her mom 's magn ifying glass to take a closer
When Daisy put the magn ifying glass down , the cater
He took out a magn ifying glass and a notebook and
the necklace and the magn ifying glass and told him that
and handed him the magn ifying glass . He
She took out a magn ifying glass and looked very carefully
He brought his magn ifying glass and looked closely .
She had a big magn ifying glass that made things look
it she had a magn ifying glass , a tiny torch
the watch with the magn ifying glass . He saw the
cried , dropping the magn ifying glass . "

MIN ACTIVATIONS
MIN = -0.191

her nest , using tw igs and leaves she found on
. He grabbed different tw igs and branches , some grass
forest and saw many tw igs on the ground . She
home . She gathered tw igs and leaves and rocks ,
nest . She collected tw igs and leaves , and the
scatter the leaves and tw igs he found lying around .
silly , Sara . Tw igs are not w ands ,
he mixed stones and tw igs in the sandbox . Then
thought that looking for tw igs and playing tag sounded fun
cut the sticks and tw igs and soon they had a
a home out of tw igs . K itty
. They drew on the walls and had so much fun
leaves , wood y tw igs , and even a few
away and quickly collected tw igs , then returned to the
to make nests with tw igs and to c aw and
grass and leaves and tw igs . He tried very hard
reached out for some tw igs and made the bed stronger
They walked along the hall ways , looking at the doors
make tree houses from tw igs . He waved it around
The nest has many tw igs . The bird is inside

INTERVAL 0.165 - 0.283
CONTAINS 0.138%

aroo !". The k ang aroo stopped jumping and hopped towards
was the sound of another whale . He looked over and
After eating the sausage , Max went to play
Once there was a t eddy bear . He was very
plan . Every week , Lib by and her mom would

INTERVAL 0.046 - 0.165
CONTAINS 14.740%

needed a regular driver who could help him go places .
always yield , because she knew it was what she needed
and wanted to show her doll to her friend , Tim
also admire the small things that are close to us ,"
couldn 't hold his frustration any longer . He yelled at

INTERVAL -0.073 - 0.046
CONTAINS 80.375%

<|endoftext|> Once upon a time , there was a boy called
blocks , and books . They liked to make up stories
man 's door . She said
" Help !" Tim and Jen scream . They are scared
It looked like a monster ! The dog bark ed at

INTERVAL -0.191 - -0.073
CONTAINS 4.747%

for a walk with his mom and learned about useful squirrel
loudly . All the other birds came to see what was
around in the pond . Everyone was having fun and the
. Every day they would stretch out in their front yard
Ben and Lily did not believe Mom . They thought
Token The
Token ID383
Feature activation-1.48e-02

Pos logit contributions
aroo+0.005
 seal+0.005
 slug+0.004
that+0.004
 pearl+0.004
Neg logit contributions
renches-0.004
 headache-0.004
 older-0.004
oken-0.003
arre-0.003
Token girl
Token ID2576
Feature activation+3.21e-02

Pos logit contributions
 older+0.023
renches+0.021
 exams+0.020
ways+0.020
 questions+0.020
Neg logit contributions
 pearl-0.024
rest-0.023
aroo-0.023
 slug-0.023
cious-0.023
Token eyed
Token ID45320
Feature activation+1.51e-02

Pos logit contributions
 pearl+0.047
rest+0.045
 seal+0.045
aroo+0.044
 globe+0.043
Neg logit contributions
renches-0.054
 older-0.053
ways-0.051
 exams-0.050
arre-0.049
Token it
Token ID340
Feature activation+7.70e-03

Pos logit contributions
 pearl+0.023
 seal+0.022
rest+0.022
aroo+0.022
 globe+0.021
Neg logit contributions
renches-0.024
 older-0.024
 exams-0.023
ways-0.023
 headache-0.022
Token hung
Token ID9174
Feature activation-1.77e-03

Pos logit contributions
 pearl+0.015
aroo+0.015
 seal+0.015
rest+0.015
 slug+0.014
Neg logit contributions
 older-0.009
renches-0.009
ways-0.009
 exams-0.008
 questions-0.008
Tokenrily
Token ID28904
Feature activation+2.83e-01

Pos logit contributions
 older+0.003
renches+0.003
ways+0.003
 exams+0.003
lights+0.003
Neg logit contributions
 pearl-0.003
rest-0.003
 seal-0.002
 slug-0.002
cious-0.002
Token.
Token ID13
Feature activation+2.50e-03

Pos logit contributions
rest+0.566
aroo+0.558
 pearl+0.553
 seal+0.533
 slug+0.531
Neg logit contributions
 older-0.332
renches-0.290
 questions-0.288
 exams-0.282
ways-0.280
Token She
Token ID1375
Feature activation-2.98e-02

Pos logit contributions
aroo+0.004
 seal+0.004
 pearl+0.004
 slug+0.004
rest+0.004
Neg logit contributions
renches-0.003
 older-0.003
 headache-0.003
oken-0.003
arre-0.003
Token thought
Token ID1807
Feature activation-2.02e-02

Pos logit contributions
 older+0.044
renches+0.043
ways+0.041
 exams+0.041
ups+0.040
Neg logit contributions
 pearl-0.049
rest-0.048
aroo-0.047
 globe-0.047
 seal-0.046
Token it
Token ID340
Feature activation+7.67e-03

Pos logit contributions
renches+0.041
 older+0.041
 exams+0.040
ways+0.039
 headache+0.038
Neg logit contributions
 pearl-0.023
 seal-0.023
rest-0.022
aroo-0.022
 slug-0.021
Token looked
Token ID3114
Feature activation-6.18e-02

Pos logit contributions
 pearl+0.011
 seal+0.011
aroo+0.011
rest+0.010
 slug+0.010
Neg logit contributions
renches-0.013
 older-0.013
ways-0.013
 exams-0.012
arre-0.012
Token both
Token ID1111
Feature activation-6.28e-02

Pos logit contributions
 pearl+0.034
rest+0.033
aroo+0.033
 seal+0.033
 slug+0.033
Neg logit contributions
 older-0.023
renches-0.023
ways-0.021
arre-0.021
 questions-0.021
Token skipped
Token ID26684
Feature activation+3.93e-02

Pos logit contributions
 older+0.079
renches+0.074
 nod+0.071
ways+0.070
arre+0.069
Neg logit contributions
 pearl-0.116
 globe-0.114
aroo-0.112
rest-0.111
 seal-0.109
Token and
Token ID290
Feature activation+8.40e-03

Pos logit contributions
 pearl+0.057
 seal+0.055
rest+0.055
aroo+0.054
 slug+0.051
Neg logit contributions
renches-0.068
 older-0.065
 exams-0.062
ways-0.062
arre-0.062
Token jumped
Token ID11687
Feature activation+4.16e-03

Pos logit contributions
 pearl+0.016
rily+0.015
rest+0.015
aroo+0.015
 globe+0.015
Neg logit contributions
 older-0.009
ways-0.009
 brushes-0.009
 exams-0.009
 Jas-0.009
Token mer
Token ID4017
Feature activation+6.51e-02

Pos logit contributions
 pearl+0.006
 seal+0.006
rest+0.006
aroo+0.006
 slug+0.006
Neg logit contributions
renches-0.007
 older-0.007
ways-0.006
 exams-0.006
arre-0.006
Tokenrily
Token ID28904
Feature activation+2.83e-01

Pos logit contributions
 pearl+0.035
 seal+0.032
rest+0.032
aroo+0.032
 slug+0.027
Neg logit contributions
renches-0.170
 older-0.166
ways-0.161
 exams-0.161
arre-0.160
Token as
Token ID355
Feature activation-3.97e-02

Pos logit contributions
 pearl+0.521
rest+0.515
aroo+0.513
 seal+0.504
 slug+0.489
Neg logit contributions
 older-0.364
renches-0.355
ways-0.333
 exams-0.330
 headache-0.325
Token they
Token ID484
Feature activation+5.30e-03

Pos logit contributions
 older+0.054
renches+0.052
ways+0.051
 exams+0.050
ups+0.047
Neg logit contributions
 pearl-0.070
rest-0.068
 seal-0.067
 slug-0.067
aroo-0.066
Token made
Token ID925
Feature activation-6.51e-02

Pos logit contributions
 pearl+0.009
aroo+0.009
 seal+0.009
rest+0.009
 globe+0.009
Neg logit contributions
ways-0.007
renches-0.007
 older-0.006
udd-0.006
ups-0.006
Token their
Token ID511
Feature activation+6.94e-02

Pos logit contributions
 older+0.094
renches+0.093
 exams+0.085
ups+0.084
 questions+0.083
Neg logit contributions
aroo-0.111
rest-0.107
 pearl-0.107
 seal-0.105
 globe-0.105
Token way
Token ID835
Feature activation+2.49e-02

Pos logit contributions
aroo+0.128
rest+0.125
 pearl+0.121
ta+0.119
 seal+0.117
Neg logit contributions
 older-0.096
renches-0.084
 questions-0.082
 exams-0.082
 meetings-0.082
Token
Token ID198
Feature activation-6.84e-03

Pos logit contributions
renches+0.009
 older+0.008
ways+0.008
 exams+0.008
arre+0.008
Neg logit contributions
 pearl-0.012
 seal-0.012
rest-0.012
aroo-0.012
 slug-0.011
TokenS
Token ID50
Feature activation+2.87e-02

Pos logit contributions
renches+0.009
arre+0.008
 older+0.008
ways+0.008
ups+0.008
Neg logit contributions
 pearl-0.013
 seal-0.013
rest-0.013
aroo-0.012
 globe-0.012
Tokenki
Token ID4106
Feature activation-7.88e-02

Pos logit contributions
 pearl+0.053
rest+0.053
aroo+0.051
 seal+0.051
 slug+0.048
Neg logit contributions
 older-0.036
renches-0.036
 exams-0.035
 headache-0.034
ways-0.033
Tokenpping
Token ID2105
Feature activation+2.37e-02

Pos logit contributions
renches+0.123
 older+0.121
arre+0.116
ways+0.113
ups+0.111
Neg logit contributions
 pearl-0.125
 seal-0.121
aroo-0.115
rest-0.114
 slug-0.112
Token mer
Token ID4017
Feature activation+6.66e-02

Pos logit contributions
 pearl+0.041
 seal+0.039
 globe+0.038
aroo+0.038
cious+0.037
Neg logit contributions
 older-0.034
renches-0.033
arre-0.032
 exams-0.031
ups-0.031
Tokenrily
Token ID28904
Feature activation+2.82e-01

Pos logit contributions
 pearl+0.096
 seal+0.094
aroo+0.093
rest+0.093
 globe+0.089
Neg logit contributions
renches-0.114
 older-0.109
ways-0.104
 exams-0.104
arre-0.103
Token along
Token ID1863
Feature activation+4.92e-02

Pos logit contributions
 pearl+0.475
aroo+0.458
rest+0.457
 seal+0.456
 globe+0.442
Neg logit contributions
renches-0.397
 older-0.396
ways-0.377
 exams-0.369
 questions-0.360
Token,
Token ID11
Feature activation-3.47e-02

Pos logit contributions
 pearl+0.084
rest+0.081
 seal+0.081
aroo+0.080
 slug+0.078
Neg logit contributions
 older-0.070
renches-0.070
ways-0.068
 exams-0.067
 questions-0.065
Token Hawk
Token ID26698
Feature activation-3.29e-02

Pos logit contributions
 older+0.050
renches+0.049
ways+0.046
 exams+0.046
arre+0.044
Neg logit contributions
 pearl-0.059
rest-0.058
aroo-0.057
 seal-0.057
 slug-0.055
Token came
Token ID1625
Feature activation-8.87e-02

Pos logit contributions
 older+0.046
renches+0.046
ways+0.043
ups+0.042
 exams+0.041
Neg logit contributions
 pearl-0.058
 seal-0.055
aroo-0.054
rest-0.054
 slug-0.054
Token across
Token ID1973
Feature activation+5.35e-02

Pos logit contributions
 older+0.134
renches+0.126
ways+0.120
 exams+0.120
 questions+0.120
Neg logit contributions
 pearl-0.149
 seal-0.144
aroo-0.144
rest-0.141
 globe-0.138
Token.
Token ID13
Feature activation+6.88e-04

Pos logit contributions
rest+0.166
 pearl+0.160
ta+0.158
Our+0.158
 seal+0.155
Neg logit contributions
 older-0.080
renches-0.075
ups-0.074
 headache-0.069
 crosses-0.068
Token He
Token ID679
Feature activation-2.49e-04

Pos logit contributions
 pearl+0.001
 seal+0.001
 slug+0.001
aroo+0.001
 ruby+0.001
Neg logit contributions
renches-0.001
 older-0.001
ways-0.001
 exams-0.001
ups-0.001
Token pur
Token ID1308
Feature activation-1.40e-01

Pos logit contributions
renches+0.000
 older+0.000
ups+0.000
ways+0.000
 exams+0.000
Neg logit contributions
rest-0.000
 pearl-0.000
aroo-0.000
 globe-0.000
 seal-0.000
Tokenred
Token ID445
Feature activation+1.35e-01

Pos logit contributions
renches+0.289
 older+0.270
ways+0.270
arre+0.269
oken+0.266
Neg logit contributions
 pearl-0.158
 seal-0.155
aroo-0.149
rest-0.144
 globe-0.139
Token hung
Token ID9174
Feature activation-2.07e-03

Pos logit contributions
 pearl+0.269
rest+0.269
aroo+0.261
 seal+0.257
 slug+0.249
Neg logit contributions
 older-0.146
ways-0.145
renches-0.142
 exams-0.136
 questions-0.135
Tokenrily
Token ID28904
Feature activation+2.81e-01

Pos logit contributions
 older+0.003
renches+0.003
 questions+0.003
ints+0.003
 louder+0.003
Neg logit contributions
 seal-0.003
rest-0.003
 pearl-0.003
cious-0.003
Our-0.003
Token and
Token ID290
Feature activation+6.52e-03

Pos logit contributions
rest+0.574
aroo+0.570
 pearl+0.569
 seal+0.548
 slug+0.548
Neg logit contributions
 older-0.308
renches-0.273
 questions-0.268
 exams-0.260
ways-0.256
Token decided
Token ID3066
Feature activation+8.76e-02

Pos logit contributions
 pearl+0.010
rest+0.010
aroo+0.010
 seal+0.010
 globe+0.010
Neg logit contributions
renches-0.010
 older-0.010
ways-0.009
 exams-0.009
 headache-0.009
Token it
Token ID340
Feature activation+3.65e-03

Pos logit contributions
rest+0.194
aroo+0.193
 pearl+0.191
 globe+0.189
 slug+0.189
Neg logit contributions
 older-0.078
ways-0.073
 exams-0.072
ups-0.071
 questions-0.071
Token would
Token ID561
Feature activation+2.31e-02

Pos logit contributions
 pearl+0.007
 seal+0.006
aroo+0.006
 globe+0.006
 coconut+0.006
Neg logit contributions
ways-0.005
ups-0.005
 older-0.005
 crosses-0.005
 questions-0.005
Token be
Token ID307
Feature activation-5.20e-02

Pos logit contributions
 pearl+0.041
 seal+0.041
rest+0.040
aroo+0.040
 slug+0.038
Neg logit contributions
renches-0.032
 older-0.030
 exams-0.028
ways-0.028
arre-0.028
Token Polly
Token ID36898
Feature activation+2.01e-02

Pos logit contributions
 pearl+0.004
 seal+0.004
aroo+0.004
rest+0.004
 slug+0.004
Neg logit contributions
renches-0.003
 older-0.002
ways-0.002
 exams-0.002
 headache-0.002
Token looked
Token ID3114
Feature activation-6.46e-02

Pos logit contributions
 pearl+0.031
rest+0.030
aroo+0.030
 seal+0.029
 globe+0.028
Neg logit contributions
 older-0.031
renches-0.031
arre-0.030
ways-0.029
 exams-0.029
Token at
Token ID379
Feature activation-2.79e-02

Pos logit contributions
 older+0.095
renches+0.090
arre+0.084
ways+0.084
 exams+0.084
Neg logit contributions
aroo-0.111
rest-0.110
 pearl-0.108
 seal-0.107
 slug-0.104
Token it
Token ID340
Feature activation+5.39e-03

Pos logit contributions
 older+0.037
 exams+0.036
 questions+0.034
renches+0.033
 ways+0.031
Neg logit contributions
rest-0.050
cious-0.050
aroo-0.049
 gob-0.047
 pearl-0.047
Token hung
Token ID9174
Feature activation-1.68e-03

Pos logit contributions
aroo+0.011
 pearl+0.011
rest+0.011
 seal+0.011
 slug+0.010
Neg logit contributions
 older-0.006
renches-0.005
ways-0.005
arre-0.005
 questions-0.005
Tokenrily
Token ID28904
Feature activation+2.80e-01

Pos logit contributions
 older+0.003
renches+0.003
ways+0.002
lights+0.002
ups+0.002
Neg logit contributions
 seal-0.002
rest-0.002
aroo-0.002
 pearl-0.002
cious-0.002
Token
Token ID6184
Feature activation+3.14e-02

Pos logit contributions
rest+0.617
aroo+0.605
 pearl+0.586
 slug+0.577
 seal+0.574
Neg logit contributions
 older-0.280
 questions-0.242
renches-0.241
ways-0.227
 exams-0.226
Token
Token ID95
Feature activation-8.44e-03

Pos logit contributions
 pearl+0.060
rest+0.059
 seal+0.059
aroo+0.058
 globe+0.056
Neg logit contributions
renches-0.039
 older-0.036
ways-0.035
 exams-0.035
 headache-0.034
Token
Token ID26391
Feature activation+5.59e-02

Pos logit contributions
renches+0.014
 older+0.013
ways+0.012
 exams+0.012
arre+0.012
Neg logit contributions
 pearl-0.013
 seal-0.013
aroo-0.013
rest-0.013
 slug-0.012
Token
Token ID447
Feature activation+1.14e-01

Pos logit contributions
 pearl+0.075
 seal+0.074
aroo+0.073
cious+0.071
rest+0.071
Neg logit contributions
renches-0.101
 older-0.098
 exams-0.092
oken-0.091
 headache-0.090
Token
Token ID250
Feature activation+4.98e-02

Pos logit contributions
 pearl+0.200
 seal+0.196
rest+0.194
aroo+0.194
 slug+0.185
Neg logit contributions
renches-0.159
 older-0.152
 exams-0.144
ways-0.142
arre-0.141
Token happy
Token ID3772
Feature activation+3.98e-02

Pos logit contributions
 pearl+0.028
 seal+0.027
cious+0.027
aroo+0.027
 slug+0.027
Neg logit contributions
 older-0.023
renches-0.019
oken-0.019
 headache-0.019
 younger-0.019
Token and
Token ID290
Feature activation+7.35e-03

Pos logit contributions
 pearl+0.069
 seal+0.067
rest+0.067
aroo+0.067
 slug+0.064
Neg logit contributions
renches-0.055
 older-0.054
 exams-0.051
ways-0.051
arre-0.050
Token sailed
Token ID33835
Feature activation+5.87e-02

Pos logit contributions
 pearl+0.013
aroo+0.012
 seal+0.012
 slug+0.012
 globe+0.012
Neg logit contributions
 older-0.010
renches-0.009
ways-0.009
ups-0.009
 headache-0.009
Token away
Token ID1497
Feature activation+4.57e-02

Pos logit contributions
 pearl+0.085
 seal+0.083
aroo+0.083
rest+0.082
 slug+0.080
Neg logit contributions
renches-0.098
 older-0.097
ways-0.092
 exams-0.091
arre-0.089
Token mer
Token ID4017
Feature activation+6.94e-02

Pos logit contributions
 pearl+0.085
aroo+0.084
 seal+0.083
rest+0.083
 slug+0.082
Neg logit contributions
 older-0.057
renches-0.054
ways-0.053
 exams-0.051
 headache-0.050
Tokenrily
Token ID28904
Feature activation+2.79e-01

Pos logit contributions
 pearl+0.077
aroo+0.075
rest+0.075
 seal+0.074
 slug+0.071
Neg logit contributions
renches-0.140
 older-0.137
ways-0.132
arre-0.130
 exams-0.130
Token.
Token ID13
Feature activation+3.22e-03

Pos logit contributions
 pearl+0.575
aroo+0.561
rest+0.552
 seal+0.551
 slug+0.548
Neg logit contributions
 older-0.288
renches-0.271
ways-0.265
 exams-0.262
 headache-0.259
Token<|endoftext|>
Token ID50256
Feature activation-4.71e-02

Pos logit contributions
 pearl+0.006
 seal+0.006
rest+0.006
aroo+0.006
 slug+0.006
Neg logit contributions
renches-0.004
 older-0.004
 exams-0.004
ways-0.004
arre-0.004
TokenOnce
Token ID7454
Feature activation+2.20e-02

Pos logit contributions
 older+0.037
ways+0.035
renches+0.035
arre+0.033
 Sometimes+0.031
Neg logit contributions
 pearl-0.110
rest-0.103
 seal-0.102
aroo-0.102
cious-0.101
Token there
Token ID612
Feature activation+3.32e-03

Pos logit contributions
 pearl+0.042
 seal+0.041
aroo+0.041
rest+0.041
 globe+0.039
Neg logit contributions
renches-0.028
 older-0.025
 exams-0.024
ways-0.024
arre-0.024
Token was
Token ID373
Feature activation+1.38e-02

Pos logit contributions
 pearl+0.006
cious+0.006
ifying+0.006
 seal+0.006
rest+0.006
Neg logit contributions
 older-0.004
 nod-0.004
 Sometimes-0.004
 One-0.004
 Despite-0.004
Token watched
Token ID7342
Feature activation-2.49e-02

Pos logit contributions
 older+0.007
renches+0.006
arre+0.006
ways+0.006
ups+0.006
Neg logit contributions
 pearl-0.008
rest-0.007
aroo-0.007
 slug-0.007
cious-0.007
Token as
Token ID355
Feature activation-5.21e-02

Pos logit contributions
renches+0.036
 older+0.036
 exams+0.033
ways+0.033
 headache+0.032
Neg logit contributions
 pearl-0.042
 seal-0.041
rest-0.041
aroo-0.040
 slug-0.039
Token she
Token ID673
Feature activation-2.05e-02

Pos logit contributions
 older+0.082
renches+0.081
 exams+0.076
ways+0.075
 headache+0.074
Neg logit contributions
 pearl-0.081
rest-0.079
 seal-0.078
aroo-0.078
 slug-0.076
Token skipped
Token ID26684
Feature activation+3.62e-02

Pos logit contributions
renches+0.031
 older+0.030
ways+0.028
 exams+0.028
arre+0.028
Neg logit contributions
 pearl-0.033
rest-0.032
aroo-0.032
 seal-0.032
 slug-0.031
Token mer
Token ID4017
Feature activation+6.38e-02

Pos logit contributions
 pearl+0.056
rest+0.054
aroo+0.053
 seal+0.053
 slug+0.052
Neg logit contributions
 older-0.057
renches-0.056
ways-0.055
 exams-0.053
 questions-0.051
Tokenrily
Token ID28904
Feature activation+2.70e-01

Pos logit contributions
rest+0.041
 pearl+0.041
aroo+0.040
 slug+0.038
 seal+0.038
Neg logit contributions
renches-0.155
 older-0.153
ways-0.148
arre-0.146
ups-0.145
Token away
Token ID1497
Feature activation+3.21e-02

Pos logit contributions
 pearl+0.512
rest+0.511
aroo+0.501
 slug+0.493
 seal+0.490
Neg logit contributions
 older-0.329
renches-0.315
ways-0.293
 exams-0.291
 questions-0.289
Token into
Token ID656
Feature activation-3.45e-02

Pos logit contributions
 pearl+0.058
 seal+0.056
rest+0.056
aroo+0.056
 slug+0.054
Neg logit contributions
renches-0.043
 older-0.042
ways-0.039
 exams-0.039
 headache-0.038
Token the
Token ID262
Feature activation-3.21e-02

