TOP ACTIVATIONS
MAX = 0.106

Le ah replied , " I have a
moon . <|endoftext|> Once upon a time , there was a
park . <|endoftext|> Once upon a time , there was a
again . <|endoftext|> Once upon a time there was a tiny
smile . <|endoftext|> Once upon a time , in a small
safe . <|endoftext|> Once upon a time , there was a
THE <|endoftext|> Once upon a time there was a little
together . <|endoftext|> Once upon a time , in a big
shore ! <|endoftext|> Once upon a time there lived a giant
ball . <|endoftext|> Once upon a time , there was a
soup . <|endoftext|> Once upon a time , there was a
end . <|endoftext|> Once upon a time , there was a
â <|endoftext|> Once upon a time there were two best
safe . <|endoftext|> Once upon a time , there was a
her . <|endoftext|> Once upon a time , there was a
make honey again . <|endoftext|> One day , a little boy
slides . <|endoftext|> Once upon a time , there was a
day !" <|endoftext|> Once upon a time there was a girl
garden . <|endoftext|> Once upon a time , there was a
them . <|endoftext|> Once upon a time there was a man

MIN ACTIVATIONS
MIN = -0.111

and always had plenty of rad ishes to eat . The
world where everyone shared their rad ishes . <|endoftext|> Once upon
other things that might make it not as good . And
and it will be bigger and prett ier than yours ."
John chose some carrots and cuc umbers to take home .
like lettuce , tomatoes and cuc umbers . He smiled as
play together without making boring or loud sounds that could hurt
can make it go faster or slower . Sam is very
, we found a rare rad ish !" they shouted ,
some buttons or rib bons or patches . Would you like
okay , Anna ? Does it hurt a lot ?"
not something you can buy or take . Love is something
more important than being clever or playing tricks . <|endoftext|> Once
we can make some cards or drawings to cheer up the
Please help me shake this rad ish , it 's too
things that could cause harm or trouble . <|endoftext|> Once upon
make sure nothing would make it slip away again . Then
hopefully you will be healthy and back to your old self
real wheat could make good bread . And Jack loved to
's important to be quiet and stay away from them .

INTERVAL 0.051 - 0.106
CONTAINS 5.333%

Tom my said ,
of the vendors and politely said , " Hello , could
not afraid ." The little girl realised that if
L ily 's mom my warned her that it might be
Once upon a time , there was a little girl named

INTERVAL -0.003 - 0.051
CONTAINS 46.596%

She waited for a quiet moment . Then she
?" asked Charles . " I 'm making a
lot . Tim had a fun day at the park with
smile . " If you guess right , you get a
later . She learned a lesson : never go on an

INTERVAL -0.057 - -0.003
CONTAINS 46.429%

Thank you for forgiving me ." The moral of
L ily learned that it 's important to take care
sat down to have their picnic . After they finished ,
in a big , dark house , there lived a stubborn
noticed the fan was dirty with dust and leaves . Lily

INTERVAL -0.111 - -0.057
CONTAINS 1.641%

isy always wanted more sun and more rain , even when
that can 't be changed or fixed . He took this
could return to the zoo and see more animals , because
fire and she quickly put it out . Tim my felt
. They start to kick and scream . They say ,
Token
Token ID220
Feature activation+7.32e-04

Pos logit contributions
ited+0.009
 saw+0.008
 heard+0.007
Hey+0.007
 year+0.007
Neg logit contributions
anks-0.012
balls-0.012
 hyd-0.011
 heal-0.011
 rid-0.011
Token
Token ID198
Feature activation+1.79e-02

Pos logit contributions
ited+0.001
 saw+0.001
 heard+0.001
 two+0.000
 little+0.000
Neg logit contributions
 hyd-0.002
anks-0.002
balls-0.002
 rid-0.002
 bleed-0.002
Token
Token ID198
Feature activation+9.74e-03

Pos logit contributions
ited+0.028
 saw+0.022
 heard+0.021
Hey+0.020
 year+0.019
Neg logit contributions
anks-0.034
balls-0.034
 heal-0.031
 hyd-0.031
 rid-0.031
TokenLe
Token ID3123
Feature activation+2.69e-02

Pos logit contributions
ited+0.016
 saw+0.013
 heard+0.012
 little+0.011
 year+0.011
Neg logit contributions
balls-0.017
anks-0.017
 hyd-0.016
 heal-0.016
fly-0.016
Tokenah
Token ID993
Feature activation+3.68e-02

Pos logit contributions
ited+0.063
 saw+0.058
 heard+0.056
Hey+0.055
 little+0.054
Neg logit contributions
anks-0.027
balls-0.025
fly-0.025
 hyd-0.025
ometer-0.024
Token replied
Token ID8712
Feature activation+1.04e-01

Pos logit contributions
ited+0.042
Hey+0.032
grown+0.030
 saw+0.028
 heard+0.028
Neg logit contributions
anks-0.076
balls-0.075
 hyd-0.075
 heal-0.074
ometer-0.073
Token,
Token ID11
Feature activation+4.20e-02

Pos logit contributions
ited+0.043
 saw+0.041
 heard+0.033
 little+0.023
 two+0.020
Neg logit contributions
anks-0.309
 hyd-0.301
balls-0.296
 rid-0.294
 bleed-0.291
Token "
Token ID366
Feature activation-1.61e-02

Pos logit contributions
 saw+0.052
ited+0.051
 heard+0.048
 little+0.044
 two+0.043
Neg logit contributions
anks-0.092
 hyd-0.085
 rid-0.085
balls-0.084
 bleed-0.083
TokenI
Token ID40
Feature activation-1.14e-02

Pos logit contributions
anks+0.036
balls+0.034
 hyd+0.033
 rid+0.033
 heal+0.033
Neg logit contributions
ited-0.020
 saw-0.019
 heard-0.018
Hey-0.016
 year-0.015
Token have
Token ID423
Feature activation+1.19e-02

Pos logit contributions
anks+0.020
 hyd+0.019
 rid+0.019
 heal+0.019
balls+0.019
Neg logit contributions
ited-0.019
 saw-0.016
 heard-0.016
Hey-0.015
 year-0.015
Token a
Token ID257
Feature activation+1.27e-02

Pos logit contributions
ited+0.017
 saw+0.013
Hey+0.013
 heard+0.012
 year+0.012
Neg logit contributions
 rid-0.023
anks-0.023
 hyd-0.023
balls-0.023
 heal-0.023
Token moon
Token ID8824
Feature activation-1.47e-02

Pos logit contributions
ited+0.016
 saw+0.014
 heard+0.013
Hey+0.012
 did+0.012
Neg logit contributions
anks-0.017
balls-0.016
 hyd-0.016
 heal-0.016
 rid-0.016
Token.
Token ID13
Feature activation-1.57e-02

Pos logit contributions
anks+0.029
balls+0.028
 rid+0.028
 bleed+0.027
 hyd+0.027
Neg logit contributions
ited-0.020
 saw-0.020
 heard-0.019
 two-0.017
 little-0.016
Token<|endoftext|>
Token ID50256
Feature activation+7.77e-02

Pos logit contributions
anks+0.034
balls+0.033
 hyd+0.032
 rid+0.032
 heal+0.032
Neg logit contributions
ited-0.020
 saw-0.018
 heard-0.017
 little-0.015
 two-0.015
TokenOnce
Token ID7454
Feature activation+9.11e-02

Pos logit contributions
ited+0.126
 saw+0.097
 heard+0.093
 conflict+0.086
 noticed+0.085
Neg logit contributions
anks-0.146
balls-0.145
 hyd-0.139
 rid-0.137
 heal-0.136
Token upon
Token ID2402
Feature activation+7.34e-02

Pos logit contributions
ited+0.102
Hey+0.071
 heard+0.070
 saw+0.068
Fine+0.058
Neg logit contributions
balls-0.209
anks-0.205
 rid-0.198
ectar-0.196
 heal-0.194
Token a
Token ID257
Feature activation+1.04e-01

Pos logit contributions
ited+0.103
 saw+0.074
Hey+0.070
 conflict+0.068
 heard+0.067
Neg logit contributions
 heal-0.149
balls-0.148
 hyd-0.145
 bleed-0.145
 rid-0.143
Token time
Token ID640
Feature activation+7.24e-02

Pos logit contributions
ited+0.080
 saw+0.062
Hey+0.060
 heard+0.058
 two+0.054
Neg logit contributions
sing-0.244
balls-0.242
 hyd-0.241
 heal-0.232
anks-0.231
Token,
Token ID11
Feature activation+8.58e-02

Pos logit contributions
ited+0.079
 saw+0.054
 heard+0.052
Hey+0.045
 noticed+0.041
Neg logit contributions
balls-0.175
anks-0.170
 hyd-0.163
 heal-0.161
sing-0.161
Token there
Token ID612
Feature activation+8.55e-02

Pos logit contributions
ited+0.140
oph+0.108
aman+0.097
 amusement+0.095
 café+0.089
Neg logit contributions
balls-0.163
fly-0.155
sing-0.154
ectar-0.151
 hyd-0.150
Token was
Token ID373
Feature activation+7.25e-02

Pos logit contributions
ited+0.144
vious+0.091
Exc+0.091
Next+0.090
 conflict+0.089
Neg logit contributions
balls-0.152
 heal-0.147
 hyd-0.144
fly-0.141
cr-0.141
Token a
Token ID257
Feature activation+6.22e-02

Pos logit contributions
ited+0.106
 saw+0.081
 heard+0.078
 sighed+0.071
Hey+0.070
Neg logit contributions
balls-0.142
 hyd-0.140
anks-0.139
ectar-0.137
 rid-0.136
Token park
Token ID3952
Feature activation+4.00e-03

Pos logit contributions
ited+0.015
 heard+0.011
 frowned+0.010
ised+0.010
 sighed+0.010
Neg logit contributions
 heal-0.019
 float-0.018
 hyd-0.018
balls-0.018
 drown-0.017
Token.
Token ID13
Feature activation-1.69e-02

Pos logit contributions
ited+0.005
 saw+0.004
 heard+0.004
Hey+0.004
 two+0.004
Neg logit contributions
anks-0.009
balls-0.008
 hyd-0.008
 heal-0.008
 rid-0.008
Token<|endoftext|>
Token ID50256
Feature activation+7.37e-02

Pos logit contributions
anks+0.039
balls+0.038
 hyd+0.037
 rid+0.037
 heal+0.036
Neg logit contributions
ited-0.018
 saw-0.017
 heard-0.016
 two-0.014
 little-0.014
TokenOnce
Token ID7454
Feature activation+8.83e-02

Pos logit contributions
ited+0.124
 saw+0.094
 heard+0.090
 conflict+0.085
 noticed+0.084
Neg logit contributions
anks-0.134
balls-0.133
 hyd-0.128
 heal-0.125
 rid-0.125
Token upon
Token ID2402
Feature activation+6.94e-02

Pos logit contributions
ited+0.095
 heard+0.065
 saw+0.065
Hey+0.065
 year+0.052
Neg logit contributions
anks-0.206
balls-0.206
 rid-0.196
 heal-0.195
ometer-0.194
Token a
Token ID257
Feature activation+1.03e-01

Pos logit contributions
ited+0.083
 saw+0.062
 heard+0.057
Hey+0.055
 year+0.051
Neg logit contributions
balls-0.153
anks-0.152
 heal-0.152
 hyd-0.152
 bleed-0.149
Token time
Token ID640
Feature activation+7.12e-02

Pos logit contributions
ited+0.076
 saw+0.056
Hey+0.053
 heard+0.052
 two+0.047
Neg logit contributions
 hyd-0.249
sing-0.247
balls-0.247
anks-0.241
 heal-0.241
Token,
Token ID11
Feature activation+8.43e-02

Pos logit contributions
ited+0.070
 saw+0.049
 heard+0.047
Hey+0.040
 noticed+0.035
Neg logit contributions
balls-0.177
anks-0.175
 hyd-0.169
 heal-0.166
sing-0.164
Token there
Token ID612
Feature activation+8.22e-02

Pos logit contributions
ited+0.126
oph+0.089
aman+0.079
Hey+0.075
 sighed+0.074
Neg logit contributions
balls-0.171
fly-0.166
ectar-0.162
 hyd-0.162
 heal-0.162
Token was
Token ID373
Feature activation+6.65e-02

Pos logit contributions
ited+0.125
 conflict+0.076
Hey+0.076
 enough+0.076
vious+0.076
Neg logit contributions
balls-0.158
 heal-0.155
 hyd-0.154
fly-0.149
anks-0.148
Token a
Token ID257
Feature activation+6.04e-02

Pos logit contributions
ited+0.094
 saw+0.072
 heard+0.069
Hey+0.062
 sighed+0.062
Neg logit contributions
balls-0.131
 hyd-0.131
anks-0.129
 rid-0.127
uggets-0.127
Token again
Token ID757
Feature activation+9.32e-03

Pos logit contributions
ited+0.035
 saw+0.033
 heard+0.032
 year+0.028
 little+0.028
Neg logit contributions
anks-0.048
balls-0.046
 hyd-0.045
 rid-0.044
 bleed-0.044
Token.
Token ID13
Feature activation-2.57e-02

Pos logit contributions
ited+0.018
 saw+0.015
Hey+0.015
 heard+0.014
 year+0.014
Neg logit contributions
anks-0.014
 hyd-0.014
balls-0.014
 rid-0.013
 heal-0.013
Token<|endoftext|>
Token ID50256
Feature activation+6.39e-02

Pos logit contributions
anks+0.057
balls+0.056
 hyd+0.055
 rid+0.054
 heal+0.054
Neg logit contributions
 saw-0.029
ited-0.028
 heard-0.026
 two-0.024
 little-0.024
TokenOnce
Token ID7454
Feature activation+7.92e-02

Pos logit contributions
ited+0.071
 saw+0.055
 heard+0.051
 noticed+0.043
 little+0.042
Neg logit contributions
anks-0.152
balls-0.149
 hyd-0.145
 rid-0.142
 heal-0.142
Token upon
Token ID2402
Feature activation+6.72e-02

Pos logit contributions
ited+0.083
 saw+0.057
 heard+0.057
Hey+0.055
 year+0.046
Neg logit contributions
anks-0.190
balls-0.189
 hyd-0.180
 rid-0.179
 heal-0.179
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.071
 saw+0.056
 heard+0.051
Hey+0.045
 year+0.044
Neg logit contributions
anks-0.158
balls-0.157
 hyd-0.155
 heal-0.153
 rid-0.152
Token time
Token ID640
Feature activation+6.83e-02

Pos logit contributions
ited+0.068
 saw+0.047
 heard+0.042
Hey+0.038
 two+0.033
Neg logit contributions
balls-0.267
 hyd-0.266
anks-0.265
sing-0.260
 heal-0.259
Token there
Token ID612
Feature activation+7.26e-02

Pos logit contributions
ited+0.053
 saw+0.040
 heard+0.037
Hey+0.028
 year+0.026
Neg logit contributions
anks-0.181
balls-0.179
 hyd-0.175
 rid-0.171
 heal-0.171
Token was
Token ID373
Feature activation+7.19e-02

Pos logit contributions
ited+0.096
 enough+0.059
 heard+0.058
 conflict+0.057
 saw+0.057
Neg logit contributions
balls-0.153
 hyd-0.150
 heal-0.149
anks-0.146
fly-0.146
Token a
Token ID257
Feature activation+6.07e-02

Pos logit contributions
ited+0.071
 saw+0.058
 heard+0.054
Hey+0.044
 year+0.044
Neg logit contributions
anks-0.176
balls-0.173
 hyd-0.170
 rid-0.167
 heal-0.166
Token tiny
Token ID7009
Feature activation+4.49e-02

Pos logit contributions
 saw+0.039
ited+0.037
 little+0.032
 heard+0.031
 year+0.024
Neg logit contributions
anks-0.174
balls-0.164
 rid-0.162
 hyd-0.160
 bleed-0.159
Token smile
Token ID8212
Feature activation+1.19e-02

Pos logit contributions
ited+0.025
 saw+0.024
 heard+0.022
 little+0.020
 two+0.019
Neg logit contributions
anks-0.050
balls-0.048
 hyd-0.047
 rid-0.046
ometers-0.045
Token.
Token ID13
Feature activation-1.14e-02

Pos logit contributions
ited+0.018
 saw+0.016
 heard+0.015
 two+0.014
 year+0.014
Neg logit contributions
anks-0.023
balls-0.022
 hyd-0.022
 rid-0.022
 heal-0.021
Token<|endoftext|>
Token ID50256
Feature activation+8.40e-02

Pos logit contributions
anks+0.029
 rid+0.028
balls+0.028
 hyd+0.027
 bleed+0.027
Neg logit contributions
 saw-0.010
 heard-0.008
 two-0.008
 little-0.008
ited-0.008
TokenOnce
Token ID7454
Feature activation+9.32e-02

Pos logit contributions
ited+0.129
 saw+0.101
 heard+0.096
Hey+0.088
 noticed+0.087
Neg logit contributions
anks-0.165
balls-0.163
 hyd-0.158
 rid-0.155
 heal-0.155
Token upon
Token ID2402
Feature activation+7.30e-02

Pos logit contributions
ited+0.098
 heard+0.067
 saw+0.066
Hey+0.065
 year+0.052
Neg logit contributions
balls-0.221
anks-0.219
 rid-0.209
 heal-0.208
 hyd-0.208
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.094
 saw+0.068
 heard+0.062
Hey+0.062
 conflict+0.058
Neg logit contributions
balls-0.156
 heal-0.156
 hyd-0.155
anks-0.154
 bleed-0.152
Token time
Token ID640
Feature activation+7.46e-02

Pos logit contributions
ited+0.072
 saw+0.052
 heard+0.047
Hey+0.046
 two+0.039
Neg logit contributions
 hyd-0.257
balls-0.255
sing-0.250
anks-0.249
 heal-0.249
Token,
Token ID11
Feature activation+8.80e-02

Pos logit contributions
ited+0.069
 saw+0.050
 heard+0.047
Hey+0.039
 noticed+0.035
Neg logit contributions
balls-0.188
anks-0.188
 hyd-0.182
 heal-0.177
 rid-0.176
Token in
Token ID287
Feature activation+7.18e-02

Pos logit contributions
ited+0.155
oph+0.117
aman+0.106
 heard+0.102
 amusement+0.102
Neg logit contributions
balls-0.157
sing-0.150
 hyd-0.149
fly-0.148
 heal-0.147
Token a
Token ID257
Feature activation+7.01e-02

Pos logit contributions
ited+0.099
 saw+0.075
 heard+0.073
Hey+0.067
 noticed+0.066
Neg logit contributions
 hyd-0.148
 heal-0.144
balls-0.144
 rid-0.140
 bleed-0.139
Token small
Token ID1402
Feature activation+5.26e-02

Pos logit contributions
ited+0.098
Hey+0.072
 heard+0.072
 enough+0.068
 saw+0.067
Neg logit contributions
 hyd-0.141
 heal-0.135
balls-0.130
sing-0.129
 drown-0.129
Token safe
Token ID3338
Feature activation-2.25e-02

Pos logit contributions
anks+0.116
balls+0.111
 hyd+0.107
 rid+0.104
 heal+0.102
Neg logit contributions
ited-0.107
 saw-0.102
 heard-0.097
 little-0.087
 two-0.087
Token.
Token ID13
Feature activation-2.21e-02

Pos logit contributions
anks+0.045
balls+0.044
 hyd+0.041
ophone+0.041
 rid+0.041
Neg logit contributions
ited-0.030
 saw-0.029
 heard-0.028
 two-0.025
 year-0.025
Token<|endoftext|>
Token ID50256
Feature activation+7.29e-02

Pos logit contributions
anks+0.057
balls+0.057
 hyd+0.056
 rid+0.056
 bleed+0.055
Neg logit contributions
 saw-0.017
ited-0.014
 little-0.014
 heard-0.014
 two-0.013
TokenOnce
Token ID7454
Feature activation+8.97e-02

Pos logit contributions
ited+0.121
 saw+0.092
 heard+0.088
 conflict+0.083
 noticed+0.082
Neg logit contributions
anks-0.134
balls-0.132
 hyd-0.128
 rid-0.125
 heal-0.124
Token upon
Token ID2402
Feature activation+6.97e-02

Pos logit contributions
ited+0.100
Hey+0.072
 heard+0.068
 saw+0.066
Fine+0.058
Neg logit contributions
balls-0.203
anks-0.203
 rid-0.194
ectar-0.194
ometer-0.192
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.093
 saw+0.067
Hey+0.062
 heard+0.061
 conflict+0.059
Neg logit contributions
 heal-0.146
balls-0.145
 hyd-0.143
 bleed-0.143
anks-0.142
Token time
Token ID640
Feature activation+7.28e-02

Pos logit contributions
ited+0.077
 saw+0.060
Hey+0.058
 heard+0.055
 two+0.051
Neg logit contributions
sing-0.240
 hyd-0.239
balls-0.238
anks-0.230
 heal-0.230
Token,
Token ID11
Feature activation+8.47e-02

Pos logit contributions
ited+0.074
 saw+0.051
 heard+0.049
Hey+0.042
 noticed+0.038
Neg logit contributions
balls-0.179
anks-0.176
 hyd-0.171
 heal-0.168
ectar-0.166
Token there
Token ID612
Feature activation+8.41e-02

Pos logit contributions
ited+0.131
oph+0.096
aman+0.087
 amusement+0.083
Fine+0.081
Neg logit contributions
balls-0.165
fly-0.161
 hyd-0.157
ectar-0.157
 heal-0.155
Token was
Token ID373
Feature activation+6.77e-02

Pos logit contributions
ited+0.136
vious+0.084
Exc+0.083
 conflict+0.083
Hey+0.082
Neg logit contributions
balls-0.155
 heal-0.152
 hyd-0.149
fly-0.146
cr-0.143
Token a
Token ID257
Feature activation+5.66e-02

Pos logit contributions
ited+0.101
 saw+0.077
 heard+0.074
 sighed+0.068
Hey+0.066
Neg logit contributions
balls-0.130
 hyd-0.129
anks-0.128
ectar-0.126
uggets-0.126
Token
Token ID198
Feature activation-1.09e-02

Pos logit contributions
anks+0.013
 hyd+0.013
balls+0.013
 rid+0.013
 bleed+0.012
Neg logit contributions
ited-0.005
 saw-0.005
 heard-0.005
 two-0.004
 little-0.004
TokenTHE
Token ID10970
Feature activation-2.14e-02

Pos logit contributions
anks+0.025
balls+0.025
 hyd+0.024
 rid+0.024
 heal+0.024
Neg logit contributions
ited-0.012
 saw-0.011
 heard-0.010
 little-0.008
 year-0.008
Token<|endoftext|>
Token ID50256
Feature activation+6.92e-02

Pos logit contributions
anks+0.054
 rid+0.052
ectar+0.052
uggets+0.051
 hyd+0.051
Neg logit contributions
 saw-0.019
 heard-0.017
 did-0.016
 thought-0.015
 two-0.014
TokenOnce
Token ID7454
Feature activation+8.87e-02

Pos logit contributions
ited+0.112
 saw+0.078
 heard+0.076
 conflict+0.075
 enough+0.073
Neg logit contributions
anks-0.130
balls-0.127
 hyd-0.125
 rid-0.121
 heal-0.120
Token upon
Token ID2402
Feature activation+7.03e-02

Pos logit contributions
ited+0.107
Hey+0.077
 heard+0.074
 saw+0.069
Fine+0.067
Neg logit contributions
balls-0.191
anks-0.187
ometer-0.183
 heal-0.181
 rid-0.181
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.106
 saw+0.075
Hey+0.072
 conflict+0.071
oph+0.068
Neg logit contributions
 heal-0.140
 bleed-0.135
balls-0.135
 hyd-0.132
 rid-0.132
Token time
Token ID640
Feature activation+7.30e-02

Pos logit contributions
ited+0.073
 saw+0.056
 heard+0.050
Hey+0.047
 two+0.042
Neg logit contributions
 hyd-0.252
balls-0.251
sing-0.249
 heal-0.245
anks-0.245
Token there
Token ID612
Feature activation+7.50e-02