Pos logit contributions
 older+0.053
renches+0.052
 exams+0.051
 questions+0.048
ways+0.048
Neg logit contributions
aroo-0.054
rest-0.054
 pearl-0.054
 seal-0.053
 slug-0.052
Token sunset
Token ID26428
Feature activation+3.60e-02

Pos logit contributions
 older+0.042
renches+0.041
 exams+0.038
 headache+0.038
ways+0.037
Neg logit contributions
rest-0.058
aroo-0.058
 pearl-0.057
 seal-0.056
cious-0.054
Token.
Token ID13
Feature activation-5.65e-03

Pos logit contributions
 pearl+0.070
 seal+0.069
rest+0.069
aroo+0.068
 slug+0.066
Neg logit contributions
renches-0.042
 older-0.041
ways-0.039
 exams-0.038
arre-0.037
Token
Token ID198
Feature activation-3.14e-03

Pos logit contributions
renches+0.005
 older+0.005
 headache+0.005
arre+0.005
 exams+0.004
Neg logit contributions
aroo-0.005
 pearl-0.005
 seal-0.005
rest-0.005
 globe-0.004
TokenHe
Token ID1544
Feature activation+4.01e-02

Pos logit contributions
renches+0.005
 older+0.005
 exams+0.005
 headache+0.005
arre+0.005
Neg logit contributions
 pearl-0.005
 seal-0.005
aroo-0.005
rest-0.005
 slug-0.004
Token pointed
Token ID6235
Feature activation+6.87e-02

Pos logit contributions
 pearl+0.063
 seal+0.062
aroo+0.060
rest+0.059
 gob+0.057
Neg logit contributions
renches-0.065
 older-0.060
 exams-0.059
 headache-0.059
arre-0.058
Token his
Token ID465
Feature activation+4.96e-02

Pos logit contributions
 pearl+0.119
rest+0.118
aroo+0.117
 seal+0.117
 slug+0.113
Neg logit contributions
renches-0.096
 older-0.095
 exams-0.088
ways-0.087
 headache-0.085
Token magn
Token ID7842
Feature activation-7.68e-03

Pos logit contributions
rest+0.088
 pearl+0.086
 seal+0.086
aroo+0.085
 slug+0.080
Neg logit contributions
renches-0.070
 older-0.068
 exams-0.064
 headache-0.063
 questions-0.062
Tokenifying
Token ID4035
Feature activation+2.66e-01

Pos logit contributions
renches+0.019
 older+0.019
ways+0.018
arre+0.018
oken+0.018
Neg logit contributions
rest-0.005
 seal-0.005
 pearl-0.005
aroo-0.005
 slug-0.004
Token glass
Token ID5405
Feature activation+4.38e-02

Pos logit contributions
 pearl+0.369
 seal+0.355
aroo+0.327
 globe+0.324
 slug+0.323
Neg logit contributions
renches-0.501
ways-0.444
 exams-0.444
oken-0.439
arre-0.437
Token at
Token ID379
Feature activation-2.37e-02

Pos logit contributions
 pearl+0.079
rest+0.078
 seal+0.078
aroo+0.077
 slug+0.073
Neg logit contributions
renches-0.058
 older-0.056
ways-0.053
 exams-0.052
 headache-0.051
Token different
Token ID1180
Feature activation-8.75e-03

Pos logit contributions
 older+0.034
renches+0.034
ways+0.032
 exams+0.032
 questions+0.031
Neg logit contributions
 pearl-0.039
rest-0.039
aroo-0.039
 seal-0.038
cious-0.038
Token classmates
Token ID28999
Feature activation-5.66e-02

Pos logit contributions
renches+0.010
 older+0.010
 exams+0.009
ways+0.009
arre+0.009
Neg logit contributions
 pearl-0.017
 seal-0.017
rest-0.017
aroo-0.017
 slug-0.016
Token and
Token ID290
Feature activation+1.13e-02

Pos logit contributions
 older+0.056
renches+0.053
 exams+0.051
ways+0.051
 questions+0.051
Neg logit contributions
 pearl-0.123
aroo-0.121
rest-0.120
 slug-0.118
 seal-0.118
Token flashlight
Token ID38371
Feature activation-2.31e-02

Pos logit contributions
 older+0.024
 headache+0.022
renches+0.022
 exams+0.021
 younger+0.021
Neg logit contributions
rest-0.030
aroo-0.027
rily-0.026
 pearl-0.026
cream-0.026
Token,
Token ID11
Feature activation-2.90e-02

Pos logit contributions
 older+0.029
ways+0.028
renches+0.027
 exams+0.026
 headache+0.026
Neg logit contributions
 pearl-0.044
rest-0.042
 slug-0.042
 seal-0.041
aroo-0.041
Token and
Token ID290
Feature activation+9.11e-03

Pos logit contributions
renches+0.035
 older+0.035
 exams+0.032
ways+0.032
 questions+0.031
Neg logit contributions
 pearl-0.056
aroo-0.055
rest-0.054
 seal-0.054
 slug-0.053
Token a
Token ID257
Feature activation-1.81e-02

Pos logit contributions
 pearl+0.016
aroo+0.016
rest+0.016
 slug+0.015
 globe+0.015
Neg logit contributions
 older-0.013
renches-0.012
 exams-0.012
 questions-0.012
ways-0.011
Token magn
Token ID7842
Feature activation-6.09e-03

Pos logit contributions
 older+0.026
renches+0.025
 headache+0.025
 exams+0.024
 questions+0.023
Neg logit contributions
rest-0.033
aroo-0.030
 pearl-0.029
rily-0.028
 seal-0.028
Tokenifying
Token ID4035
Feature activation+2.65e-01

Pos logit contributions
renches+0.014
 older+0.014
ways+0.013
oken+0.013
eth+0.013
Neg logit contributions
 slug-0.005
 pearl-0.005
 seal-0.005
rest-0.005
aroo-0.005
Token glass
Token ID5405
Feature activation+4.05e-02

Pos logit contributions
 pearl+0.411
 seal+0.402
rest+0.400
aroo+0.398
 slug+0.378
Neg logit contributions
renches-0.420
 older-0.405
 exams-0.385
ways-0.384
arre-0.380
Token.
Token ID13
Feature activation-1.81e-04

Pos logit contributions
 pearl+0.076
rest+0.076
aroo+0.075
 seal+0.075
 slug+0.072
Neg logit contributions
renches-0.049
 older-0.048
ways-0.045
 exams-0.044
ups-0.044
Token
Token ID198
Feature activation-5.18e-03

Pos logit contributions
renches+0.000
 older+0.000
 exams+0.000
ways+0.000
arre+0.000
Neg logit contributions
 pearl-0.000
 seal-0.000
rest-0.000
aroo-0.000
 slug-0.000
Token
Token ID198
Feature activation-4.90e-03

Pos logit contributions
renches+0.007
 older+0.007
 exams+0.007
ways+0.007
arre+0.007
Neg logit contributions
 pearl-0.009
 seal-0.009
rest-0.009
aroo-0.009
 slug-0.008
TokenAs
Token ID1722
Feature activation-3.09e-03

Pos logit contributions
renches+0.007
 older+0.007
 exams+0.007
ways+0.007
 headache+0.007
Neg logit contributions
 pearl-0.008
 seal-0.008
aroo-0.008
rest-0.008
 slug-0.007
Token borrowed
Token ID22546
Feature activation+4.72e-02

Pos logit contributions
 older+0.042
renches+0.041
ups+0.038
ways+0.038
 questions+0.038
Neg logit contributions
rest-0.042
 pearl-0.041
 globe-0.040
 seal-0.040
aroo-0.040
Token her
Token ID607
Feature activation-7.44e-04

Pos logit contributions
rest+0.076
 pearl+0.075
 seal+0.074
aroo+0.073
 globe+0.072
Neg logit contributions
 older-0.072
renches-0.068
 exams-0.066
 questions-0.065
 younger-0.063
Token mom
Token ID1995
Feature activation-9.54e-02

Pos logit contributions
 older+0.001
 questions+0.001
 younger+0.001
 work+0.001
 eyes+0.001
Neg logit contributions
rest-0.001
ta-0.001
Our-0.001
cream-0.001
 seal-0.001
Token's
Token ID338
Feature activation-2.57e-02

Pos logit contributions
 older+0.060
ups+0.048
 questions+0.043
ocks+0.043
renches+0.042
Neg logit contributions
 pearl-0.248
aroo-0.238
 seal-0.235
rest-0.232
 diamond-0.227
Token magn
Token ID7842
Feature activation-9.37e-03

Pos logit contributions
 older+0.035
 questions+0.033
renches+0.031
 younger+0.030
 ideas+0.030
Neg logit contributions
rest-0.048
aroo-0.046
ta-0.046
 seal-0.044
Our-0.044
Tokenifying
Token ID4035
Feature activation+2.64e-01

Pos logit contributions
renches+0.022
 older+0.022
oken+0.022
ways+0.022
arre+0.021
Neg logit contributions
rest-0.007
 seal-0.007
 pearl-0.006
aroo-0.006
 slug-0.005
Token glass
Token ID5405
Feature activation+4.16e-02

Pos logit contributions
 pearl+0.377
 seal+0.362
aroo+0.346
 slug+0.339
rest+0.336
Neg logit contributions
renches-0.477
 older-0.434
ways-0.432
 exams-0.430
arre-0.426
Token to
Token ID284
Feature activation-1.57e-02

Pos logit contributions
rest+0.081
 pearl+0.080
 seal+0.079
aroo+0.078
cious+0.075
Neg logit contributions
 older-0.048
renches-0.047
ways-0.044
 exams-0.043
ups-0.043
Token take
Token ID1011
Feature activation+1.58e-02

Pos logit contributions
renches+0.020
 older+0.019
 questions+0.018
 exams+0.018
ways+0.018
Neg logit contributions
 pearl-0.029
 seal-0.029
rest-0.029
aroo-0.028
 globe-0.027
Token a
Token ID257
Feature activation-1.54e-02

Pos logit contributions
 pearl+0.027
 seal+0.026
aroo+0.026
rest+0.026
 slug+0.025
Neg logit contributions
renches-0.023
 older-0.022
 exams-0.021
ways-0.021
arre-0.021
Token closer
Token ID5699
Feature activation-2.79e-03

Pos logit contributions
 older+0.019
 headache+0.018
 lecture+0.017
renches+0.017
 sigh+0.017
Neg logit contributions
rest-0.030
aroo-0.028
 pearl-0.027
 seal-0.027
ter-0.027
Token When
Token ID1649
Feature activation-1.29e-02

Pos logit contributions
renches+0.001
 older+0.001
 exams+0.001
ways+0.001
arre+0.001
Neg logit contributions
 pearl-0.002
rest-0.002
 seal-0.002
aroo-0.002
 slug-0.002
Token Daisy
Token ID40355
Feature activation+2.54e-02

Pos logit contributions
 older+0.019
renches+0.018
 exams+0.018
 headache+0.017
ways+0.017
Neg logit contributions
 pearl-0.021
rest-0.021
 seal-0.021
aroo-0.021
 slug-0.020
Token put
Token ID1234
Feature activation+1.59e-02

Pos logit contributions
 pearl+0.044
rest+0.042
aroo+0.042
 globe+0.041
 seal+0.041
Neg logit contributions
 older-0.036
renches-0.035
ways-0.033
 exams-0.033
 headache-0.033
Token the
Token ID262
Feature activation-2.96e-02

Pos logit contributions
 seal+0.023
rest+0.023
cious+0.023
 pearl+0.022
aroo+0.022
Neg logit contributions
 older-0.026
renches-0.025
 exams-0.025
 questions-0.025
 crosses-0.024
Token magn
Token ID7842
Feature activation-7.05e-03

Pos logit contributions
 older+0.046
renches+0.041
 questions+0.041
 exams+0.040
 headache+0.040
Neg logit contributions
rest-0.050
cious-0.046
aroo-0.045
rily-0.045
 pearl-0.044
Tokenifying
Token ID4035
Feature activation+2.64e-01

Pos logit contributions
 older+0.016
renches+0.016
ways+0.016
arre+0.016
lights+0.015
Neg logit contributions
 pearl-0.005
 seal-0.005
rest-0.005
 slug-0.005
aroo-0.005
Token glass
Token ID5405
Feature activation+4.19e-02

Pos logit contributions
rest+0.437
 pearl+0.421
aroo+0.417
 seal+0.414
cious+0.400
Neg logit contributions
 older-0.384
renches-0.378
ways-0.358
 headache-0.358
arre-0.354
Token down
Token ID866
Feature activation+2.30e-02

Pos logit contributions
 pearl+0.080
 seal+0.079
aroo+0.078
rest+0.078
 slug+0.075
Neg logit contributions
renches-0.052
 older-0.049
 exams-0.046
ways-0.045
arre-0.045
Token,
Token ID11
Feature activation-3.14e-02

Pos logit contributions
 pearl+0.045
 seal+0.044
rest+0.044
aroo+0.043
 slug+0.042
Neg logit contributions
renches-0.028
 older-0.026
 exams-0.025
ways-0.024
arre-0.024
Token the
Token ID262
Feature activation-2.76e-02

Pos logit contributions
renches+0.039
 older+0.038
 exams+0.036
ways+0.036
 headache+0.035
Neg logit contributions
 pearl-0.059
 seal-0.058
rest-0.057
aroo-0.057
 slug-0.055
Token cater
Token ID20825
Feature activation+6.92e-02

Pos logit contributions
 older+0.041
renches+0.041
 exams+0.038
 headache+0.038
ways+0.037
Neg logit contributions
rest-0.045
 pearl-0.044
aroo-0.043
 seal-0.043
cious-0.041
Token He
Token ID679
Feature activation-8.17e-04

Pos logit contributions
 seal+0.074
aroo+0.074
 globe+0.072
rest+0.069
 Gor+0.069
Neg logit contributions
renches-0.083
 older-0.078
oken-0.078
 headache-0.076
arre-0.075
Token took
Token ID1718
Feature activation+1.10e-02

Pos logit contributions
renches+0.001
 older+0.001
 exams+0.001
ways+0.001
 headache+0.001
Neg logit contributions
 pearl-0.001
 seal-0.001
rest-0.001
aroo-0.001
 slug-0.001
Token out
Token ID503
Feature activation-3.99e-02

Pos logit contributions
 pearl+0.015
 seal+0.015
rest+0.014
aroo+0.014
 slug+0.014
Neg logit contributions
renches-0.020
 older-0.019
 exams-0.018
ways-0.018
arre-0.018
Token a
Token ID257
Feature activation-1.45e-02

Pos logit contributions
renches+0.082
arre+0.077
 headache+0.076
ways+0.075
 older+0.075
Neg logit contributions
 seal-0.044
 pearl-0.044
aroo-0.040
rest-0.040
 globe-0.039
Token magn
Token ID7842
Feature activation-8.06e-03

Pos logit contributions
 older+0.020
renches+0.019
 headache+0.019
 exams+0.018
 questions+0.018
Neg logit contributions
rest-0.028
aroo-0.025
 seal-0.025
ter-0.024
 pearl-0.024
Tokenifying
Token ID4035
Feature activation+2.64e-01

Pos logit contributions
renches+0.021
 older+0.020
ways+0.020
arre+0.020
 exams+0.019
Neg logit contributions
rest-0.005
 pearl-0.004
 seal-0.004
aroo-0.004
 slug-0.004
Token glass
Token ID5405
Feature activation+4.35e-02

Pos logit contributions
 pearl+0.392
 seal+0.384
aroo+0.368
rest+0.358
 globe+0.355
Neg logit contributions
renches-0.454
 older-0.420
 exams-0.411
ways-0.406
arre-0.404
Token and
Token ID290
Feature activation+1.11e-02

Pos logit contributions
 pearl+0.081
rest+0.081
 seal+0.080
aroo+0.079
 slug+0.077
Neg logit contributions
renches-0.053
 older-0.052
ways-0.049
 exams-0.048
 headache-0.047
Token a
Token ID257
Feature activation-1.45e-02

Pos logit contributions
 pearl+0.018
 seal+0.018
rest+0.018
aroo+0.017
 slug+0.017
Neg logit contributions
renches-0.017
 older-0.016
 exams-0.015
ways-0.015
arre-0.015
Token notebook
Token ID20922
Feature activation+2.80e-02

Pos logit contributions
 older+0.021
 headache+0.020
renches+0.020
 exams+0.019
 questions+0.018
Neg logit contributions
rest-0.027
aroo-0.024
 seal-0.023
cream-0.023
rily-0.023
Token and
Token ID290
Feature activation+1.09e-02

Pos logit contributions
 pearl+0.052
rest+0.051
 seal+0.051
aroo+0.051
 slug+0.051
Neg logit contributions
 older-0.034
renches-0.034
ways-0.033
 exams-0.031
ups-0.030
Token the
Token ID262
Feature activation-2.64e-02

Pos logit contributions
 pearl+0.000
rest+0.000
aroo+0.000
 slug+0.000
 seal+0.000
Neg logit contributions
 older-0.000
 exams-0.000
 questions-0.000
renches-0.000
ups-0.000
Token necklace
Token ID38987
Feature activation+5.77e-02

Pos logit contributions
 older+0.043
renches+0.043
 exams+0.040
 headache+0.039
oken+0.039
Neg logit contributions
rest-0.040
aroo-0.039
 seal-0.038
 pearl-0.038
 slug-0.037
Token and
Token ID290
Feature activation+7.31e-03

Pos logit contributions
 pearl+0.135
 seal+0.130
rest+0.129
aroo+0.129
 slug+0.127
Neg logit contributions
 older-0.045
ways-0.039
ups-0.039
renches-0.033
lights-0.032
Token the
Token ID262
Feature activation-2.74e-02

Pos logit contributions
rest+0.014
 pearl+0.013
aroo+0.013
 seal+0.013
 globe+0.013
Neg logit contributions
 older-0.009
renches-0.009
 exams-0.008
ways-0.008
ups-0.008
Token magn
Token ID7842
Feature activation-9.42e-03

Pos logit contributions
 older+0.044
renches+0.043
 exams+0.040
 headache+0.039
 questions+0.039
Neg logit contributions
rest-0.044
aroo-0.042
 pearl-0.040
 seal-0.040
 slug-0.039
Tokenifying
Token ID4035
Feature activation+2.63e-01

Pos logit contributions
ways+0.020
renches+0.019
oms+0.019
 older+0.019
eth+0.019
Neg logit contributions
 slug-0.009
 seal-0.009
 pearl-0.009
 Gor-0.009
rest-0.009
Token glass
Token ID5405
Feature activation+4.12e-02

Pos logit contributions
 pearl+0.412
rest+0.410
aroo+0.405
 seal+0.404
 slug+0.382
Neg logit contributions
renches-0.405
 older-0.398
ways-0.373
 exams-0.373
 headache-0.372
Token and
Token ID290
Feature activation+7.18e-03

Pos logit contributions
rest+0.085
 pearl+0.083
aroo+0.083
cious+0.082
 seal+0.081
Neg logit contributions
 older-0.041
renches-0.038
ups-0.038
ways-0.037
lights-0.036
Token told
Token ID1297
Feature activation-1.22e-02

Pos logit contributions
rest+0.013
aroo+0.013
 pearl+0.012
 seal+0.012
 globe+0.012
Neg logit contributions
 older-0.010
renches-0.009
 exams-0.009
ways-0.009
ups-0.009
Token him
Token ID683
Feature activation+2.41e-04

Pos logit contributions
renches+0.020
 older+0.019
 exams+0.018
ways+0.018
arre+0.018
Neg logit contributions
 pearl-0.019
 seal-0.019
rest-0.018
aroo-0.018
 slug-0.017
Token that
Token ID326
Feature activation+6.38e-02

Pos logit contributions
 pearl+0.000
aroo+0.000
rest+0.000
 seal+0.000
 slug+0.000
Neg logit contributions
 older-0.000
renches-0.000
 exams-0.000
ways-0.000
ups-0.000
Token and
Token ID290
Feature activation+7.24e-03

Pos logit contributions
 pearl+0.098
 seal+0.097
rest+0.095
aroo+0.095
 slug+0.091
Neg logit contributions
renches-0.063
 older-0.058
 exams-0.056
ways-0.055
arre-0.055
Token handed
Token ID10158
Feature activation+6.60e-02

Pos logit contributions
 pearl+0.013
rest+0.013
 seal+0.012
 globe+0.012
aroo+0.012
Neg logit contributions
 older-0.010
renches-0.009
ways-0.009
 exams-0.009
ups-0.009
Token him
Token ID683
Feature activation-3.68e-04

Pos logit contributions
aroo+0.130
rest+0.127
 pearl+0.125
cious+0.124
 slug+0.124
Neg logit contributions
 older-0.080
 questions-0.075
 exams-0.071
renches-0.070
ups-0.069
Token the
Token ID262
Feature activation-2.44e-02

Pos logit contributions
 older+0.001
 questions+0.000
ups+0.000
 exams+0.000
renches+0.000
Neg logit contributions
 pearl-0.001
aroo-0.001
rest-0.001
 slug-0.001
 seal-0.001
Token magn
Token ID7842
Feature activation-9.90e-03

Pos logit contributions
 older+0.036
renches+0.034
 brushes+0.032
 questions+0.032
oken+0.031
Neg logit contributions
rest-0.042
aroo-0.041
cream-0.040
 Gor-0.039
 seal-0.038
Tokenifying
Token ID4035
Feature activation+2.63e-01

Pos logit contributions
ways+0.021
renches+0.021
 older+0.020
lights+0.020
eth+0.020
Neg logit contributions
 seal-0.009
rest-0.009
 slug-0.009
 Gor-0.009
 pearl-0.009
Token glass
Token ID5405
Feature activation+4.09e-02

Pos logit contributions
 pearl+0.394
 seal+0.385
aroo+0.378
rest+0.378
 slug+0.359
Neg logit contributions
renches-0.437
 older-0.418
 exams-0.401
ways-0.398
arre-0.395
Token.
Token ID13
Feature activation+5.03e-03

Pos logit contributions
rest+0.084
 pearl+0.083
aroo+0.081
 seal+0.081
cious+0.081
Neg logit contributions
 older-0.042
renches-0.039
ups-0.037
ways-0.037
 exams-0.035
Token
Token ID198
Feature activation-6.50e-03

Pos logit contributions
 seal+0.007
 pearl+0.007
aroo+0.007
rily+0.007
 ruby+0.007
Neg logit contributions
renches-0.008
oken-0.008
arre-0.007
ups-0.007
 headache-0.007
Token
Token ID198
Feature activation-6.38e-03

Pos logit contributions
renches+0.011
 older+0.010
 exams+0.010
arre+0.010
 headache+0.010
Neg logit contributions
 seal-0.010
 pearl-0.009
aroo-0.009
rest-0.009
 globe-0.009
TokenHe
Token ID1544
Feature activation+3.84e-02

Pos logit contributions
renches+0.010
 older+0.009
 headache+0.009
oken+0.009
 exams+0.008
Neg logit contributions
 seal-0.011
 pearl-0.010
aroo-0.010
cious-0.010
ta-0.010
Token She
Token ID1375
Feature activation-2.67e-02

Pos logit contributions
aroo+0.009
 seal+0.009
cious+0.009
 slug+0.008
 pearl+0.008
Neg logit contributions
renches-0.005
 headache-0.005
oken-0.005
 older-0.005
lights-0.005
Token took
Token ID1718
Feature activation+1.22e-02

Pos logit contributions
renches+0.041
 older+0.040
ways+0.037
 exams+0.037
arre+0.037
Neg logit contributions
 pearl-0.043
 seal-0.042
rest-0.042
aroo-0.042
 slug-0.040
Token out
Token ID503
Feature activation-4.00e-02

Pos logit contributions
 pearl+0.019
 seal+0.018
aroo+0.018
rest+0.018
 slug+0.017
Neg logit contributions
renches-0.020
 older-0.019
 exams-0.018
ways-0.018
arre-0.018
Token a
Token ID257
Feature activation-1.45e-02

Pos logit contributions
renches+0.074
arre+0.068
 headache+0.067
ways+0.067
 older+0.067
Neg logit contributions
 pearl-0.053
 seal-0.052
aroo-0.050
rest-0.050
 diamond-0.046
Token magn
Token ID7842
Feature activation-9.18e-03