Pos logit contributions
ited+0.069
 saw+0.049
 heard+0.047
Hey+0.038
 two+0.035
Neg logit contributions
balls-0.183
anks-0.182
 hyd-0.176
 heal-0.173
 rid-0.172
Token was
Token ID373
Feature activation+6.35e-02

Pos logit contributions
ited+0.111
 conflict+0.069
Next+0.067
 enough+0.067
Exc+0.066
Neg logit contributions
balls-0.144
 heal-0.144
 hyd-0.142
fly-0.140
 rid-0.138
Token a
Token ID257
Feature activation+5.89e-02

Pos logit contributions
ited+0.083
 saw+0.065
 heard+0.062
Hey+0.054
 year+0.053
Neg logit contributions
balls-0.133
 hyd-0.132
anks-0.130
 heal-0.129
 rid-0.129
Token little
Token ID1310
Feature activation+4.26e-02

Pos logit contributions
ited+0.032
 saw+0.028
 heard+0.022
 little+0.018
 year+0.015
Neg logit contributions
anks-0.173
balls-0.166
 hyd-0.163
 rid-0.162
 heal-0.160
Token together
Token ID1978
Feature activation+1.87e-02

Pos logit contributions
anks+0.042
 drown+0.039
 bleed+0.039
 hyd+0.039
 rid+0.039
Neg logit contributions
ited-0.025
 saw-0.024
 heard-0.023
 two-0.022
 little-0.022
Token.
Token ID13
Feature activation-1.14e-02

Pos logit contributions
ited+0.028
 saw+0.026
 heard+0.025
 year+0.022
 two+0.022
Neg logit contributions
anks-0.036
balls-0.035
 hyd-0.034
 rid-0.034
 heal-0.033
Token<|endoftext|>
Token ID50256
Feature activation+8.27e-02

Pos logit contributions
anks+0.025
balls+0.025
 hyd+0.024
 rid+0.024
 heal+0.024
Neg logit contributions
ited-0.014
 saw-0.013
 heard-0.012
 little-0.010
 two-0.010
TokenOnce
Token ID7454
Feature activation+8.93e-02

Pos logit contributions
ited+0.129
 saw+0.101
 heard+0.096
Hey+0.089
 noticed+0.088
Neg logit contributions
anks-0.159
balls-0.157
 hyd-0.152
 rid-0.150
 heal-0.150
Token upon
Token ID2402
Feature activation+7.28e-02

Pos logit contributions
ited+0.101
Hey+0.073
 heard+0.069
 saw+0.067
Fine+0.060
Neg logit contributions
balls-0.203
anks-0.201
ectar-0.192
 rid-0.192
 heal-0.190
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.102
 saw+0.073
Hey+0.069
 conflict+0.066
 heard+0.066
Neg logit contributions
 heal-0.148
balls-0.147
 hyd-0.144
 bleed-0.143
anks-0.142
Token time
Token ID640
Feature activation+7.01e-02

Pos logit contributions
ited+0.077
 saw+0.059
Hey+0.055
 heard+0.054
 two+0.049
Neg logit contributions
balls-0.244
 hyd-0.243
sing-0.243
anks-0.238
 heal-0.235
Token,
Token ID11
Feature activation+8.12e-02

Pos logit contributions
ited+0.071
 saw+0.050
 heard+0.048
Hey+0.042
 noticed+0.037
Neg logit contributions
balls-0.172
anks-0.170
 hyd-0.164
 heal-0.161
sing-0.159
Token in
Token ID287
Feature activation+6.92e-02

Pos logit contributions
ited+0.132
oph+0.103
aman+0.094
 amusement+0.091
Exc+0.085
Neg logit contributions
balls-0.151
fly-0.144
 heal-0.143
sing-0.142
 hyd-0.141
Token a
Token ID257
Feature activation+6.85e-02

Pos logit contributions
ited+0.104
 saw+0.078
 heard+0.076
Hey+0.073
 noticed+0.071
Neg logit contributions
 hyd-0.132
 heal-0.128
balls-0.128
 rid-0.125
sing-0.123
Token big
Token ID1263
Feature activation+5.29e-02

Pos logit contributions
ited+0.098
Hey+0.074
 heard+0.071
 enough+0.069
 sighed+0.068
Neg logit contributions
 hyd-0.132
 heal-0.126
 drown-0.120
balls-0.120
sing-0.120
Token shore
Token ID15191
Feature activation-1.02e-02

Pos logit contributions
anks+0.045
balls+0.044
 hyd+0.043
 rid+0.042
 heal+0.042
Neg logit contributions
ited-0.020
 saw-0.017
 heard-0.016
 two-0.014
 little-0.013
Token!
Token ID0
Feature activation-2.16e-02

Pos logit contributions
anks+0.019
balls+0.019
 hyd+0.018
 rid+0.018
 heal+0.018
Neg logit contributions
ited-0.015
 saw-0.014
 heard-0.014
 two-0.013
 little-0.012
Token<|endoftext|>
Token ID50256
Feature activation+6.92e-02

Pos logit contributions
anks+0.047
balls+0.046
 hyd+0.045
 rid+0.044
 heal+0.044
Neg logit contributions
ited-0.027
 saw-0.024
 heard-0.023
 year-0.019
Hey-0.019
TokenOnce
Token ID7454
Feature activation+8.73e-02

Pos logit contributions
ited+0.108
 saw+0.078
 heard+0.074
 conflict+0.073
 noticed+0.069
Neg logit contributions
anks-0.133
balls-0.132
 hyd-0.126
 rid-0.123
 heal-0.121
Token upon
Token ID2402
Feature activation+7.19e-02

Pos logit contributions
ited+0.107
Hey+0.079
 heard+0.073
 saw+0.069
Fine+0.069
Neg logit contributions
balls-0.185
anks-0.184
ectar-0.178
 rid-0.178
ometer-0.177
Token a
Token ID257
Feature activation+1.02e-01

Pos logit contributions
ited+0.111
 saw+0.079
Hey+0.076
 conflict+0.075
ised+0.073
Neg logit contributions
 heal-0.139
balls-0.136
 bleed-0.135
 rid-0.132
 hyd-0.131
Token time
Token ID640
Feature activation+6.98e-02

Pos logit contributions
ited+0.078
 saw+0.061
Hey+0.056
 heard+0.056
 two+0.049
Neg logit contributions
sing-0.240
 hyd-0.240
balls-0.240
 heal-0.232
anks-0.230
Token there
Token ID612
Feature activation+7.77e-02

Pos logit contributions
ited+0.082
 saw+0.056
 heard+0.054
Hey+0.048
 conflict+0.044
Neg logit contributions
balls-0.164
anks-0.157
 hyd-0.152
aro-0.151
ectar-0.151
Token lived
Token ID5615
Feature activation+7.33e-02

Pos logit contributions
ited+0.129
 conflict+0.086
 enough+0.086
Next+0.085
 weeks+0.082
Neg logit contributions
 heal-0.132
 hyd-0.131
balls-0.129
fly-0.128
cr-0.127
Token a
Token ID257
Feature activation+6.29e-02

Pos logit contributions
ited+0.087
 saw+0.066
 heard+0.062
 year+0.054
Hey+0.052
Neg logit contributions
balls-0.167
anks-0.166
 hyd-0.164
 rid-0.162
 heal-0.159
Token giant
Token ID6175
Feature activation+4.23e-02

Pos logit contributions
ited+0.042
 saw+0.037
 heard+0.031
 little+0.025
 year+0.023
Neg logit contributions
anks-0.176
balls-0.169
 hyd-0.167
 rid-0.165
 heal-0.163
Token ball
Token ID2613
Feature activation-3.30e-02

Pos logit contributions
anks+0.026
balls+0.025
 hyd+0.024
 rid+0.024
 heal+0.023
Neg logit contributions
ited-0.026
 saw-0.025
 heard-0.024
 two-0.021
 little-0.021
Token.
Token ID13
Feature activation-2.34e-02

Pos logit contributions
anks+0.066
 hyd+0.064
balls+0.063
 rid+0.062
 drown+0.060
Neg logit contributions
ited-0.046
 saw-0.044
 heard-0.042
 two-0.037
 little-0.037
Token<|endoftext|>
Token ID50256
Feature activation+8.31e-02

Pos logit contributions
anks+0.050
balls+0.049
 hyd+0.049
 rid+0.048
 heal+0.048
Neg logit contributions
ited-0.029
 saw-0.027
 heard-0.025
 little-0.023
 two-0.023
TokenOnce
Token ID7454
Feature activation+8.69e-02

Pos logit contributions
ited+0.143
 saw+0.108
 heard+0.102
 conflict+0.099
 noticed+0.097
Neg logit contributions
anks-0.147
balls-0.144
 heal-0.139
 hyd-0.139
 rid-0.137
Token upon
Token ID2402
Feature activation+7.12e-02

Pos logit contributions
ited+0.101
Hey+0.073
 heard+0.069
 saw+0.066
Fine+0.061
Neg logit contributions
balls-0.192
anks-0.191
ectar-0.185
 rid-0.183
ometer-0.181
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.095
 saw+0.069
Hey+0.063
 heard+0.062
 conflict+0.061
Neg logit contributions
 heal-0.148
balls-0.148
 bleed-0.145
anks-0.145
 hyd-0.144
Token time
Token ID640
Feature activation+7.19e-02

Pos logit contributions
ited+0.075
 saw+0.058
 heard+0.053
Hey+0.052
 two+0.048
Neg logit contributions
 hyd-0.243
balls-0.243
sing-0.243
anks-0.239
 heal-0.235
Token,
Token ID11
Feature activation+8.46e-02

Pos logit contributions
ited+0.073
 saw+0.051
 heard+0.049
Hey+0.041
 noticed+0.038
Neg logit contributions
balls-0.176
anks-0.174
 hyd-0.167
 heal-0.165
ectar-0.164
Token there
Token ID612
Feature activation+8.15e-02

Pos logit contributions
ited+0.126
oph+0.095
aman+0.087
 amusement+0.083
 realised+0.077
Neg logit contributions
balls-0.164
fly-0.162
 heal-0.160
 hyd-0.159
ectar-0.159
Token was
Token ID373
Feature activation+6.80e-02

Pos logit contributions
ited+0.127
 enough+0.079
 conflict+0.078
Exc+0.078
 weeks+0.077
Neg logit contributions
 heal-0.151
balls-0.150
 hyd-0.147
fly-0.144
anks-0.144
Token a
Token ID257
Feature activation+5.94e-02

Pos logit contributions
ited+0.106
 saw+0.079
 heard+0.077
 sighed+0.072
Hey+0.070
Neg logit contributions
balls-0.126
anks-0.125
 hyd-0.125
ectar-0.123
uggets-0.123
Token soup
Token ID17141
Feature activation-5.43e-02

Pos logit contributions
anks+0.012
balls+0.012
 hyd+0.011
 heal+0.011
ometers+0.011
Neg logit contributions
ited-0.014
 saw-0.013
 heard-0.013
 little-0.012
 two-0.012
Token.
Token ID13
Feature activation-1.93e-02

Pos logit contributions
anks+0.106
balls+0.101
 hyd+0.101
 rid+0.098
uggets+0.097
Neg logit contributions
ited-0.076
 saw-0.073
 heard-0.071
 did-0.063
 little-0.062
Token<|endoftext|>
Token ID50256
Feature activation+8.85e-02

Pos logit contributions
anks+0.046
balls+0.045
 rid+0.044
 hyd+0.044
 heal+0.044
Neg logit contributions
 saw-0.019
ited-0.018
 heard-0.017
 little-0.015
 two-0.015
TokenOnce
Token ID7454
Feature activation+9.21e-02

Pos logit contributions
ited+0.133
 saw+0.105
 heard+0.099
Hey+0.091
 noticed+0.090
Neg logit contributions
anks-0.176
balls-0.173
 hyd-0.168
 rid-0.166
 heal-0.165
Token upon
Token ID2402
Feature activation+7.64e-02

Pos logit contributions
ited+0.099
 heard+0.068
 saw+0.067
Hey+0.067
 year+0.055
Neg logit contributions
balls-0.216
anks-0.215
 rid-0.204
 hyd-0.202
 heal-0.202
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.094
 saw+0.069
 heard+0.064
Hey+0.060
 year+0.058
Neg logit contributions
balls-0.168
anks-0.166
 heal-0.166
 hyd-0.165
 bleed-0.162
Token time
Token ID640
Feature activation+7.40e-02

Pos logit contributions
ited+0.073
 saw+0.053
 heard+0.049
Hey+0.047
 two+0.041
Neg logit contributions
balls-0.254
 hyd-0.254
sing-0.249
anks-0.248
 heal-0.245
Token,
Token ID11
Feature activation+8.71e-02

Pos logit contributions
ited+0.063
 saw+0.046
 heard+0.042
Hey+0.034
 year+0.031
Neg logit contributions
anks-0.192
balls-0.192
 hyd-0.186
 heal-0.182
 rid-0.181
Token there
Token ID612
Feature activation+8.62e-02

Pos logit contributions
ited+0.142
oph+0.106
aman+0.097
 heard+0.093
 amusement+0.093
Neg logit contributions
balls-0.171
 hyd-0.160
sing-0.158
 heal-0.155
fly-0.154
Token was
Token ID373
Feature activation+7.76e-02

Pos logit contributions
ited+0.124
Exc+0.074
 enough+0.072
Hey+0.072
 heard+0.071
Neg logit contributions
balls-0.177
 hyd-0.171
 heal-0.170
anks-0.169
fly-0.164
Token a
Token ID257
Feature activation+6.38e-02

Pos logit contributions
ited+0.102
 saw+0.077
 heard+0.075
Hey+0.066
 year+0.064
Neg logit contributions
anks-0.166
balls-0.166
 hyd-0.163
 heal-0.158
 rid-0.158
Token end
Token ID886
Feature activation+4.54e-02

Pos logit contributions
ited+0.096
 saw+0.085
 heard+0.083
Hey+0.074
 noticed+0.073
Neg logit contributions
balls-0.094
 hyd-0.093
anks-0.091
 heal-0.089
 rid-0.088
Token.
Token ID13
Feature activation-2.14e-02

Pos logit contributions
 saw+0.048
 heard+0.046
 time+0.041
 year+0.040
 little+0.040
Neg logit contributions
anks-0.111
uggets-0.104
balls-0.104
 rid-0.102
 bleed-0.101
Token<|endoftext|>
Token ID50256
Feature activation+7.31e-02

Pos logit contributions
anks+0.047
balls+0.046
 hyd+0.045
 rid+0.044
 heal+0.044
Neg logit contributions
ited-0.026
 saw-0.023
 heard-0.022
Hey-0.019
 year-0.019
TokenOnce
Token ID7454
Feature activation+8.14e-02

Pos logit contributions
ited+0.122
 saw+0.093
 heard+0.089
 conflict+0.085
 noticed+0.083
Neg logit contributions
anks-0.134
balls-0.129
 hyd-0.128
 heal-0.124
 rid-0.124
Token upon
Token ID2402
Feature activation+7.03e-02

Pos logit contributions
ited+0.104
Hey+0.077
 heard+0.070
Fine+0.067
 saw+0.066
Neg logit contributions
anks-0.170
balls-0.169
ometer-0.165
ophone-0.164
ectar-0.163
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.104
 saw+0.074
Hey+0.070
 conflict+0.068
 heard+0.067
Neg logit contributions
 heal-0.141
balls-0.137
 bleed-0.136
 hyd-0.135
anks-0.133
Token time
Token ID640
Feature activation+6.96e-02

Pos logit contributions
ited+0.081
 saw+0.064
Hey+0.060
 heard+0.058
 two+0.054
Neg logit contributions
 hyd-0.237
sing-0.236
balls-0.232
 heal-0.229
anks-0.224
Token,
Token ID11
Feature activation+7.28e-02

Pos logit contributions
ited+0.071
 saw+0.051
 heard+0.048
Hey+0.041
 noticed+0.037
Neg logit contributions
balls-0.171
anks-0.169
 hyd-0.163
 heal-0.160
sing-0.158
Token there
Token ID612
Feature activation+7.60e-02

Pos logit contributions
ited+0.106
oph+0.080
aman+0.065
 weeks+0.064
 heard+0.064
Neg logit contributions
 hyd-0.143
 heal-0.142
balls-0.142
sing-0.140
fly-0.139
Token was
Token ID373
Feature activation+6.27e-02

Pos logit contributions
ited+0.126
 conflict+0.077
Exc+0.077
vious+0.076
 weeks+0.076
Neg logit contributions
 heal-0.138
 hyd-0.137
balls-0.136
fly-0.130
anks-0.129
Token a
Token ID257
Feature activation+5.96e-02

Pos logit contributions
ited+0.098
 saw+0.074
 heard+0.072
 sighed+0.067
Hey+0.065
Neg logit contributions
 hyd-0.116
balls-0.114
 heal-0.112
uggets-0.111
ectar-0.111
Tokenâ
Token ID22940
Feature activation+1.08e-02

Pos logit contributions
anks+0.070
balls+0.068
 hyd+0.067
 rid+0.066
 heal+0.065
Neg logit contributions
ited-0.057
 saw-0.053
 heard-0.051
 two-0.048
Hey-0.047
Token
Token ID26391
Feature activation-8.00e-03

Pos logit contributions
ited+0.021
 saw+0.021
 heard+0.020
 little+0.019
 two+0.019
Neg logit contributions
anks-0.015
 hyd-0.014
balls-0.013
 heal-0.013
 rid-0.013
Token<|endoftext|>
Token ID50256
Feature activation+6.56e-02

Pos logit contributions
anks+0.016
balls+0.016
 hyd+0.015
 rid+0.015
 heal+0.015
Neg logit contributions
ited-0.011
 saw-0.010
 heard-0.010
Hey-0.009
 year-0.009
TokenOnce
Token ID7454
Feature activation+8.29e-02

Pos logit contributions
ited+0.098
 saw+0.071
 heard+0.068
 conflict+0.067
 enough+0.062
Neg logit contributions
anks-0.128
balls-0.128
 hyd-0.122
 rid-0.119
ectar-0.117
Token upon
Token ID2402
Feature activation+7.32e-02

Pos logit contributions
ited+0.105
Hey+0.078
 heard+0.072
Fine+0.070
 saw+0.068
Neg logit contributions
balls-0.174
anks-0.171
ectar-0.166
ometer-0.165
 rid-0.165
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.116
 saw+0.083
 conflict+0.080
Hey+0.077
ised+0.076
Neg logit contributions
 heal-0.138
balls-0.136
 bleed-0.134
 rid-0.132
 hyd-0.131
Token time
Token ID640
Feature activation+6.62e-02

Pos logit contributions
ited+0.074
 saw+0.058
 heard+0.053
Hey+0.048
 two+0.044
Neg logit contributions
 hyd-0.248
balls-0.244
sing-0.243
 heal-0.236
anks-0.235
Token there
Token ID612
Feature activation+7.45e-02

Pos logit contributions
ited+0.062
 saw+0.044
 heard+0.041
Hey+0.034
 year+0.031
Neg logit contributions
balls-0.167
anks-0.166
 hyd-0.161
 heal-0.157
 rid-0.157
Token were
Token ID547
Feature activation+6.62e-02

Pos logit contributions
ited+0.111
 enough+0.076
 conflict+0.073
 weeks+0.071
Next+0.069
Neg logit contributions
 hyd-0.139
 heal-0.138
balls-0.135
fly-0.134
cr-0.133
Token two
Token ID734
Feature activation+7.38e-02

Pos logit contributions
ited+0.066
 saw+0.040
 heard+0.038
 sighed+0.030
Hey+0.029
Neg logit contributions
balls-0.164
 hyd-0.163
ectar-0.160
anks-0.159
uggets-0.159
Token best
Token ID1266
Feature activation+6.04e-02

Pos logit contributions
ited+0.123
 saw+0.107
 heard+0.104
Hey+0.099
 did+0.093
Neg logit contributions
anks-0.128
balls-0.128
 hyd-0.123
 rid-0.118
 heal-0.117
Token safe
Token ID3338
Feature activation-2.10e-02

Pos logit contributions
anks+0.060
balls+0.058
 hyd+0.056
 rid+0.055
 heal+0.054
Neg logit contributions
ited-0.044
 saw-0.041
 heard-0.039
 little-0.035
 two-0.035
Token.
Token ID13
Feature activation-1.75e-02

Pos logit contributions
anks+0.043
balls+0.041
 rid+0.040
 bleed+0.040
 hyd+0.040
Neg logit contributions
ited-0.027
 saw-0.027
 heard-0.026
 two-0.023
 enough-0.022
Token<|endoftext|>
Token ID50256
Feature activation+8.55e-02

Pos logit contributions
anks+0.034
 rid+0.034
balls+0.033
 hyd+0.033
 bleed+0.033
Neg logit contributions
 saw-0.024
ited-0.023
 heard-0.022
 little-0.021
 two-0.020
TokenOnce
Token ID7454
Feature activation+9.12e-02

Pos logit contributions
ited+0.141
 saw+0.108
 heard+0.102
 conflict+0.096
Hey+0.096
Neg logit contributions
anks-0.157
balls-0.155
 hyd-0.149
 heal-0.149
 rid-0.147
Token upon
Token ID2402
Feature activation+7.15e-02

Pos logit contributions
ited+0.104
Hey+0.073
 heard+0.070
 saw+0.067
Fine+0.061
Neg logit contributions
balls-0.206
anks-0.203
ometer-0.194
ectar-0.193
 rid-0.192
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.101
 saw+0.071
Hey+0.067
 conflict+0.066
 heard+0.064
Neg logit contributions
 heal-0.145
balls-0.145
 bleed-0.141
 hyd-0.141
anks-0.140
Token time
Token ID640
Feature activation+7.23e-02

Pos logit contributions
ited+0.078
 saw+0.058
 heard+0.054
Hey+0.053
 two+0.049
Neg logit contributions
 hyd-0.243
balls-0.242
sing-0.240
anks-0.234
 heal-0.233
Token,
Token ID11
Feature activation+8.28e-02

Pos logit contributions
ited+0.074
 saw+0.052
 heard+0.050
Hey+0.043
 noticed+0.039
Neg logit contributions
balls-0.177
anks-0.175
 hyd-0.169
 heal-0.166
ectar-0.164
Token there
Token ID612
Feature activation+8.08e-02

Pos logit contributions
ited+0.137
oph+0.107
aman+0.097
 amusement+0.093
 realised+0.092
Neg logit contributions
balls-0.152
 hyd-0.144
 heal-0.144
sing-0.144
ectar-0.142
Token was
Token ID373
Feature activation+7.05e-02

Pos logit contributions
ited+0.124
 enough+0.077
Exc+0.077
 conflict+0.076
 weeks+0.075
Neg logit contributions
balls-0.153
 heal-0.151
 hyd-0.150
anks-0.145
fly-0.142
Token a
Token ID257
Feature activation+6.21e-02

Pos logit contributions
ited+0.106
 saw+0.079
 heard+0.077
 sighed+0.071
Hey+0.070
Neg logit contributions
balls-0.136
 hyd-0.134
anks-0.134
ectar-0.132
 heal-0.130
Token her
Token ID607
Feature activation+2.83e-02

Pos logit contributions
ited+0.039
 saw+0.033
 heard+0.031
Hey+0.029
 noticed+0.027
Neg logit contributions
 rid-0.054
anks-0.054
 heal-0.054
balls-0.054
 hyd-0.053
Token.
Token ID13
Feature activation-1.39e-02

Pos logit contributions
ited+0.041
 saw+0.039
 heard+0.037
 little+0.035
 two+0.034
Neg logit contributions
anks-0.055
 hyd-0.053
balls-0.053
 rid-0.051
 bleed-0.051
Token<|endoftext|>
Token ID50256
Feature activation+8.04e-02

Pos logit contributions
anks+0.034
 hyd+0.034
 rid+0.034
balls+0.034
 bleed+0.033
Neg logit contributions
 saw-0.012
ited-0.010
 heard-0.010
 little-0.010
 two-0.010
TokenOnce
Token ID7454
Feature activation+8.55e-02

Pos logit contributions
ited+0.128
 saw+0.099
 heard+0.094
 conflict+0.087
 noticed+0.086
Neg logit contributions
anks-0.154
balls-0.151
 hyd-0.146
 rid-0.144
 heal-0.144
Token upon
Token ID2402
Feature activation+6.99e-02

Pos logit contributions
ited+0.096
Hey+0.067
 heard+0.064
 saw+0.063
Fine+0.055
Neg logit contributions
balls-0.195
anks-0.194
 rid-0.186
ectar-0.185
 heal-0.184
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.102
 saw+0.072
Hey+0.068
 conflict+0.068
ised+0.065
Neg logit contributions
 heal-0.140
balls-0.138
 bleed-0.136
 hyd-0.135
 rid-0.133
Token time
Token ID640
Feature activation+7.18e-02

Pos logit contributions
ited+0.075
 saw+0.057
 heard+0.052
Hey+0.051
 two+0.046
Neg logit contributions
 hyd-0.247
balls-0.245
sing-0.243
anks-0.238
 heal-0.238
Token,
Token ID11
Feature activation+7.99e-02

Pos logit contributions
ited+0.069
 saw+0.049
 heard+0.046
Hey+0.039
 noticed+0.035
Neg logit contributions
balls-0.180
anks-0.179
 hyd-0.173
 heal-0.169
 rid-0.168
Token there
Token ID612
Feature activation+8.09e-02

Pos logit contributions
ited+0.128
oph+0.100
aman+0.087
 amusement+0.085
 weeks+0.081
Neg logit contributions
balls-0.152
 hyd-0.145
fly-0.145
sing-0.145
 heal-0.144
Token was
Token ID373
Feature activation+6.73e-02

Pos logit contributions
ited+0.126
Exc+0.078
 conflict+0.078
 enough+0.077
 weeks+0.076
Neg logit contributions
balls-0.153
 heal-0.151
 hyd-0.150
fly-0.145
anks-0.144
Token a
Token ID257
Feature activation+5.95e-02

Pos logit contributions
ited+0.105
 saw+0.079
 heard+0.077
 sighed+0.072
Hey+0.070
Neg logit contributions
balls-0.124
 hyd-0.123
anks-0.121
ectar-0.120
uggets-0.120
Token make
Token ID787
Feature activation+6.82e-04

Pos logit contributions
balls+0.040
ometers+0.039
anks+0.039
ometer+0.039
ophone+0.037
Neg logit contributions
 saw-0.032
 heard-0.030
ited-0.029
 little-0.028
 two-0.027
Token honey
Token ID12498
Feature activation-4.56e-02

Pos logit contributions
 two+0.001
 saw+0.001
 little+0.001
 heard+0.001
 enough+0.000
Neg logit contributions
 hyd-0.002
anks-0.002
ometer-0.002
ophone-0.002
ometers-0.002
Token again
Token ID757
Feature activation-1.30e-02

Pos logit contributions
aro+0.095
ometers+0.094
ometer+0.091
ophone+0.090
anks+0.089
Neg logit contributions
 heard-0.053
 saw-0.051
 two-0.049
 little-0.047
 enough-0.047
Token.
Token ID13
Feature activation-1.66e-02

Pos logit contributions
anks+0.029
balls+0.029
ometer+0.028
ophone+0.028
ometers+0.028
Neg logit contributions
 saw-0.015
 heard-0.015
ited-0.014
 little-0.013
 two-0.013
Token<|endoftext|>
Token ID50256
Feature activation+8.11e-02

Pos logit contributions
 bleed+0.046
uries+0.046
sing+0.045
 rid+0.045
 hyd+0.045
Neg logit contributions
 saw-0.010
 two-0.009
 little-0.008
 heard-0.006
,-0.006
TokenOne
Token ID3198
Feature activation+1.01e-01

Pos logit contributions
ited+0.111
 saw+0.091
 heard+0.086
Hey+0.077
 year+0.076
Neg logit contributions
anks-0.170
balls-0.167
 hyd-0.163
 rid-0.160
 heal-0.159
Token day
Token ID1110
Feature activation+6.62e-02

Pos logit contributions
ited+0.155
Hey+0.136
 saw+0.126
 heard+0.124
Exc+0.124
Neg logit contributions
 hyd-0.169
 heal-0.168
sing-0.164
balls-0.163
ectar-0.162
Token,
Token ID11
Feature activation+7.22e-02

Pos logit contributions
ited+0.042
 saw+0.033
 heard+0.029
 two+0.020
 little+0.020
Neg logit contributions
anks-0.186
balls-0.181
 hyd-0.179
 rid-0.176
 heal-0.175
Token a
Token ID257
Feature activation+7.68e-02

Pos logit contributions
ited+0.080
 saw+0.062
 heard+0.060
 year+0.051
Hey+0.050
Neg logit contributions
balls-0.168
anks-0.167
 hyd-0.164
 rid-0.162
 heal-0.160
Token little
Token ID1310
Feature activation+5.23e-02

Pos logit contributions
ited+0.047
 saw+0.042
 heard+0.035
 little+0.029
 two+0.024
Neg logit contributions
anks-0.219
balls-0.210
 hyd-0.207
 rid-0.206
 heal-0.202
Token boy
Token ID2933
Feature activation+5.17e-02

Pos logit contributions
ited+0.051
 saw+0.045
 heard+0.041
 little+0.035
 year+0.034
Neg logit contributions
anks-0.129
balls-0.125
 hyd-0.123
 rid-0.121
 heal-0.120
Token slides
Token ID19392
Feature activation-1.26e-02

Pos logit contributions
balls+0.071
anks+0.069
 hyd+0.068
 rid+0.067
 heal+0.067
Neg logit contributions
ited-0.057
 saw-0.044
 heard-0.044
 year-0.043
Hey-0.040
Token.
Token ID13
Feature activation-2.33e-02

Pos logit contributions
anks+0.024
balls+0.023
 hyd+0.023
 rid+0.022
 heal+0.022
Neg logit contributions
ited-0.020
 saw-0.018
 heard-0.017
Hey-0.016
 year-0.015
Token<|endoftext|>
Token ID50256
Feature activation+7.88e-02

Pos logit contributions
anks+0.051
balls+0.050
 hyd+0.049
 rid+0.048
 heal+0.047
Neg logit contributions
ited-0.029
 saw-0.026
 heard-0.025
 little-0.022
 two-0.021
TokenOnce
Token ID7454
Feature activation+9.04e-02

Pos logit contributions
ited+0.127
 saw+0.098
 heard+0.093
 conflict+0.086
 noticed+0.086
Neg logit contributions
anks-0.150
balls-0.149
 hyd-0.142
 rid-0.140
 heal-0.139
Token upon
Token ID2402
Feature activation+7.31e-02

Pos logit contributions
ited+0.102
Hey+0.073
 heard+0.070
 saw+0.069
Fine+0.058
Neg logit contributions
balls-0.206
anks-0.205
 rid-0.195
 heal-0.195
ectar-0.194
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.099
 saw+0.070
Hey+0.066
 heard+0.064
 conflict+0.063
Neg logit contributions
 heal-0.151
balls-0.150
 hyd-0.148
 bleed-0.147
anks-0.147
Token time
Token ID640
Feature activation+7.01e-02

Pos logit contributions
ited+0.076
 saw+0.057
Hey+0.056
 heard+0.053
 two+0.048
Neg logit contributions
 hyd-0.242
balls-0.242
sing-0.241
anks-0.235
 heal-0.234
Token,
Token ID11
Feature activation+8.36e-02

Pos logit contributions
ited+0.074
 saw+0.051
 heard+0.049
Hey+0.044
 noticed+0.038
Neg logit contributions
balls-0.171
anks-0.167
 hyd-0.161
 heal-0.158
sing-0.157
Token there
Token ID612
Feature activation+8.38e-02

Pos logit contributions
ited+0.136
oph+0.099
 amusement+0.091
aman+0.090
Fine+0.086
Neg logit contributions
balls-0.160
fly-0.151
ectar-0.148
sing-0.148
 hyd-0.146
Token was
Token ID373
Feature activation+7.26e-02

Pos logit contributions
ited+0.144
vious+0.096
Hey+0.092
Exc+0.092
 conflict+0.090
Neg logit contributions
balls-0.142
 heal-0.140
 hyd-0.135
 float-0.132
fly-0.130
Token a
Token ID257
Feature activation+6.03e-02

Pos logit contributions
ited+0.099
 saw+0.076
 heard+0.074
Hey+0.066
 sighed+0.064
Neg logit contributions
balls-0.148
anks-0.147
 hyd-0.147
ectar-0.143
 heal-0.142
Token day
Token ID1110
Feature activation+2.00e-04

Pos logit contributions
ited+0.018
 saw+0.016
 heard+0.016
 year+0.014
Hey+0.014
Neg logit contributions
anks-0.026
balls-0.025
 hyd-0.025
 rid-0.024
 heal-0.024
Token!"
Token ID2474
Feature activation-9.05e-03

Pos logit contributions
ited+0.000
 saw+0.000
 heard+0.000
 year+0.000
 two+0.000
Neg logit contributions
anks-0.000
balls-0.000
 hyd-0.000
 rid-0.000
 bleed-0.000
Token<|endoftext|>
Token ID50256
Feature activation+6.75e-02

Pos logit contributions
anks+0.019
balls+0.019
 hyd+0.018
 rid+0.018
 heal+0.018
Neg logit contributions
ited-0.012
 saw-0.011
 heard-0.011
 two-0.009
 little-0.009
TokenOnce
Token ID7454
Feature activation+8.28e-02

Pos logit contributions
ited+0.095
 saw+0.072
 heard+0.067
 conflict+0.062
 noticed+0.060
Neg logit contributions
anks-0.140
balls-0.139
 hyd-0.135
 rid-0.131
 heal-0.129
Token upon
Token ID2402
Feature activation+6.97e-02

Pos logit contributions
ited+0.090
Hey+0.064
 heard+0.062
 saw+0.062
Fine+0.051
Neg logit contributions
anks-0.189
balls-0.188
 rid-0.181
ectar-0.180
 heal-0.180
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.102
 saw+0.073
Hey+0.068
 conflict+0.068
 heard+0.065
Neg logit contributions
 heal-0.139
balls-0.136
 bleed-0.135
 hyd-0.135
 rid-0.132
Token time
Token ID640
Feature activation+6.91e-02

Pos logit contributions
ited+0.076
 saw+0.059
Hey+0.054
 heard+0.053
 two+0.047
Neg logit contributions
 hyd-0.243
sing-0.241
balls-0.240
 heal-0.233
anks-0.231
Token there
Token ID612
Feature activation+7.48e-02

Pos logit contributions
ited+0.074
 saw+0.051
 heard+0.049
Hey+0.043
 noticed+0.038
Neg logit contributions
balls-0.167
anks-0.163
 hyd-0.157
 heal-0.154
sing-0.154
Token was
Token ID373
Feature activation+6.92e-02

Pos logit contributions
ited+0.119
 enough+0.079
 conflict+0.079
Next+0.077
 weeks+0.077
Neg logit contributions
 heal-0.130
 hyd-0.130
balls-0.129
fly-0.128
cr-0.125
Token a
Token ID257
Feature activation+6.05e-02

Pos logit contributions
ited+0.085
 saw+0.067
 heard+0.063
Hey+0.056
 year+0.053
Neg logit contributions
anks-0.150
balls-0.150
 hyd-0.148
 rid-0.145
 heal-0.144
Token girl
Token ID2576
Feature activation+4.06e-02

Pos logit contributions
ited+0.032
 saw+0.028
 heard+0.022
 little+0.018
 year+0.015
Neg logit contributions
anks-0.178
balls-0.171
 rid-0.168
 hyd-0.168
 heal-0.165
Token garden
Token ID11376
Feature activation-1.35e-02

Pos logit contributions
ited+0.000
 saw+0.000
 heard+0.000
Hey+0.000
 two+0.000
Neg logit contributions
anks-0.000
balls-0.000
 hyd-0.000
 rid-0.000
 heal-0.000
Token.
Token ID13
Feature activation-1.11e-02

Pos logit contributions
 bleed+0.031
 drown+0.030
anks+0.030
ric+0.030
ometers+0.029
Neg logit contributions
 saw-0.014
 heard-0.013
 year-0.012
 little-0.012
 two-0.012
Token<|endoftext|>
Token ID50256
Feature activation+8.60e-02

Pos logit contributions
anks+0.027
balls+0.027
 hyd+0.027
 rid+0.027
 bleed+0.027
Neg logit contributions
 saw-0.009
 heard-0.008
ited-0.008
 two-0.007
 little-0.007
TokenOnce
Token ID7454
Feature activation+9.22e-02

Pos logit contributions
ited+0.132
 saw+0.104
 heard+0.098
Hey+0.091
 noticed+0.089
Neg logit contributions
anks-0.167
balls-0.165
 hyd-0.160
 heal-0.158
 rid-0.158
Token upon
Token ID2402
Feature activation+7.06e-02

Pos logit contributions
ited+0.105
Hey+0.073
 heard+0.071
 saw+0.070
Fine+0.058
Neg logit contributions
balls-0.211
anks-0.209
 rid-0.198
ometer-0.197
 heal-0.197
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.092
 saw+0.066
 heard+0.060
Hey+0.060
 conflict+0.058
Neg logit contributions
balls-0.150
 heal-0.149
 hyd-0.148
anks-0.146
 bleed-0.145
Token time
Token ID640
Feature activation+7.33e-02

Pos logit contributions
ited+0.077
 saw+0.058
 heard+0.054
Hey+0.053
 two+0.047
Neg logit contributions
 hyd-0.243
balls-0.242
sing-0.241
anks-0.238
 heal-0.235
Token,
Token ID11
Feature activation+8.48e-02

Pos logit contributions
ited+0.070
 saw+0.050
 heard+0.048
Hey+0.040
 noticed+0.036
Neg logit contributions
balls-0.183
anks-0.182
 hyd-0.175
 heal-0.172
 rid-0.170
Token there
Token ID612
Feature activation+8.30e-02

Pos logit contributions
ited+0.134
oph+0.102
aman+0.091
 amusement+0.088
 realised+0.085
Neg logit contributions
balls-0.167
fly-0.157
 hyd-0.157
sing-0.155
 heal-0.155
Token was
Token ID373
Feature activation+7.21e-02

Pos logit contributions
ited+0.120
Exc+0.071
Hey+0.071
 heard+0.071
 conflict+0.070
Neg logit contributions
balls-0.168
 heal-0.163
 hyd-0.161
anks-0.159
fly-0.155
Token a
Token ID257
Feature activation+6.19e-02

Pos logit contributions
ited+0.098
 saw+0.075
 heard+0.072
Hey+0.064
 sighed+0.063
Neg logit contributions
balls-0.149
anks-0.148
 hyd-0.148
 heal-0.143
 rid-0.143
Token them
Token ID606
Feature activation-2.36e-02

Pos logit contributions
anks+0.023
balls+0.022
 hyd+0.021
ometer+0.021
 bleed+0.021
Neg logit contributions
ited-0.012
 saw-0.011
 heard-0.011
 two-0.010
 little-0.009
Token.
Token ID13
Feature activation-1.89e-02

Pos logit contributions
anks+0.049
 hyd+0.047
balls+0.046
 bleed+0.046
 rid+0.045
Neg logit contributions
 saw-0.030
ited-0.030
 heard-0.028
 two-0.026
 little-0.025
Token<|endoftext|>
Token ID50256
Feature activation+7.20e-02

Pos logit contributions
anks+0.040
balls+0.038
 hyd+0.038
 heal+0.038
 bleed+0.038
Neg logit contributions
ited-0.024
 saw-0.023
 heard-0.021
 two-0.019
 little-0.019
TokenOnce
Token ID7454
Feature activation+8.07e-02

Pos logit contributions
ited+0.119
 saw+0.089
 heard+0.086
 conflict+0.082
 noticed+0.079
Neg logit contributions
anks-0.134
balls-0.132
 hyd-0.127
 rid-0.124
 heal-0.123
Token upon
Token ID2402
Feature activation+6.85e-02

Pos logit contributions
ited+0.082
 saw+0.057
 heard+0.057
Hey+0.053
 year+0.046
Neg logit contributions
anks-0.193
balls-0.193
 rid-0.185
 hyd-0.184
 heal-0.181
Token a
Token ID257
Feature activation+1.01e-01

Pos logit contributions
ited+0.096
 saw+0.069
Hey+0.065
 conflict+0.063
 heard+0.063
Neg logit contributions
 heal-0.139
balls-0.139
 bleed-0.136
 hyd-0.136
 rid-0.134
Token time
Token ID640
Feature activation+7.15e-02

Pos logit contributions
ited+0.075
 saw+0.057
 heard+0.052
Hey+0.052
 two+0.045
Neg logit contributions
 hyd-0.243
balls-0.242
sing-0.242
 heal-0.235
anks-0.234
Token there
Token ID612
Feature activation+7.49e-02

Pos logit contributions
ited+0.075
 saw+0.051
 heard+0.050
Hey+0.044
 noticed+0.038
Neg logit contributions
balls-0.174
anks-0.170
 hyd-0.165
 heal-0.162
sing-0.161
Token was
Token ID373
Feature activation+6.82e-02

Pos logit contributions
ited+0.117
 enough+0.078
 conflict+0.078
Next+0.075
 weeks+0.074
Neg logit contributions
 hyd-0.134
 heal-0.133
balls-0.131
fly-0.131
cr-0.127
Token a
Token ID257
Feature activation+6.10e-02

Pos logit contributions
ited+0.085
 saw+0.067
 heard+0.064
Hey+0.056
 year+0.054
Neg logit contributions
balls-0.146
anks-0.146
 hyd-0.146
 rid-0.141
 heal-0.141
Token man
Token ID582
Feature activation+3.01e-02

Pos logit contributions
ited+0.031
 saw+0.026
 heard+0.021
 little+0.015
 year+0.013
Neg logit contributions
anks-0.179
balls-0.173
 hyd-0.171
 rid-0.170
 heal-0.168
Token and
Token ID290
Feature activation-6.45e-02

Pos logit contributions
anks+0.091
balls+0.088
 bleed+0.088
 drown+0.086
 rid+0.086
Neg logit contributions
 saw-0.034
ited-0.033
 heard-0.031
 little-0.028
 year-0.028
Token always
Token ID1464
Feature activation-3.74e-02

Pos logit contributions
anks+0.140
balls+0.137
 hyd+0.135
 rid+0.133
 heal+0.132
Neg logit contributions
ited-0.082
 saw-0.070
 heard-0.067
Hey-0.060
 year-0.059
Token had
Token ID550
Feature activation-1.71e-02

Pos logit contributions
 hyd+0.082
anks+0.081
 rid+0.079
 heal+0.079
balls+0.079
Neg logit contributions
ited-0.045
 saw-0.036
 heard-0.034
Hey-0.033
 year-0.032
Token plenty
Token ID6088
Feature activation-5.38e-03

Pos logit contributions
anks+0.036
balls+0.035
 hyd+0.035
 rid+0.035
 heal+0.034
Neg logit contributions
ited-0.023
 saw-0.019
 heard-0.018
Hey-0.017
 year-0.016
Token of
Token ID286
Feature activation-4.44e-02

Pos logit contributions
anks+0.014
balls+0.013
 hyd+0.013
 rid+0.013
 heal+0.013
Neg logit contributions
ited-0.005
 saw-0.004
 heard-0.004
 year-0.003
 two-0.003
Token rad
Token ID2511
Feature activation-1.11e-01

Pos logit contributions
anks+0.087
balls+0.085
 hyd+0.083
 rid+0.080
 heal+0.080
Neg logit contributions
ited-0.066
 saw-0.060
 heard-0.057
 year-0.051
 little-0.051
Tokenishes
Token ID5614
Feature activation-8.63e-02

Pos logit contributions
anks+0.136
sing+0.133
balls+0.130
 heal+0.129
 bleed+0.128
Neg logit contributions
ited-0.237
 saw-0.228
 heard-0.212
 two-0.204
 enough-0.200
Token to
Token ID284
Feature activation-5.73e-02

Pos logit contributions
anks+0.183
sing+0.177
 bleed+0.175
balls+0.175
 hyd+0.171
Neg logit contributions
ited-0.102
 saw-0.100
 heard-0.094
 two-0.084
 enough-0.082
Token eat
Token ID4483
Feature activation-1.85e-02

Pos logit contributions
anks+0.092
balls+0.089
ometer+0.084
 hyd+0.082
ophone+0.082
Neg logit contributions
ited-0.104
 saw-0.101
 heard-0.096
 little-0.089
 two-0.089
Token.
Token ID13
Feature activation-4.91e-02

Pos logit contributions
anks+0.038
balls+0.036
sing+0.035
 bleed+0.035
 hyd+0.035
Neg logit contributions
ited-0.025
 saw-0.024
 heard-0.022
 little-0.021
 year-0.021
Token The
Token ID383
Feature activation+2.61e-02

Pos logit contributions
anks+0.125
balls+0.125
 bleed+0.122
 rid+0.121
 hyd+0.120
Neg logit contributions
 saw-0.038
ited-0.037
 heard-0.032
 little-0.031
 two-0.031
Token world
Token ID995
Feature activation-2.87e-02

Pos logit contributions
anks+0.003
balls+0.003
 hyd+0.003
 rid+0.003
 bleed+0.003
Neg logit contributions
ited-0.001
 saw-0.001
 heard-0.001
 little-0.001
 year-0.001
Token where
Token ID810
Feature activation-2.03e-02

Pos logit contributions
anks+0.061
balls+0.060
 hyd+0.059
 heal+0.059
 rid+0.058
Neg logit contributions
ited-0.037
 saw-0.031
 heard-0.029
Hey-0.026
 two-0.026
Token everyone
Token ID2506
Feature activation-3.99e-02

Pos logit contributions
anks+0.040
balls+0.039
 hyd+0.038
 rid+0.037
 heal+0.037
Neg logit contributions
ited-0.029
 saw-0.027
 heard-0.026
 two-0.023
 little-0.023
Token shared
Token ID4888
Feature activation+1.04e-03

Pos logit contributions
anks+0.093
balls+0.091
 hyd+0.090
 rid+0.088
 heal+0.087
Neg logit contributions
ited-0.043
 saw-0.037
 heard-0.035
 year-0.029
Hey-0.029
Token their
Token ID511
Feature activation-4.48e-02

Pos logit contributions
ited+0.001
 saw+0.001
 heard+0.001
Hey+0.001
 year+0.001
Neg logit contributions
anks-0.002
balls-0.002
 hyd-0.002
 heal-0.002
 rid-0.002
Token rad
Token ID2511
Feature activation-1.10e-01

Pos logit contributions
anks+0.075
balls+0.073
 hyd+0.071
 rid+0.070
 heal+0.069
Neg logit contributions
ited-0.079
 saw-0.072
 heard-0.069
Hey-0.063
 year-0.062
Tokenishes
Token ID5614
Feature activation-8.31e-02

Pos logit contributions
anks+0.136
balls+0.133
 bleed+0.126
sing+0.125
 heal+0.124
Neg logit contributions
ited-0.233
 saw-0.224
 heard-0.208
 little-0.205
 two-0.204
Token.
Token ID13
Feature activation-5.11e-02

Pos logit contributions
anks+0.193
balls+0.189
 bleed+0.185
sing+0.184
 hyd+0.182
Neg logit contributions
ited-0.084
 saw-0.079
 heard-0.073
 two-0.068
 little-0.067
Token<|endoftext|>
Token ID50256
Feature activation+4.59e-02

Pos logit contributions
balls+0.132
anks+0.132
 bleed+0.129
 rid+0.129
 hyd+0.127
Neg logit contributions
 saw-0.038
ited-0.034
 two-0.033
 heard-0.032
 little-0.031
TokenOnce
Token ID7454
Feature activation+6.77e-02

Pos logit contributions
ited+0.070
 saw+0.055
 heard+0.052
 conflict+0.047
 noticed+0.047
Neg logit contributions
anks-0.091
balls-0.088
 hyd-0.086
 rid-0.084
 heal-0.084
Token upon
Token ID2402
Feature activation+4.39e-02