Pos logit contributions
 older+0.020
renches+0.019
 headache+0.019
 questions+0.017
 exams+0.017
Neg logit contributions
rest-0.028
aroo-0.025
ter-0.024
 seal-0.024
rily-0.024
Tokenifying
Token ID4035
Feature activation+2.63e-01

Pos logit contributions
renches+0.024
 older+0.024
ways+0.023
arre+0.023
oken+0.023
Neg logit contributions
 pearl-0.005
 seal-0.005
rest-0.005
aroo-0.004
 slug-0.004
Token glass
Token ID5405
Feature activation+4.06e-02

Pos logit contributions
 pearl+0.376
 seal+0.367
aroo+0.350
 globe+0.339
rest+0.334
Neg logit contributions
renches-0.475
 older-0.431
 exams-0.430
ways-0.425
arre-0.424
Token and
Token ID290
Feature activation+8.25e-03

Pos logit contributions
 pearl+0.077
rest+0.076
 seal+0.076
aroo+0.075
 slug+0.072
Neg logit contributions
renches-0.049
 older-0.048
ways-0.045
 exams-0.044
arre-0.043
Token looked
Token ID3114
Feature activation-6.37e-02

Pos logit contributions
 pearl+0.013
 seal+0.013
rest+0.013
aroo+0.013
 slug+0.012
Neg logit contributions
renches-0.013
 older-0.012
 exams-0.012
ways-0.012
arre-0.011
Token very
Token ID845
Feature activation+3.48e-02

Pos logit contributions
renches+0.109
 older+0.108
ways+0.102
 exams+0.101
 questions+0.099
Neg logit contributions
 pearl-0.089
rest-0.089
aroo-0.089
 seal-0.088
 slug-0.085
Token carefully
Token ID7773
Feature activation-4.63e-02

Pos logit contributions
 pearl+0.048
 seal+0.044
 diamond+0.042
rest+0.041
rily+0.041
Neg logit contributions
renches-0.063
arre-0.061
 headache-0.059
irens-0.058
 exams-0.058
Token
Token ID198
Feature activation-5.81e-03

Pos logit contributions
renches+0.009
 older+0.009
 exams+0.008
ways+0.008
arre+0.008
Neg logit contributions
 pearl-0.010
 seal-0.010
aroo-0.010
rest-0.010
 slug-0.009
TokenHe
Token ID1544
Feature activation+4.06e-02

Pos logit contributions
renches+0.010
 older+0.010
 headache+0.009
 exams+0.009
oken+0.009
Neg logit contributions
 seal-0.009
aroo-0.009
 pearl-0.008
rest-0.008
 slug-0.008
Token brought
Token ID3181
Feature activation+1.07e-02

Pos logit contributions
 pearl+0.062
 seal+0.061
aroo+0.060
rest+0.060
 slug+0.057
Neg logit contributions
renches-0.067
 older-0.063
 exams-0.062
arre-0.061
 headache-0.061
Token his
Token ID465
Feature activation+4.61e-02

Pos logit contributions
 pearl+0.017
 seal+0.016
rest+0.016
cious+0.016
aroo+0.016
Neg logit contributions
 older-0.016
renches-0.016
 exams-0.015
 questions-0.015
 younger-0.014
Token magn
Token ID7842
Feature activation-7.86e-03

Pos logit contributions
rest+0.094
 pearl+0.090
aroo+0.089
rily+0.088
ta+0.087
Neg logit contributions
 older-0.056
 exams-0.047
 younger-0.047
 headache-0.046
 questions-0.046
Tokenifying
Token ID4035
Feature activation+2.63e-01

Pos logit contributions
renches+0.019
 older+0.019
ways+0.019
arre+0.019
oken+0.018
Neg logit contributions
rest-0.005
 pearl-0.005
 seal-0.005
aroo-0.004
 slug-0.004
Token glass
Token ID5405
Feature activation+4.34e-02

Pos logit contributions
 pearl+0.359
 seal+0.346
aroo+0.323
 coconut+0.317
 slug+0.314
Neg logit contributions
renches-0.500
 exams-0.451
ways-0.446
arre-0.442
oken-0.439
Token and
Token ID290
Feature activation+6.13e-03

Pos logit contributions
 pearl+0.074
 seal+0.072
rest+0.071
aroo+0.071
 slug+0.068
Neg logit contributions
renches-0.063
 older-0.061
 exams-0.057
ways-0.057
arre-0.056
Token looked
Token ID3114
Feature activation-6.32e-02

Pos logit contributions
 pearl+0.011
rest+0.011
aroo+0.010
 seal+0.010
 globe+0.010
Neg logit contributions
renches-0.008
 older-0.008
ways-0.008
 exams-0.008
ups-0.008
Token closely
Token ID7173
Feature activation+3.25e-02

Pos logit contributions
renches+0.105
 older+0.105
ways+0.098
 questions+0.096
 exams+0.096
Neg logit contributions
rest-0.091
 pearl-0.090
aroo-0.090
 seal-0.089
 slug-0.086
Token.
Token ID13
Feature activation+1.54e-03

Pos logit contributions
 pearl+0.065
rest+0.065
aroo+0.065
 seal+0.064
 slug+0.063
Neg logit contributions
renches-0.036
 older-0.036
ways-0.033
 questions-0.032
 exams-0.031
Token She
Token ID1375
Feature activation-3.27e-02

Pos logit contributions
 pearl+0.004
 seal+0.004
aroo+0.004
rest+0.004
 slug+0.004
Neg logit contributions
renches-0.003
 older-0.002
 exams-0.002
ways-0.002
arre-0.002
Token had
Token ID550
Feature activation+8.60e-02

Pos logit contributions
renches+0.061
 older+0.058
 exams+0.056
 headache+0.056
arre+0.056
Neg logit contributions
 seal-0.042
 pearl-0.042
rest-0.040
aroo-0.040
 slug-0.038
Token a
Token ID257
Feature activation-1.45e-02

Pos logit contributions
 pearl+0.132
 seal+0.129
aroo+0.129
rest+0.129
 slug+0.122
Neg logit contributions
renches-0.139
 older-0.135
 exams-0.129
ways-0.128
arre-0.125
Token big
Token ID1263
Feature activation+1.19e-02

Pos logit contributions
renches+0.021
 older+0.021
 headache+0.020
 exams+0.020
arre+0.019
Neg logit contributions
rest-0.025
 pearl-0.024
aroo-0.024
 seal-0.023
 slug-0.022
Token magn
Token ID7842
Feature activation-1.00e-02

Pos logit contributions
 pearl+0.021
 seal+0.020
 globe+0.019
 diamond+0.019
 slug+0.019
Neg logit contributions
renches-0.018
ups-0.018
ways-0.017
arre-0.016
 older-0.016
Tokenifying
Token ID4035
Feature activation+2.62e-01

Pos logit contributions
renches+0.026
 older+0.026
ways+0.025
arre+0.024
 exams+0.024
Neg logit contributions
 pearl-0.005
rest-0.005
 seal-0.005
aroo-0.005
 slug-0.005
Token glass
Token ID5405
Feature activation+4.21e-02

Pos logit contributions
 pearl+0.353
 seal+0.347
aroo+0.317
 globe+0.316
 diamond+0.311
Neg logit contributions
renches-0.501
ups-0.455
 exams-0.454
arre-0.446
 older-0.443
Token that
Token ID326
Feature activation+6.33e-02

Pos logit contributions
 pearl+0.076
rest+0.076
 seal+0.074
aroo+0.074
 slug+0.072
Neg logit contributions
renches-0.054
 older-0.053
ways-0.050
 exams-0.050
 headache-0.048
Token made
Token ID925
Feature activation-6.31e-02

Pos logit contributions
 pearl+0.097
 seal+0.095
rest+0.095
aroo+0.094
 slug+0.089
Neg logit contributions
renches-0.102
 older-0.098
 exams-0.093
arre-0.093
ways-0.092
Token things
Token ID1243
Feature activation-6.53e-02

Pos logit contributions
renches+0.102
arre+0.095
 older+0.095
 headache+0.092
 exams+0.092
Neg logit contributions
 pearl-0.095
rest-0.095
 seal-0.094
aroo-0.092
 globe-0.087
Token look
Token ID804
Feature activation-6.34e-02

Pos logit contributions
renches+0.099
 older+0.095
 exams+0.090
ways+0.090
 headache+0.088
Neg logit contributions
 pearl-0.107
 seal-0.104
aroo-0.104
rest-0.103
 slug-0.099
Token it
Token ID340
Feature activation+4.06e-03

Pos logit contributions
 older+0.050
renches+0.050
 exams+0.048
 headache+0.046
ways+0.046
Neg logit contributions
 pearl-0.057
rest-0.057
 seal-0.056
aroo-0.055
 slug-0.053
Token she
Token ID673
Feature activation-1.36e-02

Pos logit contributions
 pearl+0.007
rest+0.007
 seal+0.007
aroo+0.007
 globe+0.007
Neg logit contributions
renches-0.006
 older-0.005
 exams-0.005
arre-0.005
 headache-0.005
Token had
Token ID550
Feature activation+8.50e-02

Pos logit contributions
renches+0.023
 older+0.023
ways+0.022
 exams+0.021
arre+0.021
Neg logit contributions
 pearl-0.019
aroo-0.019
rest-0.019
 seal-0.019
 slug-0.018
Token a
Token ID257
Feature activation-1.57e-02

Pos logit contributions
aroo+0.133
 pearl+0.131
 seal+0.129
rest+0.128
 slug+0.125
Neg logit contributions
renches-0.132
 older-0.132
ways-0.125
 exams-0.124
ups-0.122
Token magn
Token ID7842
Feature activation-8.78e-03

Pos logit contributions
 older+0.021
 headache+0.021
renches+0.020
 exams+0.019
 younger+0.018
Neg logit contributions
rest-0.031
aroo-0.028
cream-0.027
rily-0.026
ter-0.026
Tokenifying
Token ID4035
Feature activation+2.62e-01

Pos logit contributions
renches+0.021
 older+0.021
ways+0.021
oken+0.020
arre+0.020
Neg logit contributions
 pearl-0.006
 seal-0.006
rest-0.005
 slug-0.005
aroo-0.005
Token glass
Token ID5405
Feature activation+4.19e-02

Pos logit contributions
 pearl+0.379
 seal+0.370
aroo+0.353
rest+0.345
 globe+0.342
Neg logit contributions
renches-0.464
 older-0.424
 exams-0.423
ways-0.416
ups-0.415
Token,
Token ID11
Feature activation-3.56e-02

Pos logit contributions
rest+0.077
 pearl+0.076
aroo+0.074
 seal+0.074
 slug+0.071
Neg logit contributions
 older-0.053
renches-0.052
ways-0.050
ups-0.048
 exams-0.048
Token a
Token ID257
Feature activation-1.55e-02

Pos logit contributions
renches+0.048
 older+0.048
 exams+0.045
ways+0.044
 brushes+0.042
Neg logit contributions
rest-0.062
 pearl-0.062
aroo-0.062
 seal-0.061
 globe-0.058
Token tiny
Token ID7009
Feature activation+2.94e-02

Pos logit contributions
 older+0.022
 headache+0.021
renches+0.019
 younger+0.019
 exams+0.019
Neg logit contributions
rest-0.031
cream-0.027
aroo-0.027
rily-0.026
that-0.025
Token torch
Token ID28034
Feature activation-3.10e-02

Pos logit contributions
rest+0.046
aroo+0.042
 pearl+0.041
 seal+0.041
cious+0.039
Neg logit contributions
 older-0.047
renches-0.047
 headache-0.044
 exams-0.044
ways-0.043
Token the
Token ID262
Feature activation-2.51e-02

Pos logit contributions
renches+0.044
 older+0.044
 exams+0.041
ways+0.040
 headache+0.039
Neg logit contributions
 pearl-0.047
rest-0.045
 seal-0.045
cious-0.045
aroo-0.044
Token watch
Token ID2342
Feature activation-3.48e-02

Pos logit contributions
renches+0.037
 older+0.036
 exams+0.034
ways+0.033
 headache+0.033
Neg logit contributions
 pearl-0.042
rest-0.041
 seal-0.041
aroo-0.041
 slug-0.039
Token with
Token ID351
Feature activation+9.45e-03

Pos logit contributions
 older+0.042
renches+0.042
ways+0.040
 exams+0.038
ups+0.037
Neg logit contributions
 pearl-0.068
 seal-0.065
rest-0.065
aroo-0.064
 slug-0.063
Token the
Token ID262
Feature activation-2.47e-02

Pos logit contributions
 pearl+0.014
 seal+0.013
rest+0.013
aroo+0.013
 slug+0.013
Neg logit contributions
renches-0.016
 older-0.016
 exams-0.015
ways-0.015
 headache-0.015
Token magn
Token ID7842
Feature activation-9.57e-03

Pos logit contributions
renches+0.037
 older+0.034
arre+0.033
ways+0.033
 exams+0.033
Neg logit contributions
 pearl-0.042
 seal-0.041
aroo-0.040
rest-0.039
 globe-0.038
Tokenifying
Token ID4035
Feature activation+2.61e-01

Pos logit contributions
renches+0.025
 older+0.024
ways+0.024
arre+0.023
oken+0.023
Neg logit contributions
 pearl-0.005
 seal-0.005
rest-0.005
aroo-0.004
 slug-0.004
Token glass
Token ID5405
Feature activation+4.13e-02

Pos logit contributions
 pearl+0.379
 seal+0.372
aroo+0.360
rest+0.353
 globe+0.344
Neg logit contributions
renches-0.454
 older-0.423
 exams-0.414
ways-0.408
arre-0.407
Token.
Token ID13
Feature activation+9.94e-04

Pos logit contributions
 pearl+0.077
rest+0.076
 seal+0.076
aroo+0.075
 slug+0.072
Neg logit contributions
renches-0.051
 older-0.050
ways-0.047
 exams-0.046
 headache-0.045
Token He
Token ID679
Feature activation-4.13e-03

Pos logit contributions
 slug+0.001
 seal+0.001
 pearl+0.001
aroo+0.001
 globe+0.001
Neg logit contributions
arre-0.002
renches-0.002
ups-0.002
 older-0.002
 headache-0.002
Token saw
Token ID2497
Feature activation-7.41e-03

Pos logit contributions
renches+0.007
 older+0.006
 exams+0.006
ways+0.006
arre+0.006
Neg logit contributions
 pearl-0.006
 seal-0.006
aroo-0.006
rest-0.006
 slug-0.006
Token the
Token ID262
Feature activation-2.31e-02

Pos logit contributions
 older+0.011
renches+0.011
 exams+0.010
ways+0.010
ups+0.010
Neg logit contributions
 pearl-0.012
rest-0.012
cious-0.012
 seal-0.011
aroo-0.011
Token cried
Token ID16896
Feature activation-1.27e-01

Pos logit contributions
renches+0.061
 headache+0.058
oken+0.058
lights+0.057
ways+0.055
Neg logit contributions
 seal-0.052
cious-0.049
 pearl-0.049
 slug-0.049
 gob-0.048
Token,
Token ID11
Feature activation-3.26e-02

Pos logit contributions
arre+0.249
renches+0.245
aj+0.244
ways+0.241
 meetings+0.237
Neg logit contributions
ter-0.153
aroo-0.136
 seal-0.131
rest-0.128
 slug-0.124
Token dropping
Token ID12047
Feature activation+1.10e-01

Pos logit contributions
 older+0.052
renches+0.051
 exams+0.048
ways+0.048
 headache+0.047
Neg logit contributions
 pearl-0.051
rest-0.049
 seal-0.048
aroo-0.048
 globe-0.046
Token the
Token ID262
Feature activation-2.07e-02

Pos logit contributions
 seal+0.147
aroo+0.144
 pearl+0.143
rest+0.140
 slug+0.135
Neg logit contributions
renches-0.204
arre-0.191
ways-0.189
 older-0.189
 headache-0.184
Token magn
Token ID7842
Feature activation-1.29e-02

Pos logit contributions
 older+0.033
renches+0.032
 exams+0.030
 headache+0.030
 questions+0.029
Neg logit contributions
rest-0.032
 pearl-0.032
 seal-0.031
aroo-0.031
rily-0.029
Tokenifying
Token ID4035
Feature activation+2.61e-01

Pos logit contributions
renches+0.031
 older+0.031
ways+0.031
arre+0.030
oken+0.030
Neg logit contributions
 pearl-0.009
 seal-0.008
rest-0.008
 slug-0.007
aroo-0.007
Token glass
Token ID5405
Feature activation+3.85e-02

Pos logit contributions
 pearl+0.360
 seal+0.357
aroo+0.340
 globe+0.330
rest+0.326
Neg logit contributions
renches-0.480
 older-0.432
 exams-0.432
arre-0.428
ways-0.426
Token.
Token ID13
Feature activation+7.80e-03

Pos logit contributions
 pearl+0.072
rest+0.071
 seal+0.070
aroo+0.069
cious+0.067
Neg logit contributions
 older-0.048
renches-0.048
ways-0.044
 exams-0.043
 headache-0.043
Token
Token ID198
Feature activation-6.70e-03

Pos logit contributions
aroo+0.010
 seal+0.010
 pearl+0.010
 slug+0.010
 globe+0.010
Neg logit contributions
renches-0.014
 older-0.013
oken-0.013
 exams-0.013
arre-0.013
Token
Token ID198
Feature activation-6.43e-03

Pos logit contributions
renches+0.011
 older+0.011
 exams+0.010
 headache+0.010
arre+0.010
Neg logit contributions
aroo-0.010
 seal-0.009
 pearl-0.009
rest-0.009
 slug-0.009
Token"
Token ID1
Feature activation+5.36e-02

Pos logit contributions
renches+0.011
 older+0.011
 headache+0.010
 exams+0.010
oken+0.010
Neg logit contributions
aroo-0.009
 seal-0.009
 pearl-0.009
rest-0.009
cious-0.008
Token her
Token ID607
Feature activation-8.78e-03

Pos logit contributions
renches+0.147
 older+0.140
arre+0.137
 headache+0.137
 exams+0.134
Neg logit contributions
 pearl-0.120
 seal-0.115
rest-0.112
aroo-0.110
 globe-0.109
Token nest
Token ID16343
Feature activation-4.78e-02

Pos logit contributions
 older+0.011
renches+0.011
 exams+0.010
 headache+0.010
ways+0.010
Neg logit contributions
rest-0.017
 pearl-0.016
aroo-0.016
 seal-0.016
 slug-0.015
Token,
Token ID11
Feature activation-3.76e-02

Pos logit contributions
renches+0.074
 older+0.070
 exams+0.068
 headache+0.067
arre+0.066
Neg logit contributions
 pearl-0.076
rest-0.074
 seal-0.074
aroo-0.074
 globe-0.070
Token using
Token ID1262
Feature activation+1.81e-02

Pos logit contributions
renches+0.055
 older+0.053
 exams+0.050
ways+0.050
arre+0.049
Neg logit contributions
 pearl-0.064
 seal-0.062
rest-0.062
aroo-0.061
 slug-0.059
Token tw
Token ID665
Feature activation+1.66e-02

Pos logit contributions
 pearl+0.030
 seal+0.030
rest+0.030
aroo+0.030
 slug+0.028
Neg logit contributions
renches-0.026
 older-0.025
 exams-0.024
ways-0.024
 headache-0.023
Tokenigs
Token ID9235
Feature activation-1.90e-01

Pos logit contributions
 pearl+0.037
aroo+0.036
rest+0.036
rily+0.035
cious+0.034
Neg logit contributions
 older-0.014
 headache-0.013
renches-0.013
 exams-0.013
arre-0.012
Token and
Token ID290
Feature activation+4.03e-03

Pos logit contributions
renches+0.234
 older+0.228
ways+0.215
 exams+0.210
arre+0.208
Neg logit contributions
 pearl-0.361
 seal-0.355
aroo-0.348
rest-0.347
 slug-0.338
Token leaves
Token ID5667
Feature activation-1.25e-01

Pos logit contributions
 pearl+0.007
 seal+0.007
rest+0.007
aroo+0.007
 slug+0.007
Neg logit contributions
renches-0.005
 older-0.005
 exams-0.005
ways-0.005
arre-0.005
Token she
Token ID673
Feature activation-2.16e-02

Pos logit contributions
renches+0.154
 older+0.150
ways+0.144
 exams+0.137
ups+0.136
Neg logit contributions
 pearl-0.239
 seal-0.234
rest-0.228
aroo-0.227
 slug-0.223
Token found
Token ID1043
Feature activation-1.22e-02

Pos logit contributions
renches+0.036
 older+0.035
ways+0.034
 exams+0.033
ups+0.033
Neg logit contributions
 pearl-0.032
 seal-0.032
aroo-0.031
rest-0.030
 slug-0.030
Token on
Token ID319
Feature activation+6.82e-02

Pos logit contributions
 older+0.017
renches+0.017
ups+0.017
ways+0.016
 questions+0.016
Neg logit contributions
 seal-0.021
 pearl-0.021
aroo-0.020
 slug-0.020
rest-0.019
Token.
Token ID13
Feature activation+1.05e-03

Pos logit contributions
renches+0.066
 older+0.063
 exams+0.060
ways+0.060
arre+0.060
Neg logit contributions
 pearl-0.062
 seal-0.061
rest-0.060
aroo-0.059
 slug-0.057
Token He
Token ID679
Feature activation-3.70e-03

Pos logit contributions
 pearl+0.002
 seal+0.002
rest+0.002
aroo+0.002
 slug+0.002
Neg logit contributions
renches-0.001
 older-0.001
 exams-0.001
ways-0.001
arre-0.001
Token grabbed
Token ID13176
Feature activation+5.57e-02

Pos logit contributions
renches+0.006
 older+0.006
ways+0.005
 exams+0.005
arre+0.005
Neg logit contributions
 pearl-0.006
rest-0.006
 seal-0.006
aroo-0.006
 slug-0.005
Token different
Token ID1180
Feature activation-9.45e-03

Pos logit contributions
 pearl+0.086
 seal+0.085
rest+0.084
aroo+0.084
cious+0.080
Neg logit contributions
renches-0.091
 older-0.088
 exams-0.084
ways-0.081
 questions-0.080
Token tw
Token ID665
Feature activation+1.84e-02

Pos logit contributions
renches+0.012
 older+0.011
 headache+0.011
arre+0.011
ways+0.011
Neg logit contributions
 pearl-0.018
 seal-0.017
rest-0.017
aroo-0.016
 slug-0.016
Tokenigs
Token ID9235
Feature activation-1.90e-01

Pos logit contributions
 pearl+0.042
rest+0.041
aroo+0.040
 seal+0.040
cious+0.039
Neg logit contributions
 older-0.015
renches-0.015
 headache-0.014
arre-0.013
 exams-0.013
Token and
Token ID290
Feature activation+6.89e-03

Pos logit contributions
 older+0.196
renches+0.194
ways+0.183
arre+0.174
ups+0.170
Neg logit contributions
 pearl-0.394
 seal-0.387
aroo-0.380
rest-0.379
 slug-0.373
Token branches
Token ID13737
Feature activation-1.18e-01

Pos logit contributions
 pearl+0.012
 seal+0.012
rest+0.012
aroo+0.012
 slug+0.011
Neg logit contributions
renches-0.010
 older-0.009
 exams-0.009
ways-0.009
arre-0.009
Token,
Token ID11
Feature activation-3.52e-02

Pos logit contributions
 older+0.125
renches+0.125
ways+0.122
ups+0.112
 exams+0.107
Neg logit contributions
 pearl-0.242
 seal-0.236
rest-0.229
 slug-0.228
aroo-0.227
Token some
Token ID617
Feature activation-7.08e-02

Pos logit contributions
renches+0.049
 older+0.047
 exams+0.044
ways+0.044
arre+0.043
Neg logit contributions
 pearl-0.062
 seal-0.061
rest-0.060
aroo-0.060
 slug-0.057
Token grass
Token ID8701
Feature activation-2.68e-03