Pos logit contributions
ited+0.074
 heard+0.051
Hey+0.051
 saw+0.050
Fine+0.040
Neg logit contributions
balls-0.157
anks-0.156
 rid-0.149
ectar-0.149
ometer-0.148
Token other
Token ID584
Feature activation-1.87e-02

Pos logit contributions
balls+0.125
anks+0.124
 hyd+0.123
 rid+0.121
 heal+0.119
Neg logit contributions
ited-0.120
 saw-0.099
 heard-0.096
Hey-0.094
 year-0.092
Token things
Token ID1243
Feature activation-6.28e-02

Pos logit contributions
anks+0.035
balls+0.034
 hyd+0.033
 rid+0.033
 heal+0.032
Neg logit contributions
ited-0.030
 saw-0.026
 heard-0.024
Hey-0.023
 year-0.022
Token that
Token ID326
Feature activation-3.78e-02

Pos logit contributions
anks+0.133
balls+0.126
 hyd+0.124
uggets+0.120
sing+0.120
Neg logit contributions
ited-0.082
 saw-0.080
 heard-0.075
 did-0.066
 little-0.065
Token might
Token ID1244
Feature activation-5.62e-02

Pos logit contributions
anks+0.090
balls+0.087
 hyd+0.083
 bleed+0.082
 rid+0.081
Neg logit contributions
 saw-0.040
ited-0.038
 heard-0.036
 did-0.032
 two-0.030
Token make
Token ID787
Feature activation-4.20e-02

Pos logit contributions
anks+0.101
balls+0.097
 hyd+0.092
uggets+0.090
 rid+0.089
Neg logit contributions
ited-0.093
 saw-0.090
 heard-0.086
 little-0.077
 two-0.077
Token it
Token ID340
Feature activation-1.07e-01

Pos logit contributions
anks+0.089
balls+0.087
 hyd+0.085
sing+0.081
ometer+0.081
Neg logit contributions
ited-0.052
 saw-0.050
 heard-0.048
 little-0.044
 two-0.044
Token not
Token ID407
Feature activation-6.18e-02

Pos logit contributions
anks+0.191
balls+0.187
 hyd+0.179
uggets+0.178
aro+0.174
Neg logit contributions
 saw-0.167
ited-0.166
 heard-0.160
 little-0.149
 two-0.146
Token as
Token ID355
Feature activation-6.50e-02

Pos logit contributions
anks+0.121
balls+0.117
 hyd+0.113
 rid+0.108
ometer+0.107
Neg logit contributions
ited-0.091
 saw-0.087
 heard-0.083
 little-0.076
 two-0.074
Token good
Token ID922
Feature activation-5.02e-02

Pos logit contributions
anks+0.131
balls+0.127
 hyd+0.125
 rid+0.122
 heal+0.121
Neg logit contributions
ited-0.092
 saw-0.083
 heard-0.079
 year-0.070
Hey-0.070
Token.
Token ID13
Feature activation-5.08e-02

Pos logit contributions
anks+0.100
 hyd+0.099
balls+0.099
ophone+0.097
 rid+0.096
Neg logit contributions
ited-0.063
 saw-0.060
 heard-0.057
 little-0.054
 two-0.054
Token And
Token ID843
Feature activation-2.95e-02

Pos logit contributions
anks+0.116
balls+0.114
 hyd+0.112
 rid+0.111
 bleed+0.109
Neg logit contributions
ited-0.056
 saw-0.053
 heard-0.048
 two-0.044
 little-0.043
Token and
Token ID290
Feature activation-8.42e-02

Pos logit contributions
anks+0.120
balls+0.117
 hyd+0.114
 rid+0.113
 bleed+0.109
Neg logit contributions
ited-0.105
 saw-0.100
 heard-0.097
 little-0.089
 two-0.088
Token it
Token ID340
Feature activation-6.42e-02

Pos logit contributions
anks+0.148
balls+0.141
 hyd+0.139
 rid+0.133
 heal+0.130
Neg logit contributions
ited-0.137
 saw-0.134
 heard-0.127
 two-0.117
 little-0.116
Token will
Token ID481
Feature activation-5.99e-02

Pos logit contributions
anks+0.140
uggets+0.139
ectar+0.131
 hyd+0.130
balls+0.130
Neg logit contributions
 saw-0.079
ited-0.078
 heard-0.075
 did-0.071
 little-0.066
Token be
Token ID307
Feature activation-7.47e-02

Pos logit contributions
anks+0.112
balls+0.106
uggets+0.102
ometer+0.101
ometers+0.101
Neg logit contributions
 saw-0.092
ited-0.090
 heard-0.087
 little-0.082
 two-0.082
Token bigger
Token ID5749
Feature activation-4.29e-02

Pos logit contributions
anks+0.154
balls+0.150
 hyd+0.146
 rid+0.142
 heal+0.141
Neg logit contributions
ited-0.103
 saw-0.094
 heard-0.089
 little-0.079
 two-0.079
Token and
Token ID290
Feature activation-1.06e-01

Pos logit contributions
anks+0.083
balls+0.081
 hyd+0.079
 rid+0.078
 heal+0.077
Neg logit contributions
ited-0.063
 saw-0.058
 heard-0.055
 two-0.050
 year-0.049
Token prett
Token ID46442
Feature activation-3.74e-02

Pos logit contributions
anks+0.178
balls+0.170
 hyd+0.165
 rid+0.160
 heal+0.158
Neg logit contributions
ited-0.188
 saw-0.176
 heard-0.169
 two-0.154
 little-0.154
Tokenier
Token ID959
Feature activation-5.13e-02

Pos logit contributions
anks+0.061
balls+0.058
 hyd+0.057
 rid+0.056
 heal+0.055
Neg logit contributions
ited-0.068
 saw-0.064
 heard-0.061
 little-0.055
Hey-0.055
Token than
Token ID621
Feature activation-5.86e-02

Pos logit contributions
anks+0.075
balls+0.073
 hyd+0.070
 rid+0.069
 bleed+0.068
Neg logit contributions
ited-0.097
 saw-0.093
 heard-0.090
 two-0.084
 little-0.084
Token yours
Token ID12431
Feature activation-4.16e-02

Pos logit contributions
anks+0.101
balls+0.096
 hyd+0.096
 rid+0.090
 bleed+0.090
Neg logit contributions
ited-0.099
 saw-0.093
 heard-0.090
 two-0.083
 little-0.082
Token."
Token ID526
Feature activation-3.06e-02

Pos logit contributions
anks+0.072
 hyd+0.070
balls+0.069
 bleed+0.068
sing+0.068
Neg logit contributions
 saw-0.064
ited-0.062
 heard-0.062
 two-0.058
 did-0.056
Token John
Token ID1757
Feature activation-8.25e-03

Pos logit contributions
ited+0.026
 conflict+0.020
Hey+0.019
 year+0.019
Next+0.018
Neg logit contributions
flies-0.041
 hyd-0.040
fly-0.040
 heal-0.039
balls-0.039
Token chose
Token ID7690
Feature activation-7.40e-03

Pos logit contributions
balls+0.018
anks+0.018
 hyd+0.017
 rid+0.017
 heal+0.017
Neg logit contributions
ited-0.009
Hey-0.007
 conflict-0.007
 saw-0.006
 year-0.006
Token some
Token ID617
Feature activation-8.42e-03

Pos logit contributions
anks+0.015
balls+0.015
 rid+0.015
 hyd+0.015
 heal+0.014
Neg logit contributions
ited-0.010
 saw-0.008
 heard-0.008
Hey-0.007
 year-0.007
Token carrots
Token ID34397
Feature activation-6.15e-02

Pos logit contributions
anks+0.015
balls+0.015
 rid+0.015
 hyd+0.015
 heal+0.015
Neg logit contributions
ited-0.013
 saw-0.012
 heard-0.011
Hey-0.010
 noticed-0.010
Token and
Token ID290
Feature activation-5.62e-02

Pos logit contributions
anks+0.150
balls+0.143
 hyd+0.142
 rid+0.141
 bleed+0.140
Neg logit contributions
ited-0.059
 saw-0.054
 heard-0.050
 two-0.042
 little-0.041
Token cuc
Token ID38421
Feature activation-1.05e-01

Pos logit contributions
balls+0.111
anks+0.111
 hyd+0.108
 rid+0.107
 heal+0.106
Neg logit contributions
ited-0.079
 saw-0.068
 heard-0.066
 year-0.060
Hey-0.060
Tokenumbers
Token ID17024
Feature activation-3.14e-02

Pos logit contributions
anks+0.155
balls+0.149
 hyd+0.149
 heal+0.143
 rid+0.142
Neg logit contributions
ited-0.205
 saw-0.187
 heard-0.180
 two-0.168
Hey-0.167
Token to
Token ID284
Feature activation-3.57e-02

Pos logit contributions
anks+0.063
balls+0.062
 hyd+0.061
 rid+0.059
 heal+0.059
Neg logit contributions
ited-0.045
 saw-0.039
 heard-0.037
Hey-0.034
 year-0.033
Token take
Token ID1011
Feature activation-3.20e-03

Pos logit contributions
anks+0.058
balls+0.055
 hyd+0.054
 rid+0.051
 bleed+0.051
Neg logit contributions
ited-0.065
 saw-0.061
 heard-0.058
 two-0.054
 little-0.054
Token home
Token ID1363
Feature activation-1.02e-02

Pos logit contributions
anks+0.007
balls+0.007
 hyd+0.007
 rid+0.006
 heal+0.006
Neg logit contributions
ited-0.004
 saw-0.004
 heard-0.004
 year-0.003
Hey-0.003
Token.
Token ID13
Feature activation-4.99e-02

Pos logit contributions
anks+0.020
balls+0.019
 hyd+0.019
 bleed+0.019
sing+0.019
Neg logit contributions
ited-0.014
 saw-0.013
 heard-0.013
 two-0.012
 little-0.012
Token like
Token ID588
Feature activation-2.71e-02

Pos logit contributions
anks+0.151
balls+0.145
 hyd+0.143
 rid+0.139
 bleed+0.139
Neg logit contributions
 saw-0.080
ited-0.080
 heard-0.075
 two-0.065
 did-0.065
Token lettuce
Token ID39406
Feature activation-6.82e-02

Pos logit contributions
anks+0.056
balls+0.053
 hyd+0.053
 rid+0.051
 bleed+0.051
Neg logit contributions
ited-0.035
 saw-0.035
 heard-0.033
 two-0.030
 little-0.030
Token,
Token ID11
Feature activation-5.34e-02

Pos logit contributions
anks+0.177
balls+0.171
 hyd+0.171
 rid+0.169
 bleed+0.166
Neg logit contributions
ited-0.053
 saw-0.049
 heard-0.043
 two-0.035
 little-0.034
Token tomatoes
Token ID23972
Feature activation-5.60e-02

Pos logit contributions
anks+0.098
balls+0.093
 hyd+0.092
 rid+0.089
 heal+0.088
Neg logit contributions
ited-0.084
 saw-0.080
 heard-0.076
 little-0.070
 two-0.070
Token and
Token ID290
Feature activation-7.90e-02

Pos logit contributions
anks+0.151
balls+0.146
 hyd+0.146
 rid+0.144
 bleed+0.142
Neg logit contributions
ited-0.038
 saw-0.034
 heard-0.030
 two-0.023
 little-0.022
Token cuc
Token ID38421
Feature activation-1.05e-01

Pos logit contributions
anks+0.132
balls+0.125
 hyd+0.123
 rid+0.119
 heal+0.117
Neg logit contributions
ited-0.139
 saw-0.132
 heard-0.126
 two-0.116
 little-0.116
Tokenumbers
Token ID17024
Feature activation-4.81e-02

Pos logit contributions
anks+0.154
balls+0.149
 hyd+0.148
 heal+0.143
 bleed+0.141
Neg logit contributions
ited-0.205
 saw-0.187
 heard-0.181
 two-0.168
Hey-0.167
Token.
Token ID13
Feature activation-5.88e-02

Pos logit contributions
anks+0.093
balls+0.090
 hyd+0.088
 rid+0.086
 heal+0.085
Neg logit contributions
ited-0.071
 saw-0.065
 heard-0.062
 two-0.055
 year-0.055
Token He
Token ID679
Feature activation-2.29e-02

Pos logit contributions
anks+0.124
 hyd+0.119
balls+0.118
 rid+0.117
 bleed+0.117
Neg logit contributions
ited-0.073
 saw-0.071
 heard-0.066
 two-0.064
 little-0.061
Token smiled
Token ID13541
Feature activation+9.86e-03

Pos logit contributions
anks+0.049
balls+0.049
 hyd+0.048
 rid+0.047
 heal+0.047
Neg logit contributions
ited-0.028
 saw-0.023
 heard-0.022
Hey-0.020
 year-0.020
Token as
Token ID355
Feature activation-2.38e-03

Pos logit contributions
ited+0.009
 saw+0.008
 heard+0.007
Hey+0.006
 year+0.006
Neg logit contributions
anks-0.025
balls-0.024
 hyd-0.024
 rid-0.023
 heal-0.023
Token play
Token ID711
Feature activation+1.68e-02

Pos logit contributions
anks+0.046
balls+0.045
 hyd+0.043
 rid+0.042
 heal+0.042
Neg logit contributions
ited-0.042
 saw-0.038
 heard-0.037
 two-0.033
 year-0.033
Token together
Token ID1978
Feature activation-5.14e-03

Pos logit contributions
ited+0.030
 saw+0.028
 heard+0.027
 little+0.024
 two+0.024
Neg logit contributions
anks-0.028
balls-0.027
 hyd-0.026
 rid-0.025
 heal-0.025
Token without
Token ID1231
Feature activation-3.23e-02

Pos logit contributions
anks+0.011
balls+0.011
 hyd+0.011
 rid+0.011
 heal+0.011
Neg logit contributions
ited-0.006
 saw-0.005
 heard-0.005
Hey-0.004
 two-0.004
Token making
Token ID1642
Feature activation-2.03e-02

Pos logit contributions
anks+0.057
balls+0.056
 rid+0.056
 hyd+0.055
 heal+0.055
Neg logit contributions
ited-0.053
 saw-0.047
 heard-0.045
Hey-0.042
 year-0.042
Token boring
Token ID14262
Feature activation-3.27e-02

Pos logit contributions
anks+0.043
balls+0.042
 hyd+0.041
 rid+0.039
 heal+0.039
Neg logit contributions
ited-0.026
 saw-0.024
 heard-0.023
 little-0.021
 two-0.021
Token or
Token ID393
Feature activation-1.05e-01

Pos logit contributions
anks+0.066
balls+0.065
 hyd+0.064
 rid+0.063
 heal+0.061
Neg logit contributions
ited-0.043
 saw-0.041
 heard-0.039
 little-0.035
 two-0.034
Token loud
Token ID7812
Feature activation-4.31e-02

Pos logit contributions
anks+0.193
balls+0.186
 hyd+0.183
 rid+0.177
 heal+0.176
Neg logit contributions
ited-0.166
 saw-0.153
 heard-0.147
 little-0.131
 two-0.131
Token sounds
Token ID5238
Feature activation-2.54e-02

Pos logit contributions
anks+0.074
balls+0.071
 hyd+0.070
 rid+0.068
 heal+0.067
Neg logit contributions
ited-0.072
 saw-0.068
 heard-0.066
 little-0.060
 two-0.059
Token that
Token ID326
Feature activation-3.75e-02

Pos logit contributions
anks+0.055
 hyd+0.052
balls+0.052
uggets+0.051
sing+0.051
Neg logit contributions
 saw-0.030
ited-0.029
 heard-0.028
 little-0.027
 two-0.025
Token could
Token ID714
Feature activation-5.62e-02

Pos logit contributions
anks+0.086
balls+0.082
uggets+0.082
 hyd+0.081
sing+0.078
Neg logit contributions
 saw-0.043
 heard-0.040
ited-0.040
 did-0.037
 time-0.034
Token hurt
Token ID5938
Feature activation-2.79e-02

Pos logit contributions
anks+0.099
balls+0.092
ometer+0.091
 hyd+0.091
uggets+0.091
Neg logit contributions
 saw-0.094
ited-0.090
 heard-0.090
 two-0.082
 little-0.082
Token can
Token ID460
Feature activation-5.61e-02

Pos logit contributions
anks+0.122
balls+0.115
uggets+0.114
 hyd+0.114
ectar+0.114
Neg logit contributions
 saw-0.066
 heard-0.060
 did-0.057
ited-0.056
 two-0.055
Token make
Token ID787
Feature activation-3.68e-02

Pos logit contributions
anks+0.100
 hyd+0.093
ectar+0.092
uggets+0.091
balls+0.090
Neg logit contributions
 saw-0.096
 heard-0.090
ited-0.086
 two-0.082
 little-0.080
Token it
Token ID340
Feature activation-9.97e-02

Pos logit contributions
anks+0.079
 hyd+0.078
balls+0.078
 rid+0.074
sing+0.073
Neg logit contributions
 saw-0.042
ited-0.041
 heard-0.039
 little-0.037
 two-0.037
Token go
Token ID467
Feature activation-4.43e-02

Pos logit contributions
anks+0.192
balls+0.188
uggets+0.186
aro+0.186
ectar+0.181
Neg logit contributions
 saw-0.143
 heard-0.134
 little-0.129
ited-0.128
 two-0.126
Token faster
Token ID5443
Feature activation-3.14e-02

Pos logit contributions
anks+0.089
balls+0.085
 hyd+0.081
ectar+0.081
uggets+0.080
Neg logit contributions
 saw-0.062
ited-0.061
 heard-0.058
 little-0.054
 time-0.053
Token or
Token ID393
Feature activation-1.04e-01

Pos logit contributions
anks+0.060
balls+0.058
 hyd+0.057
 rid+0.056
 heal+0.055
Neg logit contributions
ited-0.048
 saw-0.043
 heard-0.042
 year-0.037
 two-0.037
Token slower
Token ID13611
Feature activation-3.74e-02

Pos logit contributions
anks+0.195
balls+0.183
 hyd+0.182
ectar+0.176
uggets+0.175
Neg logit contributions
 saw-0.160
ited-0.158
 heard-0.151
 two-0.139
 little-0.138
Token.
Token ID13
Feature activation-6.15e-02

Pos logit contributions
anks+0.068
balls+0.066
 hyd+0.065
 rid+0.063
 heal+0.062
Neg logit contributions
ited-0.059
 saw-0.055
 heard-0.053
 two-0.048
 little-0.048
Token Sam
Token ID3409
Feature activation+1.04e-02

Pos logit contributions
anks+0.128
balls+0.126
 hyd+0.124
 rid+0.122
 bleed+0.119
Neg logit contributions
ited-0.076
 saw-0.074
 heard-0.069
 two-0.068
Hey-0.065
Token is
Token ID318
Feature activation-1.90e-03

Pos logit contributions
ited+0.013
 saw+0.011
 heard+0.010
 year+0.010
Hey+0.009
Neg logit contributions
anks-0.022
balls-0.021
 hyd-0.021
 heal-0.021
 rid-0.021
Token very
Token ID845
Feature activation-1.82e-02

Pos logit contributions
anks+0.004
balls+0.004
 hyd+0.004
 rid+0.004
 heal+0.004
Neg logit contributions
ited-0.003
 saw-0.002
 heard-0.002
 year-0.002
 two-0.002
Token,
Token ID11
Feature activation+1.87e-03

Pos logit contributions
anks+0.019
balls+0.019
 hyd+0.018
 heal+0.018
 rid+0.018
Neg logit contributions
ited-0.028
 saw-0.025
 heard-0.024
Hey-0.023
 year-0.023
Token we
Token ID356
Feature activation-3.10e-02

Pos logit contributions
ited+0.003
 saw+0.002
 heard+0.002
 year+0.002
Hey+0.002
Neg logit contributions
anks-0.004
balls-0.004
 hyd-0.004
 rid-0.004
 heal-0.004
Token found
Token ID1043
Feature activation+3.08e-02

Pos logit contributions
anks+0.056
 hyd+0.055
 rid+0.055
balls+0.055
 heal+0.054
Neg logit contributions
ited-0.049
 saw-0.041
 heard-0.039
 year-0.037
Hey-0.037
Token a
Token ID257
Feature activation-2.40e-03

Pos logit contributions
ited+0.031
 saw+0.027
 heard+0.025
 two+0.022
 little+0.021
Neg logit contributions
anks-0.075
balls-0.073
 hyd-0.071
 heal-0.070
 rid-0.070
Token rare
Token ID4071
Feature activation-1.71e-02

Pos logit contributions
anks+0.005
balls+0.005
 hyd+0.005
 rid+0.005
 heal+0.004
Neg logit contributions
ited-0.003
 saw-0.003
 heard-0.003
 little-0.003
 year-0.003
Token rad
Token ID2511
Feature activation-1.04e-01

Pos logit contributions
anks+0.029
balls+0.028
 hyd+0.027
 rid+0.026
 heal+0.026
Neg logit contributions
ited-0.030
 saw-0.028
 heard-0.027
 year-0.024
 little-0.024
Tokenish
Token ID680
Feature activation-6.49e-02

Pos logit contributions
anks+0.132
balls+0.125
sing+0.118
 bleed+0.118
 heal+0.118
Neg logit contributions
ited-0.223
 saw-0.215
 heard-0.200
 two-0.192
 little-0.190
Token!"
Token ID2474
Feature activation-2.80e-02

Pos logit contributions
anks+0.145
 bleed+0.137
 heal+0.133
 rid+0.133
sing+0.133
Neg logit contributions
 saw-0.074
ited-0.070
 heard-0.067
 two-0.061
 year-0.060
Token they
Token ID484
Feature activation-6.82e-03

Pos logit contributions
anks+0.060
balls+0.058
 hyd+0.057
 rid+0.056
 heal+0.056
Neg logit contributions
ited-0.036
 saw-0.033
 heard-0.031
 two-0.028
 little-0.027
Token shouted
Token ID17293
Feature activation+3.83e-02

Pos logit contributions
anks+0.016
balls+0.015
 rid+0.015
 hyd+0.015
 heal+0.015
Neg logit contributions
ited-0.007
 saw-0.006
 heard-0.005
Hey-0.005
 year-0.005
Token,
Token ID11
Feature activation+8.11e-04

Pos logit contributions
 saw+0.031
 little+0.030
 heard+0.029
 two+0.028
,+0.027
Neg logit contributions
anks-0.096
 bleed-0.091
ometers-0.090
 affect-0.090
ophone-0.089
Token some
Token ID617
Feature activation-2.10e-02

Pos logit contributions
anks+0.028
balls+0.027
 hyd+0.027
 rid+0.026
 heal+0.026
Neg logit contributions
ited-0.018
 saw-0.016
 heard-0.015
 two-0.013
Hey-0.013
Token buttons
Token ID12163
Feature activation-5.39e-02

Pos logit contributions
anks+0.041
balls+0.039
 hyd+0.039
 rid+0.038
 heal+0.037
Neg logit contributions
ited-0.031
 saw-0.028
 heard-0.027
 two-0.024
 year-0.024
Token or
Token ID393
Feature activation-1.01e-01

Pos logit contributions
anks+0.120
balls+0.117
 hyd+0.115
 rid+0.113
 heal+0.113
Neg logit contributions
ited-0.066
 saw-0.056
 heard-0.053
 year-0.047
Hey-0.046
Token rib
Token ID12183
Feature activation-5.92e-02

Pos logit contributions
anks+0.175
balls+0.170
 hyd+0.167
 rid+0.162
 heal+0.160
Neg logit contributions
ited-0.171
 saw-0.157
 heard-0.151
 two-0.137
 year-0.136
Tokenbons
Token ID23461
Feature activation-4.79e-02

Pos logit contributions
anks+0.067
balls+0.065
 hyd+0.063
 rid+0.060
 heal+0.059
Neg logit contributions
ited-0.135
 saw-0.127
 heard-0.123
 two-0.115
Hey-0.115
Token or
Token ID393
Feature activation-1.04e-01

Pos logit contributions
anks+0.089
balls+0.087
 hyd+0.085
 rid+0.083
 heal+0.083
Neg logit contributions
ited-0.077
 saw-0.067
 heard-0.065
Hey-0.059
 year-0.059
Token patches
Token ID16082
Feature activation-5.52e-02

Pos logit contributions
anks+0.179
balls+0.174
 hyd+0.170
 rid+0.166
 heal+0.165
Neg logit contributions
ited-0.176
 saw-0.161
 heard-0.154
 year-0.140
Hey-0.140
Token.
Token ID13
Feature activation-7.67e-02

Pos logit contributions
anks+0.103
balls+0.100
 hyd+0.098
 rid+0.096
 heal+0.094
Neg logit contributions
ited-0.085
 saw-0.078
 heard-0.075
 two-0.068
 year-0.067
Token Would
Token ID10928
Feature activation-2.57e-03

Pos logit contributions
anks+0.123
balls+0.121
 rid+0.117
 bleed+0.114
 hyd+0.110
Neg logit contributions
 saw-0.126
 two-0.125
 heard-0.119
ited-0.119
 little-0.118
Token you
Token ID345
Feature activation-3.65e-02

Pos logit contributions
anks+0.004
 hyd+0.004
 bleed+0.004
 rid+0.004
balls+0.004
Neg logit contributions
 saw-0.004
ited-0.004
 heard-0.004
 two-0.004
 little-0.004
Token like
Token ID588
Feature activation-2.66e-03

Pos logit contributions
anks+0.064
 heal+0.061
 rid+0.061
balls+0.060
 hyd+0.060
Neg logit contributions
ited-0.062
 saw-0.053
 heard-0.051
Hey-0.050
 year-0.048
Token okay
Token ID8788
Feature activation-8.30e-03

Pos logit contributions
anks+0.098
balls+0.097
 rid+0.096
 hyd+0.095
 bleed+0.091
Neg logit contributions
ited-0.110
 saw-0.097
 heard-0.094
Hey-0.092
 year-0.092
Token,
Token ID11
Feature activation-3.23e-02

Pos logit contributions
anks+0.024
balls+0.023
 rid+0.023
 hyd+0.023
 bleed+0.022
Neg logit contributions
ited-0.004
 saw-0.004
 heard-0.003
 little-0.003
 two-0.002
Token Anna
Token ID11735
Feature activation-9.42e-03

Pos logit contributions
balls+0.058
anks+0.058
 hyd+0.056
 heal+0.055
 rid+0.054
Neg logit contributions
ited-0.053
 saw-0.045
 heard-0.044
 year-0.040
Hey-0.039
Token?
Token ID30
Feature activation-7.65e-02

Pos logit contributions
anks+0.018
balls+0.017
 hyd+0.017
 rid+0.016
 heal+0.016
Neg logit contributions
ited-0.015
 saw-0.013
 heard-0.013
 year-0.011
Hey-0.011
Token Does
Token ID8314
Feature activation-4.58e-03

Pos logit contributions
anks+0.142
balls+0.133
 rid+0.131
 heal+0.125
 hyd+0.125
Neg logit contributions
 saw-0.118
ited-0.109
 heard-0.108
Hey-0.104
 two-0.104
Token it
Token ID340
Feature activation-1.03e-01

Pos logit contributions
anks+0.010
 hyd+0.010
balls+0.009
 rid+0.009
sing+0.009
Neg logit contributions
 saw-0.005
ited-0.005
 heard-0.005
 little-0.005
 two-0.005
Token hurt
Token ID5938
Feature activation-3.27e-02

Pos logit contributions
anks+0.191
balls+0.185
 hyd+0.184
uggets+0.182
aro+0.177
Neg logit contributions
 saw-0.160
ited-0.153
 heard-0.152
 did-0.141
 little-0.137
Token a
Token ID257
Feature activation-2.83e-02

Pos logit contributions
anks+0.066
balls+0.064
 hyd+0.062
 rid+0.060
 heal+0.059
Neg logit contributions
ited-0.046
 saw-0.043
 heard-0.041
 little-0.036
 two-0.036
Token lot
Token ID1256
Feature activation-4.32e-02

Pos logit contributions
anks+0.058
balls+0.054
 hyd+0.053
ometers+0.053
 rid+0.051
Neg logit contributions
ited-0.039
 saw-0.039
 heard-0.037
 little-0.035
 year-0.034
Token?"
Token ID1701
Feature activation-2.50e-02

Pos logit contributions
anks+0.075
 heal+0.072
 hyd+0.071
balls+0.071
 rid+0.070
Neg logit contributions
ited-0.075
 saw-0.066
 heard-0.063
 noticed-0.057
Hey-0.057
Token
Token ID198
Feature activation-2.32e-03

Pos logit contributions
anks+0.058
balls+0.056
 hyd+0.056
 rid+0.054
 heal+0.054
Neg logit contributions
ited-0.028
 saw-0.026
 heard-0.024
 two-0.021
 little-0.021
Token not
Token ID407
Feature activation-4.18e-02

Pos logit contributions
anks+0.080
balls+0.078
 hyd+0.076
 rid+0.075
 heal+0.074
Neg logit contributions
ited-0.059
 saw-0.054
 heard-0.052
 two-0.046
 little-0.046
Token something
Token ID1223
Feature activation-4.50e-02

Pos logit contributions
anks+0.090
balls+0.087
 hyd+0.085
 rid+0.082
 heal+0.081
Neg logit contributions
ited-0.051
 saw-0.050
 heard-0.047
 little-0.045
 two-0.044
Token you
Token ID345
Feature activation-5.73e-02

Pos logit contributions
anks+0.110
 hyd+0.103
balls+0.102
sing+0.101
 bleed+0.101
Neg logit contributions
 saw-0.045
 little-0.042
 heard-0.039
ited-0.039
 time-0.038
Token can
Token ID460
Feature activation-6.63e-02

Pos logit contributions
anks+0.106
balls+0.100
 heal+0.098
 hyd+0.098
 rid+0.098
Neg logit contributions
ited-0.092
 saw-0.078
 heard-0.076
 year-0.075
Hey-0.072
Token buy
Token ID2822
Feature activation-2.36e-02

Pos logit contributions
anks+0.124
 hyd+0.121
balls+0.120
 rid+0.115
ometers+0.114
Neg logit contributions
ited-0.097
 saw-0.096
 heard-0.091
 two-0.084
 little-0.082
Token or
Token ID393
Feature activation-1.03e-01

Pos logit contributions
anks+0.044
balls+0.042
 hyd+0.042
 rid+0.040
 heal+0.040
Neg logit contributions
ited-0.037
 saw-0.034
 heard-0.033
 two-0.030
 little-0.029
Token take
Token ID1011
Feature activation-1.19e-02

Pos logit contributions
anks+0.167
balls+0.162
 hyd+0.159
 rid+0.155
 heal+0.154
Neg logit contributions
ited-0.187
 saw-0.170
 heard-0.164
 year-0.151
Hey-0.150
Token.
Token ID13
Feature activation-5.91e-02

Pos logit contributions
anks+0.023
balls+0.023
 rid+0.022
 hyd+0.022
 heal+0.022
Neg logit contributions
ited-0.018
 saw-0.015
 heard-0.015
Hey-0.013
 year-0.013
Token Love
Token ID5896
Feature activation+1.51e-02

Pos logit contributions
anks+0.118
balls+0.118
 rid+0.112
 hyd+0.111
 heal+0.110
Neg logit contributions
 saw-0.075
ited-0.074
 two-0.071
 heard-0.070
Hey-0.068
Token is
Token ID318
Feature activation-2.70e-02

Pos logit contributions
ited+0.016
 saw+0.013
 heard+0.012
Hey+0.010
 year+0.010
Neg logit contributions
anks-0.037
balls-0.036
 hyd-0.035
 heal-0.035
 rid-0.035
Token something
Token ID1223
Feature activation-2.72e-02

Pos logit contributions
anks+0.058
balls+0.056
 hyd+0.056
 bleed+0.054
 rid+0.054
Neg logit contributions
ited-0.032
 saw-0.031
 heard-0.029
 two-0.027
 little-0.027
Token more
Token ID517
Feature activation-3.26e-02

Pos logit contributions
anks+0.080
balls+0.077
 hyd+0.076
 rid+0.073
 heal+0.073
Neg logit contributions
ited-0.067
 saw-0.062
 heard-0.060
 two-0.054
 little-0.054
Token important
Token ID1593
Feature activation-1.21e-02

Pos logit contributions
anks+0.063
balls+0.062
 hyd+0.059
 bleed+0.058
sing+0.057
Neg logit contributions
ited-0.047
 saw-0.046
 heard-0.043
 little-0.042
 two-0.040
Token than
Token ID621
Feature activation-4.54e-02

Pos logit contributions
anks+0.018
balls+0.017
 hyd+0.017
 rid+0.016
 heal+0.016
Neg logit contributions
ited-0.024
 saw-0.022
 heard-0.022
 year-0.020
 two-0.020
Token being
Token ID852
Feature activation-5.91e-02

Pos logit contributions
anks+0.082
balls+0.079
 hyd+0.078
 rid+0.074
 bleed+0.074
Neg logit contributions
ited-0.072
 saw-0.068
 heard-0.065
 two-0.060
 little-0.060
Token clever
Token ID14169
Feature activation-1.85e-02

Pos logit contributions
anks+0.120
balls+0.117
 hyd+0.115
 rid+0.114
 heal+0.112
Neg logit contributions
ited-0.082
 saw-0.073
 heard-0.069
 year-0.062
Hey-0.062
Token or
Token ID393
Feature activation-1.03e-01

Pos logit contributions
anks+0.040
balls+0.039
 hyd+0.038
 rid+0.038
 heal+0.038
Neg logit contributions
ited-0.024
 saw-0.020
 heard-0.018
Hey-0.016
 year-0.016
Token playing
Token ID2712
Feature activation-3.00e-02

Pos logit contributions
anks+0.193
balls+0.186
 hyd+0.182
 rid+0.176
 heal+0.175
Neg logit contributions
ited-0.159
 saw-0.148
 heard-0.141
 two-0.128
 little-0.128
Token tricks
Token ID15910
Feature activation-6.37e-02

Pos logit contributions
anks+0.055
balls+0.053
 hyd+0.052
 rid+0.051
 bleed+0.050
Neg logit contributions
ited-0.046
 saw-0.044
 heard-0.042
 little-0.039
 two-0.038
Token.
Token ID13
Feature activation-5.48e-02

Pos logit contributions
anks+0.132
 hyd+0.129
balls+0.128
 rid+0.125
 bleed+0.124
Neg logit contributions
ited-0.081
 saw-0.076
 heard-0.073
 two-0.066
 little-0.064
Token<|endoftext|>
Token ID50256
Feature activation+3.91e-02

Pos logit contributions
balls+0.156
anks+0.150
 rid+0.147
 bleed+0.147
sing+0.143
Neg logit contributions
 saw-0.028
 year-0.026
 two-0.025
Once-0.025
 heard-0.023
TokenOnce
Token ID7454
Feature activation+6.14e-02

Pos logit contributions
ited+0.054
 saw+0.044
 heard+0.041
 noticed+0.036
 little+0.036
Neg logit contributions
anks-0.083
balls-0.080
 hyd-0.078
 rid-0.077
 heal-0.077
Token we
Token ID356
Feature activation-5.59e-02

Pos logit contributions
anks+0.137
balls+0.134
 hyd+0.131
 rid+0.129
 heal+0.129
Neg logit contributions
ited-0.127
 saw-0.113
 heard-0.108
Hey-0.099
 year-0.099
Token can
Token ID460
Feature activation-4.86e-02

Pos logit contributions
anks+0.116
balls+0.113
 hyd+0.110
 rid+0.107
 heal+0.105
Neg logit contributions
ited-0.076
 saw-0.071
 heard-0.067
 two-0.059
 did-0.059
Token make
Token ID787
Feature activation-2.32e-02

Pos logit contributions
anks+0.091
balls+0.090
 hyd+0.088
 rid+0.085
ectar+0.082
Neg logit contributions
ited-0.074
 saw-0.071
 heard-0.067
 two-0.062
 little-0.060
Token some
Token ID617
Feature activation-2.71e-02

Pos logit contributions
anks+0.051
 hyd+0.050
balls+0.050
sing+0.048
fly+0.048
Neg logit contributions
 saw-0.025
 two-0.024
ited-0.024
 heard-0.023
 little-0.023
Token cards
Token ID4116
Feature activation-8.55e-02

Pos logit contributions
anks+0.055
balls+0.054
 hyd+0.053
 rid+0.052
 heal+0.051
Neg logit contributions
ited-0.037
 saw-0.034
 heard-0.032
 little-0.029
 year-0.029
Token or
Token ID393
Feature activation-1.03e-01

Pos logit contributions
anks+0.170
balls+0.164
 hyd+0.163
 bleed+0.159
sing+0.159
Neg logit contributions
 saw-0.112
ited-0.110
 heard-0.106
 two-0.098
 little-0.095
Token drawings
Token ID23388
Feature activation-5.85e-02

Pos logit contributions
anks+0.174
balls+0.170
 heal+0.167
 hyd+0.166
 rid+0.165
Neg logit contributions
ited-0.182
 saw-0.156
 heard-0.150
 year-0.142
Hey-0.142
Token to
Token ID284
Feature activation-5.96e-02

Pos logit contributions
anks+0.104
balls+0.101
 hyd+0.101
 rid+0.098
sing+0.096
Neg logit contributions
ited-0.091
 saw-0.087
 heard-0.084
 two-0.077
 little-0.076
Token cheer
Token ID14042
Feature activation-3.23e-02

Pos logit contributions
anks+0.096
balls+0.093
 hyd+0.091
 rid+0.087
 bleed+0.085
Neg logit contributions
ited-0.107
 saw-0.102
 heard-0.098
 little-0.089
 two-0.089
Token up
Token ID510
Feature activation-2.73e-02

Pos logit contributions
anks+0.068
 hyd+0.066
balls+0.066
 bleed+0.064
 rid+0.064
Neg logit contributions
ited-0.039
 saw-0.038
 heard-0.036
 little-0.033
 two-0.033
Token the
Token ID262
Feature activation-1.77e-02

Pos logit contributions
anks+0.053
balls+0.051
 hyd+0.051
 rid+0.049
 heal+0.049
Neg logit contributions
ited-0.041
 saw-0.037
 heard-0.035
Hey-0.031
 year-0.031
TokenPlease
Token ID5492
Feature activation-2.54e-02

Pos logit contributions
anks+0.093
 hyd+0.093
balls+0.092
 rid+0.091
 heal+0.090
Neg logit contributions
ited-0.057
 saw-0.051
 heard-0.048
 little-0.042
 year-0.042
Token help
Token ID1037
Feature activation+2.22e-02

Pos logit contributions
anks+0.051
balls+0.049
 hyd+0.048
 rid+0.047
 bleed+0.047
Neg logit contributions
ited-0.035
 saw-0.033
 heard-0.032
 little-0.029
 two-0.028
Token me
Token ID502
Feature activation-3.27e-02

Pos logit contributions
ited+0.032
 saw+0.031
 heard+0.030
 little+0.028
 two+0.028
Neg logit contributions
anks-0.043
balls-0.041
 hyd-0.040
 bleed-0.039
 rid-0.039
Token shake
Token ID13279
Feature activation-2.22e-03

Pos logit contributions
anks+0.063
balls+0.061
ometers+0.059
ometer+0.059
 hyd+0.058
Neg logit contributions
 saw-0.044
ited-0.043
 heard-0.042
 little-0.041
 two-0.041
Token this
Token ID428
Feature activation-1.65e-02

Pos logit contributions
anks+0.005
balls+0.005
 hyd+0.005
 rid+0.004
 heal+0.004
Neg logit contributions
ited-0.003
 saw-0.003
 heard-0.002
Hey-0.002
 year-0.002
Token rad
Token ID2511
Feature activation-1.03e-01

Pos logit contributions
anks+0.029
balls+0.028
 hyd+0.028
 rid+0.027
 heal+0.027
Neg logit contributions
ited-0.027
 saw-0.025
 heard-0.024
Hey-0.021
 year-0.021
Tokenish
Token ID680
Feature activation-5.92e-02

Pos logit contributions
anks+0.124
balls+0.120
 hyd+0.111
sing+0.109
 heal+0.109
Neg logit contributions
ited-0.229
 saw-0.217
 heard-0.206
 two-0.195
 noticed-0.191
Token,
Token ID11
Feature activation-3.49e-02

Pos logit contributions
anks+0.112
balls+0.107
 hyd+0.105
 rid+0.103
 bleed+0.101
Neg logit contributions
ited-0.089
 saw-0.083
 heard-0.080
 two-0.074
 little-0.072
Token it
Token ID340
Feature activation-4.07e-02

Pos logit contributions
anks+0.063
balls+0.059
 hyd+0.058
 rid+0.056
 bleed+0.056
Neg logit contributions
ited-0.055
 saw-0.055
 heard-0.052
 little-0.049
 two-0.049
Token's
Token ID338
Feature activation-2.29e-02

Pos logit contributions
anks+0.089
uggets+0.085
balls+0.083
ectar+0.080
aro+0.079
Neg logit contributions
 saw-0.053
ited-0.053
 heard-0.049
 did-0.046
 two-0.044
Token too
Token ID1165
Feature activation-3.97e-02

Pos logit contributions
anks+0.048
balls+0.046
 hyd+0.045
 rid+0.044
 heal+0.043
Neg logit contributions
ited-0.031
 saw-0.028
 heard-0.027
 two-0.024
 little-0.024
Token things
Token ID1243
Feature activation-4.72e-02

Pos logit contributions
anks+0.034
balls+0.033
 hyd+0.032
 rid+0.032
 heal+0.031
Neg logit contributions
ited-0.036
 saw-0.033
 heard-0.032
 year-0.029
Hey-0.029
Token that
Token ID326
Feature activation-3.53e-02

Pos logit contributions
anks+0.119
balls+0.116
 hyd+0.115
sing+0.114
 rid+0.112
Neg logit contributions
 saw-0.037
ited-0.034
 heard-0.032
 little-0.029
 two-0.028
Token could
Token ID714
Feature activation-5.08e-02

Pos logit contributions
anks+0.074
balls+0.073
uggets+0.071
 hyd+0.070
sing+0.067
Neg logit contributions
 saw-0.048
 heard-0.043
 did-0.043
ited-0.041
 two-0.039
Token cause
Token ID2728
Feature activation-4.14e-02

Pos logit contributions
anks+0.096
balls+0.092
ophone+0.090
 hyd+0.090
ometer+0.089
Neg logit contributions
 saw-0.076
ited-0.074
 heard-0.072
 two-0.068
 little-0.067
Token harm
Token ID4419
Feature activation-1.80e-02

Pos logit contributions
anks+0.076
 rid+0.073
balls+0.072
 heal+0.072
 hyd+0.071
Neg logit contributions
ited-0.067
 saw-0.058
 heard-0.055
Hey-0.051
 year-0.051
Token or
Token ID393
Feature activation-1.03e-01

Pos logit contributions
anks+0.037
balls+0.037
 hyd+0.036
 rid+0.035
sing+0.034
Neg logit contributions
ited-0.023
 saw-0.022
 heard-0.021
 little-0.019
 two-0.019
Token trouble
Token ID5876
Feature activation-4.08e-02

Pos logit contributions
anks+0.176
balls+0.170
 hyd+0.166
 rid+0.158
 heal+0.155
Neg logit contributions
ited-0.170
 saw-0.165
 heard-0.158
 little-0.144
 two-0.144
Token.
Token ID13
Feature activation-4.00e-02

Pos logit contributions
anks+0.078
balls+0.077
 hyd+0.076
sing+0.073
uggets+0.073
Neg logit contributions
 saw-0.055
ited-0.054
 heard-0.053
 two-0.048
 little-0.048
Token<|endoftext|>
Token ID50256
Feature activation+7.21e-02

Pos logit contributions
balls+0.093
anks+0.091
 rid+0.089
 hyd+0.088
 bleed+0.087
Neg logit contributions
 saw-0.040
ited-0.040
 heard-0.037
 little-0.036
 two-0.034
TokenOnce
Token ID7454
Feature activation+8.31e-02

Pos logit contributions
ited+0.116
 saw+0.091
 heard+0.086
 noticed+0.079
Hey+0.079
Neg logit contributions
anks-0.137
balls-0.134
 hyd-0.129
 heal-0.129
 rid-0.128
Token upon
Token ID2402
Feature activation+6.34e-02

Pos logit contributions
ited+0.093
Hey+0.068
 heard+0.065
 saw+0.062
Fine+0.054
Neg logit contributions
anks-0.188
balls-0.188
 rid-0.179
 heal-0.179
ometer-0.178
Token make
Token ID787
Feature activation-8.51e-03

Pos logit contributions
anks+0.039
 hyd+0.038
 heal+0.038
 rid+0.038
balls+0.037
Neg logit contributions
ited-0.048
 saw-0.042
 heard-0.041
 year-0.039
Hey-0.039
Token sure
Token ID1654
Feature activation+7.06e-03

Pos logit contributions
anks+0.018
balls+0.018
 hyd+0.018
 rid+0.017
 heal+0.017
Neg logit contributions
ited-0.010
 saw-0.010
 heard-0.009
 little-0.008
 two-0.008
Token nothing
Token ID2147
Feature activation-1.29e-02

Pos logit contributions
ited+0.010
 saw+0.009
 heard+0.008
Hey+0.008
 year+0.008
Neg logit contributions
anks-0.014
balls-0.014
 hyd-0.014
 rid-0.013
 heal-0.013
Token would
Token ID561
Feature activation-5.27e-02

Pos logit contributions
anks+0.027
balls+0.026
 hyd+0.026
 rid+0.026
 heal+0.026
Neg logit contributions
ited-0.017
 saw-0.015
 heard-0.014
Hey-0.013
 year-0.013
Token make
Token ID787
Feature activation-3.17e-02

Pos logit contributions
anks+0.080
balls+0.077
 hyd+0.076
 rid+0.074
 heal+0.074
Neg logit contributions
ited-0.100
 saw-0.092
 heard-0.089
Hey-0.082
 year-0.082
Token it
Token ID340
Feature activation-1.02e-01

Pos logit contributions
anks+0.073
balls+0.073
 hyd+0.071
sing+0.069
 rid+0.068
Neg logit contributions
ited-0.032
 saw-0.031
 heard-0.029
 little-0.027
 two-0.027
Token slip
Token ID13819
Feature activation-3.54e-02

Pos logit contributions
anks+0.177
balls+0.176
 hyd+0.168
uggets+0.163
aro+0.163
Neg logit contributions
ited-0.167
 saw-0.164
 heard-0.158
 little-0.147
 two-0.143
Token away
Token ID1497
Feature activation-2.50e-02

Pos logit contributions
anks+0.070
balls+0.068
 hyd+0.067
 rid+0.065
 heal+0.064
Neg logit contributions
ited-0.051
 saw-0.047
 heard-0.045
 two-0.039
 year-0.039
Token again
Token ID757
Feature activation-2.28e-02

Pos logit contributions
anks+0.047
balls+0.046
 hyd+0.045
 rid+0.044
 heal+0.043
Neg logit contributions
ited-0.038
 saw-0.035
 heard-0.033
 two-0.030
 little-0.030
Token.
Token ID13
Feature activation-4.43e-02

Pos logit contributions
anks+0.039
balls+0.038
 hyd+0.038
 rid+0.037
 heal+0.037
Neg logit contributions
ited-0.039
 saw-0.035
 heard-0.034
Hey-0.031
 year-0.030
Token Then
Token ID3244
Feature activation+1.55e-02

Pos logit contributions
anks+0.097
balls+0.095
 hyd+0.094
 rid+0.092
 bleed+0.091
Neg logit contributions
ited-0.053
 saw-0.049
 heard-0.046
 two-0.042
 little-0.041
Token hopefully
Token ID11481
Feature activation-2.26e-02

Pos logit contributions
anks+0.141
balls+0.133
 hyd+0.125
ophone+0.123
ectar+0.123
Neg logit contributions
 saw-0.127
ited-0.123
 heard-0.118
 little-0.109
 two-0.108
Token you
Token ID345
Feature activation-5.45e-02

Pos logit contributions
anks+0.039
balls+0.038
 hyd+0.037
 rid+0.036
 heal+0.035
Neg logit contributions
ited-0.038
 saw-0.036
 heard-0.034
 little-0.031
 two-0.031
Token will
Token ID481
Feature activation-2.91e-02