Pos logit contributions
renches+0.101
 older+0.098
 exams+0.092
ways+0.091
 headache+0.090
Neg logit contributions
 pearl-0.121
 seal-0.119
rest-0.118
aroo-0.117
 slug-0.112
Token forest
Token ID8222
Feature activation-6.47e-02

Pos logit contributions
 older+0.052
renches+0.049
 exams+0.046
 headache+0.045
 younger+0.045
Neg logit contributions
rest-0.058
aroo-0.056
 pearl-0.055
 seal-0.055
cious-0.053
Token and
Token ID290
Feature activation+4.95e-03

Pos logit contributions
renches+0.096
 older+0.085
 exams+0.084
 headache+0.084
oken+0.083
Neg logit contributions
 pearl-0.111
 seal-0.110
aroo-0.108
rest-0.106
 globe-0.105
Token saw
Token ID2497
Feature activation-1.20e-02

Pos logit contributions
 pearl+0.008
rest+0.008
aroo+0.008
 seal+0.008
 slug+0.008
Neg logit contributions
 older-0.007
ways-0.007
renches-0.007
 exams-0.006
ups-0.006
Token many
Token ID867
Feature activation-6.18e-03

Pos logit contributions
 older+0.017
renches+0.016
 exams+0.015
 questions+0.015
ups+0.015
Neg logit contributions
rest-0.020
 pearl-0.020
 seal-0.020
aroo-0.020
 globe-0.020
Token tw
Token ID665
Feature activation+1.65e-02

Pos logit contributions
renches+0.010
 older+0.009
ways+0.009
arre+0.009
 exams+0.009
Neg logit contributions
 pearl-0.010
 seal-0.010
rest-0.010
aroo-0.010
 slug-0.009
Tokenigs
Token ID9235
Feature activation-1.89e-01

Pos logit contributions
 pearl+0.036
aroo+0.035
rest+0.035
 perched+0.033
cious+0.033
Neg logit contributions
 older-0.016
renches-0.015
 headache-0.015
 exams-0.015
arre-0.014
Token on
Token ID319
Feature activation+6.73e-02

Pos logit contributions
 older+0.211
renches+0.199
ways+0.198
ups+0.184
arre+0.184
Neg logit contributions
 pearl-0.378
 seal-0.370
aroo-0.364
 slug-0.358
rest-0.358
Token the
Token ID262
Feature activation-3.37e-02

Pos logit contributions
 pearl+0.110
rest+0.110
 seal+0.108
aroo+0.106
cious+0.102
Neg logit contributions
 older-0.099
renches-0.099
ways-0.094
 exams-0.093
 questions-0.088
Token ground
Token ID2323
Feature activation-2.83e-02

Pos logit contributions
 older+0.041
renches+0.040
ups+0.036
 younger+0.036
 exams+0.035
Neg logit contributions
 pearl-0.063
rest-0.063
aroo-0.063
 seal-0.062
 slug-0.059
Token.
Token ID13
Feature activation+1.43e-03

Pos logit contributions
renches+0.032
 older+0.032
ways+0.030
 exams+0.029
ups+0.028
Neg logit contributions
 pearl-0.056
 seal-0.055
rest-0.054
aroo-0.054
 slug-0.052
Token She
Token ID1375
Feature activation-3.80e-02

Pos logit contributions
 seal+0.003
 pearl+0.003
aroo+0.003
 slug+0.003
rest+0.003
Neg logit contributions
renches-0.002
 older-0.001
 headache-0.001
arre-0.001
oken-0.001
Token home
Token ID1363
Feature activation+6.22e-03

Pos logit contributions
 older+0.042
renches+0.040
 exams+0.038
 headache+0.037
 questions+0.037
Neg logit contributions
rest-0.056
aroo-0.055
 pearl-0.054
 seal-0.053
 slug-0.051
Token.
Token ID13
Feature activation+7.37e-04

Pos logit contributions
 pearl+0.012
 seal+0.012
aroo+0.012
rest+0.012
 slug+0.012
Neg logit contributions
 older-0.007
renches-0.007
ways-0.007
 exams-0.007
ups-0.007
Token She
Token ID1375
Feature activation-3.83e-02

Pos logit contributions
 pearl+0.002
rest+0.002
 seal+0.002
aroo+0.001
 globe+0.001
Neg logit contributions
renches-0.001
 older-0.001
 exams-0.001
ways-0.001
arre-0.001
Token gathered
Token ID9272
Feature activation+3.28e-02

Pos logit contributions
renches+0.065
 older+0.063
 exams+0.060
ways+0.060
arre+0.059
Neg logit contributions
 pearl-0.056
rest-0.054
 seal-0.054
aroo-0.054
 slug-0.051
Token tw
Token ID665
Feature activation+1.56e-02

Pos logit contributions
 pearl+0.053
 seal+0.052
rest+0.052
aroo+0.052
 slug+0.049
Neg logit contributions
renches-0.050
 older-0.048
 exams-0.046
ways-0.045
arre-0.044
Tokenigs
Token ID9235
Feature activation-1.88e-01

Pos logit contributions
 pearl+0.038
aroo+0.037
rest+0.037
 seal+0.037
rily+0.035
Neg logit contributions
 older-0.011
renches-0.010
 headache-0.010
 exams-0.009
arre-0.009
Token and
Token ID290
Feature activation+4.40e-03

Pos logit contributions
renches+0.222
 older+0.218
ways+0.203
 exams+0.197
arre+0.196
Neg logit contributions
 pearl-0.366
 seal-0.360
rest-0.354
aroo-0.353
 slug-0.344
Token leaves
Token ID5667
Feature activation-1.23e-01

Pos logit contributions
 pearl+0.008
 seal+0.008
rest+0.008
aroo+0.008
 slug+0.007
Neg logit contributions
renches-0.006
 older-0.006
 exams-0.005
ways-0.005
arre-0.005
Token and
Token ID290
Feature activation+4.77e-03

Pos logit contributions
renches+0.149
 older+0.146
ways+0.139
 exams+0.132
ups+0.131
Neg logit contributions
 pearl-0.237
 seal-0.232
rest-0.228
aroo-0.226
 slug-0.222
Token rocks
Token ID12586
Feature activation-7.38e-02

Pos logit contributions
 pearl+0.008
 seal+0.008
rest+0.008
aroo+0.008
 slug+0.008
Neg logit contributions
renches-0.007
 older-0.006
 exams-0.006
ways-0.006
arre-0.006
Token,
Token ID11
Feature activation-3.57e-02

Pos logit contributions
renches+0.090
 older+0.087
ways+0.083
 exams+0.079
ups+0.079
Neg logit contributions
 pearl-0.142
 seal-0.139
rest-0.137
aroo-0.136
 slug-0.134
Token nest
Token ID16343
Feature activation-4.25e-02

Pos logit contributions
rest+0.109
aroo+0.109
 pearl+0.108
 seal+0.105
 slug+0.105
Neg logit contributions
 older-0.078
renches-0.074
 headache-0.067
 exams-0.067
ups-0.066
Token.
Token ID13
Feature activation-1.60e-03

Pos logit contributions
renches+0.059
 older+0.056
 exams+0.053
ways+0.053
arre+0.052
Neg logit contributions
 pearl-0.075
 seal-0.073
rest-0.073
aroo-0.073
 slug-0.069
Token She
Token ID1375
Feature activation-3.95e-02

Pos logit contributions
renches+0.002
 older+0.002
oken+0.002
 headache+0.002
arre+0.002
Neg logit contributions
 seal-0.003
aroo-0.003
 pearl-0.003
rest-0.003
 Gor-0.003
Token collected
Token ID7723
Feature activation+4.61e-02

Pos logit contributions
renches+0.056
 older+0.056
ways+0.052
ups+0.051
arre+0.050
Neg logit contributions
 pearl-0.066
rest-0.066
 seal-0.063
aroo-0.063
 slug-0.063
Token tw
Token ID665
Feature activation+1.52e-02

Pos logit contributions
rest+0.079
aroo+0.078
 seal+0.078
 pearl+0.077
cious+0.077
Neg logit contributions
 older-0.064
renches-0.063
 questions-0.057
 exams-0.056
ups-0.056
Tokenigs
Token ID9235
Feature activation-1.88e-01

Pos logit contributions
 pearl+0.038
aroo+0.037
 seal+0.037
rest+0.037
 slug+0.036
Neg logit contributions
renches-0.010
 older-0.010
 exams-0.008
 headache-0.008
arre-0.008
Token and
Token ID290
Feature activation+4.41e-03

Pos logit contributions
renches+0.197
 older+0.197
ways+0.183
arre+0.172
 exams+0.172
Neg logit contributions
 pearl-0.384
 seal-0.379
rest-0.374
aroo-0.371
 slug-0.365
Token leaves
Token ID5667
Feature activation-1.22e-01

Pos logit contributions
 pearl+0.008
rest+0.008
 seal+0.008
aroo+0.008
 globe+0.008
Neg logit contributions
renches-0.006
 older-0.005
ways-0.005
ups-0.005
 exams-0.005
Token,
Token ID11
Feature activation-3.74e-02

Pos logit contributions
renches+0.134
 older+0.132
ways+0.128
ups+0.118
 exams+0.115
Neg logit contributions
 pearl-0.246
 seal-0.241
rest-0.236
 slug-0.233
aroo-0.233
Token and
Token ID290
Feature activation+5.17e-03

Pos logit contributions
renches+0.049
 older+0.047
ways+0.044
 exams+0.044
arre+0.043
Neg logit contributions
 pearl-0.069
 seal-0.067
rest-0.067
aroo-0.067
 slug-0.064
Token the
Token ID262
Feature activation-3.05e-02

Pos logit contributions
 pearl+0.009
aroo+0.009
 seal+0.008
rest+0.008
 globe+0.008
Neg logit contributions
 older-0.007
renches-0.007
ways-0.007
ups-0.007
 exams-0.007
Token scatter
Token ID41058
Feature activation+1.90e-02

Pos logit contributions
renches+0.037
 older+0.034
arre+0.034
ways+0.033
ups+0.033
Neg logit contributions
 seal-0.038
 pearl-0.038
aroo-0.037
rest-0.036
 slug-0.035
Token the
Token ID262
Feature activation-3.08e-02

Pos logit contributions
 pearl+0.033
rest+0.033
aroo+0.033
 seal+0.032
cious+0.032
Neg logit contributions
renches-0.026
 older-0.025
 exams-0.024
ways-0.023
arre-0.023
Token leaves
Token ID5667
Feature activation-1.19e-01

Pos logit contributions
renches+0.042
 older+0.042
 exams+0.038
arre+0.037
 headache+0.037
Neg logit contributions
 pearl-0.054
rest-0.053
aroo-0.052
 seal-0.052
cious-0.050
Token and
Token ID290
Feature activation+1.02e-02

Pos logit contributions
 older+0.187
renches+0.179
ways+0.177
ups+0.166
 exams+0.164
Neg logit contributions
 pearl-0.190
 seal-0.183
rest-0.177
aroo-0.174
cious-0.174
Token tw
Token ID665
Feature activation+1.66e-02

Pos logit contributions
 pearl+0.018
rest+0.018
aroo+0.018
 seal+0.018
 slug+0.017
Neg logit contributions
renches-0.014
 older-0.013
 exams-0.013
ways-0.013
 headache-0.012
Tokenigs
Token ID9235
Feature activation-1.87e-01

Pos logit contributions
 pearl+0.038
aroo+0.037
rest+0.037
 seal+0.036
rily+0.036
Neg logit contributions
 older-0.014
renches-0.013
 exams-0.013
 headache-0.012
 questions-0.012
Token he
Token ID339
Feature activation+1.93e-02

Pos logit contributions
 older+0.241
renches+0.238
ways+0.230
arre+0.215
ups+0.215
Neg logit contributions
 pearl-0.342
 seal-0.332
aroo-0.326
 slug-0.323
rest-0.320
Token found
Token ID1043
Feature activation-1.14e-02

Pos logit contributions
 pearl+0.031
 seal+0.029
aroo+0.029
rest+0.029
 slug+0.028
Neg logit contributions
renches-0.030
 older-0.029
ways-0.028
 exams-0.028
ups-0.027
Token lying
Token ID9105
Feature activation+3.68e-02

Pos logit contributions
 older+0.015
ups+0.015
renches+0.015
 questions+0.014
ways+0.014
Neg logit contributions
aroo-0.020
ta-0.020
 seal-0.020
cious-0.019
 pearl-0.019
Token around
Token ID1088
Feature activation-4.68e-02

Pos logit contributions
 pearl+0.053
 seal+0.053
aroo+0.052
cious+0.051
rest+0.051
Neg logit contributions
 older-0.060
ways-0.058
renches-0.056
 questions-0.054
 exams-0.054
Token.
Token ID13
Feature activation+1.71e-03

Pos logit contributions
 older+0.052
ways+0.050
 exams+0.050
renches+0.049
 questions+0.049
Neg logit contributions
 pearl-0.092
aroo-0.090
cious-0.089
rest-0.089
 seal-0.088
Token silly
Token ID14397
Feature activation-2.09e-02

Pos logit contributions
 pearl+0.079
aroo+0.079
 seal+0.076
rest+0.076
 globe+0.075
Neg logit contributions
 older-0.072
renches-0.067
 exams-0.065
 younger-0.062
 questions-0.061
Token,
Token ID11
Feature activation-3.51e-02

Pos logit contributions
renches+0.037
ways+0.034
oken+0.034
 older+0.033
 headache+0.033
Neg logit contributions
 seal-0.029
 pearl-0.028
rest-0.028
aroo-0.027
 slug-0.027
Token Sara
Token ID24799
Feature activation+5.70e-02

Pos logit contributions
 older+0.039
renches+0.036
 exams+0.036
 headache+0.034
 questions+0.033
Neg logit contributions
rest-0.072
 pearl-0.070
aroo-0.068
cious-0.067
 seal-0.066
Token.
Token ID13
Feature activation-7.82e-03

Pos logit contributions
 pearl+0.101
 seal+0.097
aroo+0.096
rest+0.096
 slug+0.092
Neg logit contributions
renches-0.078
 older-0.078
 exams-0.072
arre-0.072
ways-0.071
Token Tw
Token ID1815
Feature activation-2.47e-02

Pos logit contributions
renches+0.015
 older+0.014
ways+0.014
arre+0.014
 headache+0.014
Neg logit contributions
 seal-0.009
aroo-0.009
 pearl-0.009
rest-0.009
 slug-0.009
Tokenigs
Token ID9235
Feature activation-1.87e-01

Pos logit contributions
renches+0.029
 older+0.027
ways+0.026
ups+0.026
 exams+0.026
Neg logit contributions
 pearl-0.049
 seal-0.048
aroo-0.048
rest-0.047
 globe-0.046
Token are
Token ID389
Feature activation+4.71e-02

Pos logit contributions
renches+0.283
 older+0.272
 exams+0.259
ways+0.257
arre+0.254
Neg logit contributions
 pearl-0.303
 seal-0.297
rest-0.293
aroo-0.293
 slug-0.279
Token not
Token ID407
Feature activation-1.20e-01

Pos logit contributions
 pearl+0.082
 seal+0.080
aroo+0.080
rest+0.079
 slug+0.077
Neg logit contributions
renches-0.066
 older-0.066
 exams-0.061
ways-0.059
 headache-0.059
Token w
Token ID266
Feature activation-5.82e-02

Pos logit contributions
 older+0.181
renches+0.175
 exams+0.164
arre+0.159
 headache+0.158
Neg logit contributions
 pearl-0.195
aroo-0.192
 seal-0.191
rest-0.187
 globe-0.185
Tokenands
Token ID1746
Feature activation-8.01e-02

Pos logit contributions
 older+0.061
ways+0.060
renches+0.058
 exams+0.057
 headache+0.056
Neg logit contributions
rest-0.113
 seal-0.111
aroo-0.110
 pearl-0.110
 slug-0.104
Token,
Token ID11
Feature activation-3.72e-02

Pos logit contributions
renches+0.101
 older+0.098
 exams+0.092
ways+0.092
arre+0.090
Neg logit contributions
 pearl-0.152
 seal-0.147
aroo-0.146
rest-0.144
 slug-0.141
Token he
Token ID339
Feature activation+2.24e-02

Pos logit contributions
 pearl+0.026
rest+0.025
 seal+0.025
aroo+0.025
 slug+0.024
Neg logit contributions
renches-0.013
 older-0.013
ways-0.012
 exams-0.012
arre-0.012
Token mixed
Token ID7668
Feature activation+1.29e-02

Pos logit contributions
 pearl+0.039
rest+0.038
aroo+0.038
 slug+0.037
 seal+0.037
Neg logit contributions
 older-0.031
renches-0.030
ways-0.030
 exams-0.028
arre-0.028
Token stones
Token ID14966
Feature activation-5.87e-02

Pos logit contributions
aroo+0.023
rest+0.022
 seal+0.022
 pearl+0.022
cious+0.022
Neg logit contributions
renches-0.018
 older-0.017
ways-0.016
ups-0.016
 exams-0.015
Token and
Token ID290
Feature activation+1.10e-02

Pos logit contributions
 older+0.068
renches+0.068
ways+0.063
ups+0.060
 exams+0.058
Neg logit contributions
 pearl-0.113
rest-0.113
 seal-0.112
aroo-0.112
 slug-0.109
Token tw
Token ID665
Feature activation+1.74e-02

Pos logit contributions
aroo+0.020
rest+0.020
 pearl+0.019
 seal+0.019
 globe+0.018
Neg logit contributions
renches-0.015
 older-0.014
ways-0.014
 exams-0.013
ups-0.013
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.044
 seal+0.043
aroo+0.042
rest+0.042
 slug+0.041
Neg logit contributions
renches-0.011
 older-0.011
 exams-0.010
 headache-0.009
arre-0.009
Token in
Token ID287
Feature activation+2.24e-03

Pos logit contributions
renches+0.213
 older+0.211
ways+0.198
ups+0.187
arre+0.186
Neg logit contributions
 pearl-0.369
 seal-0.362
aroo-0.357
rest-0.354
 slug-0.350
Token the
Token ID262
Feature activation-2.36e-02

Pos logit contributions
aroo+0.004
rest+0.004
cious+0.004
 seal+0.004
 pearl+0.004
Neg logit contributions
 older-0.003
renches-0.003
 exams-0.003
ways-0.003
 questions-0.003
Token sandbox
Token ID35204
Feature activation-2.22e-02

Pos logit contributions
renches+0.031
 older+0.031
ways+0.029
 exams+0.028
 headache+0.028
Neg logit contributions
rest-0.043
aroo-0.042
 pearl-0.042
 seal-0.040
rily-0.039
Token.
Token ID13
Feature activation+4.02e-03

Pos logit contributions
renches+0.028
 older+0.027
ways+0.025
 exams+0.025
arre+0.024
Neg logit contributions
 pearl-0.042
 seal-0.042
aroo-0.041
rest-0.041
 slug-0.040
Token Then
Token ID3244
Feature activation-4.86e-03

Pos logit contributions
 seal+0.007
 pearl+0.007
aroo+0.007
rest+0.007
 slug+0.007
Neg logit contributions
renches-0.005
 older-0.005
 headache-0.005
arre-0.005
 exams-0.005
Token thought
Token ID1807
Feature activation-2.49e-02

Pos logit contributions
renches+0.082
 older+0.077
 exams+0.076
ways+0.074
 headache+0.073
Neg logit contributions
 seal-0.065
 pearl-0.064
rest-0.060
aroo-0.060
 gob-0.059
Token that
Token ID326
Feature activation+5.66e-02

Pos logit contributions
renches+0.044
 older+0.043
 exams+0.041
ways+0.041
arre+0.040
Neg logit contributions
 pearl-0.034
 seal-0.034
rest-0.033
aroo-0.033
 slug-0.031
Token looking
Token ID2045
Feature activation-1.99e-02

Pos logit contributions
 pearl+0.095
aroo+0.092
 seal+0.092
rest+0.091
cious+0.087
Neg logit contributions
renches-0.082
 older-0.082
 exams-0.079
 headache-0.075
arre-0.075
Token for
Token ID329
Feature activation-2.77e-02

Pos logit contributions
renches+0.028
 older+0.028
 exams+0.026
arre+0.025
ways+0.025
Neg logit contributions
 pearl-0.034
aroo-0.034
 seal-0.034
rest-0.033
 slug-0.032
Token tw
Token ID665
Feature activation+1.60e-02

Pos logit contributions
renches+0.039
 older+0.039
 exams+0.036
 questions+0.035
oken+0.035
Neg logit contributions
aroo-0.047
 pearl-0.047
rest-0.046
 seal-0.046
 globe-0.045
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.035
rest+0.035
aroo+0.034
 seal+0.033
cious+0.033
Neg logit contributions
 older-0.014
renches-0.014
 headache-0.013
arre-0.013
 younger-0.013
Token and
Token ID290
Feature activation+2.58e-03

Pos logit contributions
 older+0.192
renches+0.186
ways+0.177
arre+0.169
ups+0.168
Neg logit contributions
 pearl-0.389
 seal-0.376
aroo-0.373
 slug-0.366
rest-0.363
Token playing
Token ID2712
Feature activation+5.43e-03

Pos logit contributions
 pearl+0.005
 seal+0.005
rest+0.005
aroo+0.005
 slug+0.005
Neg logit contributions
renches-0.003
 older-0.003
 exams-0.003
ways-0.003
 headache-0.003
Token tag
Token ID7621
Feature activation+2.71e-02

Pos logit contributions
 pearl+0.009
rest+0.008
 seal+0.008
 slug+0.008
aroo+0.008
Neg logit contributions
renches-0.008
 older-0.008
 exams-0.008
ways-0.008
 questions-0.008
Token sounded
Token ID14846
Feature activation-6.53e-02

Pos logit contributions
 pearl+0.055
aroo+0.053
rest+0.053
 seal+0.053
 slug+0.052
Neg logit contributions
 older-0.029
renches-0.029
ways-0.029
ups-0.027
 questions-0.027
Token fun
Token ID1257
Feature activation-9.87e-02

Pos logit contributions
 older+0.089
renches+0.088
ways+0.082
 exams+0.082
 headache+0.080
Neg logit contributions
 pearl-0.115
aroo-0.114
 seal-0.113
rest-0.111
 slug-0.108
Token cut
Token ID2005
Feature activation-3.10e-02

Pos logit contributions
renches+0.023
 older+0.022
 exams+0.022
arre+0.021
ways+0.021
Neg logit contributions
 pearl-0.029
aroo-0.029
 seal-0.028
rest-0.028
 globe-0.027
Token the
Token ID262
Feature activation-2.78e-02

Pos logit contributions
renches+0.053
 older+0.052
 exams+0.048
ways+0.048
 questions+0.047
Neg logit contributions
 pearl-0.045
 seal-0.045
rest-0.043
aroo-0.043
cious-0.042
Token sticks
Token ID16461
Feature activation-1.08e-01

Pos logit contributions
 older+0.042
renches+0.041
 headache+0.037
 younger+0.036
 exams+0.036
Neg logit contributions
rest-0.046
 pearl-0.045
 seal-0.045
aroo-0.045
cious-0.043
Token and
Token ID290
Feature activation+9.93e-03

Pos logit contributions
renches+0.135
 older+0.134
ways+0.123
arre+0.120
 exams+0.120
Neg logit contributions
 pearl-0.202
 seal-0.200
rest-0.196
aroo-0.195
 slug-0.190
Token tw
Token ID665
Feature activation+1.76e-02

Pos logit contributions
 pearl+0.017
 seal+0.017
aroo+0.016
rest+0.016
 slug+0.016
Neg logit contributions
renches-0.014
 older-0.014
 exams-0.013
ways-0.013
arre-0.013
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.041
aroo+0.041
rest+0.040
 seal+0.039
rily+0.039
Neg logit contributions
 older-0.014
 exams-0.013
renches-0.013
 headache-0.012
ways-0.012
Token and
Token ID290
Feature activation+9.57e-03