Pos logit contributions
anks+0.101
balls+0.098
 hyd+0.095
 rid+0.093
 heal+0.091
Neg logit contributions
ited-0.085
 saw-0.080
 heard-0.077
 two-0.069
 little-0.068
Token be
Token ID307
Feature activation-5.86e-02

Pos logit contributions
anks+0.053
balls+0.052
 hyd+0.048
 rid+0.048
ometers+0.047
Neg logit contributions
ited-0.045
 saw-0.044
 heard-0.043
 two-0.039
 little-0.039
Token healthy
Token ID5448
Feature activation-3.17e-02

Pos logit contributions
anks+0.110
balls+0.107
 hyd+0.103
 rid+0.100
 heal+0.100
Neg logit contributions
ited-0.092
 saw-0.085
 heard-0.082
 two-0.074
 little-0.073
Token and
Token ID290
Feature activation-1.02e-01

Pos logit contributions
anks+0.055
 hyd+0.054
balls+0.053
 heal+0.052
 rid+0.051
Neg logit contributions
ited-0.055
 saw-0.048
 heard-0.046
Hey-0.044
 did-0.041
Token back
Token ID736
Feature activation-3.42e-02

Pos logit contributions
anks+0.178
balls+0.172
 hyd+0.168
 rid+0.163
 heal+0.161
Neg logit contributions
ited-0.171
 saw-0.158
 heard-0.151
 two-0.137
 little-0.136
Token to
Token ID284
Feature activation-2.45e-02

Pos logit contributions
anks+0.062
balls+0.060
 hyd+0.059
 rid+0.058
 heal+0.057
Neg logit contributions
ited-0.054
 saw-0.050
 heard-0.048
 two-0.044
 little-0.043
Token your
Token ID534
Feature activation-1.06e-02

Pos logit contributions
anks+0.047
balls+0.045
ometer+0.044
ometers+0.043
ophone+0.042
Neg logit contributions
 saw-0.036
ited-0.036
 heard-0.034
 two-0.032
 little-0.032
Token old
Token ID1468
Feature activation-3.22e-02

Pos logit contributions
anks+0.019
balls+0.018
 hyd+0.018
 rid+0.018
 bleed+0.017
Neg logit contributions
ited-0.017
 saw-0.016
 heard-0.015
 little-0.014
 year-0.014
Token self
Token ID2116
Feature activation-9.10e-03

Pos logit contributions
anks+0.058
balls+0.057
 hyd+0.056
 rid+0.055
 bleed+0.053
Neg logit contributions
ited-0.050
 saw-0.048
 heard-0.046
 little-0.042
 two-0.042
Token real
Token ID1103
Feature activation-2.63e-02

Pos logit contributions
 heal+0.021
 rid+0.020
 hyd+0.020
balls+0.020
anks+0.020
Neg logit contributions
ited-0.017
 saw-0.014
 heard-0.013
Hey-0.013
 year-0.013
Token wheat
Token ID17135
Feature activation-8.06e-02

Pos logit contributions
anks+0.053
balls+0.051
 hyd+0.050
 rid+0.049
 bleed+0.048
Neg logit contributions
ited-0.036
 saw-0.035
 heard-0.033
 little-0.030
 two-0.030
Token could
Token ID714
Feature activation-5.89e-02

Pos logit contributions
anks+0.174
balls+0.168
 hyd+0.165
 rid+0.162
 heal+0.160
Neg logit contributions
ited-0.102
 saw-0.093
 heard-0.088
 two-0.076
 year-0.076
Token make
Token ID787
Feature activation-4.39e-02

Pos logit contributions
anks+0.105
balls+0.100
ophone+0.095
 hyd+0.095
ometer+0.094
Neg logit contributions
ited-0.094
 saw-0.092
 heard-0.088
 two-0.082
 little-0.081
Token good
Token ID922
Feature activation-5.54e-02

Pos logit contributions
anks+0.097
balls+0.094
 hyd+0.091
sing+0.089
 bleed+0.088
Neg logit contributions
ited-0.050
 saw-0.049
 heard-0.046
 little-0.044
 two-0.043
Token bread
Token ID8509
Feature activation-1.02e-01

Pos logit contributions
anks+0.103
balls+0.100
 hyd+0.097
 rid+0.095
 heal+0.094
Neg logit contributions
ited-0.087
 saw-0.080
 heard-0.077
 two-0.069
 little-0.069
Token.
Token ID13
Feature activation-5.44e-02

Pos logit contributions
anks+0.217
ophone+0.205
sing+0.203
balls+0.202
 bleed+0.201
Neg logit contributions
 saw-0.115
 heard-0.108
ited-0.108
 two-0.102
 little-0.100
Token And
Token ID843
Feature activation-3.18e-02

Pos logit contributions
anks+0.117
balls+0.115
 hyd+0.112
 rid+0.111
 heal+0.109
Neg logit contributions
ited-0.068
 saw-0.063
 heard-0.059
 two-0.053
 little-0.053
Token Jack
Token ID3619
Feature activation+8.84e-03

Pos logit contributions
anks+0.062
balls+0.061
 hyd+0.060
 rid+0.059
 heal+0.058
Neg logit contributions
ited-0.048
 saw-0.041
 heard-0.039
Hey-0.036
 year-0.035
Token loved
Token ID6151
Feature activation+1.96e-02

Pos logit contributions
ited+0.010
Hey+0.008
 year+0.007
 conflict+0.007
Next+0.007
Neg logit contributions
 hyd-0.019
 heal-0.018
anks-0.018
fly-0.018
balls-0.018
Token to
Token ID284
Feature activation-2.46e-02

Pos logit contributions
ited+0.026
 saw+0.024
 heard+0.023
 little+0.021
 two+0.021
Neg logit contributions
anks-0.041
balls-0.039
 hyd-0.039
 rid-0.038
 heal-0.037
Token's
Token ID338
Feature activation-2.06e-02

Pos logit contributions
uggets+0.144
anks+0.141
ectar+0.133
balls+0.132
sing+0.132
Neg logit contributions
 saw-0.063
 heard-0.058
 did-0.058
ited-0.053
 two-0.052
Token important
Token ID1593
Feature activation-7.19e-03

Pos logit contributions
anks+0.038
balls+0.036
 hyd+0.035
 rid+0.034
 heal+0.033
Neg logit contributions
ited-0.032
 saw-0.031
 heard-0.030
 two-0.028
 little-0.028
Token to
Token ID284
Feature activation-2.85e-02

Pos logit contributions
anks+0.015
balls+0.014
 hyd+0.014
 rid+0.014
 heal+0.014
Neg logit contributions
ited-0.009
 saw-0.009
 heard-0.009
 two-0.008
 little-0.007
Token be
Token ID307
Feature activation-4.64e-02

Pos logit contributions
anks+0.049
balls+0.048
 hyd+0.047
 rid+0.045
ometer+0.044
Neg logit contributions
ited-0.047
 saw-0.046
 heard-0.043
 two-0.040
 little-0.040
Token quiet
Token ID5897
Feature activation-1.82e-02

Pos logit contributions
anks+0.101
balls+0.098
 hyd+0.095
 rid+0.093
 heal+0.092
Neg logit contributions
ited-0.058
 saw-0.055
 heard-0.052
 two-0.046
 little-0.046
Token and
Token ID290
Feature activation-1.02e-01

Pos logit contributions
anks+0.035
balls+0.034
 hyd+0.034
 heal+0.033
 rid+0.033
Neg logit contributions
ited-0.029
 saw-0.024
 heard-0.023
Hey-0.022
 year-0.021
Token stay
Token ID2652
Feature activation-3.61e-02

Pos logit contributions
anks+0.194
balls+0.187
 hyd+0.181
 rid+0.176
 heal+0.173
Neg logit contributions
ited-0.153
 saw-0.145
 heard-0.138
 little-0.124
 two-0.123
Token away
Token ID1497
Feature activation-2.65e-02

Pos logit contributions
anks+0.068
balls+0.066
ometers+0.061
 hyd+0.061
ometer+0.060
Neg logit contributions
ited-0.056
 saw-0.054
 heard-0.051
 little-0.047
 two-0.047
Token from
Token ID422
Feature activation-1.75e-02

Pos logit contributions
anks+0.051
balls+0.050
 hyd+0.047
ophone+0.047
aro+0.046
Neg logit contributions
ited-0.036
 saw-0.036
 heard-0.034
 little-0.033
 two-0.032
Token them
Token ID606
Feature activation-5.19e-02

Pos logit contributions
anks+0.034
balls+0.033
 hyd+0.033
 rid+0.032
 heal+0.032
Neg logit contributions
ited-0.026
 saw-0.023
 heard-0.022
Hey-0.020
 year-0.020
Token.
Token ID13
Feature activation-5.44e-02

Pos logit contributions
anks+0.097
balls+0.095
 hyd+0.093
 rid+0.090
 bleed+0.088
Neg logit contributions
ited-0.078
 saw-0.074
 heard-0.071
 little-0.065
 two-0.065
Token
Token ID220
Feature activation-1.74e-02

Pos logit contributions
anks+0.017
balls+0.017
 rid+0.016
 hyd+0.016
 heal+0.016
Neg logit contributions
ited-0.013
 saw-0.013
 heard-0.012
 did-0.011
 two-0.011
Token
Token ID198
Feature activation+9.32e-03

Pos logit contributions
 hyd+0.043
anks+0.042
balls+0.041
 bleed+0.041
 rid+0.041
Neg logit contributions
ited-0.014
 saw-0.014
 heard-0.013
 two-0.013
 little-0.011
Token
Token ID198
Feature activation-4.82e-03

Pos logit contributions
ited+0.015
 saw+0.013
 heard+0.012
 enough+0.011
 noticed+0.011
Neg logit contributions
balls-0.017
anks-0.017
 heal-0.016
 rid-0.016
 hyd-0.015
TokenTom
Token ID13787
Feature activation+2.99e-02

Pos logit contributions
 hyd+0.009
balls+0.009
uries+0.009
 heal+0.009
anks+0.009
Neg logit contributions
ited-0.006
 saw-0.006
 heard-0.005
 enough-0.004
 did-0.004
Tokenmy
Token ID1820
Feature activation+2.95e-02

Pos logit contributions
ited+0.042
 heard+0.034
 saw+0.034
Hey+0.033
 year+0.033
Neg logit contributions
anks-0.058
balls-0.055
ophone-0.055
 hyd-0.055
fly-0.054
Token said
Token ID531
Feature activation+8.82e-02

Pos logit contributions
ited+0.027
 saw+0.019
 heard+0.019
 year+0.017
Hey+0.017
Neg logit contributions
anks-0.071
balls-0.071
 hyd-0.069
 heal-0.068
 rid-0.068
Token,
Token ID11
Feature activation+1.92e-02

Pos logit contributions
ited+0.067
 saw+0.055
 heard+0.049
 year+0.037
Hey+0.037
Neg logit contributions
anks-0.235
balls-0.231
 hyd-0.228
 rid-0.224
 heal-0.223
Token
Token ID6184
Feature activation+1.82e-03

Pos logit contributions
ited+0.029
 saw+0.026
 heard+0.025
 year+0.022
Hey+0.022
Neg logit contributions
anks-0.037
balls-0.036
 hyd-0.035
 rid-0.034
 heal-0.034
Token
Token ID95
Feature activation-8.78e-03

Pos logit contributions
ited+0.003
 saw+0.003
Hey+0.003
 heard+0.003
 did+0.003
Neg logit contributions
balls-0.003
anks-0.003
 rid-0.003
fly-0.003
 bleed-0.003
Token
Token ID26391
Feature activation-9.28e-03

Pos logit contributions
balls+0.014
anks+0.014
 rid+0.013
 bleed+0.013
fly+0.013
Neg logit contributions
ited-0.016
 saw-0.015
 heard-0.014
Hey-0.014
 did-0.013
Token
Token ID129
Feature activation+1.24e-03

Pos logit contributions
anks+0.015
balls+0.015
 rid+0.014
 hyd+0.014
 heal+0.014
Neg logit contributions
ited-0.016
 saw-0.016
 heard-0.015
 did-0.014
 year-0.013
Token of
Token ID286
Feature activation+3.73e-03

Pos logit contributions
ited+0.036
 heard+0.032
 saw+0.030
Hey+0.030
 did+0.029
Neg logit contributions
anks-0.057
 hyd-0.057
 heal-0.055
balls-0.055
ectar-0.054
Token the
Token ID262
Feature activation-1.13e-03

Pos logit contributions
ited+0.006
Fine+0.005
 saw+0.005
 heard+0.005
Hey+0.005
Neg logit contributions
 rid-0.007
anks-0.006
uggets-0.006
 hyd-0.006
 heal-0.006
Token vendors
Token ID17192
Feature activation-1.81e-02

Pos logit contributions
 hyd+0.002
balls+0.002
anks+0.002
 rid+0.002
 heal+0.002
Neg logit contributions
ited-0.002
 heard-0.001
 saw-0.001
ara-0.001
 did-0.001
Token and
Token ID290
Feature activation-2.24e-02

Pos logit contributions
 hyd+0.035
anks+0.035
balls+0.035
 heal+0.034
ometers+0.033
Neg logit contributions
ited-0.025
Hey-0.022
 heard-0.019
ha-0.018
 conflict-0.018
Token politely
Token ID34313
Feature activation+5.70e-03

Pos logit contributions
balls+0.045
 rid+0.044
 hyd+0.044
 heal+0.042
anks+0.042
Neg logit contributions
ited-0.030
 year-0.022
 End-0.021
Hey-0.021
 time-0.021
Token said
Token ID531
Feature activation+6.56e-02

Pos logit contributions
ited+0.007
Hey+0.006
 End+0.006
 year+0.006
 conflict+0.005
Neg logit contributions
 hyd-0.011
anks-0.011
balls-0.011
 heal-0.010
 rid-0.010
Token,
Token ID11
Feature activation+4.40e-03

Pos logit contributions
ited+0.096
 saw+0.074
 heard+0.070
Hey+0.063
 year+0.063
Neg logit contributions
balls-0.138
anks-0.136
 heal-0.131
 hyd-0.128
 rid-0.128
Token "
Token ID366
Feature activation-2.31e-02

Pos logit contributions
ited+0.006
 saw+0.006
 heard+0.005
 two+0.005
 little+0.005
Neg logit contributions
anks-0.009
balls-0.009
 hyd-0.008
 rid-0.008
 heal-0.008
TokenHello
Token ID15496
Feature activation+2.97e-02

Pos logit contributions
 hyd+0.043
anks+0.041
balls+0.041
cr+0.040
 rid+0.040
Neg logit contributions
ited-0.032
 saw-0.029
 heard-0.027
 did-0.026
 conflict-0.025
Token,
Token ID11
Feature activation+8.10e-03

Pos logit contributions
ited+0.043
 heard+0.031
 sighed+0.030
 saw+0.030
 End+0.028
Neg logit contributions
anks-0.060
balls-0.057
 hyd-0.056
flies-0.054
ometers-0.054
Token could
Token ID714
Feature activation-4.39e-03

Pos logit contributions
ited+0.013
 heard+0.011
 saw+0.010
 sighed+0.009
ised+0.009
Neg logit contributions
anks-0.014
balls-0.014
 hyd-0.014
 heal-0.014
ometers-0.013
Token not
Token ID407
Feature activation-3.55e-02

Pos logit contributions
anks+0.063
balls+0.061
 hyd+0.060
 rid+0.058
 heal+0.058
Neg logit contributions
ited-0.046
 saw-0.042
 heard-0.040
 two-0.036
 little-0.036
Token afraid
Token ID7787
Feature activation-8.12e-03

Pos logit contributions
anks+0.071
balls+0.069
 hyd+0.068
 rid+0.066
 heal+0.065
Neg logit contributions
ited-0.051
 saw-0.047
 heard-0.044
 little-0.039
 two-0.039
Token."
Token ID526
Feature activation-2.65e-02

Pos logit contributions
anks+0.014
balls+0.013
 hyd+0.013
 rid+0.012
 heal+0.012
Neg logit contributions
ited-0.014
 saw-0.013
 heard-0.013
 little-0.012
 year-0.012
Token
Token ID198
Feature activation+3.39e-04

Pos logit contributions
anks+0.070
balls+0.069
 bleed+0.067
sing+0.066
ophone+0.066
Neg logit contributions
 saw-0.021
ited-0.019
 little-0.018
 heard-0.018
 two-0.017
Token
Token ID198
Feature activation-1.52e-02

Pos logit contributions
ited+0.000
 saw+0.000
 heard+0.000
 year+0.000
 two+0.000
Neg logit contributions
anks-0.001
balls-0.001
 hyd-0.001
 rid-0.001
 heal-0.001
TokenThe
Token ID464
Feature activation+6.36e-02

Pos logit contributions
balls+0.034
 hyd+0.034
anks+0.034
 rid+0.034
 heal+0.034
Neg logit contributions
ited-0.017
 saw-0.015
 heard-0.014
Hey-0.011
 enough-0.011
Token little
Token ID1310
Feature activation+3.92e-02

Pos logit contributions
ited+0.066
Hey+0.045
 saw+0.045
aw+0.044
Tom+0.043
Neg logit contributions
balls-0.141
 heal-0.140
 hyd-0.139
 rid-0.138
sing-0.134
Token girl
Token ID2576
Feature activation+1.04e-02

Pos logit contributions
ited+0.050
 heard+0.039
Tom+0.039
 saw+0.039
When+0.039
Neg logit contributions
 hyd-0.076
 rid-0.074
 heal-0.069
flies-0.069
fly-0.069
Token realised
Token ID19169
Feature activation+6.86e-02

Pos logit contributions
ited+0.012
 saw+0.008
Hey+0.008
 heard+0.008
 year+0.008
Neg logit contributions
 hyd-0.023
balls-0.023
anks-0.022
 heal-0.022
 rid-0.022
Token that
Token ID326
Feature activation+3.00e-02

Pos logit contributions
ited+0.102
 saw+0.082
 heard+0.077
Hey+0.074
 year+0.067
Neg logit contributions
 heal-0.133
 hyd-0.132
anks-0.131
balls-0.131
 rid-0.130
Token if
Token ID611
Feature activation+1.89e-02

Pos logit contributions
ited+0.040
 saw+0.034
 heard+0.032
Hey+0.029
 year+0.028
Neg logit contributions
anks-0.063
balls-0.062
 hyd-0.062
 rid-0.061
 heal-0.060
TokenL
Token ID43
Feature activation+3.68e-02

Pos logit contributions
ited+0.001
 enough+0.001
 saw+0.001
 heard+0.001
joy+0.001
Neg logit contributions
balls-0.001
 heal-0.001
 hyd-0.001
ophone-0.001
tr-0.001
Tokenily
Token ID813
Feature activation+4.17e-02

Pos logit contributions
ited+0.068
 saw+0.064
 heard+0.060
 noticed+0.056
 arrived+0.055
Neg logit contributions
anks-0.058
ectar-0.055
balls-0.051
ophone-0.051
uggets-0.050
Token's
Token ID338
Feature activation+6.76e-02

Pos logit contributions
ited+0.070
grown+0.060
Hey+0.059
joy+0.056
he+0.056
Neg logit contributions
 hyd-0.062
balls-0.060
gy-0.059
 fet-0.056
 heal-0.056
Token mom
Token ID1995
Feature activation+3.23e-02

Pos logit contributions
ited+0.136
 saw+0.115
Hey+0.115
 heard+0.113
 noticed+0.103
Neg logit contributions
balls-0.101
anks-0.100
 hyd-0.094
sing-0.091
 heal-0.090
Tokenmy
Token ID1820
Feature activation+2.84e-02

Pos logit contributions
ited+0.059
Hey+0.058
 amusement+0.055
joy+0.054
 trouble+0.054
Neg logit contributions
 heal-0.036
fly-0.036
ophone-0.035
balls-0.035
 hyd-0.034
Token warned
Token ID7728
Feature activation+7.93e-02

Pos logit contributions
ited+0.037
Hey+0.032
 time+0.030
 conflict+0.030
Next+0.029
Neg logit contributions
 hyd-0.050
 heal-0.049
balls-0.049
fly-0.049
flies-0.046
Token her
Token ID607
Feature activation+4.93e-02

Pos logit contributions
ited+0.135
 saw+0.115
 heard+0.104
Hey+0.102
 noticed+0.100
Neg logit contributions
 heal-0.141
 rid-0.132
anks-0.132
balls-0.130
 bleed-0.128
Token that
Token ID326
Feature activation+4.66e-02

Pos logit contributions
ited+0.079
 saw+0.070
 heard+0.067
Hey+0.061
 year+0.060
Neg logit contributions
anks-0.092
balls-0.090
 hyd-0.088
 heal-0.086
 rid-0.086
Token it
Token ID340
Feature activation-3.23e-04

Pos logit contributions
ited+0.060
 saw+0.056
 heard+0.053
 two+0.047
 little+0.047
Neg logit contributions
anks-0.099
balls-0.095
 hyd-0.093
 rid-0.092
 heal-0.090
Token might
Token ID1244
Feature activation-1.85e-02

Pos logit contributions
anks+0.001
balls+0.001
 hyd+0.001
 rid+0.001
 heal+0.001
Neg logit contributions
ited-0.000
 saw-0.000
 heard-0.000
 two-0.000
 did-0.000
Token be
Token ID307
Feature activation-4.07e-02

Pos logit contributions
anks+0.035
balls+0.034
ophone+0.034
uggets+0.032
ometers+0.032
Neg logit contributions
 saw-0.027
ited-0.026
 heard-0.026
 time-0.024
 two-0.024
TokenOnce
Token ID7454
Feature activation+7.05e-02

Pos logit contributions
ited+0.073
 saw+0.056
 heard+0.053
 conflict+0.049
 noticed+0.048
Neg logit contributions
anks-0.094
balls-0.091
 hyd-0.089
 rid-0.087
 heal-0.087
Token upon
Token ID2402
Feature activation+5.11e-02

Pos logit contributions
ited+0.079
Hey+0.055
 heard+0.053
 saw+0.051
Fine+0.045
Neg logit contributions
balls-0.161
anks-0.159
ometer-0.153
 rid-0.153
ectar-0.152
Token a
Token ID257
Feature activation+8.07e-02

Pos logit contributions
ited+0.074
 saw+0.054
Hey+0.051
 conflict+0.049
 heard+0.049
Neg logit contributions
 heal-0.102
balls-0.101
 bleed-0.100
 hyd-0.098
 rid-0.098
Token time
Token ID640
Feature activation+5.34e-02

Pos logit contributions
ited+0.061
 saw+0.050
Hey+0.048
 heard+0.045
 two+0.042
Neg logit contributions
 hyd-0.188
sing-0.188
balls-0.185
 heal-0.181
anks-0.177
Token,
Token ID11
Feature activation+6.42e-02

Pos logit contributions
ited+0.050
 saw+0.035
 heard+0.033
Hey+0.027
 year+0.025
Neg logit contributions
balls-0.135
anks-0.134
 hyd-0.130
 heal-0.128
 rid-0.127
Token there
Token ID612
Feature activation+6.66e-02

Pos logit contributions
ited+0.092
oph+0.068
aman+0.056
 amusement+0.056
 conflict+0.056
Neg logit contributions
balls-0.133
fly-0.129
 hyd-0.126
sing-0.125
 heal-0.124
Token was
Token ID373
Feature activation+5.29e-02

Pos logit contributions
ited+0.108
vious+0.066
 conflict+0.066
Exc+0.065
Hey+0.065
Neg logit contributions
balls-0.124
 heal-0.122
 hyd-0.121
cr-0.114
fly-0.114
Token a
Token ID257
Feature activation+4.43e-02

Pos logit contributions
ited+0.077
 saw+0.058
 heard+0.057
 sighed+0.051
Hey+0.050
Neg logit contributions
balls-0.105
 hyd-0.102
anks-0.100
uggets-0.100
 rid-0.100
Token little
Token ID1310
Feature activation+2.67e-02

Pos logit contributions
ited+0.025
 saw+0.019
 heard+0.016
 year+0.010
 little+0.010
Neg logit contributions
anks-0.127
balls-0.124
 hyd-0.123
 rid-0.121
 heal-0.120
Token girl
Token ID2576
Feature activation+2.44e-02