Pos logit contributions
renches+0.241
 older+0.234
ways+0.219
arre+0.214
 exams+0.213
Neg logit contributions
 pearl-0.345
 seal-0.338
aroo-0.331
rest-0.329
 slug-0.322
Token soon
Token ID2582
Feature activation-3.96e-03

Pos logit contributions
 pearl+0.017
 seal+0.016
aroo+0.016
rest+0.016
 slug+0.015
Neg logit contributions
renches-0.014
 older-0.013
 exams-0.012
arre-0.012
ways-0.012
Token they
Token ID484
Feature activation+1.65e-03

Pos logit contributions
 older+0.006
renches+0.006
 exams+0.006
ways+0.006
ups+0.006
Neg logit contributions
aroo-0.006
 pearl-0.006
rest-0.006
 seal-0.006
 slug-0.006
Token had
Token ID550
Feature activation+8.66e-02

Pos logit contributions
 pearl+0.003
aroo+0.003
 seal+0.003
 slug+0.003
rest+0.003
Neg logit contributions
renches-0.002
 older-0.002
ways-0.002
arre-0.002
ups-0.002
Token a
Token ID257
Feature activation-1.44e-02

Pos logit contributions
aroo+0.142
 pearl+0.140
 seal+0.140
rest+0.138
 slug+0.133
Neg logit contributions
renches-0.131
 older-0.128
ways-0.120
arre-0.119
 exams-0.119
Token a
Token ID257
Feature activation-1.47e-02

Pos logit contributions
renches+0.076
 older+0.074
 exams+0.070
ways+0.070
 headache+0.068
Neg logit contributions
 pearl-0.073
 seal-0.072
aroo-0.072
rest-0.072
 slug-0.067
Token home
Token ID1363
Feature activation+9.45e-03

Pos logit contributions
 older+0.021
renches+0.021
 headache+0.020
 exams+0.019
arre+0.019
Neg logit contributions
rest-0.026
aroo-0.025
 pearl-0.024
 seal-0.024
cious-0.023
Token out
Token ID503
Feature activation-4.43e-02

Pos logit contributions
 pearl+0.018
 seal+0.017
aroo+0.017
 slug+0.017
rest+0.017
Neg logit contributions
 older-0.012
ways-0.012
renches-0.011
 exams-0.011
ups-0.011
Token of
Token ID286
Feature activation+4.17e-02

Pos logit contributions
 older+0.050
renches+0.047
ways+0.045
 exams+0.045
 questions+0.045
Neg logit contributions
 pearl-0.091
 seal-0.089
 slug-0.087
rest-0.086
aroo-0.086
Token tw
Token ID665
Feature activation+1.73e-02

Pos logit contributions
rest+0.071
aroo+0.069
 seal+0.068
cious+0.067
 pearl+0.067
Neg logit contributions
 older-0.061
renches-0.059
 exams-0.056
 questions-0.055
 headache-0.055
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.034
aroo+0.032
 seal+0.031
rest+0.031
rily+0.031
Neg logit contributions
renches-0.021
 older-0.021
 exams-0.020
 headache-0.019
ways-0.019
Token.
Token ID13
Feature activation+4.05e-03

Pos logit contributions
 older+0.201
renches+0.187
ways+0.183
arre+0.175
ups+0.172
Neg logit contributions
 pearl-0.380
 seal-0.368
aroo-0.365
rest-0.365
 slug-0.362
Token
Token ID198
Feature activation-3.37e-03

Pos logit contributions
 seal+0.007
 pearl+0.007
aroo+0.007
 slug+0.007
rest+0.007
Neg logit contributions
renches-0.005
 older-0.005
 headache-0.005
oken-0.005
 exams-0.005
Token
Token ID198
Feature activation-3.48e-03

Pos logit contributions
renches+0.005
 older+0.005
 exams+0.004
ways+0.004
arre+0.004
Neg logit contributions
 pearl-0.006
 seal-0.006
rest-0.006
aroo-0.006
 slug-0.005
TokenK
Token ID42
Feature activation+4.89e-02

Pos logit contributions
renches+0.006
 older+0.005
 exams+0.005
ways+0.005
arre+0.005
Neg logit contributions
 pearl-0.005
 seal-0.005
rest-0.005
aroo-0.005
 slug-0.005
Tokenitty
Token ID9760
Feature activation+2.92e-02

Pos logit contributions
 pearl+0.070
 seal+0.069
rest+0.067
aroo+0.067
 slug+0.063
Neg logit contributions
renches-0.085
 older-0.082
 exams-0.078
ways-0.078
arre-0.077
Token.
Token ID13
Feature activation-1.20e-02

Pos logit contributions
 older+0.021
renches+0.019
 questions+0.017
ways+0.017
 headache+0.017
Neg logit contributions
aroo-0.039
rest-0.039
 pearl-0.039
 seal-0.038
 slug-0.037
Token They
Token ID1119
Feature activation-4.50e-02

Pos logit contributions
 older+0.013
renches+0.013
ways+0.012
 exams+0.012
arre+0.012
Neg logit contributions
 pearl-0.024
rest-0.024
 seal-0.023
aroo-0.023
 globe-0.022
Token drew
Token ID9859
Feature activation-5.53e-02

Pos logit contributions
renches+0.068
 older+0.066
ways+0.063
arre+0.062
 exams+0.062
Neg logit contributions
 pearl-0.073
rest-0.070
aroo-0.070
 seal-0.069
 slug-0.067
Token on
Token ID319
Feature activation+5.59e-02

Pos logit contributions
 older+0.077
renches+0.076
ways+0.070
 questions+0.070
 exams+0.069
Neg logit contributions
 pearl-0.096
aroo-0.096
rest-0.096
 slug-0.093
 seal-0.093
Token the
Token ID262
Feature activation-3.17e-02

Pos logit contributions
rest+0.094
 pearl+0.092
aroo+0.091
 seal+0.086
 slug+0.086
Neg logit contributions
renches-0.082
 older-0.081
 exams-0.079
ways-0.077
 questions-0.075
Token walls
Token ID7714
Feature activation-1.86e-01

Pos logit contributions
 older+0.044
renches+0.043
 exams+0.040
 questions+0.039
 headache+0.039
Neg logit contributions
rest-0.057
aroo-0.054
 pearl-0.053
 slug-0.051
 seal-0.051
Token and
Token ID290
Feature activation+6.12e-03

Pos logit contributions
renches+0.241
 older+0.224
 exams+0.215
arre+0.212
 headache+0.211
Neg logit contributions
 pearl-0.345
 seal-0.340
rest-0.336
aroo-0.335
 globe-0.323
Token had
Token ID550
Feature activation+8.36e-02

Pos logit contributions
 pearl+0.010
 seal+0.010
rest+0.010
aroo+0.010
 slug+0.009
Neg logit contributions
renches-0.009
 older-0.009
 exams-0.009
ways-0.009
arre-0.008
Token so
Token ID523
Feature activation+3.89e-02

Pos logit contributions
 pearl+0.137
aroo+0.135
 seal+0.134
rest+0.134
 slug+0.127
Neg logit contributions
renches-0.126
 older-0.122
 exams-0.116
ways-0.116
arre-0.113
Token much
Token ID881
Feature activation+1.23e-02

Pos logit contributions
 pearl+0.067
 seal+0.066
rest+0.066
aroo+0.065
 globe+0.062
Neg logit contributions
renches-0.055
 older-0.053
 exams-0.050
ways-0.050
arre-0.049
Token fun
Token ID1257
Feature activation-9.98e-02

Pos logit contributions
 pearl+0.021
 diamond+0.020
 seal+0.020
 globe+0.019
 ruby+0.019
Neg logit contributions
renches-0.017
arre-0.017
eth-0.016
 headache-0.016
 exams-0.015
Token leaves
Token ID5667
Feature activation-1.20e-01

Pos logit contributions
 pearl+0.040
aroo+0.040
 seal+0.040
rest+0.039
 slug+0.038
Neg logit contributions
renches-0.025
 older-0.025
ways-0.023
 exams-0.023
arre-0.022
Token,
Token ID11
Feature activation-3.59e-02

Pos logit contributions
 older+0.155
renches+0.152
ways+0.144
ups+0.135
arre+0.134
Neg logit contributions
 pearl-0.218
 seal-0.216
aroo-0.213
rest-0.210
 slug-0.208
Token wood
Token ID4898
Feature activation-7.63e-02

Pos logit contributions
 older+0.052
renches+0.051
 exams+0.048
ways+0.047
 questions+0.046
Neg logit contributions
aroo-0.060
 pearl-0.059
rest-0.059
 seal-0.059
 globe-0.056
Tokeny
Token ID88
Feature activation+1.41e-01

Pos logit contributions
renches+0.089
 older+0.088
ways+0.082
 exams+0.078
ups+0.078
Neg logit contributions
 pearl-0.149
 seal-0.146
rest-0.145
aroo-0.143
 slug-0.140
Token tw
Token ID665
Feature activation+1.79e-02

Pos logit contributions
 pearl+0.269
rest+0.268
aroo+0.268
 seal+0.265
 slug+0.252
Neg logit contributions
renches-0.170
 older-0.169
ways-0.159
 exams-0.157
 questions-0.151
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.043
aroo+0.042
rest+0.042
 seal+0.041
rily+0.041
Neg logit contributions
 older-0.013
renches-0.012
 headache-0.012
 exams-0.011
arre-0.011
Token,
Token ID11
Feature activation-3.61e-02

Pos logit contributions
renches+0.201
 older+0.200
ways+0.189
arre+0.178
ups+0.174
Neg logit contributions
 pearl-0.377
 seal-0.370
aroo-0.365
rest-0.360
 slug-0.358
Token and
Token ID290
Feature activation+6.26e-03

Pos logit contributions
renches+0.044
 older+0.044
ways+0.040
 exams+0.039
 questions+0.038
Neg logit contributions
 pearl-0.068
aroo-0.068
rest-0.068
 seal-0.067
 slug-0.065
Token even
Token ID772
Feature activation-1.24e-02

Pos logit contributions
 pearl+0.010
aroo+0.010
rest+0.010
 seal+0.010
 slug+0.010
Neg logit contributions
renches-0.009
 older-0.009
 exams-0.008
ways-0.008
 questions-0.008
Token a
Token ID257
Feature activation-1.66e-02

Pos logit contributions
 older+0.018
renches+0.018
 exams+0.017
ways+0.017
 questions+0.016
Neg logit contributions
 pearl-0.021
aroo-0.021
 seal-0.020
rest-0.020
 globe-0.020
Token few
Token ID1178
Feature activation+3.13e-02

Pos logit contributions
 older+0.025
renches+0.024
 headache+0.023
 exams+0.023
ways+0.023
Neg logit contributions
rest-0.028
aroo-0.026
 pearl-0.026
 seal-0.026
 slug-0.024
Token away
Token ID1497
Feature activation+3.77e-02

Pos logit contributions
 older+0.017
renches+0.017
ways+0.016
 exams+0.016
arre+0.015
Neg logit contributions
 pearl-0.017
rest-0.016
 seal-0.016
aroo-0.016
 slug-0.015
Token and
Token ID290
Feature activation+2.61e-03

Pos logit contributions
 pearl+0.070
 seal+0.069
rest+0.069
aroo+0.068
 slug+0.066
Neg logit contributions
renches-0.047
 older-0.047
ways-0.043
 exams-0.043
arre-0.042
Token quickly
Token ID2952
Feature activation-2.86e-02

Pos logit contributions
 pearl+0.004
rest+0.004
 globe+0.004
 seal+0.004
cious+0.004
Neg logit contributions
 older-0.004
ways-0.003
renches-0.003
ups-0.003
 exams-0.003
Token collected
Token ID7723
Feature activation+4.56e-02

Pos logit contributions
renches+0.043
 older+0.042
ways+0.041
ups+0.039
 headache+0.038
Neg logit contributions
 pearl-0.045
rest-0.045
aroo-0.044
 seal-0.044
 globe-0.043
Token tw
Token ID665
Feature activation+1.74e-02

Pos logit contributions
aroo+0.077
rest+0.077
cious+0.077
 seal+0.076
 pearl+0.075
Neg logit contributions
 older-0.064
renches-0.063
 questions-0.057
 exams-0.056
ups-0.056
Tokenigs
Token ID9235
Feature activation-1.86e-01

Pos logit contributions
 pearl+0.044
aroo+0.043
rest+0.042
 seal+0.042
 slug+0.041
Neg logit contributions
renches-0.011
 older-0.010
 exams-0.009
arre-0.009
 headache-0.009
Token,
Token ID11
Feature activation-3.68e-02

Pos logit contributions
 older+0.204
ways+0.192
renches+0.192
ups+0.184
arre+0.173
Neg logit contributions
 pearl-0.372
 seal-0.365
aroo-0.359
 slug-0.357
rest-0.355
Token then
Token ID788
Feature activation+1.24e-02

Pos logit contributions
 older+0.050
renches+0.050
ways+0.046
 exams+0.045
ups+0.045
Neg logit contributions
 pearl-0.065
aroo-0.063
 seal-0.063
rest-0.063
 slug-0.061
Token returned
Token ID4504
Feature activation+9.76e-02

Pos logit contributions
rest+0.019
 pearl+0.019
 seal+0.019
aroo+0.019
 globe+0.019
Neg logit contributions
 older-0.019
renches-0.018
ups-0.017
ways-0.017
 crosses-0.017
Token to
Token ID284
Feature activation-1.95e-02

Pos logit contributions
rest+0.178
 pearl+0.176
 seal+0.175
aroo+0.174
 slug+0.169
Neg logit contributions
 older-0.123
renches-0.118
ways-0.116
ups-0.113
 exams-0.112
Token the
Token ID262
Feature activation-3.01e-02

Pos logit contributions
 older+0.027
 younger+0.024
 meetings+0.023
 questions+0.023
renches+0.023
Neg logit contributions
rest-0.036
aroo-0.035
cious-0.034
 globe-0.034
 slug-0.034
Token to
Token ID284
Feature activation-1.87e-02

Pos logit contributions
renches+0.094
 older+0.091
 exams+0.087
ways+0.087
arre+0.086
Neg logit contributions
 pearl-0.071
 seal-0.069
rest-0.068
aroo-0.068
 slug-0.065
Token make
Token ID787
Feature activation-4.79e-02

Pos logit contributions
renches+0.029
 older+0.027
ways+0.026
arre+0.026
 exams+0.025
Neg logit contributions
 pearl-0.031
 seal-0.031
aroo-0.030
rest-0.030
 globe-0.029
Token nests
Token ID44382
Feature activation-1.34e-01

Pos logit contributions
renches+0.071
 older+0.070
 exams+0.065
 questions+0.064
ups+0.063
Neg logit contributions
aroo-0.079
 pearl-0.079
rest-0.078
 seal-0.077
 globe-0.075
Token with
Token ID351
Feature activation+8.66e-03

Pos logit contributions
renches+0.183
 older+0.181
ways+0.169
 exams+0.166
arre+0.163
Neg logit contributions
 pearl-0.235
 seal-0.230
aroo-0.228
rest-0.226
 slug-0.221
Token tw
Token ID665
Feature activation+1.74e-02

Pos logit contributions
aroo+0.014
rest+0.014
cious+0.014
 pearl+0.014
 slug+0.014
Neg logit contributions
 older-0.012
renches-0.012
 questions-0.012
 exams-0.011
 chores-0.011
Tokenigs
Token ID9235
Feature activation-1.85e-01

Pos logit contributions
 pearl+0.038
rest+0.037
aroo+0.036
cious+0.035
 seal+0.035
Neg logit contributions
 older-0.016
 headache-0.015
renches-0.015
 exams-0.015
arre-0.014
Token and
Token ID290
Feature activation+6.65e-03

Pos logit contributions
renches+0.209
 older+0.205
ways+0.190
 exams+0.186
arre+0.184
Neg logit contributions
 pearl-0.370
 seal-0.362
aroo-0.359
rest-0.357
 slug-0.349
Token to
Token ID284
Feature activation-1.79e-02

Pos logit contributions
 pearl+0.012
 seal+0.012
rest+0.011
aroo+0.011
 slug+0.011
Neg logit contributions
renches-0.009
 older-0.009
 exams-0.008
ways-0.008
arre-0.008
Token c
Token ID269
Feature activation+1.00e-01

Pos logit contributions
renches+0.025
 older+0.024
 exams+0.023
ways+0.023
arre+0.023
Neg logit contributions
 pearl-0.031
 seal-0.030
aroo-0.030
rest-0.030
 slug-0.029
Tokenaw
Token ID707
Feature activation-3.80e-02

Pos logit contributions
 pearl+0.193
 seal+0.185
cream+0.184
rily+0.184
rest+0.182
Neg logit contributions
 older-0.120
 exams-0.116
renches-0.111
 younger-0.106
 headache-0.101
Token and
Token ID290
Feature activation+6.53e-03

Pos logit contributions
renches+0.042
ways+0.042
 older+0.041
ups+0.039
 exams+0.037
Neg logit contributions
 pearl-0.077
 seal-0.074
rest-0.074
aroo-0.074
 slug-0.073
Token grass
Token ID8701
Feature activation+1.31e-03

Pos logit contributions
rest+0.078
 pearl+0.078
aroo+0.077
 seal+0.077
 slug+0.074
Neg logit contributions
renches-0.051
 older-0.048
 exams-0.044
ways-0.043
 headache-0.042
Token and
Token ID290
Feature activation+4.39e-03

Pos logit contributions
 pearl+0.003
rest+0.003
aroo+0.003
 seal+0.003
cious+0.003
Neg logit contributions
 older-0.001
ways-0.001
renches-0.001
ups-0.001
lights-0.001
Token leaves
Token ID5667
Feature activation-1.21e-01

Pos logit contributions
 pearl+0.008
rest+0.008
aroo+0.008
 seal+0.008
 slug+0.007
Neg logit contributions
renches-0.006
 older-0.006
 exams-0.005
ways-0.005
ups-0.005
Token and
Token ID290
Feature activation+4.51e-03

Pos logit contributions
renches+0.135
 older+0.132
ways+0.126
ups+0.118
 exams+0.114
Neg logit contributions
 pearl-0.246
 seal-0.239
rest-0.237
aroo-0.234
 slug-0.232
Token tw
Token ID665
Feature activation+1.71e-02

Pos logit contributions
 pearl+0.009
rest+0.008
aroo+0.008
 seal+0.008
 slug+0.008
Neg logit contributions
renches-0.006
 older-0.005
ways-0.005
 exams-0.005
ups-0.005
Tokenigs
Token ID9235
Feature activation-1.85e-01

Pos logit contributions
 pearl+0.040
aroo+0.039
 seal+0.038
rest+0.038
rily+0.037
Neg logit contributions
 older-0.014
 exams-0.013
 headache-0.013
renches-0.013
arre-0.012
Token.
Token ID13
Feature activation-4.76e-03

Pos logit contributions
renches+0.209
 older+0.205
ways+0.192
ups+0.183
arre+0.182
Neg logit contributions
 pearl-0.371
 seal-0.362
aroo-0.357
rest-0.354
 slug-0.351
Token He
Token ID679
Feature activation-3.15e-03

Pos logit contributions
renches+0.006
 older+0.006
 exams+0.006
ways+0.006
arre+0.006
Neg logit contributions
 pearl-0.009
 seal-0.008
rest-0.008
aroo-0.008
 slug-0.008
Token tried
Token ID3088
Feature activation-4.13e-02

Pos logit contributions
renches+0.005
 older+0.005
ways+0.005
 exams+0.005
ups+0.005
Neg logit contributions
 pearl-0.005
 seal-0.005
rest-0.005
aroo-0.005
 slug-0.004
Token very
Token ID845
Feature activation+3.31e-02

Pos logit contributions
renches+0.066
 older+0.062
ways+0.060
arre+0.060
 headache+0.059
Neg logit contributions
 pearl-0.065
 seal-0.062
rest-0.060
aroo-0.059
 globe-0.058
Token hard
Token ID1327
Feature activation-2.34e-02

Pos logit contributions
 pearl+0.054
aroo+0.053
rest+0.051
 seal+0.051
 coconut+0.050
Neg logit contributions
renches-0.050
arre-0.049
 headache-0.046
 exams-0.046
 older-0.046
Token reached
Token ID4251
Feature activation-1.40e-02

Pos logit contributions
 pearl+0.050
rest+0.048
aroo+0.047
 seal+0.047
 globe+0.046
Neg logit contributions
 older-0.046
renches-0.046
arre-0.044
ways-0.043
ups-0.043
Token out
Token ID503
Feature activation-4.65e-02

Pos logit contributions
 older+0.021
renches+0.020
 questions+0.019
 exams+0.019
 younger+0.018
Neg logit contributions
rest-0.023
 pearl-0.023
 seal-0.023
aroo-0.023
 slug-0.022
Token for
Token ID329
Feature activation-2.84e-02

Pos logit contributions
renches+0.065
 older+0.064
 exams+0.060
ways+0.060
arre+0.058
Neg logit contributions
 pearl-0.080
 seal-0.079
rest-0.079
aroo-0.078
 slug-0.075
Token some
Token ID617
Feature activation-7.36e-02

Pos logit contributions
renches+0.043
 older+0.043
 exams+0.041
 questions+0.040
ways+0.039
Neg logit contributions
rest-0.044
 globe-0.043
 seal-0.043
aroo-0.043
 pearl-0.043
Token tw
Token ID665
Feature activation+1.79e-02

Pos logit contributions
renches+0.110
 older+0.107
 exams+0.099
oken+0.097
 questions+0.097
Neg logit contributions
 pearl-0.121
 seal-0.120
aroo-0.120
rest-0.120
 globe-0.116
Tokenigs
Token ID9235
Feature activation-1.85e-01

Pos logit contributions
 pearl+0.039
aroo+0.037
rest+0.037
 seal+0.036
rily+0.036
Neg logit contributions
renches-0.018
 older-0.017
 headache-0.016
 exams-0.016
arre-0.016
Token and
Token ID290
Feature activation+4.03e-03

Pos logit contributions
 older+0.197
renches+0.189
ways+0.181
arre+0.173
ups+0.172
Neg logit contributions
 pearl-0.380
 seal-0.372
aroo-0.365
rest-0.365
 slug-0.360
Token made
Token ID925
Feature activation-6.90e-02

Pos logit contributions
 pearl+0.007
rest+0.007
 seal+0.007
aroo+0.007
 globe+0.007
Neg logit contributions
renches-0.006
 older-0.006
ways-0.005
ups-0.005
 exams-0.005
Token the
Token ID262
Feature activation-3.38e-02

Pos logit contributions
renches+0.108
 older+0.108
 exams+0.099
ups+0.099
 questions+0.097
Neg logit contributions
 pearl-0.108
 seal-0.105
aroo-0.104
 globe-0.104
rest-0.101
Token bed
Token ID3996
Feature activation-6.19e-02

Pos logit contributions
renches+0.051
 older+0.051
 exams+0.047
 headache+0.046
arre+0.046
Neg logit contributions
 pearl-0.054
rest-0.053
aroo-0.053
 seal-0.053
 globe-0.050
Token stronger
Token ID7387
Feature activation-5.92e-04

Pos logit contributions
renches+0.093
 older+0.087
 exams+0.085
arre+0.083
oken+0.081
Neg logit contributions
 pearl-0.100
rest-0.099
 seal-0.098
aroo-0.096
 globe-0.094
Token They
Token ID1119
Feature activation-4.53e-02

Pos logit contributions
renches+0.009
 older+0.008
 exams+0.008
ups+0.008
oken+0.008
Neg logit contributions
 seal-0.013
 pearl-0.013
aroo-0.013
 slug-0.012
 globe-0.012
Token walked
Token ID6807
Feature activation-4.06e-02

Pos logit contributions
renches+0.079
 older+0.074
 exams+0.072
arre+0.071
 headache+0.071
Neg logit contributions
 seal-0.066
 pearl-0.064
rest-0.062
aroo-0.061
 gob-0.060
Token along
Token ID1863
Feature activation+4.30e-02

Pos logit contributions
renches+0.077
 older+0.071
arre+0.070
 headache+0.070
 exams+0.070
Neg logit contributions
 seal-0.053
 pearl-0.053
aroo-0.051
rest-0.051
 globe-0.049
Token the
Token ID262
Feature activation-2.71e-02

Pos logit contributions
 pearl+0.067
 seal+0.065
rest+0.065
aroo+0.064
 slug+0.061
Neg logit contributions
renches-0.069
 older-0.066
 exams-0.063
ways-0.063
arre-0.062
Token hall
Token ID6899
Feature activation-5.88e-02

Pos logit contributions
renches+0.032
 older+0.030
arre+0.028
 exams+0.028
ways+0.028
Neg logit contributions
 pearl-0.054
 seal-0.053
rest-0.052
aroo-0.052
 globe-0.051
Tokenways
Token ID1322
Feature activation-1.85e-01

Pos logit contributions
renches+0.099
 headache+0.091
oken+0.087
arre+0.085
 Deal+0.084
Neg logit contributions
 globe-0.086
 seal-0.085
rest-0.083
cious-0.082
aroo-0.081
Token,
Token ID11
Feature activation-3.84e-02

Pos logit contributions
renches+0.269
 headache+0.243
 older+0.242
arre+0.240
 exams+0.239
Neg logit contributions
 seal-0.313
rest-0.310
 pearl-0.307
aroo-0.305
 globe-0.295
Token looking
Token ID2045
Feature activation-2.10e-02

Pos logit contributions
renches+0.072
 older+0.066
arre+0.066
 exams+0.066
 headache+0.065
Neg logit contributions
 seal-0.050
 pearl-0.049
aroo-0.047
rest-0.047
 globe-0.047
Token at
Token ID379
Feature activation-3.12e-02

Pos logit contributions
renches+0.038
arre+0.035
 exams+0.035
 headache+0.034
 older+0.034
Neg logit contributions
 pearl-0.028
 seal-0.028
rest-0.027
aroo-0.026
 globe-0.025
Token the
Token ID262
Feature activation-2.63e-02

Pos logit contributions
 older+0.045
renches+0.042
 exams+0.041
 questions+0.040
 meetings+0.039
Neg logit contributions
 pearl-0.054
aroo-0.054
cious-0.053
rest-0.053
 slug-0.052
Token doors
Token ID8215
Feature activation-1.51e-01

Pos logit contributions
renches+0.038
 older+0.035
arre+0.034
ways+0.034
 exams+0.033
Neg logit contributions
 pearl-0.046
 seal-0.045
rest-0.044
aroo-0.043
 globe-0.043
Token make
Token ID787
Feature activation-4.48e-02

Pos logit contributions
 pearl+0.082
 seal+0.081
aroo+0.080
rest+0.080
 slug+0.076
Neg logit contributions
renches-0.069
 older-0.065
 exams-0.062
ways-0.062
arre-0.062
Token tree
Token ID5509
Feature activation+1.27e-02

Pos logit contributions
renches+0.066
 older+0.064
 exams+0.061
ways+0.060
arre+0.059
Neg logit contributions
 pearl-0.075
 seal-0.073
aroo-0.073
rest-0.072
 slug-0.070
Tokenhouses
Token ID20089
Feature activation-9.34e-02

Pos logit contributions
 pearl+0.022
 seal+0.022
rest+0.021
aroo+0.021
 globe+0.020
Neg logit contributions
renches-0.018
 older-0.017
 exams-0.016
ways-0.016
 headache-0.016
Token from
Token ID422
Feature activation+9.69e-02

Pos logit contributions
renches+0.129
 older+0.125
ways+0.117
 exams+0.117
arre+0.115
Neg logit contributions
 pearl-0.165
 seal-0.161
aroo-0.159
rest-0.158
 slug-0.154
Token tw
Token ID665
Feature activation+1.92e-02

Pos logit contributions
cious+0.171
rest+0.167
aroo+0.164
 pearl+0.163
 slug+0.162
Neg logit contributions
 older-0.141
renches-0.129
 exams-0.128
 questions-0.123
 younger-0.118
Tokenigs
Token ID9235
Feature activation-1.85e-01

Pos logit contributions
 pearl+0.040
aroo+0.038
 seal+0.038
rest+0.038
rily+0.037
Neg logit contributions
 older-0.019
renches-0.018
 headache-0.018
 exams-0.017
 nod-0.017
Token.
Token ID13
Feature activation+3.13e-04

Pos logit contributions
 older+0.207
renches+0.204
ways+0.191
arre+0.186
 exams+0.184
Neg logit contributions
 pearl-0.371
 seal-0.360
aroo-0.357
 slug-0.354
rest-0.352
Token He
Token ID679
Feature activation-5.73e-03

Pos logit contributions
 pearl+0.001
 seal+0.001
aroo+0.001
rest+0.001
 slug+0.001
Neg logit contributions
renches-0.000
 older-0.000
 headache-0.000
 exams-0.000
ways-0.000
Token waved
Token ID26834
Feature activation+3.65e-02

Pos logit contributions
renches+0.008
 older+0.008
ways+0.008
 exams+0.008
ups+0.008
Neg logit contributions
 pearl-0.009
rest-0.009
aroo-0.009
 seal-0.009
 slug-0.009
Token it
Token ID340
Feature activation+4.77e-03

Pos logit contributions
rest+0.057
 pearl+0.056
aroo+0.055
 slug+0.054
 globe+0.053
Neg logit contributions
 older-0.057
renches-0.055
 exams-0.053
ways-0.052
 questions-0.052
Token around
Token ID1088
Feature activation-4.91e-02

Pos logit contributions
 pearl+0.006
 seal+0.006
aroo+0.006
rest+0.006
 slug+0.006
Neg logit contributions
renches-0.009
 older-0.009
ways-0.008
 exams-0.008
arre-0.008
Token The
Token ID383
Feature activation-2.40e-02

Pos logit contributions
 seal+0.013
 pearl+0.013
aroo+0.013
rest+0.012
 globe+0.012
Neg logit contributions
renches-0.008
 older-0.008
 exams-0.007
arre-0.007
 headache-0.007
Token nest
Token ID16343
Feature activation-4.18e-02

Pos logit contributions
renches+0.043
 older+0.042
 exams+0.040
 headache+0.040
ways+0.039
Neg logit contributions
 pearl-0.032
rest-0.032
 seal-0.031
aroo-0.031
 slug-0.029
Token has
Token ID468
Feature activation+4.48e-02

Pos logit contributions
renches+0.077
 exams+0.072
 older+0.072
arre+0.071
ways+0.070
Neg logit contributions
 pearl-0.055
 seal-0.055
aroo-0.054
rest-0.053
 globe-0.052
Token many
Token ID867
Feature activation-2.60e-03

Pos logit contributions
 pearl+0.069
 seal+0.068
aroo+0.067
rest+0.067
 slug+0.064
Neg logit contributions
renches-0.071
 older-0.069
 exams-0.066
ways-0.065
arre-0.064
Token tw
Token ID665
Feature activation+1.62e-02

Pos logit contributions
renches+0.004
arre+0.004
 older+0.004
 exams+0.004
ways+0.004
Neg logit contributions
 pearl-0.004
 seal-0.004
rest-0.004
aroo-0.004
 diamond-0.004
Tokenigs
Token ID9235
Feature activation-1.85e-01

Pos logit contributions
 pearl+0.036
aroo+0.035
rest+0.035
 seal+0.034
cious+0.033
Neg logit contributions
 older-0.015
renches-0.015
 exams-0.014
arre-0.014
 headache-0.013
Token.
Token ID13
Feature activation+7.08e-03

Pos logit contributions
 older+0.186
renches+0.177
ways+0.170
ups+0.160
arre+0.154
Neg logit contributions
 pearl-0.393
aroo-0.382
 seal-0.382
 slug-0.375
rest-0.373
Token The
Token ID383
Feature activation-2.39e-02

Pos logit contributions
 seal+0.012
aroo+0.012
 pearl+0.011
 globe+0.011
rest+0.011
Neg logit contributions
renches-0.010
 exams-0.010
 older-0.010
arre-0.010
 headache-0.009
Token bird
Token ID6512
Feature activation-2.70e-03

Pos logit contributions
renches+0.046
 older+0.043
ways+0.043
arre+0.043
 exams+0.042
Neg logit contributions
 pearl-0.031
 seal-0.030
aroo-0.029
rest-0.029
 slug-0.027
Token is
Token ID318
Feature activation-1.00e-02

Pos logit contributions
renches+0.004
 older+0.004
 exams+0.004
ways+0.004
arre+0.004
Neg logit contributions
 pearl-0.004
 seal-0.004
rest-0.004
aroo-0.004
 slug-0.004
Token inside
Token ID2641
Feature activation+5.17e-02

Pos logit contributions
 older+0.015
renches+0.013
 younger+0.012
oken+0.012
 headache+0.012
Neg logit contributions
 seal-0.017
aroo-0.017
 pearl-0.017
rest-0.017
 slug-0.017
Tokenaroo
Token ID38049
Feature activation+1.96e-01

Pos logit contributions
 pearl+0.072
Our+0.057
 ruby+0.056
rily+0.052
cream+0.052
Neg logit contributions
 older-0.188
ap-0.182
 exams-0.181
 nod-0.178
 questions-0.176
Token!".
Token ID48220
Feature activation+7.35e-02

Pos logit contributions
 pearl+0.359
rest+0.349
 seal+0.329
 diamond+0.324
 slug+0.323
Neg logit contributions
 older-0.256
renches-0.230
 exams-0.227
ways-0.227
ups-0.223
Token The
Token ID383
Feature activation-1.48e-02

Pos logit contributions
 pearl+0.130
rest+0.126
 seal+0.126
aroo+0.124
 slug+0.119
Neg logit contributions
renches-0.102
 older-0.098
ways-0.092
 exams-0.092
 headache-0.090
Token k
Token ID479
Feature activation+1.12e-01

Pos logit contributions
 older+0.024
 younger+0.021
 nod+0.020
 wind+0.020
 crosses+0.019
Neg logit contributions
 pearl-0.026
cious-0.024
cream-0.024
rest-0.023
ifying-0.023
Tokenang
Token ID648
Feature activation+8.65e-02

Pos logit contributions
 pearl+0.150
rest+0.136
 seal+0.129
cious+0.129
 gob+0.128
Neg logit contributions
renches-0.203
 older-0.201
 exams-0.191
arre-0.191
ways-0.189
Tokenaroo
Token ID38049
Feature activation+1.96e-01

Pos logit contributions
 pearl+0.069
 ruby+0.052
cious+0.050
 gob+0.048
ifying+0.045
Neg logit contributions
 older-0.199
 nod-0.190
 questions-0.189
 crosses-0.188
 shakes-0.187
Token stopped
Token ID5025
Feature activation-1.24e-02

Pos logit contributions
 pearl+0.285
 seal+0.284
aroo+0.276
rest+0.271
 slug+0.262
Neg logit contributions
renches-0.341
 older-0.318
 exams-0.313
ways-0.306
 headache-0.304
Token jumping
Token ID14284
Feature activation+4.13e-02

Pos logit contributions
renches+0.019
 older+0.018
 exams+0.017
ways+0.017
arre+0.017
Neg logit contributions
 pearl-0.021
 seal-0.020
rest-0.020
aroo-0.020
 slug-0.019
Token and
Token ID290
Feature activation+7.71e-03

Pos logit contributions
 pearl+0.088
rest+0.086
aroo+0.084
 seal+0.084
cious+0.082
Neg logit contributions
 older-0.042
renches-0.038
ways-0.037
arre-0.036
 questions-0.034
Token hopped
Token ID45230
Feature activation+1.62e-02

Pos logit contributions
rest+0.014
 pearl+0.014
aroo+0.013
 globe+0.013
 seal+0.013
Neg logit contributions
 older-0.010
ways-0.009
 nod-0.009
 headache-0.009
 exams-0.009
Token towards
Token ID3371
Feature activation+7.12e-02

Pos logit contributions
 pearl+0.025
rest+0.023
aroo+0.023
 seal+0.023
 globe+0.022
Neg logit contributions
 older-0.026
renches-0.025
 questions-0.024
ways-0.024
arre-0.023
Token was
Token ID373
Feature activation+1.67e-02

Pos logit contributions
 pearl+0.022
 seal+0.021
rest+0.021
aroo+0.021
 globe+0.019
Neg logit contributions
renches-0.032
 older-0.031
ways-0.030
 exams-0.030
arre-0.029
Token the
Token ID262
Feature activation-2.82e-02

Pos logit contributions
 pearl+0.024
 seal+0.024
rest+0.024
aroo+0.024
 slug+0.022
Neg logit contributions
renches-0.028
 older-0.027
 exams-0.026
ways-0.026
 headache-0.025
Token sound
Token ID2128
Feature activation-9.92e-02

Pos logit contributions
renches+0.041
 older+0.040
 exams+0.037
 headache+0.037
ways+0.037
Neg logit contributions
 pearl-0.048
rest-0.047
 seal-0.046
aroo-0.046
 slug-0.044
Token of
Token ID286
Feature activation+3.99e-02

Pos logit contributions
 older+0.115
renches+0.110
ways+0.107
 exams+0.102
 headache+0.102
Neg logit contributions
 pearl-0.198
 seal-0.188
aroo-0.187
rest-0.184
 slug-0.183
Token another
Token ID1194
Feature activation-2.58e-02

Pos logit contributions
rest+0.079
aroo+0.078
 pearl+0.077
cious+0.076
 seal+0.076
Neg logit contributions
 older-0.046
renches-0.045
 exams-0.042
 headache-0.041
 questions-0.040
Token whale
Token ID22206
Feature activation+1.71e-01

Pos logit contributions
renches+0.034
 older+0.034
 exams+0.031
 headache+0.031
ways+0.031
Neg logit contributions
 pearl-0.046
rest-0.045
aroo-0.045
 seal-0.044
 slug-0.042
Token.
Token ID13
Feature activation-3.03e-04

Pos logit contributions
 pearl+0.318
rest+0.311
 slug+0.304
 ruby+0.300
aroo+0.298
Neg logit contributions
 older-0.210
ways-0.180
ups-0.179
 headache-0.172
 thunder-0.170
Token He
Token ID679
Feature activation-2.91e-03

Pos logit contributions
renches+0.000
 older+0.000
 exams+0.000
ways+0.000
arre+0.000
Neg logit contributions
 pearl-0.001
 seal-0.001
rest-0.001
aroo-0.001
 slug-0.000
Token looked
Token ID3114
Feature activation-6.55e-02

Pos logit contributions
 older+0.004
renches+0.004
ways+0.004
ups+0.004
 questions+0.004
Neg logit contributions
rest-0.005
 pearl-0.005
 globe-0.005
aroo-0.005
 slug-0.005
Token over
Token ID625
Feature activation+1.48e-03

Pos logit contributions
 older+0.108
renches+0.104
ways+0.098
arre+0.097
 exams+0.097
Neg logit contributions
rest-0.098
aroo-0.097
 pearl-0.097
 seal-0.094
 slug-0.091
Token and
Token ID290
Feature activation+8.48e-03

Pos logit contributions
rest+0.003
aroo+0.003
 pearl+0.003
 seal+0.003
 slug+0.003
Neg logit contributions
 older-0.002
renches-0.002
 exams-0.002
ways-0.002
 questions-0.002
Token
Token ID198
Feature activation-7.58e-03

Pos logit contributions
renches+0.003
 older+0.003
 exams+0.003
ways+0.003
oken+0.003
Neg logit contributions
 pearl-0.003
 seal-0.003
aroo-0.003
rest-0.003
 slug-0.003
Token
Token ID198
Feature activation-7.31e-03

Pos logit contributions
renches+0.009
 older+0.008
arre+0.008
 exams+0.008
ways+0.007
Neg logit contributions
 pearl-0.016
 seal-0.015
rest-0.015
aroo-0.015
 diamond-0.014
TokenAfter
Token ID3260
Feature activation+1.88e-02

Pos logit contributions
renches+0.009
arre+0.008
 older+0.008
 exams+0.008
ways+0.008
Neg logit contributions
 pearl-0.015
 seal-0.014
rest-0.014
aroo-0.013
 slug-0.013
Token eating
Token ID6600
Feature activation+1.27e-01

Pos logit contributions
rest+0.032
rily+0.032
 pearl+0.031
aroo+0.031
 globe+0.031
Neg logit contributions
renches-0.026
 older-0.026
 exams-0.025
 headache-0.025
 meetings-0.024
Token the
Token ID262
Feature activation-3.01e-02

Pos logit contributions
cious+0.258
 pearl+0.257
rest+0.256
 seal+0.250
aroo+0.247
Neg logit contributions
 older-0.138
renches-0.134
 headache-0.127
 exams-0.126
 early-0.121
Token sausage
Token ID31889
Feature activation+1.79e-01

Pos logit contributions
 older+0.059
 chores+0.054
renches+0.052
 headache+0.052
arre+0.051
Neg logit contributions
aroo-0.041
cious-0.041
rest-0.039
 pearl-0.037
 ruby-0.036
Token,
Token ID11
Feature activation-3.08e-02

Pos logit contributions
 pearl+0.428
 seal+0.416
rest+0.412
aroo+0.408
ifying+0.404
Neg logit contributions
 older-0.123
renches-0.120
arre-0.107
ways-0.105
 headache-0.100
Token Max
Token ID5436
Feature activation-4.71e-02

Pos logit contributions
renches+0.035
 older+0.033
oken+0.030
ways+0.030
arre+0.030
Neg logit contributions
 seal-0.061
 pearl-0.061
aroo-0.061
rest-0.060
 globe-0.058
Token went
Token ID1816
Feature activation-1.07e-02

Pos logit contributions
renches+0.072
 older+0.070
 exams+0.066
ways+0.066
arre+0.066
Neg logit contributions
 pearl-0.076
 seal-0.074
rest-0.073
aroo-0.073
 slug-0.069
Token to
Token ID284
Feature activation-1.84e-02

Pos logit contributions
renches+0.023
 exams+0.020
arre+0.020
oken+0.020
 headache+0.020
Neg logit contributions
 seal-0.012
 pearl-0.012
rest-0.011
aroo-0.011
 globe-0.010
Token play
Token ID711
Feature activation-3.91e-02

Pos logit contributions
renches+0.023
 older+0.022
ways+0.020
 exams+0.020
arre+0.020
Neg logit contributions
 pearl-0.035
 seal-0.035
rest-0.034
aroo-0.034
 slug-0.033
TokenOnce
Token ID7454
Feature activation+2.45e-02

Pos logit contributions
 older+0.039
renches+0.037
ways+0.036
arre+0.035
 exams+0.033
Neg logit contributions
 pearl-0.101
rest-0.095
 seal-0.095
aroo-0.094
cious-0.092
Token there
Token ID612
Feature activation+5.57e-03

Pos logit contributions
 pearl+0.046
 seal+0.046
aroo+0.046
rest+0.045
 globe+0.044
Neg logit contributions
renches-0.031
 older-0.028
ways-0.027
 exams-0.027
arre-0.027
Token was
Token ID373
Feature activation+1.46e-02

Pos logit contributions
 pearl+0.010
cious+0.010
 seal+0.010
ifying+0.010
rest+0.010
Neg logit contributions
 older-0.007
 nod-0.006
 Sometimes-0.006
 One-0.006
 Despite-0.006
Token a
Token ID257
Feature activation-1.63e-02

Pos logit contributions
 pearl+0.027
aroo+0.027
rest+0.026
 seal+0.026
 globe+0.026
Neg logit contributions
 older-0.018
renches-0.018
oken-0.016
 exams-0.016
arre-0.016
Token t
Token ID256
Feature activation+6.59e-02

Pos logit contributions
 older+0.021
renches+0.020
 headache+0.019
 exams+0.018
 younger+0.018
Neg logit contributions
rest-0.031
 pearl-0.029
aroo-0.029
rily-0.028
cream-0.027
Tokeneddy
Token ID21874
Feature activation+1.96e-01

Pos logit contributions
 pearl+0.066
rest+0.063
 seal+0.061
 slug+0.060
aroo+0.059
Neg logit contributions
 older-0.136
renches-0.134
ups-0.129
 exams-0.128
 headache-0.127
Token bear
Token ID6842
Feature activation+3.35e-02

Pos logit contributions
 pearl+0.330
cious+0.328
rest+0.325
ifying+0.314
 coconut+0.314
Neg logit contributions
 older-0.267
 headache-0.244
renches-0.234
 exams-0.231
 questions-0.231
Token.
Token ID13
Feature activation+2.99e-03

Pos logit contributions
 pearl+0.069
rest+0.064
rily+0.063
 slug+0.063
 ruby+0.060
Neg logit contributions
 older-0.035
ways-0.033
renches-0.032
 exams-0.032
udd-0.032
Token He
Token ID679
Feature activation+2.07e-03

Pos logit contributions
 pearl+0.006
rest+0.006
 seal+0.006
aroo+0.006
cious+0.005
Neg logit contributions
renches-0.003
 older-0.003
 exams-0.003
ways-0.003
 headache-0.003
Token was
Token ID373
Feature activation+1.48e-02

Pos logit contributions
rily+0.004
 pearl+0.004
aroo+0.004
ifying+0.004
rest+0.003
Neg logit contributions
renches-0.003
 older-0.002
 exams-0.002
 crosses-0.002
 headache-0.002
Token very
Token ID845
Feature activation+3.51e-02

Pos logit contributions
 pearl+0.027
 globe+0.026
aroo+0.025
rest+0.025
 slug+0.025
Neg logit contributions
 older-0.023
 younger-0.019
renches-0.019
 exams-0.018
oken-0.018
Token plan
Token ID1410
Feature activation-3.42e-02

Pos logit contributions
renches+0.021
 older+0.020
 exams+0.019
ways+0.019
arre+0.019
Neg logit contributions
 pearl-0.030
 seal-0.030
rest-0.030
aroo-0.030
 slug-0.028
Token.
Token ID13
Feature activation+7.50e-04

Pos logit contributions
renches+0.050
 older+0.047
arre+0.045
 headache+0.045
ways+0.044
Neg logit contributions
 pearl-0.057
rest-0.056
 seal-0.056
aroo-0.055
 globe-0.053
Token Every
Token ID3887
Feature activation-7.35e-03

Pos logit contributions
 pearl+0.002
aroo+0.002
 seal+0.002
rest+0.002
 slug+0.002
Neg logit contributions
renches-0.000
 older-0.000
 exams-0.000
ways-0.000
arre-0.000
Token week
Token ID1285
Feature activation+7.07e-02

Pos logit contributions
 older+0.011
renches+0.011
 exams+0.011
 headache+0.011
arre+0.010
Neg logit contributions
 pearl-0.012
 seal-0.011
aroo-0.011
rest-0.011
 slug-0.011
Token,
Token ID11
Feature activation-3.50e-02

Pos logit contributions
 pearl+0.135
 seal+0.133
aroo+0.131
rest+0.131
 slug+0.126
Neg logit contributions
renches-0.087
 older-0.084
 exams-0.079
ways-0.078
arre-0.076
Token Lib
Token ID7980
Feature activation+1.82e-01

Pos logit contributions
 older+0.041
 exams+0.039
ways+0.038
 meetings+0.037
 questions+0.037
Neg logit contributions
 pearl-0.068
 slug-0.068
aroo-0.067
rest-0.065
 globe-0.064
Tokenby
Token ID1525
Feature activation+3.93e-02

Pos logit contributions
 pearl+0.312
 seal+0.306
rest+0.302
aroo+0.300
 slug+0.287
Neg logit contributions
renches-0.263
 older-0.250
 exams-0.239
ways-0.237
arre-0.234
Token and
Token ID290
Feature activation+5.15e-03

Pos logit contributions
 pearl+0.059
 seal+0.056
rest+0.056
aroo+0.056
 slug+0.054
Neg logit contributions
renches-0.066
 older-0.063
 exams-0.061
ways-0.060
arre-0.060
Token her
Token ID607
Feature activation-2.12e-03

Pos logit contributions
 pearl+0.009
aroo+0.009
 seal+0.009
rest+0.009
 slug+0.009
Neg logit contributions
 older-0.007
renches-0.007
 exams-0.007
 questions-0.006
ways-0.006
Token mom
Token ID1995
Feature activation-9.53e-02