Pos logit contributions
ited+0.022
 saw+0.020
 heard+0.017
 little+0.015
 year+0.014
Neg logit contributions
anks-0.071
balls-0.068
 hyd-0.066
 rid-0.066
 heal-0.065
Token named
Token ID3706
Feature activation+3.38e-02

Pos logit contributions
 saw+0.031
ited+0.029
 heard+0.028
 year+0.025
 little+0.025
Neg logit contributions
anks-0.057
balls-0.055
 rid-0.051
 bleed-0.051
ometers-0.051
Token
Token ID198
Feature activation-1.60e-03

Pos logit contributions
anks+0.110
balls+0.108
 hyd+0.107
 rid+0.104
 heal+0.103
Neg logit contributions
ited-0.054
 saw-0.049
 heard-0.046
 two-0.040
Hey-0.039
Token
Token ID198
Feature activation-1.33e-02

Pos logit contributions
anks+0.003
balls+0.003
 heal+0.003
 hyd+0.003
 rid+0.003
Neg logit contributions
ited-0.002
 saw-0.002
 heard-0.002
Hey-0.002
 year-0.002
TokenShe
Token ID3347
Feature activation+5.92e-03

Pos logit contributions
balls+0.024
 hyd+0.023
 heal+0.023
anks+0.023
 bleed+0.023
Neg logit contributions
ited-0.021
 saw-0.018
 heard-0.017
 enough-0.016
 noticed-0.015
Token waited
Token ID13488
Feature activation+4.60e-02

Pos logit contributions
ited+0.009
 End+0.006
 year+0.006
ised+0.006
Today+0.006
Neg logit contributions
 fet-0.011
 hyd-0.010
balls-0.010
ophone-0.010
 rid-0.010
Token for
Token ID329
Feature activation+2.25e-02

Pos logit contributions
ited+0.092
Hey+0.077
 saw+0.077
 heard+0.073
 year+0.070
Neg logit contributions
balls-0.072
anks-0.069
 heal-0.066
 hyd-0.066
 rid-0.064
Token a
Token ID257
Feature activation+1.60e-02

Pos logit contributions
ited+0.030
 saw+0.024
 heard+0.022
Hey+0.021
 year+0.020
Neg logit contributions
anks-0.048
balls-0.048
 rid-0.048
 heal-0.047
 hyd-0.046
Token quiet
Token ID5897
Feature activation+2.57e-02

Pos logit contributions
ited+0.018
 saw+0.016
 heard+0.016
 little+0.014
 year+0.013
Neg logit contributions
anks-0.037
balls-0.036
 hyd-0.035
 rid-0.034
 heal-0.034
Token moment
Token ID2589
Feature activation+2.31e-02

Pos logit contributions
ited+0.026
 saw+0.018
Hey+0.017
 heard+0.015
Next+0.015
Neg logit contributions
 heal-0.060
 hyd-0.060
balls-0.059
anks-0.059
 rid-0.059
Token.
Token ID13
Feature activation-3.68e-02

Pos logit contributions
ited+0.036
 saw+0.031
 heard+0.029
Hey+0.029
 noticed+0.026
Neg logit contributions
anks-0.043
balls-0.043
 hyd-0.041
 heal-0.040
 rid-0.040
Token Then
Token ID3244
Feature activation+1.73e-02

Pos logit contributions
anks+0.068
balls+0.064
 hyd+0.061
 heal+0.061
ometers+0.059
Neg logit contributions
ited-0.064
 saw-0.054
 heard-0.052
 year-0.049
Hey-0.047
Token she
Token ID673
Feature activation-8.26e-04

Pos logit contributions
ited+0.030
 saw+0.021
 year+0.021
 heard+0.021
Hey+0.021
Neg logit contributions
balls-0.032
anks-0.031
 heal-0.030
 rid-0.030
 hyd-0.030
Token?"
Token ID1701
Feature activation+1.82e-02

Pos logit contributions
anks+0.114
balls+0.111
 hyd+0.109
 rid+0.106
 bleed+0.104
Neg logit contributions
ited-0.081
 saw-0.077
 heard-0.073
 two-0.066
 little-0.065
Token asked
Token ID1965
Feature activation+8.27e-02

Pos logit contributions
ited+0.025
 saw+0.018
 heard+0.017
 conflict+0.017
Hey+0.016
Neg logit contributions
 heal-0.035
 hyd-0.035
balls-0.034
fly-0.034
anks-0.034
Token Charles
Token ID7516
Feature activation+2.76e-02

Pos logit contributions
ited+0.170
 saw+0.144
 heard+0.137
 conflict+0.132
Hey+0.130
Neg logit contributions
balls-0.113
anks-0.112
 heal-0.108
 rid-0.106
 hyd-0.105
Token.
Token ID13
Feature activation-9.20e-03

Pos logit contributions
ited+0.039
 saw+0.037
 heard+0.035
 little+0.031
 two+0.031
Neg logit contributions
anks-0.056
balls-0.054
 hyd-0.054
 rid-0.052
 heal-0.051
Token
Token ID198
Feature activation+1.07e-02

Pos logit contributions
anks+0.021
balls+0.020
 hyd+0.020
 rid+0.020
 heal+0.019
Neg logit contributions
ited-0.011
 saw-0.010
 heard-0.009
 two-0.008
 year-0.008
Token
Token ID198
Feature activation-3.96e-04

Pos logit contributions
ited+0.014
 saw+0.012
 heard+0.011
Hey+0.010
 did+0.009
Neg logit contributions
anks-0.023
balls-0.023
 hyd-0.022
 heal-0.022
 rid-0.022
Token"
Token ID1
Feature activation-9.80e-03

Pos logit contributions
balls+0.001
 hyd+0.001
 heal+0.001
anks+0.001
fly+0.001
Neg logit contributions
ited-0.001
 saw-0.000
 heard-0.000
 did-0.000
 conflict-0.000
TokenI
Token ID40
Feature activation-1.13e-02

Pos logit contributions
anks+0.021
balls+0.021
 hyd+0.020
 rid+0.020
 heal+0.020
Neg logit contributions
ited-0.012
 saw-0.011
 heard-0.010
 little-0.009
 year-0.009
Token'm
Token ID1101
Feature activation+2.00e-03

Pos logit contributions
 hyd+0.018
 rid+0.017
anks+0.017
balls+0.017
 heal+0.017
Neg logit contributions
ited-0.020
 saw-0.017
 heard-0.017
Hey-0.017
 year-0.016
Token making
Token ID1642
Feature activation+1.47e-02

Pos logit contributions
 saw+0.003
ited+0.003
 two+0.003
 heard+0.003
 little+0.003
Neg logit contributions
anks-0.004
balls-0.004
 hyd-0.004
 heal-0.004
ophone-0.003
Token a
Token ID257
Feature activation-2.18e-02

Pos logit contributions
ited+0.015
 saw+0.014
 heard+0.013
 two+0.012
 little+0.012
Neg logit contributions
anks-0.034
 hyd-0.034
balls-0.034
 rid-0.032
sing-0.032
Token lot
Token ID1256
Feature activation+3.48e-03

Pos logit contributions
ited+0.003
 saw+0.003
 heard+0.003
 little+0.002
 year+0.002
Neg logit contributions
anks-0.007
balls-0.007
 hyd-0.007
 rid-0.007
 bleed-0.007
Token.
Token ID13
Feature activation-2.73e-02

Pos logit contributions
ited+0.005
 saw+0.005
 heard+0.005
 year+0.004
 two+0.004
Neg logit contributions
anks-0.007
balls-0.006
 hyd-0.006
 rid-0.006
 heal-0.006
Token Tim
Token ID5045
Feature activation+2.24e-02

Pos logit contributions
anks+0.054
 rid+0.053
balls+0.053
 hyd+0.052
 bleed+0.052
Neg logit contributions
 saw-0.037
ited-0.036
 heard-0.033
 little-0.032
 two-0.032
Token had
Token ID550
Feature activation+2.27e-02

Pos logit contributions
ited+0.028
Hey+0.022
 conflict+0.021
 year+0.021
ised+0.021
Neg logit contributions
 hyd-0.043
balls-0.042
 heal-0.042
anks-0.041
ophone-0.041
Token a
Token ID257
Feature activation+2.02e-02

Pos logit contributions
ited+0.032
Hey+0.024
 saw+0.023
 year+0.022
 heard+0.021
Neg logit contributions
balls-0.047
 rid-0.046
anks-0.046
 hyd-0.046
 heal-0.044
Token fun
Token ID1257
Feature activation+1.93e-02

Pos logit contributions
ited+0.029
 saw+0.024
 heard+0.023
Hey+0.021
 year+0.020
Neg logit contributions
balls-0.040
anks-0.040
 hyd-0.039
 rid-0.038
 heal-0.038
Token day
Token ID1110
Feature activation+1.13e-03

Pos logit contributions
ited+0.024
 saw+0.020
 heard+0.020
Hey+0.018
 two+0.017
Neg logit contributions
anks-0.042
balls-0.041
 hyd-0.041
 rid-0.039
 heal-0.039
Token at
Token ID379
Feature activation+1.14e-02

Pos logit contributions
ited+0.001
 saw+0.001
 heard+0.001
 two+0.001
 year+0.001
Neg logit contributions
anks-0.003
balls-0.003
 heal-0.003
 rid-0.003
 bleed-0.003
Token the
Token ID262
Feature activation+8.70e-03

Pos logit contributions
ited+0.012
 saw+0.010
 heard+0.009
Hey+0.008
 year+0.007
Neg logit contributions
anks-0.028
balls-0.027
 hyd-0.027
 rid-0.026
 heal-0.026
Token park
Token ID3952
Feature activation-3.49e-04

Pos logit contributions
ited+0.013
ised+0.009
 heard+0.009
 sighed+0.009
 frowned+0.009
Neg logit contributions
 heal-0.017
anks-0.016
 drown-0.016
 float-0.015
 hyd-0.015
Token with
Token ID351
Feature activation+3.28e-04

Pos logit contributions
anks+0.001
balls+0.001
 hyd+0.001
 rid+0.001
 heal+0.001
Neg logit contributions
ited-0.001
 saw-0.001
 heard-0.000
Hey-0.000
 year-0.000
Token smile
Token ID8212
Feature activation+1.53e-02

Pos logit contributions
ited+0.038
 saw+0.031
 heard+0.030
Hey+0.030
 year+0.028
Neg logit contributions
balls-0.043
 hyd-0.043
anks-0.043
 heal-0.042
 rid-0.042
Token.
Token ID13
Feature activation-1.25e-02

Pos logit contributions
ited+0.028
 saw+0.025
 heard+0.024
Hey+0.022
 year+0.022
Neg logit contributions
anks-0.025
balls-0.024
 hyd-0.024
 heal-0.023
 rid-0.023
Token "
Token ID366
Feature activation-1.69e-02

Pos logit contributions
anks+0.025
balls+0.024
 hyd+0.023
 rid+0.023
 heal+0.023
Neg logit contributions
ited-0.019
 saw-0.016
 heard-0.015
 year-0.014
Hey-0.014
TokenIf
Token ID1532
Feature activation+1.71e-02

Pos logit contributions
balls+0.034
anks+0.034
 hyd+0.033
 rid+0.033
 heal+0.032
Neg logit contributions
ited-0.023
 saw-0.021
 heard-0.020
 little-0.017
 year-0.017
Token you
Token ID345
Feature activation-2.19e-02

Pos logit contributions
ited+0.038
 saw+0.034
 heard+0.033
Hey+0.031
 year+0.031
Neg logit contributions
anks-0.021
balls-0.021
 hyd-0.020
 heal-0.019
 rid-0.019
Token guess
Token ID4724
Feature activation+9.09e-03

Pos logit contributions
anks+0.036
balls+0.035
 hyd+0.035
 rid+0.034
 heal+0.034
Neg logit contributions
ited-0.038
 saw-0.035
 heard-0.034
 year-0.031
Hey-0.031
Token right
Token ID826
Feature activation+1.97e-02

Pos logit contributions
ited+0.012
 saw+0.011
 heard+0.010
 year+0.009
Hey+0.009
Neg logit contributions
anks-0.019
balls-0.019
 hyd-0.018
 rid-0.018
 heal-0.018
Token,
Token ID11
Feature activation-1.33e-02

Pos logit contributions
ited+0.029
 saw+0.020
Hey+0.019
 heard+0.019
oph+0.019
Neg logit contributions
balls-0.040
anks-0.039
 hyd-0.039
ometer-0.038
 rid-0.038
Token you
Token ID345
Feature activation-1.39e-02

Pos logit contributions
anks+0.019
balls+0.019
 hyd+0.018
 heal+0.018
 rid+0.018
Neg logit contributions
ited-0.027
 saw-0.024
 heard-0.023
 year-0.022
Hey-0.022
Token get
Token ID651
Feature activation-7.03e-03

Pos logit contributions
anks+0.026
balls+0.025
 hyd+0.025
 heal+0.024
 rid+0.024
Neg logit contributions
ited-0.022
 saw-0.019
 heard-0.019
Hey-0.017
 year-0.017
Token a
Token ID257
Feature activation+1.53e-03

Pos logit contributions
anks+0.014
balls+0.014
 rid+0.014
 hyd+0.013
 heal+0.013
Neg logit contributions
ited-0.010
 saw-0.009
 heard-0.008
 year-0.008
Hey-0.008
Token later
Token ID1568
Feature activation+1.04e-02

Pos logit contributions
anks+0.033
balls+0.032
 hyd+0.032
 rid+0.031
 heal+0.031
Neg logit contributions
ited-0.027
 saw-0.024
 heard-0.023
Hey-0.021
 year-0.021
Token.
Token ID13
Feature activation-3.64e-02

Pos logit contributions
ited+0.018
 saw+0.015
Hey+0.015
 heard+0.014
 conflict+0.013
Neg logit contributions
balls-0.018
anks-0.017
 heal-0.017
 hyd-0.017
 rid-0.016
Token She
Token ID1375
Feature activation+1.50e-03

Pos logit contributions
anks+0.079
balls+0.077
 hyd+0.076
 rid+0.074
 heal+0.074
Neg logit contributions
ited-0.046
 saw-0.041
 heard-0.039
 two-0.034
 year-0.034
Token learned
Token ID4499
Feature activation+4.62e-02

Pos logit contributions
ited+0.002
Hey+0.002
 conflict+0.002
 year+0.002
 saw+0.002
Neg logit contributions
 hyd-0.003
balls-0.003
 heal-0.003
 rid-0.003
 bleed-0.003
Token a
Token ID257
Feature activation+1.03e-02

Pos logit contributions
ited+0.049
 saw+0.042
 heard+0.039
Hey+0.033
 year+0.033
Neg logit contributions
anks-0.111
balls-0.108
 hyd-0.106
 rid-0.105
 heal-0.104
Token lesson
Token ID11483
Feature activation+2.23e-02

Pos logit contributions
 little+0.009
 saw+0.009
 heard+0.008
ited+0.008
 year+0.008
Neg logit contributions
anks-0.026
 hyd-0.025
balls-0.025
uggets-0.024
aro-0.024
Token:
Token ID25
Feature activation-2.16e-02

Pos logit contributions
ited+0.042
 saw+0.031
Hey+0.031
Quick+0.029
Uh+0.029
Neg logit contributions
anks-0.039
balls-0.037
 heal-0.035
ometers-0.035
 hyd-0.035
Token never
Token ID1239
Feature activation-3.46e-03

Pos logit contributions
anks+0.044
balls+0.042
 hyd+0.041
 rid+0.041
 bleed+0.040
Neg logit contributions
ited-0.029
 saw-0.028
 heard-0.026
 two-0.024
 little-0.024
Token go
Token ID467
Feature activation-4.24e-03

Pos logit contributions
anks+0.006
balls+0.006
 hyd+0.006
 rid+0.006
 heal+0.005
Neg logit contributions
ited-0.006
 saw-0.005
 heard-0.005
 year-0.005
Hey-0.005
Token on
Token ID319
Feature activation-3.96e-03

Pos logit contributions
anks+0.008
balls+0.008
 hyd+0.008
 rid+0.007
 heal+0.007
Neg logit contributions
ited-0.007
 saw-0.006
 heard-0.006
Hey-0.005
 year-0.005
Token an
Token ID281
Feature activation+1.22e-02

Pos logit contributions
anks+0.009
balls+0.008
 hyd+0.008
 heal+0.008
 rid+0.008
Neg logit contributions
ited-0.005
 saw-0.004
 heard-0.004
Hey-0.004
 year-0.003
Token Thank
Token ID6952
Feature activation+6.87e-04

Pos logit contributions
anks+0.109
 rid+0.103
balls+0.100
 bleed+0.100
aro+0.097
Neg logit contributions
Hi-0.073
 saw-0.072
Hey-0.070
 two-0.070
 did-0.068
Token you
Token ID345
Feature activation+3.90e-03

Pos logit contributions
ited+0.001
 saw+0.001
 two+0.001
 heard+0.001
Hey+0.001
Neg logit contributions
anks-0.001
sing-0.001
ometer-0.001
 hyd-0.001
ometers-0.001
Token for
Token ID329
Feature activation-6.26e-03

Pos logit contributions
ited+0.004
 saw+0.004
 heard+0.004
 two+0.003
 little+0.003
Neg logit contributions
anks-0.009
 hyd-0.009
 bleed-0.009
 rid-0.009
balls-0.009
Token forgiving
Token ID43486
Feature activation-3.77e-03

Pos logit contributions
anks+0.011
balls+0.010
 hyd+0.010
 rid+0.010
 heal+0.010
Neg logit contributions
ited-0.011
 saw-0.010
 heard-0.010
 two-0.009
 little-0.009
Token me
Token ID502
Feature activation-2.54e-02

Pos logit contributions
anks+0.009
balls+0.008
 hyd+0.008
 rid+0.008
 heal+0.008
Neg logit contributions
ited-0.005
 saw-0.004
 heard-0.003
 year-0.003
 noticed-0.003
Token."
Token ID526
Feature activation-3.10e-02

Pos logit contributions
anks+0.046
balls+0.044
 bleed+0.043
 hyd+0.042
 rid+0.042
Neg logit contributions
ited-0.037
 saw-0.036
 heard-0.036
 two-0.035
 little-0.034
Token
Token ID198
Feature activation+4.30e-03

Pos logit contributions
anks+0.073
 hyd+0.073
 rid+0.072
 affect+0.072
 bleed+0.071
Neg logit contributions
 saw-0.031
 two-0.026
 heard-0.026
 little-0.025
ited-0.024
Token
Token ID198
Feature activation-1.01e-02

Pos logit contributions
ited+0.005
 saw+0.004
 heard+0.004
 year+0.004
 two+0.004
Neg logit contributions
anks-0.010
balls-0.010
 hyd-0.009
 rid-0.009
 heal-0.009
TokenThe
Token ID464
Feature activation+3.47e-02

Pos logit contributions
anks+0.021
balls+0.021
 hyd+0.020
 rid+0.020
 heal+0.020
Neg logit contributions
ited-0.013
 saw-0.012
 heard-0.011
 year-0.010
Hey-0.010
Token moral
Token ID6573
Feature activation+4.13e-02

Pos logit contributions
ited+0.042
 saw+0.035
 heard+0.033
 year+0.028
Hey+0.028
Neg logit contributions
anks-0.078
balls-0.078
 hyd-0.076
 rid-0.074
 heal-0.074
Token of
Token ID286
Feature activation+4.38e-02

Pos logit contributions
ited+0.059
 saw+0.043
 heard+0.040
 noticed+0.039
 spotted+0.034
Neg logit contributions
anks-0.087
 heal-0.085
 hyd-0.084
 rid-0.084
balls-0.083
Token
Token ID198
Feature activation-2.22e-02

Pos logit contributions
anks+0.034
balls+0.033
 hyd+0.032
 rid+0.032
 heal+0.032
Neg logit contributions
ited-0.019
 saw-0.017
 heard-0.016
Hey-0.013
 year-0.013
TokenL
Token ID43
Feature activation+1.38e-02

Pos logit contributions
balls+0.047
anks+0.047
 hyd+0.046
 rid+0.045
 heal+0.045
Neg logit contributions
ited-0.028
 saw-0.024
 heard-0.023
Hey-0.020
 two-0.019
Tokenily
Token ID813
Feature activation+3.09e-02

Pos logit contributions
ited+0.025
 saw+0.024
 heard+0.022
 noticed+0.021
 arrived+0.021
Neg logit contributions
anks-0.021
ectar-0.020
ophone-0.020
uggets-0.019
balls-0.019
Token learned
Token ID4499
Feature activation+4.41e-02

Pos logit contributions
ited+0.040
grown+0.033
Hey+0.032
 year+0.030
 time+0.030
Neg logit contributions
 hyd-0.059
 heal-0.058
balls-0.057
 fet-0.055
anks-0.055
Token that
Token ID326
Feature activation+2.66e-02

Pos logit contributions
ited+0.068
 saw+0.055
 heard+0.050
Hey+0.046
 noticed+0.044
Neg logit contributions
anks-0.090
 heal-0.089
balls-0.088
 rid-0.086
 hyd-0.085
Token it
Token ID340
Feature activation-2.63e-02

Pos logit contributions
ited+0.031
 saw+0.030
 heard+0.028
 two+0.026
 year+0.025
Neg logit contributions
anks-0.059
balls-0.057
 hyd-0.055
 rid-0.055
ophone-0.054
Token's
Token ID338
Feature activation-3.32e-03

Pos logit contributions
anks+0.058
balls+0.056
 hyd+0.056
uggets+0.055
sing+0.054
Neg logit contributions
 saw-0.030
ited-0.030
 heard-0.029
 two-0.026
 did-0.026
Token important
Token ID1593
Feature activation+2.06e-02

Pos logit contributions
anks+0.005
 hyd+0.005
balls+0.005
 heal+0.005
 rid+0.005
Neg logit contributions
ited-0.006
 saw-0.006
 heard-0.005
Hey-0.005
 year-0.005
Token to
Token ID284
Feature activation-1.46e-02

Pos logit contributions
ited+0.034
 saw+0.029
 heard+0.028
Hey+0.026
 year+0.025
Neg logit contributions
anks-0.038
balls-0.038
 hyd-0.036
 rid-0.036
 heal-0.036
Token take
Token ID1011
Feature activation-1.82e-03

Pos logit contributions
anks+0.023
 hyd+0.022
 heal+0.022
balls+0.022
 rid+0.022
Neg logit contributions
ited-0.027
 saw-0.024
 heard-0.024
Hey-0.022
 year-0.022
Token care
Token ID1337
Feature activation-2.09e-04

Pos logit contributions
anks+0.003
 rid+0.003
 hyd+0.003
balls+0.003
 heal+0.003
Neg logit contributions
ited-0.003
 saw-0.003
 heard-0.003
Hey-0.003
 year-0.003
Token sat
Token ID3332
Feature activation+3.96e-02

Pos logit contributions
 hyd+0.017
balls+0.017
 rid+0.016
anks+0.016
 heal+0.016
Neg logit contributions
ited-0.011
Hey-0.008
 saw-0.008
 year-0.008
 heard-0.008
Token down
Token ID866
Feature activation+2.45e-02

Pos logit contributions
ited+0.087
 heard+0.073
 saw+0.073
Hey+0.073
 year+0.070
Neg logit contributions
 hyd-0.054
balls-0.052
anks-0.051
 rid-0.048
 heal-0.045
Token to
Token ID284
Feature activation-3.29e-03

Pos logit contributions
ited+0.042
 saw+0.035
 heard+0.034
Hey+0.032
 year+0.031
Neg logit contributions
anks-0.044
 hyd-0.043
balls-0.043
 rid-0.042
 heal-0.040
Token have
Token ID423
Feature activation-5.27e-03

Pos logit contributions
anks+0.005
balls+0.005
 hyd+0.005
 rid+0.005
 heal+0.005
Neg logit contributions
ited-0.006
 saw-0.005
 heard-0.005
 year-0.005
Hey-0.005
Token their
Token ID511
Feature activation-1.82e-02