Pos logit contributions
 older+0.002
renches+0.002
 exams+0.002
 headache+0.002
 questions+0.002
Neg logit contributions
 pearl-0.005
rest-0.005
aroo-0.004
 seal-0.004
 slug-0.004
Token would
Token ID561
Feature activation+2.64e-02

Pos logit contributions
renches+0.140
 older+0.138
arre+0.129
 exams+0.128
ways+0.127
Neg logit contributions
 pearl-0.163
aroo-0.158
 seal-0.156
rest-0.153
 slug-0.152
Token needed
Token ID2622
Feature activation+1.84e-02

Pos logit contributions
 pearl+0.037
aroo+0.034
 seal+0.034
 globe+0.033
rest+0.033
Neg logit contributions
renches-0.041
 older-0.040
ways-0.038
ups-0.038
 exams-0.037
Token a
Token ID257
Feature activation-1.35e-02

Pos logit contributions
 pearl+0.030
 seal+0.029
rest+0.029
aroo+0.029
 slug+0.027
Neg logit contributions
renches-0.028
 older-0.027
ways-0.026
 exams-0.026
arre-0.026
Token regular
Token ID3218
Feature activation+4.83e-02

Pos logit contributions
renches+0.019
 older+0.019
 headache+0.017
 exams+0.017
arre+0.017
Neg logit contributions
rest-0.023
 pearl-0.023
aroo-0.023
 seal-0.023
 slug-0.022
Token driver
Token ID4639
Feature activation-4.77e-02

Pos logit contributions
 pearl+0.076
 seal+0.075
aroo+0.074
rest+0.074
 slug+0.070
Neg logit contributions
renches-0.076
 older-0.073
 exams-0.070
ways-0.069
arre-0.068
Token who
Token ID508
Feature activation+1.25e-01

Pos logit contributions
 older+0.050
 questions+0.045
 exams+0.044
ups+0.043
renches+0.042
Neg logit contributions
 pearl-0.105
 slug-0.100
rily-0.098
 seal-0.096
aroo-0.096
Token could
Token ID714
Feature activation+5.20e-02

Pos logit contributions
 pearl+0.199
 slug+0.185
 seal+0.182
aroo+0.181
rily+0.178
Neg logit contributions
renches-0.198
 older-0.193
 exams-0.187
ways-0.181
ups-0.180
Token help
Token ID1037
Feature activation-7.72e-02

Pos logit contributions
 pearl+0.098
 seal+0.096
rest+0.095
aroo+0.095
 slug+0.092
Neg logit contributions
renches-0.066
 older-0.063
 exams-0.059
ways-0.058
 headache-0.058
Token him
Token ID683
Feature activation-2.13e-03

Pos logit contributions
renches+0.117
 older+0.114
 exams+0.109
ways+0.104
 questions+0.103
Neg logit contributions
 pearl-0.130
 seal-0.125
aroo-0.125
rest-0.122
 slug-0.119
Token go
Token ID467
Feature activation+1.53e-02

Pos logit contributions
 older+0.002
renches+0.002
 exams+0.002
 questions+0.002
ups+0.002
Neg logit contributions
 pearl-0.004
aroo-0.004
 seal-0.004
 slug-0.004
rest-0.004
Token places
Token ID4113
Feature activation-4.19e-02

Pos logit contributions
 pearl+0.028
aroo+0.027
 seal+0.026
rest+0.026
 slug+0.026
Neg logit contributions
 older-0.021
renches-0.021
ways-0.019
 exams-0.019
ups-0.019
Token.
Token ID13
Feature activation+3.70e-03

Pos logit contributions
 older+0.044
renches+0.043
ways+0.040
 exams+0.039
ups+0.037
Neg logit contributions
 pearl-0.087
aroo-0.086
 slug-0.085
 seal-0.085
rest-0.084
Token always
Token ID1464
Feature activation-6.77e-02

Pos logit contributions
renches+0.021
 older+0.020
 exams+0.019
ways+0.019
 headache+0.018
Neg logit contributions
 pearl-0.026
 seal-0.025
rest-0.025
aroo-0.025
 slug-0.024
Token yield
Token ID7800
Feature activation+2.01e-02

Pos logit contributions
renches+0.090
 older+0.086
 exams+0.081
ways+0.080
 headache+0.079
Neg logit contributions
 pearl-0.123
 seal-0.121
aroo-0.120
rest-0.120
 slug-0.115
Token,
Token ID11
Feature activation-3.28e-02

Pos logit contributions
 pearl+0.034
 seal+0.033
rest+0.033
aroo+0.032
 slug+0.031
Neg logit contributions
renches-0.030
 older-0.028
 exams-0.027
ways-0.027
arre-0.027
Token because
Token ID780
Feature activation-1.60e-02

Pos logit contributions
renches+0.042
 older+0.041
 exams+0.038
ways+0.037
 headache+0.037
Neg logit contributions
 pearl-0.061
 seal-0.060
aroo-0.060
rest-0.060
 slug-0.057
Token she
Token ID673
Feature activation-1.17e-02

Pos logit contributions
 older+0.021
 exams+0.020
renches+0.019
 questions+0.018
 headache+0.018
Neg logit contributions
 seal-0.029
cious-0.029
 pearl-0.029
aroo-0.028
rest-0.028
Token knew
Token ID2993
Feature activation+6.48e-02

Pos logit contributions
renches+0.020
 older+0.019
 exams+0.019
arre+0.019
 headache+0.018
Neg logit contributions
 pearl-0.016
aroo-0.016
 seal-0.016
rest-0.015
 slug-0.015
Token it
Token ID340
Feature activation+7.36e-03

Pos logit contributions
 seal+0.103
 pearl+0.101
aroo+0.100
rest+0.098
 slug+0.097
Neg logit contributions
 older-0.101
 exams-0.097
renches-0.097
 questions-0.093
ups-0.093
Token was
Token ID373
Feature activation+1.83e-02

Pos logit contributions
 pearl+0.011
 seal+0.011
aroo+0.010
 slug+0.010
rest+0.009
Neg logit contributions
 older-0.012
renches-0.012
 exams-0.012
ways-0.012
ups-0.012
Token what
Token ID644
Feature activation+7.89e-02

Pos logit contributions
 seal+0.033
aroo+0.033
 pearl+0.033
 globe+0.032
rest+0.031
Neg logit contributions
 older-0.025
renches-0.024
 exams-0.022
 headache-0.022
 questions-0.022
Token she
Token ID673
Feature activation-1.13e-02

Pos logit contributions
 pearl+0.144
 seal+0.142
aroo+0.140
rest+0.140
 globe+0.134
Neg logit contributions
renches-0.102
 older-0.100
 exams-0.097
ways-0.092
 headache-0.092
Token needed
Token ID2622
Feature activation+1.98e-02

Pos logit contributions
 older+0.016
renches+0.016
 questions+0.015
 exams+0.015
arre+0.015
Neg logit contributions
aroo-0.019
 pearl-0.019
rily-0.019
 seal-0.018
 globe-0.018
Token and
Token ID290
Feature activation+9.32e-03

Pos logit contributions
 pearl+0.076
 seal+0.075
rest+0.074
aroo+0.074
 slug+0.071
Neg logit contributions
renches-0.066
 older-0.062
 exams-0.058
ways-0.058
arre-0.058
Token wanted
Token ID2227
Feature activation+5.61e-02

Pos logit contributions
 pearl+0.016
aroo+0.016
rest+0.015
 seal+0.015
 globe+0.015
Neg logit contributions
 older-0.013
renches-0.013
ways-0.012
 exams-0.012
 headache-0.012
Token to
Token ID284
Feature activation-1.54e-02

Pos logit contributions
aroo+0.112
 pearl+0.110
 seal+0.109
rest+0.109
 globe+0.107
Neg logit contributions
 older-0.066
 exams-0.063
renches-0.062
 questions-0.059
ups-0.058
Token show
Token ID905
Feature activation-1.47e-02

Pos logit contributions
renches+0.022
 older+0.021
ways+0.020
arre+0.020
 exams+0.020
Neg logit contributions
 pearl-0.027
 seal-0.026
rest-0.026
aroo-0.026
 slug-0.025
Token her
Token ID607
Feature activation+2.59e-05

Pos logit contributions
renches+0.025
 older+0.024
ways+0.023
arre+0.023
 exams+0.023
Neg logit contributions
 pearl-0.021
 seal-0.021
rest-0.021
aroo-0.020
 slug-0.019
Token doll
Token ID3654
Feature activation+6.52e-02

Pos logit contributions
 pearl+0.000
aroo+0.000
rest+0.000
cious+0.000
 seal+0.000
Neg logit contributions
 older-0.000
 exams-0.000
renches-0.000
 headache-0.000
 questions-0.000
Token to
Token ID284
Feature activation-1.51e-02

Pos logit contributions
 pearl+0.124
aroo+0.121
 seal+0.118
rest+0.116
 slug+0.115
Neg logit contributions
 older-0.085
renches-0.079
 exams-0.078
ways-0.075
ups-0.075
Token her
Token ID607
Feature activation+2.24e-04

Pos logit contributions
renches+0.020
 older+0.019
 exams+0.018
ways+0.017
 headache+0.017
Neg logit contributions
 pearl-0.028
 seal-0.027
aroo-0.027
rest-0.027
 slug-0.026
Token friend
Token ID1545
Feature activation-4.73e-02

Pos logit contributions
 pearl+0.001
rest+0.001
aroo+0.001
 seal+0.000
cious+0.000
Neg logit contributions
 older-0.000
renches-0.000
 exams-0.000
 headache-0.000
 younger-0.000
Token,
Token ID11
Feature activation-3.11e-02

Pos logit contributions
 older+0.047
renches+0.046
 exams+0.042
ways+0.040
arre+0.040
Neg logit contributions
 pearl-0.104
 seal-0.100
aroo-0.099
rest-0.099
 globe-0.096
Token Tim
Token ID5045
Feature activation-2.01e-02

Pos logit contributions
 older+0.028
renches+0.023
 exams+0.022
 younger+0.022
 headache+0.021
Neg logit contributions
rest-0.071
 pearl-0.071
aroo-0.069
 globe-0.068
cious-0.068
Token also
Token ID635
Feature activation+4.47e-02

Pos logit contributions
renches+0.000
 older+0.000
 exams+0.000
ways+0.000
arre+0.000
Neg logit contributions
 pearl-0.001
 seal-0.001
aroo-0.001
rest-0.000
 diamond-0.000
Token admire
Token ID21099
Feature activation+4.68e-03

Pos logit contributions
 pearl+0.080
 seal+0.079
rest+0.077
aroo+0.077
 slug+0.074
Neg logit contributions
renches-0.061
 older-0.057
ways-0.055
 exams-0.055
arre-0.054
Token the
Token ID262
Feature activation-2.52e-02

Pos logit contributions
aroo+0.006
cious+0.006
 seal+0.006
rest+0.006
 slug+0.006
Neg logit contributions
 older-0.008
 exams-0.008
renches-0.008
 questions-0.008
ways-0.007
Token small
Token ID1402
Feature activation-8.50e-03

Pos logit contributions
renches+0.047
 older+0.045
 exams+0.043
ways+0.043
arre+0.043
Neg logit contributions
 pearl-0.033
 seal-0.032
rest-0.031
aroo-0.031
 globe-0.030
Token things
Token ID1243
Feature activation-6.63e-02

Pos logit contributions
renches+0.016
 older+0.015
 exams+0.015
ways+0.015
arre+0.014
Neg logit contributions
 pearl-0.011
 seal-0.011
rest-0.011
aroo-0.011
 slug-0.010
Token that
Token ID326
Feature activation+6.09e-02

Pos logit contributions
 older+0.102
renches+0.096
ways+0.092
 exams+0.092
ups+0.090
Neg logit contributions
 pearl-0.107
 slug-0.106
 seal-0.102
aroo-0.101
cious-0.099
Token are
Token ID389
Feature activation+5.42e-02

Pos logit contributions
 pearl+0.103
aroo+0.100
 seal+0.099
rest+0.097
 slug+0.096
Neg logit contributions
renches-0.086
 older-0.085
 exams-0.082
ways-0.078
 headache-0.078
Token close
Token ID1969
Feature activation-7.54e-02

Pos logit contributions
aroo+0.093
 seal+0.091
 slug+0.090
 pearl+0.089
rest+0.088
Neg logit contributions
 older-0.081
renches-0.075
 questions-0.069
ways-0.069
 younger-0.068
Token to
Token ID284
Feature activation-1.60e-02

Pos logit contributions
 older+0.107
renches+0.104
ways+0.096
 exams+0.095
 questions+0.094
Neg logit contributions
 pearl-0.130
aroo-0.129
 seal-0.127
rest-0.127
 slug-0.125
Token us
Token ID514
Feature activation-3.06e-02

Pos logit contributions
renches+0.021
 older+0.020
 questions+0.020
 exams+0.019
oken+0.018
Neg logit contributions
aroo-0.030
 pearl-0.029
 seal-0.029
 slug-0.029
rest-0.028
Token,"
Token ID553
Feature activation+9.24e-02

Pos logit contributions
 older+0.042
renches+0.041
 exams+0.037
 questions+0.037
ways+0.037
Neg logit contributions
 pearl-0.055
aroo-0.053
 seal-0.053
rest-0.052
 slug-0.052
Token couldn
Token ID3521
Feature activation+8.58e-02

Pos logit contributions
 pearl+0.042
aroo+0.041
 seal+0.040
 globe+0.040
rest+0.040
Neg logit contributions
 older-0.039
renches-0.038
ways-0.037
ups-0.037
arre-0.037
Token't
Token ID470
Feature activation-4.08e-02

Pos logit contributions
 pearl+0.151
 seal+0.149
rest+0.149
aroo+0.149
 slug+0.141
Neg logit contributions
renches-0.118
 older-0.111
 exams-0.105
ways-0.105
arre-0.103
Token hold
Token ID1745
Feature activation+2.27e-02

Pos logit contributions
renches+0.063
 older+0.057
arre+0.056
aj+0.056
ways+0.055
Neg logit contributions
 seal-0.068
rest-0.068
 pearl-0.066
aroo-0.065
 gob-0.062
Token his
Token ID465
Feature activation+4.82e-02

Pos logit contributions
 pearl+0.034
aroo+0.034
 seal+0.033
rest+0.033
 globe+0.032
Neg logit contributions
 older-0.039
 exams-0.037
renches-0.037
 questions-0.036
ups-0.034
Token frustration
Token ID14285
Feature activation-1.04e-01

Pos logit contributions
 pearl+0.092
 seal+0.090
rest+0.088
aroo+0.088
 slug+0.085
Neg logit contributions
renches-0.061
 older-0.056
ways-0.054
arre-0.054
 exams-0.052
Token any
Token ID597
Feature activation+5.64e-02

Pos logit contributions
 older+0.123
renches+0.117
 exams+0.110
ways+0.109
 headache+0.105
Neg logit contributions
 pearl-0.206
aroo-0.204
 seal-0.203
rest-0.197
 slug-0.195
Token longer
Token ID2392
Feature activation-1.48e-01

Pos logit contributions
 pearl+0.110
 seal+0.108
rest+0.106
aroo+0.106
 slug+0.102
Neg logit contributions
renches-0.069
 older-0.064
ways-0.061
 exams-0.061
arre-0.061
Token.
Token ID13
Feature activation+3.11e-03

Pos logit contributions
 older+0.165
renches+0.161
 exams+0.152
ways+0.150
 questions+0.142
Neg logit contributions
 pearl-0.301
aroo-0.296
 seal-0.292
rest-0.287
 slug-0.284
Token He
Token ID679
Feature activation+1.08e-04

Pos logit contributions
 seal+0.005
 pearl+0.005
aroo+0.005
rest+0.005
 slug+0.005
Neg logit contributions
renches-0.004
 older-0.004
 headache-0.004
oken-0.004
 exams-0.004
Token yelled
Token ID22760
Feature activation-2.47e-02

Pos logit contributions
 pearl+0.000
aroo+0.000
rest+0.000
 seal+0.000
 globe+0.000
Neg logit contributions
renches-0.000
 older-0.000
ways-0.000
 exams-0.000
arre-0.000
Token at
Token ID379
Feature activation-2.91e-02

Pos logit contributions
renches+0.034
 older+0.032
 exams+0.031
ways+0.030
arre+0.030
Neg logit contributions
 pearl-0.044
 seal-0.043
rest-0.043
aroo-0.043
 slug-0.041
Token<|endoftext|>
Token ID50256
Feature activation-4.48e-02

Pos logit contributions
 pearl+0.045
rest+0.044
 seal+0.043
aroo+0.043
 slug+0.041
Neg logit contributions
renches-0.027
 older-0.026
 exams-0.025
arre-0.024
ways-0.024
TokenOnce
Token ID7454
Feature activation+2.33e-02

Pos logit contributions
 older+0.038
renches+0.036
arre+0.036
ways+0.036
 exams+0.033
Neg logit contributions
 pearl-0.102
rest-0.096
 seal-0.096
aroo-0.095
 diamond-0.093
Token upon
Token ID2402
Feature activation-2.47e-02

Pos logit contributions
 pearl+0.046
 seal+0.045
aroo+0.045
rest+0.044
 globe+0.043
Neg logit contributions
renches-0.028
 older-0.026
 exams-0.024
ways-0.024
arre-0.024
Token a
Token ID257
Feature activation-1.44e-02

Pos logit contributions
 older+0.029
renches+0.028
 questions+0.027
 exams+0.026
 headache+0.026
Neg logit contributions
 pearl-0.047
aroo-0.047
 slug-0.045
rest-0.045
 seal-0.045
Token time
Token ID640
Feature activation-2.57e-02

Pos logit contributions
renches+0.019
 older+0.017
ways+0.017
ups+0.017
 exams+0.016
Neg logit contributions
 seal-0.026
 pearl-0.026
rest-0.025
cious-0.025
aroo-0.025
Token,
Token ID11
Feature activation-3.55e-02

Pos logit contributions
renches+0.023
 older+0.022
 exams+0.021
ways+0.020
arre+0.020
Neg logit contributions
 pearl-0.057
 seal-0.056
rest-0.056
aroo-0.055
 slug-0.054
Token there
Token ID612
Feature activation+6.96e-03

Pos logit contributions
 older+0.035
 Sometimes+0.033
 One+0.033
ap+0.030
renches+0.027
Neg logit contributions
 pearl-0.081
 gob-0.074
cream-0.074
rest-0.074
 seal-0.073
Token was
Token ID373
Feature activation+1.58e-02

Pos logit contributions
 pearl+0.015
rily+0.014
ifying+0.014
cious+0.014
cream+0.014
Neg logit contributions
 Sometimes-0.007
 One-0.006
 older-0.006
 nod-0.006
 exams-0.006
Token a
Token ID257
Feature activation-1.43e-02

Pos logit contributions
 pearl+0.032
aroo+0.031
rest+0.031
 globe+0.031
 seal+0.030
Neg logit contributions
 older-0.018
renches-0.017
oken-0.015
 thunder-0.015
 exams-0.014
Token boy
Token ID2933
Feature activation+2.54e-02

Pos logit contributions
 older+0.019
renches+0.018
 headache+0.017
 exams+0.016
 younger+0.016
Neg logit contributions
rest-0.027
 pearl-0.026
aroo-0.025
rily-0.024
cream-0.024
Token called
Token ID1444
Feature activation-4.43e-02

Pos logit contributions
 pearl+0.042
rest+0.040
 seal+0.040
aroo+0.039
 slug+0.038
Neg logit contributions
renches-0.038
 older-0.037
 exams-0.035
ways-0.034
 headache-0.034
Token blocks
Token ID7021
Feature activation-7.85e-02

Pos logit contributions
renches+0.053
arre+0.049
oken+0.049
 headache+0.049
 older+0.048
Neg logit contributions
 seal-0.067
 pearl-0.065
 globe-0.063
aroo-0.063
rest-0.062
Token,
Token ID11
Feature activation-3.76e-02

Pos logit contributions
renches+0.092
 older+0.085
 exams+0.081
 headache+0.080
arre+0.080
Neg logit contributions
 pearl-0.154
 seal-0.154
aroo-0.151
rest-0.151
 globe-0.147
Token and
Token ID290
Feature activation+7.87e-03

Pos logit contributions
renches+0.055
 older+0.052
arre+0.052
 headache+0.051
oken+0.050
Neg logit contributions
 seal-0.062
 pearl-0.061
 globe-0.059
rest-0.059
aroo-0.059
Token books
Token ID3835
Feature activation-5.60e-02

Pos logit contributions
 pearl+0.014
 seal+0.014
rest+0.013
aroo+0.013
 slug+0.013
Neg logit contributions
renches-0.011
 older-0.010
arre-0.010
ways-0.010
 exams-0.010
Token.
Token ID13
Feature activation-3.06e-03

Pos logit contributions
renches+0.062
 older+0.061
ways+0.059
 exams+0.056
ups+0.054
Neg logit contributions
 pearl-0.115
 seal-0.111
aroo-0.111
 slug-0.110
rest-0.110
Token They
Token ID1119
Feature activation-4.17e-02

Pos logit contributions
renches+0.004
 older+0.003
 exams+0.003
 headache+0.003
ways+0.003
Neg logit contributions
 seal-0.006
 pearl-0.006
aroo-0.005
 slug-0.005
cious-0.005
Token liked
Token ID8288
Feature activation+2.10e-02

Pos logit contributions
renches+0.078
 older+0.075
 exams+0.072
arre+0.072
 headache+0.072
Neg logit contributions
 seal-0.053
 pearl-0.053
rest-0.051
aroo-0.051
 globe-0.048
Token to
Token ID284
Feature activation-2.00e-02

Pos logit contributions
aroo+0.044
 pearl+0.044
rest+0.043
 slug+0.042
cious+0.042
Neg logit contributions
 older-0.023
 exams-0.023
renches-0.022
 questions-0.021
 meetings-0.020
Token make
Token ID787
Feature activation-4.49e-02

Pos logit contributions
arre+0.036
renches+0.035
 older+0.034
oken+0.034
ups+0.034
Neg logit contributions
 seal-0.028
aroo-0.026
 pearl-0.026
 globe-0.025
 diamond-0.025
Token up
Token ID510
Feature activation+7.07e-05

Pos logit contributions
renches+0.064
 headache+0.060
 older+0.059
ways+0.059
arre+0.058
Neg logit contributions
 seal-0.075
 pearl-0.075
rest-0.072
aroo-0.070
 globe-0.069
Token stories
Token ID3923
Feature activation-2.39e-02

Pos logit contributions
 seal+0.000
 pearl+0.000
rest+0.000
 globe+0.000
aroo+0.000
Neg logit contributions
renches-0.000
 older-0.000
arre-0.000
oken-0.000
 headache-0.000
Token man
Token ID582
Feature activation+1.75e-03

Pos logit contributions
renches+0.006
 older+0.005
oken+0.005
arre+0.005
 exams+0.005
Neg logit contributions
 pearl-0.006
 seal-0.006
aroo-0.006
rest-0.006
 globe-0.005
Token's
Token ID338
Feature activation-2.55e-02

Pos logit contributions
 pearl+0.003
aroo+0.003
rest+0.003
 seal+0.003
 slug+0.003
Neg logit contributions
 older-0.002
renches-0.002
 questions-0.002
ways-0.002
 exams-0.002
Token door
Token ID3420
Feature activation-1.11e-01

Pos logit contributions
renches+0.026
 older+0.024
 exams+0.022
ways+0.022
arre+0.022
Neg logit contributions
 pearl-0.055
 seal-0.054
rest-0.054
aroo-0.053
 slug-0.052
Token.
Token ID13
Feature activation+4.03e-03

Pos logit contributions
renches+0.149
 older+0.139
 exams+0.133
arre+0.131
ways+0.130
Neg logit contributions
 pearl-0.201
 seal-0.199
rest-0.197
aroo-0.196
 globe-0.187
Token She
Token ID1375
Feature activation-2.93e-02