Pos logit contributions
anks+0.010
 hyd+0.010
balls+0.010
 rid+0.010
 heal+0.010
Neg logit contributions
ited-0.008
 saw-0.006
 heard-0.006
Hey-0.006
 year-0.005
Token picnic
Token ID35715
Feature activation-8.95e-03

Pos logit contributions
anks+0.031
balls+0.029
 hyd+0.029
 rid+0.028
 heal+0.028
Neg logit contributions
ited-0.032
 saw-0.030
 heard-0.028
 year-0.026
 two-0.026
Token.
Token ID13
Feature activation-3.35e-02

Pos logit contributions
anks+0.017
balls+0.016
 hyd+0.016
 rid+0.016
 heal+0.016
Neg logit contributions
ited-0.014
 saw-0.013
 heard-0.012
 two-0.011
 year-0.011
Token After
Token ID2293
Feature activation+1.58e-02

Pos logit contributions
anks+0.067
balls+0.065
 hyd+0.064
 rid+0.063
 heal+0.062
Neg logit contributions
ited-0.048
 saw-0.044
 heard-0.041
 two-0.037
 little-0.037
Token they
Token ID484
Feature activation-2.73e-03

Pos logit contributions
ited+0.030
Hey+0.022
 saw+0.022
 heard+0.021
Fine+0.020
Neg logit contributions
balls-0.029
 hyd-0.026
anks-0.026
aro-0.025
 heal-0.025
Token finished
Token ID5201
Feature activation+1.83e-02

Pos logit contributions
 hyd+0.005
 rid+0.005
cr+0.005
aro+0.005
 heal+0.005
Neg logit contributions
ited-0.003
Fine-0.003
Hey-0.003
ised-0.003
Next-0.003
Token,
Token ID11
Feature activation+8.66e-03

Pos logit contributions
ited+0.029
 saw+0.024
 heard+0.023
Hey+0.022
 year+0.021
Neg logit contributions
anks-0.035
balls-0.035
 hyd-0.034
 heal-0.033
 rid-0.032
Token in
Token ID287
Feature activation-6.04e-03

Pos logit contributions
anks+0.003
balls+0.003
 hyd+0.003
 rid+0.003
 heal+0.003
Neg logit contributions
ited-0.002
 saw-0.002
 heard-0.001
 two-0.001
 year-0.001
Token a
Token ID257
Feature activation+9.56e-03

Pos logit contributions
balls+0.012
 hyd+0.012
anks+0.012
 rid+0.012
 heal+0.012
Neg logit contributions
ited-0.008
 saw-0.006
 heard-0.006
Hey-0.006
oph-0.005
Token big
Token ID1263
Feature activation+4.56e-04

Pos logit contributions
ited+0.015
Hey+0.011
 saw+0.011
 heard+0.011
Fine+0.011
Neg logit contributions
 hyd-0.017
 heal-0.017
balls-0.016
 rid-0.016
sing-0.016
Token,
Token ID11
Feature activation-2.14e-02

Pos logit contributions
ited+0.001
Next+0.001
When+0.001
Fine+0.001
Hey+0.001
Neg logit contributions
 hyd-0.001
 heal-0.001
sing-0.001
fly-0.001
 float-0.001
Token dark
Token ID3223
Feature activation-9.62e-04

Pos logit contributions
 hyd+0.042
balls+0.041
anks+0.041
 heal+0.040
 rid+0.040
Neg logit contributions
ited-0.031
 saw-0.026
 heard-0.025
 year-0.022
Hey-0.022
Token house
Token ID2156
Feature activation-2.11e-02

Pos logit contributions
 hyd+0.002
fly+0.002
anks+0.002
 heal+0.002
balls+0.002
Neg logit contributions
ited-0.002
 saw-0.001
 heard-0.001
Hey-0.001
 did-0.001
Token,
Token ID11
Feature activation-2.82e-02

Pos logit contributions
 hyd+0.032
balls+0.032
anks+0.032
 heal+0.030
flies+0.029
Neg logit contributions
ited-0.041
Hey-0.033
 saw-0.032
 heard-0.031
oph-0.031
Token there
Token ID612
Feature activation+5.54e-03

Pos logit contributions
balls+0.060
 hyd+0.060
anks+0.060
 heal+0.057
 rid+0.057
Neg logit contributions
ited-0.037
 saw-0.029
 heard-0.028
Hey-0.026
 year-0.026
Token lived
Token ID5615
Feature activation+1.87e-02

Pos logit contributions
ited+0.007
Hey+0.004
 conflict+0.004
 saw+0.004
Exc+0.004
Neg logit contributions
balls-0.012
anks-0.011
 heal-0.011
 hyd-0.011
flies-0.011
Token a
Token ID257
Feature activation+1.56e-02

Pos logit contributions
ited+0.025
 saw+0.018
 heard+0.017
 conflict+0.015
Hey+0.015
Neg logit contributions
balls-0.040
anks-0.039
 hyd-0.039
 rid-0.039
 heal-0.038
Token stubborn
Token ID24444
Feature activation+1.89e-02

Pos logit contributions
ited+0.017
 saw+0.013
 heard+0.013
Hey+0.011
 two+0.010
Neg logit contributions
balls-0.036
 hyd-0.036
anks-0.036
 heal-0.035
 rid-0.035
Token noticed
Token ID6810
Feature activation+5.62e-02

Pos logit contributions
balls+0.014
 rid+0.013
cr+0.013
 hyd+0.013
 bleed+0.013
Neg logit contributions
ited-0.006
 year-0.004
Fine-0.004
Next-0.004
Hey-0.004
Token the
Token ID262
Feature activation+8.16e-03

Pos logit contributions
ited+0.095
 saw+0.063
 year+0.063
 heard+0.059
Hey+0.057
Neg logit contributions
balls-0.106
 rid-0.104
 hyd-0.103
 heal-0.100
aro-0.099
Token fan
Token ID4336
Feature activation-5.35e-02

Pos logit contributions
ited+0.017
 saw+0.015
 heard+0.015
Hey+0.014
 noticed+0.013
Neg logit contributions
balls-0.011
anks-0.011
 hyd-0.011
 rid-0.011
 heal-0.011
Token was
Token ID373
Feature activation-1.27e-02

Pos logit contributions
anks+0.108
balls+0.106
 hyd+0.105
 heal+0.102
 rid+0.102
Neg logit contributions
ited-0.075
 saw-0.065
 heard-0.061
Hey-0.057
 year-0.055
Token dirty
Token ID11841
Feature activation-4.96e-03

Pos logit contributions
 hyd+0.021
balls+0.021
anks+0.021
 rid+0.021
 heal+0.020
Neg logit contributions
ited-0.022
 saw-0.019
 heard-0.018
Hey-0.018
 year-0.017
Token with
Token ID351
Feature activation-2.84e-02

Pos logit contributions
anks+0.010
 hyd+0.010
balls+0.010
 heal+0.009
 rid+0.009
Neg logit contributions
ited-0.007
 saw-0.006
 heard-0.006
Hey-0.005
 year-0.005
Token dust
Token ID8977
Feature activation-2.74e-02

Pos logit contributions
 rid+0.052
 hyd+0.051
balls+0.051
anks+0.050
 heal+0.050
Neg logit contributions
ited-0.046
 saw-0.038
 heard-0.037
Hey-0.036
 year-0.035
Token and
Token ID290
Feature activation-6.95e-02

Pos logit contributions
anks+0.051
balls+0.050
 hyd+0.049
 rid+0.048
 heal+0.048
Neg logit contributions
ited-0.043
 saw-0.038
 heard-0.036
Hey-0.033
 year-0.033
Token leaves
Token ID5667
Feature activation-3.88e-02

Pos logit contributions
balls+0.126
anks+0.124
 hyd+0.123
 rid+0.122
 heal+0.119
Neg logit contributions
ited-0.115
 saw-0.095
 heard-0.091
Hey-0.090
 year-0.088
Token.
Token ID13
Feature activation-4.82e-02

Pos logit contributions
anks+0.068
balls+0.067
 hyd+0.065
 rid+0.064
 heal+0.064
Neg logit contributions
ited-0.065
 saw-0.058
 heard-0.055
Hey-0.052
 year-0.051
Token Lily
Token ID20037
Feature activation+1.98e-02

Pos logit contributions
anks+0.089
balls+0.086
 hyd+0.084
 rid+0.083
 heal+0.082
Neg logit contributions
ited-0.075
 saw-0.071
 heard-0.067
 little-0.062
 two-0.061
Tokenisy
Token ID13560
Feature activation-1.86e-02

Pos logit contributions
anks+0.017
balls+0.017
 hyd+0.016
 rid+0.016
 heal+0.016
Neg logit contributions
ited-0.020
 saw-0.019
 heard-0.018
 year-0.017
 two-0.016
Token always
Token ID1464
Feature activation-1.09e-03

Pos logit contributions
 hyd+0.041
balls+0.040
anks+0.040
 heal+0.039
 rid+0.039
Neg logit contributions
ited-0.021
Hey-0.016
 saw-0.014
 heard-0.014
 two-0.014
Token wanted
Token ID2227
Feature activation+4.02e-02

Pos logit contributions
 hyd+0.002
anks+0.002
balls+0.002
 rid+0.002
 heal+0.002
Neg logit contributions
ited-0.001
 saw-0.001
 heard-0.001
Hey-0.001
 year-0.001
Token more
Token ID517
Feature activation-9.86e-03

Pos logit contributions
ited+0.059
 saw+0.052
 heard+0.049
Hey+0.045
 year+0.044
Neg logit contributions
anks-0.080
balls-0.078
 hyd-0.076
 rid-0.075
 heal-0.074
Token sun
Token ID4252
Feature activation-4.83e-02

Pos logit contributions
anks+0.021
balls+0.021
 hyd+0.020
 rid+0.020
 heal+0.020
Neg logit contributions
ited-0.013
 saw-0.011
 heard-0.011
 year-0.010
 little-0.010
Token and
Token ID290
Feature activation-5.76e-02

Pos logit contributions
 drown+0.097
anks+0.096
balls+0.096
 hyd+0.095
 rid+0.095
Neg logit contributions
ited-0.058
 saw-0.058
 heard-0.056
 two-0.051
 enough-0.050
Token more
Token ID517
Feature activation-1.35e-02

Pos logit contributions
anks+0.107
balls+0.106
 hyd+0.105
 heal+0.105
 rid+0.105
Neg logit contributions
ited-0.091
 saw-0.074
 heard-0.071
Hey-0.070
 year-0.067
Token rain
Token ID6290
Feature activation-4.17e-02

Pos logit contributions
anks+0.027
balls+0.027
 hyd+0.026
 rid+0.025
 heal+0.025
Neg logit contributions
ited-0.019
 saw-0.018
 heard-0.017
 little-0.016
 year-0.016
Token,
Token ID11
Feature activation-2.93e-02

Pos logit contributions
anks+0.077
 hyd+0.075
balls+0.074
 rid+0.072
uggets+0.072
Neg logit contributions
 saw-0.061
ited-0.061
 heard-0.058
 did-0.053
 little-0.050
Token even
Token ID772
Feature activation-1.18e-02

Pos logit contributions
anks+0.066
balls+0.064
 hyd+0.063
 rid+0.062
 heal+0.061
Neg logit contributions
ited-0.035
 saw-0.031
 heard-0.029
 two-0.025
 year-0.025
Token when
Token ID618
Feature activation-1.10e-02

Pos logit contributions
anks+0.025
 rid+0.024
balls+0.024
 hyd+0.024
 heal+0.024
Neg logit contributions
ited-0.016
 saw-0.013
 heard-0.012
Hey-0.011
 year-0.011
Token that
Token ID326
Feature activation-2.38e-02

Pos logit contributions
anks+0.073
balls+0.070
 hyd+0.069
uggets+0.068
 rid+0.066
Neg logit contributions
ited-0.044
 saw-0.044
 heard-0.041
 did-0.038
 little-0.035
Token can
Token ID460
Feature activation-4.08e-02

Pos logit contributions
anks+0.050
balls+0.048
uggets+0.047
sing+0.045
 hyd+0.045
Neg logit contributions
 saw-0.033
 heard-0.031
ited-0.030
 did-0.029
 two-0.028
Token't
Token ID470
Feature activation-3.29e-02

Pos logit contributions
anks+0.080
uggets+0.079
ectar+0.075
 hyd+0.075
balls+0.074
Neg logit contributions
 saw-0.061
 heard-0.056
ited-0.054
 two-0.054
 little-0.051
Token be
Token ID307
Feature activation-4.49e-02

Pos logit contributions
anks+0.067
balls+0.062
uggets+0.062
ometer+0.060
ectar+0.060
Neg logit contributions
 saw-0.050
 heard-0.047
ited-0.044
 little-0.043
 two-0.042
Token changed
Token ID3421
Feature activation-1.63e-02

Pos logit contributions
anks+0.091
balls+0.089
 hyd+0.088
 rid+0.086
 heal+0.085
Neg logit contributions
ited-0.063
 saw-0.056
 heard-0.053
Hey-0.047
 year-0.047
Token or
Token ID393
Feature activation-8.01e-02

Pos logit contributions
anks+0.031
balls+0.030
 hyd+0.030
 rid+0.029
 heal+0.029
Neg logit contributions
ited-0.025
 saw-0.022
 heard-0.021
Hey-0.019
 year-0.019
Token fixed
Token ID5969
Feature activation+4.93e-03

Pos logit contributions
anks+0.146
balls+0.139
 hyd+0.136
 rid+0.131
 heal+0.128
Neg logit contributions
ited-0.125
 saw-0.121
 heard-0.116
 two-0.105
 little-0.104
Token.
Token ID13
Feature activation-3.60e-02

Pos logit contributions
ited+0.007
 saw+0.007
 heard+0.006
 two+0.006
 little+0.006
Neg logit contributions
anks-0.010
balls-0.009
 hyd-0.009
 rid-0.009
 heal-0.009
Token He
Token ID679
Feature activation-8.11e-03

Pos logit contributions
balls+0.076
 bleed+0.073
 hyd+0.073
anks+0.072
 rid+0.072
Neg logit contributions
ited-0.041
 saw-0.041
 two-0.040
 little-0.039
 heard-0.038
Token took
Token ID1718
Feature activation+2.85e-02

Pos logit contributions
anks+0.018
 hyd+0.018
balls+0.018
 heal+0.017
 rid+0.017
Neg logit contributions
ited-0.009
 saw-0.007
 heard-0.007
Hey-0.007
 year-0.007
Token this
Token ID428
Feature activation+1.23e-02

Pos logit contributions
ited+0.034
 saw+0.029
 heard+0.027
 year+0.023
Hey+0.023
Neg logit contributions
anks-0.064
balls-0.063
 rid-0.062
 hyd-0.062
 heal-0.061
Token could
Token ID714
Feature activation-2.48e-02

Pos logit contributions
anks+0.054
balls+0.053
 hyd+0.052
 rid+0.052
 heal+0.051
Neg logit contributions
ited-0.020
 saw-0.015
 heard-0.014
Hey-0.013
 year-0.012
Token return
Token ID1441
Feature activation+9.71e-04

Pos logit contributions
anks+0.052
 hyd+0.049
balls+0.049
uggets+0.048
ectar+0.048
Neg logit contributions
 saw-0.035
 heard-0.032
 little-0.030
ited-0.030
 did-0.030
Token to
Token ID284
Feature activation-2.53e-02

Pos logit contributions
ited+0.001
 saw+0.001
 heard+0.001
 two+0.001
 year+0.001
Neg logit contributions
anks-0.002
balls-0.002
 hyd-0.002
 rid-0.002
 heal-0.002
Token the
Token ID262
Feature activation-1.22e-02

Pos logit contributions
anks+0.060
balls+0.058
 hyd+0.058
 rid+0.057
 heal+0.057
Neg logit contributions
ited-0.027
 saw-0.023
 heard-0.022
Hey-0.018
 year-0.018
Token zoo
Token ID26626
Feature activation-2.50e-02

Pos logit contributions
balls+0.023
anks+0.023
 hyd+0.023
 rid+0.022
 heal+0.022
Neg logit contributions
ited-0.019
 saw-0.015
 heard-0.015
Hey-0.013
 year-0.013
Token and
Token ID290
Feature activation-6.84e-02

Pos logit contributions
anks+0.046
balls+0.045
 hyd+0.044
 rid+0.043
 heal+0.043
Neg logit contributions
ited-0.040
 saw-0.035
 heard-0.034
Hey-0.031
 two-0.030
Token see
Token ID766
Feature activation+1.30e-02

Pos logit contributions
anks+0.123
balls+0.119
 hyd+0.117
 rid+0.113
 heal+0.112
Neg logit contributions
ited-0.110
 saw-0.104
 heard-0.099
 two-0.089
 little-0.089
Token more
Token ID517
Feature activation-3.28e-02

Pos logit contributions
ited+0.018
 saw+0.015
 heard+0.014
Hey+0.013
 year+0.013
Neg logit contributions
anks-0.027
balls-0.026
 heal-0.026
 rid-0.026
 hyd-0.026
Token animals
Token ID4695
Feature activation-5.67e-02

Pos logit contributions
anks+0.065
balls+0.064
 hyd+0.063
 rid+0.061
 heal+0.061
Neg logit contributions
ited-0.047
 saw-0.042
 heard-0.040
Hey-0.036
 year-0.036
Token,
Token ID11
Feature activation-4.92e-02

Pos logit contributions
anks+0.113
balls+0.109
 hyd+0.107
 rid+0.105
 bleed+0.103
Neg logit contributions
ited-0.080
 saw-0.076
 heard-0.072
 two-0.063
 year-0.063
Token because
Token ID780
Feature activation-2.67e-02

Pos logit contributions
balls+0.089
anks+0.089
 hyd+0.088
 heal+0.086
 rid+0.085
Neg logit contributions
ited-0.080
 saw-0.065
 heard-0.062
Hey-0.061
 year-0.059
Token fire
Token ID2046
Feature activation-4.59e-02

Pos logit contributions
anks+0.018
balls+0.017
 hyd+0.017
 rid+0.017
 heal+0.017
Neg logit contributions
ited-0.015
 saw-0.013
 heard-0.013
Hey-0.011
 year-0.011
Token and
Token ID290
Feature activation-4.82e-02

Pos logit contributions
anks+0.099
balls+0.096
 hyd+0.094
 rid+0.094
 heal+0.092
Neg logit contributions
ited-0.057
 saw-0.052
 heard-0.049
 two-0.043
 year-0.043
Token she
Token ID673
Feature activation-2.33e-02

Pos logit contributions
 hyd+0.093
balls+0.091
 heal+0.091
 rid+0.090
anks+0.089
Neg logit contributions
ited-0.078
Hey-0.058
 year-0.056
 heard-0.054
 saw-0.053
Token quickly
Token ID2952
Feature activation-2.36e-03

Pos logit contributions
anks+0.056
balls+0.055
 hyd+0.055
 heal+0.054
 rid+0.054
Neg logit contributions
ited-0.024
 saw-0.017
 heard-0.016
 year-0.015
Hey-0.015
Token put
Token ID1234
Feature activation-2.64e-03

Pos logit contributions
 hyd+0.005
anks+0.005
 heal+0.005
 bleed+0.004
 rid+0.004
Neg logit contributions
ited-0.003
Hey-0.003
 year-0.003
 End-0.002
Next-0.002
Token it
Token ID340
Feature activation-5.87e-02

Pos logit contributions
anks+0.006
balls+0.006
 hyd+0.006
 heal+0.006
sing+0.006
Neg logit contributions
ited-0.003
 saw-0.003
 heard-0.003
 two-0.002
 little-0.002
Token out
Token ID503
Feature activation-2.18e-02

Pos logit contributions
anks+0.119
balls+0.113
 hyd+0.111
 rid+0.109
 heal+0.108
Neg logit contributions
ited-0.082
 saw-0.077
 heard-0.073
 two-0.065
 little-0.064
Token.
Token ID13
Feature activation-4.33e-02

Pos logit contributions
anks+0.037
 rid+0.037
balls+0.037
 hyd+0.037
 heal+0.036
Neg logit contributions
ited-0.037
 saw-0.032
 heard-0.030
Hey-0.030
 year-0.028
Token Tim
Token ID5045
Feature activation+2.07e-02

Pos logit contributions
anks+0.087
 hyd+0.085
 rid+0.084
balls+0.084
 heal+0.083
Neg logit contributions
 saw-0.056
ited-0.055
 heard-0.050
 two-0.048
 little-0.048
Tokenmy
Token ID1820
Feature activation+1.55e-02

Pos logit contributions
ited+0.036
Hey+0.034
Next+0.031
 minutes+0.031
ised+0.031
Neg logit contributions
gy-0.029
ophone-0.028
anks-0.028
fly-0.027
 hyd-0.027
Token felt
Token ID2936
Feature activation+2.20e-02

Pos logit contributions
ited+0.020
Hey+0.015
grown+0.014
 year+0.013
Hi+0.013
Neg logit contributions
 hyd-0.032
balls-0.032
 heal-0.031
anks-0.030
 rid-0.030
Token.
Token ID13
Feature activation-4.96e-02

Pos logit contributions
ited+0.019
 saw+0.017
 heard+0.017
 two+0.015
 year+0.015
Neg logit contributions
anks-0.022
balls-0.021
 hyd-0.021
 rid-0.021
 heal-0.020
Token They
Token ID1119
Feature activation-2.05e-02

Pos logit contributions
anks+0.082
balls+0.079
 hyd+0.077
 rid+0.076
 heal+0.075
Neg logit contributions
ited-0.087
 saw-0.083
 heard-0.079
 two-0.074
Hey-0.073
Token start
Token ID923
Feature activation+3.12e-03

Pos logit contributions
 heal+0.031
anks+0.031
 bleed+0.031
 hyd+0.031
 rid+0.030
Neg logit contributions
ited-0.037
 saw-0.031
 heard-0.031
 year-0.031
Hey-0.030
Token to
Token ID284
Feature activation-2.50e-02

Pos logit contributions
ited+0.005
 saw+0.004
 heard+0.004
Hey+0.004
 year+0.004
Neg logit contributions
anks-0.006
balls-0.006
 hyd-0.006
 rid-0.006
 heal-0.006
Token kick
Token ID4829
Feature activation+1.02e-02

Pos logit contributions
anks+0.036
balls+0.034
 hyd+0.034
 heal+0.033
 rid+0.033
Neg logit contributions
ited-0.050
 saw-0.046
 heard-0.044
Hey-0.042
 year-0.041
Token and
Token ID290
Feature activation-6.21e-02

Pos logit contributions
ited+0.014
 saw+0.012
 heard+0.011
 year+0.010
Hey+0.010
Neg logit contributions
anks-0.021
balls-0.021
 hyd-0.021
 rid-0.020
 heal-0.020
Token scream
Token ID8196
Feature activation+9.42e-03

Pos logit contributions
anks+0.112
ectar+0.104
 hyd+0.102
balls+0.101
ometers+0.100
Neg logit contributions
 saw-0.102
ited-0.102
 heard-0.098
 two-0.090
 did-0.089
Token.
Token ID13
Feature activation-4.71e-02

Pos logit contributions
ited+0.012
 saw+0.012
 heard+0.011
 little+0.010
 two+0.010
Neg logit contributions
anks-0.019
 hyd-0.018
balls-0.018
 rid-0.018
 bleed-0.018
Token They
Token ID1119
Feature activation-2.22e-02

Pos logit contributions
anks+0.083
balls+0.080
 hyd+0.079
 rid+0.078
 heal+0.077
Neg logit contributions
ited-0.076
 saw-0.073
 heard-0.070
 two-0.065
 little-0.064
Token say
Token ID910
Feature activation+4.92e-02

Pos logit contributions
anks+0.040
balls+0.038
 hyd+0.038
 rid+0.037
 heal+0.036
Neg logit contributions
ited-0.037
 saw-0.033
 heard-0.032
 year-0.029
Hey-0.029
Token,
Token ID11
Feature activation-6.37e-03

Pos logit contributions
ited+0.060
 saw+0.055
 heard+0.052
 two+0.045
 little+0.045
Neg logit contributions
anks-0.108
balls-0.105
 hyd-0.104
 rid-0.102
 heal-0.101