Pos logit contributions
 seal+0.008
aroo+0.008
 pearl+0.008
rest+0.008
 Gor+0.008
Neg logit contributions
renches-0.005
oken-0.004
arre-0.004
 older-0.004
 headache-0.004
Token said
Token ID531
Feature activation+1.68e-03

Pos logit contributions
renches+0.046
 older+0.045
ways+0.042
 exams+0.042
arre+0.042
Neg logit contributions
 pearl-0.046
rest-0.045
 seal-0.045
aroo-0.045
 slug-0.043
Token
Token ID6184
Feature activation+3.45e-02

Pos logit contributions
 pearl+0.004
rest+0.004
 seal+0.004
 slug+0.004
aroo+0.004
Neg logit contributions
 older-0.002
renches-0.001
 exams-0.001
 questions-0.001
ways-0.001
Token
Token ID95
Feature activation-6.51e-03

Pos logit contributions
 pearl+0.050
 seal+0.050
aroo+0.049
rest+0.048
 slug+0.046
Neg logit contributions
renches-0.058
 older-0.056
arre-0.053
 exams-0.053
 headache-0.053
Token
Token ID26391
Feature activation+6.03e-02

Pos logit contributions
renches+0.011
 older+0.011
 exams+0.010
ways+0.010
 headache+0.010
Neg logit contributions
 pearl-0.009
 seal-0.009
rest-0.009
aroo-0.009
 slug-0.009
Token
Token ID129
Feature activation-4.38e-02

Pos logit contributions
cious+0.078
 globe+0.075
 seal+0.071
aroo+0.070
 pearl+0.068
Neg logit contributions
 older-0.116
renches-0.115
oken-0.109
 younger-0.105
 questions-0.104
Token
Token ID241
Feature activation+1.48e-02

Pos logit contributions
renches+0.069
 older+0.067
 exams+0.063
ways+0.063
arre+0.062
Neg logit contributions
 pearl-0.070
 seal-0.067
aroo-0.067
rest-0.066
 globe-0.064
Token"
Token ID1
Feature activation+5.48e-02

Pos logit contributions
renches+0.007
 older+0.007
ways+0.007
ups+0.006
 exams+0.006
Neg logit contributions
 pearl-0.009
 seal-0.008
rest-0.008
 globe-0.008
 diamond-0.008
TokenHelp
Token ID22087
Feature activation-2.27e-03

Pos logit contributions
 pearl+0.099
rily+0.090
rest+0.090
 globe+0.090
 seal+0.089
Neg logit contributions
renches-0.074
ups-0.069
 older-0.069
 crosses-0.067
 exams-0.067
Token!"
Token ID2474
Feature activation+6.65e-02

Pos logit contributions
renches+0.003
 older+0.003
ways+0.003
arre+0.003
 exams+0.003
Neg logit contributions
 pearl-0.004
 seal-0.004
aroo-0.004
rest-0.004
 slug-0.003
Token Tim
Token ID5045
Feature activation-2.02e-02

Pos logit contributions
aroo+0.094
 seal+0.093
 pearl+0.085
 slug+0.083
rest+0.082
Neg logit contributions
renches-0.119
arre-0.118
oken-0.111
 older-0.110
udd-0.110
Token and
Token ID290
Feature activation+1.26e-02

Pos logit contributions
renches+0.040
 headache+0.038
lights+0.038
oken+0.037
icker+0.037
Neg logit contributions
 seal-0.025
aroo-0.024
 Gor-0.022
 gob-0.022
 slug-0.022
Token Jen
Token ID13364
Feature activation+5.16e-03

Pos logit contributions
 seal+0.023
aroo+0.022
 pearl+0.022
rest+0.022
 slug+0.021
Neg logit contributions
renches-0.017
 older-0.016
arre-0.016
 exams-0.015
ways-0.015
Token scream
Token ID8196
Feature activation-2.33e-02

Pos logit contributions
aroo+0.007
 seal+0.007
 gob+0.007
rest+0.006
 perched+0.006
Neg logit contributions
renches-0.009
lights-0.009
aj-0.009
arre-0.009
 exams-0.008
Token.
Token ID13
Feature activation+1.15e-02

Pos logit contributions
renches+0.045
 exams+0.041
arre+0.041
ways+0.041
azes+0.040
Neg logit contributions
 seal-0.028
rest-0.027
rily-0.026
aroo-0.026
cious-0.025
Token They
Token ID1119
Feature activation-3.59e-02

Pos logit contributions
aroo+0.021
 seal+0.020
 pearl+0.019
 globe+0.019
 slug+0.019
Neg logit contributions
renches-0.016
 older-0.015
arre-0.014
 exams-0.014
oken-0.014
Token are
Token ID389
Feature activation+5.31e-02

Pos logit contributions
renches+0.058
 older+0.055
ways+0.053
 exams+0.053
arre+0.053
Neg logit contributions
 seal-0.057
aroo-0.055
 pearl-0.054
rest-0.053
 slug-0.051
Token scared
Token ID12008
Feature activation-5.93e-02

Pos logit contributions
 seal+0.091
rest+0.089
aroo+0.089
 pearl+0.088
 slug+0.084
Neg logit contributions
renches-0.078
ways-0.072
 exams-0.070
arre-0.070
 older-0.068
Token It
Token ID632
Feature activation+2.34e-02

Pos logit contributions
 seal+0.011
 pearl+0.011
aroo+0.010
 slug+0.010
rest+0.010
Neg logit contributions
renches-0.010
 older-0.010
arre-0.009
oken-0.009
 exams-0.009
Token looked
Token ID3114
Feature activation-6.06e-02

Pos logit contributions
 pearl+0.029
 seal+0.028
rest+0.027
aroo+0.027
 slug+0.025
Neg logit contributions
renches-0.044
 older-0.044
ways-0.042
 exams-0.041
 headache-0.041
Token like
Token ID588
Feature activation+6.57e-02

Pos logit contributions
renches+0.090
 older+0.089
 exams+0.082
ways+0.082
arre+0.081
Neg logit contributions
 pearl-0.100
aroo-0.098
rest-0.098
 seal-0.097
 globe-0.092
Token a
Token ID257
Feature activation-1.27e-02

Pos logit contributions
rest+0.110
aroo+0.110
rily+0.109
cious+0.109
 pearl+0.108
Neg logit contributions
 older-0.092
renches-0.086
 thunder-0.084
 exams-0.084
 headache-0.080
Token monster
Token ID9234
Feature activation+3.79e-02

Pos logit contributions
 older+0.018
 headache+0.017
renches+0.017
 exams+0.016
arre+0.016
Neg logit contributions
rest-0.023
aroo-0.021
 pearl-0.021
rily-0.020
cious-0.020
Token!
Token ID0
Feature activation+2.60e-02

Pos logit contributions
 pearl+0.087
rily+0.084
rest+0.080
cious+0.078
 coconut+0.078
Neg logit contributions
 older-0.029
ways-0.028
 thunder-0.027
 crosses-0.027
 headache-0.027
Token The
Token ID383
Feature activation-1.24e-02

Pos logit contributions
 seal+0.042
aroo+0.041
 pearl+0.041
 slug+0.040
rest+0.039
Neg logit contributions
renches-0.041
 older-0.039
oken-0.037
 exams-0.037
arre-0.037
Token dog
Token ID3290
Feature activation+5.51e-02

Pos logit contributions
renches+0.024
 older+0.022
ways+0.022
arre+0.021
 exams+0.021
Neg logit contributions
 seal-0.017
 pearl-0.016
aroo-0.016
rest-0.016
 globe-0.015
Token bark
Token ID21405
Feature activation-4.53e-02

Pos logit contributions
 pearl+0.081
rest+0.077
aroo+0.074
 seal+0.074
 globe+0.072
Neg logit contributions
renches-0.091
 older-0.091
ways-0.087
 exams-0.085
arre-0.085
Tokened
Token ID276
Feature activation+1.23e-01

Pos logit contributions
renches+0.093
ways+0.092
 older+0.091
arre+0.091
udd+0.088
Neg logit contributions
 pearl-0.050
aroo-0.044
 seal-0.043
 globe-0.041
rest-0.041
Token at
Token ID379
Feature activation-2.49e-02

Pos logit contributions
 pearl+0.229
 seal+0.225
rest+0.222
aroo+0.222
 slug+0.213
Neg logit contributions
renches-0.160
 older-0.151
 exams-0.144
ways-0.142
arre-0.141
Token for
Token ID329
Feature activation-1.83e-02

Pos logit contributions
 pearl+0.031
 seal+0.030
rest+0.029
aroo+0.028
 globe+0.026
Neg logit contributions
renches-0.073
 older-0.070
ways-0.068
 exams-0.068
arre-0.067
Token a
Token ID257
Feature activation-1.27e-02

Pos logit contributions
renches+0.028
 older+0.027
 exams+0.026
ways+0.025
 headache+0.025
Neg logit contributions
 pearl-0.029
aroo-0.029
 seal-0.029
rest-0.028
 globe-0.028
Token walk
Token ID2513
Feature activation-5.82e-02

Pos logit contributions
renches+0.016
 older+0.016
 headache+0.015
 exams+0.015
ways+0.015
Neg logit contributions
 pearl-0.023
rest-0.023
aroo-0.023
 seal-0.022
 slug-0.022
Token with
Token ID351
Feature activation+1.22e-02

Pos logit contributions
 older+0.073
renches+0.072
ways+0.071
 exams+0.067
ups+0.064
Neg logit contributions
 pearl-0.109
rest-0.104
 seal-0.103
aroo-0.103
 slug-0.102
Token his
Token ID465
Feature activation+5.04e-02

Pos logit contributions
 pearl+0.017
aroo+0.016
 seal+0.016
rest+0.016
 globe+0.016
Neg logit contributions
renches-0.021
 older-0.021
 exams-0.020
ways-0.020
 headache-0.020
Token mom
Token ID1995
Feature activation-9.30e-02

Pos logit contributions
 pearl+0.108
rest+0.108
 seal+0.107
aroo+0.106
cious+0.102
Neg logit contributions
 older-0.049
renches-0.049
 exams-0.043
 headache-0.043
ways-0.041
Token and
Token ID290
Feature activation+1.14e-02

Pos logit contributions
 older+0.079
 questions+0.062
ups+0.059
 exams+0.058
arre+0.057
Neg logit contributions
 pearl-0.223
aroo-0.222
rest-0.212
 seal-0.211
 slug-0.210
Token learned
Token ID4499
Feature activation-4.75e-02

Pos logit contributions
 pearl+0.023
rest+0.023
aroo+0.023
rily+0.022
 slug+0.022
Neg logit contributions
 older-0.013
 exams-0.012
renches-0.011
 headache-0.011
 questions-0.011
Token about
Token ID546
Feature activation-9.17e-03

Pos logit contributions
 older+0.076
renches+0.075
 exams+0.072
 questions+0.070
ups+0.069
Neg logit contributions
 pearl-0.072
aroo-0.072
rest-0.070
 seal-0.070
 slug-0.068
Token useful
Token ID4465
Feature activation+1.95e-02

Pos logit contributions
 older+0.013
renches+0.013
 exams+0.012
 meetings+0.012
 questions+0.012
Neg logit contributions
aroo-0.015
rest-0.015
 pearl-0.015
 seal-0.015
cious-0.015
Token squirrel
Token ID33039
Feature activation+6.71e-02

Pos logit contributions
 pearl+0.032
 seal+0.031
rest+0.030
 diamond+0.029
aroo+0.029
Neg logit contributions
renches-0.030
ways-0.029
 older-0.029
arre-0.029
 headache-0.027
Token loudly
Token ID23112
Feature activation-2.68e-02

Pos logit contributions
renches+0.004
 older+0.004
oken+0.004
arre+0.004
 exams+0.004
Neg logit contributions
rest-0.004
 seal-0.004
 pearl-0.004
aroo-0.004
 globe-0.004
Token.
Token ID13
Feature activation+4.81e-03

Pos logit contributions
renches+0.033
 older+0.032
 exams+0.030
ways+0.030
arre+0.029
Neg logit contributions
 pearl-0.051
 seal-0.050
rest-0.050
aroo-0.050
 slug-0.048
Token All
Token ID1439
Feature activation-8.03e-02

Pos logit contributions
 slug+0.007
 seal+0.007
aroo+0.007
 gob+0.007
 Gor+0.007
Neg logit contributions
renches-0.008
oken-0.008
ups-0.007
lights-0.007
 exams-0.007
Token the
Token ID262
Feature activation-2.89e-02

Pos logit contributions
renches+0.137
 older+0.130
ways+0.125
 exams+0.125
arre+0.124
Neg logit contributions
 pearl-0.118
 seal-0.116
rest-0.115
aroo-0.115
 slug-0.108
Token other
Token ID584
Feature activation+5.98e-02

Pos logit contributions
renches+0.051
ways+0.050
 exams+0.048
 older+0.047
arre+0.047
Neg logit contributions
 seal-0.041
 pearl-0.041
rest-0.039
aroo-0.038
 slug-0.037
Token birds
Token ID10087
Feature activation-8.45e-02

Pos logit contributions
 seal+0.094
rest+0.093
 pearl+0.090
aroo+0.089
 slug+0.084
Neg logit contributions
renches-0.100
 headache-0.093
ways-0.093
 exams-0.093
arre-0.089
Token came
Token ID1625
Feature activation-8.66e-02

Pos logit contributions
 older+0.124
renches+0.123
arre+0.116
ways+0.115
 headache+0.113
Neg logit contributions
 pearl-0.140
rest-0.131
 seal-0.130
 slug-0.130
aroo-0.128
Token to
Token ID284
Feature activation-1.59e-02

Pos logit contributions
renches+0.124
 older+0.115
 exams+0.112
ways+0.110
 headache+0.109
Neg logit contributions
 pearl-0.150
 seal-0.149
rest-0.147
aroo-0.147
 slug-0.139
Token see
Token ID766
Feature activation-3.19e-02

Pos logit contributions
renches+0.023
 older+0.023
 exams+0.021
 headache+0.021
ways+0.021
Neg logit contributions
 pearl-0.027
rest-0.026
aroo-0.026
 seal-0.026
 slug-0.025
Token what
Token ID644
Feature activation+7.91e-02

Pos logit contributions
 older+0.040
renches+0.038
 questions+0.036
 Sometimes+0.036
 exams+0.035
Neg logit contributions
 pearl-0.064
cious-0.060
 seal-0.060
 globe-0.059
rest-0.059
Token was
Token ID373
Feature activation+1.41e-02

Pos logit contributions
 pearl+0.109
 seal+0.103
rest+0.102
aroo+0.101
 globe+0.097
Neg logit contributions
renches-0.141
 older-0.138
 exams-0.133
ways-0.130
arre-0.129
Token around
Token ID1088
Feature activation-4.45e-02

Pos logit contributions
 older+0.093
renches+0.091
ways+0.087
arre+0.084
 exams+0.083
Neg logit contributions
 pearl-0.093
rest-0.091
aroo-0.089
 seal-0.088
 slug-0.087
Token in
Token ID287
Feature activation+2.39e-03

Pos logit contributions
 older+0.061
 exams+0.057
 questions+0.057
renches+0.057
 early+0.054
Neg logit contributions
rest-0.079
 pearl-0.077
cious-0.074
aroo-0.074
cream-0.073
Token the
Token ID262
Feature activation-2.37e-02

Pos logit contributions
rest+0.005
aroo+0.005
cious+0.005
cream+0.005
 pearl+0.005
Neg logit contributions
 older-0.003
 questions-0.003
 ways-0.002
 exams-0.002
 ideas-0.002
Token pond
Token ID16723
Feature activation+4.56e-02

Pos logit contributions
 older+0.033
renches+0.029
 younger+0.028
arre+0.027
 exams+0.026
Neg logit contributions
rest-0.044
aroo-0.044
 pearl-0.042
rily-0.041
cious-0.041
Token.
Token ID13
Feature activation+2.69e-03

Pos logit contributions
 pearl+0.084
 seal+0.082
rest+0.082
aroo+0.082
 slug+0.079
Neg logit contributions
renches-0.058
 older-0.057
ways-0.053
 exams-0.052
arre-0.052
Token Everyone
Token ID11075
Feature activation-9.75e-02

Pos logit contributions
 pearl+0.005
 seal+0.005
rest+0.005
aroo+0.005
 slug+0.005
Neg logit contributions
renches-0.003
 older-0.003
 exams-0.003
ways-0.003
arre-0.003
Token was
Token ID373
Feature activation+1.79e-02

Pos logit contributions
 crossed+0.096
 older+0.093
ups+0.091
 brushes+0.086
 shook+0.086
Neg logit contributions
ifying-0.193
stack-0.192
rily-0.190
 Gor-0.188
 pearl-0.187
Token having
Token ID1719
Feature activation+6.46e-02

Pos logit contributions
 pearl+0.033
aroo+0.032
 globe+0.031
rest+0.031
 slug+0.031
Neg logit contributions
 older-0.027
 younger-0.022
renches-0.022
arre-0.021
oken-0.020
Token fun
Token ID1257
Feature activation-9.45e-02

Pos logit contributions
aroo+0.110
 pearl+0.108
rest+0.107
 seal+0.105
 globe+0.104
Neg logit contributions
 older-0.094
renches-0.094
 exams-0.087
arre-0.085
ways-0.084
Token and
Token ID290
Feature activation+6.58e-03

Pos logit contributions
renches+0.126
 older+0.120
 exams+0.114
ways+0.114
arre+0.112
Neg logit contributions
 pearl-0.173
 seal-0.169
aroo-0.168
rest-0.167
 slug-0.160
Token the
Token ID262
Feature activation-2.36e-02

Pos logit contributions
 pearl+0.010
aroo+0.010
 seal+0.010
rest+0.010
 globe+0.010
Neg logit contributions
renches-0.010
 older-0.010
ways-0.010
 exams-0.009
arre-0.009
Token.
Token ID13
Feature activation-6.85e-03

Pos logit contributions
 pearl+0.083
aroo+0.079
 seal+0.079
rest+0.078
 slug+0.077
Neg logit contributions
 older-0.042
renches-0.042
ways-0.039
 exams-0.038
ups-0.037
Token Every
Token ID3887
Feature activation-1.16e-02

Pos logit contributions
renches+0.009
 older+0.008
oken+0.008
arre+0.008
ups+0.008
Neg logit contributions
 seal-0.013
 pearl-0.012
aroo-0.012
 globe-0.012
rest-0.012
Token day
Token ID1110
Feature activation+5.95e-02

Pos logit contributions
 older+0.015
 headache+0.015
 exams+0.013
renches+0.013
 lecture+0.013
Neg logit contributions
 pearl-0.020
rily-0.020
rest-0.020
aroo-0.020
cious-0.020
Token they
Token ID484
Feature activation+2.37e-03

Pos logit contributions
 pearl+0.118
rest+0.115
 seal+0.115
aroo+0.115
 slug+0.111
Neg logit contributions
renches-0.069
 older-0.067
 exams-0.063
ways-0.062
 headache-0.060
Token would
Token ID561
Feature activation+2.55e-02

Pos logit contributions
 pearl+0.004
aroo+0.003
 seal+0.003
rest+0.003
 slug+0.003
Neg logit contributions
renches-0.004
 older-0.004
 exams-0.004
ways-0.004
arre-0.004
Token stretch
Token ID7539
Feature activation-7.37e-02

Pos logit contributions
 pearl+0.044
 seal+0.044
rest+0.043
aroo+0.043
 globe+0.041
Neg logit contributions
renches-0.036
 older-0.034
ways-0.033
 exams-0.033
arre-0.032
Token out
Token ID503
Feature activation-4.15e-02

Pos logit contributions
 older+0.096
renches+0.091
 headache+0.088
 exams+0.088
 questions+0.087
Neg logit contributions
rest-0.134
 pearl-0.131
 slug-0.128
aroo-0.127
 seal-0.125
Token in
Token ID287
Feature activation-1.30e-03

Pos logit contributions
renches+0.077
arre+0.070
 older+0.070
 headache+0.069
ways+0.069
Neg logit contributions
 pearl-0.054
 seal-0.054
rest-0.052
aroo-0.052
 globe-0.050
Token their
Token ID511
Feature activation+6.74e-02

Pos logit contributions
 older+0.002
renches+0.002
 exams+0.002
 headache+0.002
 questions+0.002
Neg logit contributions
aroo-0.002
rest-0.002
 pearl-0.002
 seal-0.002
 slug-0.002
Token front
Token ID2166
Feature activation-8.99e-03

Pos logit contributions
 pearl+0.120
rest+0.118
 seal+0.118
aroo+0.118
 slug+0.112
Neg logit contributions
renches-0.091
 older-0.088
 exams-0.083
ways-0.082
 headache-0.081
Token yard
Token ID12699
Feature activation-4.68e-02

Pos logit contributions
renches+0.013
arre+0.012
 headache+0.011
oken+0.011
 older+0.011
Neg logit contributions
 pearl-0.015
 seal-0.015
 globe-0.014
aroo-0.014
 diamond-0.014
Token
Token ID198
Feature activation-8.15e-03

Pos logit contributions
renches+0.013
 older+0.012
 exams+0.012
ways+0.012
arre+0.012
Neg logit contributions
 pearl-0.013
 seal-0.013
aroo-0.013
rest-0.013
 slug-0.012
TokenBen
Token ID11696
Feature activation-9.64e-03

Pos logit contributions
renches+0.011
 older+0.011
ways+0.010
 exams+0.010
 headache+0.010
Neg logit contributions
 pearl-0.015
aroo-0.014
 seal-0.014
 slug-0.014
rest-0.014
Token and
Token ID290
Feature activation+8.19e-03

Pos logit contributions
renches+0.016
 older+0.014
 exams+0.014
ways+0.014
oken+0.014
Neg logit contributions
 seal-0.015
 pearl-0.015
 slug-0.015
aroo-0.015
rest-0.015
Token Lily
Token ID20037
Feature activation-3.67e-02

Pos logit contributions
 seal+0.014
 pearl+0.014
aroo+0.014
 slug+0.014
rest+0.014
Neg logit contributions
renches-0.012
ways-0.011
arre-0.011
oken-0.010
 older-0.010
Token did
Token ID750
Feature activation-1.21e-02

Pos logit contributions
renches+0.063
 exams+0.060
 older+0.059
ways+0.055
oken+0.055
Neg logit contributions
 seal-0.056
 pearl-0.054
 gob-0.052
aroo-0.051
cious-0.051
Token not
Token ID407
Feature activation-1.14e-01

Pos logit contributions
renches+0.012
 older+0.011
 exams+0.011
ways+0.011
arre+0.010
Neg logit contributions
 pearl-0.026
 seal-0.026
rest-0.025
aroo-0.025
 slug-0.024
Token believe
Token ID1975
Feature activation+4.71e-02

Pos logit contributions
renches+0.193
ways+0.185
lights+0.183
aj+0.175
azes+0.173
Neg logit contributions
rest-0.184
 seal-0.183
 pearl-0.168
aroo-0.160
 gob-0.155
Token Mom
Token ID11254
Feature activation-6.87e-02

Pos logit contributions
 pearl+0.083
rest+0.081
 seal+0.080
cious+0.080
aroo+0.079
Neg logit contributions
 older-0.063
renches-0.061
 exams-0.060
 questions-0.057
lights-0.057
Token.
Token ID13
Feature activation+7.21e-03

Pos logit contributions
 older+0.069
renches+0.066
ups+0.061
 questions+0.060
ways+0.059
Neg logit contributions
 pearl-0.148
 seal-0.143
aroo-0.143
rest-0.138
 slug-0.136
Token They
Token ID1119
Feature activation-4.17e-02

Pos logit contributions
 seal+0.014
 pearl+0.013
 slug+0.013
 globe+0.013
aroo+0.012
Neg logit contributions
renches-0.009
oken-0.008
 older-0.008
 exams-0.008
ups-0.008
Token thought
Token ID1807
Feature activation-1.88e-02

Pos logit contributions
renches+0.067
 older+0.064
 exams+0.061
ways+0.060
arre+0.060
Neg logit contributions
 pearl-0.065
 seal-0.064
rest-0.063
aroo-0.063
 slug-0.060