Token | flashlight |
Token ID | 38371 |
Feature activation | +8.45e-02 |
Pos logit contributions | |
most | +0.100 |
biggest | +0.100 |
fastest | +0.098 |
highest | +0.097 |
truly | +0.094 |
Neg logit contributions | |
Lee | -0.105 |
ens | -0.104 |
Sue | -0.099 |
handing | -0.099 |
Sophie | -0.099 |
Token | , |
Token ID | 11 |
Feature activation | -7.95e-02 |
Pos logit contributions | |
biggest | +0.162 |
most | +0.161 |
fastest | +0.156 |
highest | +0.156 |
best | +0.152 |
Neg logit contributions | |
Lee | -0.124 |
ens | -0.124 |
handing | -0.119 |
Sue | -0.116 |
Sophie | -0.113 |
Token | and |
Token ID | 290 |
Feature activation | -6.33e-02 |
Pos logit contributions | |
ens | +0.126 |
Lee | +0.123 |
riet | +0.119 |
kids | +0.117 |
handing | +0.116 |
Neg logit contributions | |
most | -0.145 |
biggest | -0.145 |
fastest | -0.141 |
highest | -0.140 |
best | -0.137 |
Token | a |
Token ID | 257 |
Feature activation | +5.37e-02 |
Pos logit contributions | |
ens | +0.117 |
Lee | +0.114 |
riet | +0.111 |
kids | +0.110 |
handing | +0.108 |
Neg logit contributions | |
most | -0.101 |
biggest | -0.100 |
fastest | -0.095 |
highest | -0.095 |
best | -0.094 |
Token | magn |
Token ID | 7842 |
Feature activation | +5.35e-02 |
Pos logit contributions | |
most | +0.090 |
biggest | +0.089 |
fastest | +0.088 |
highest | +0.087 |
truly | +0.084 |
Neg logit contributions | |
Lee | -0.089 |
ens | -0.088 |
handing | -0.084 |
Sue | -0.084 |
Sophie | -0.083 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.30e-01 |
Pos logit contributions | |
most | +0.049 |
biggest | +0.046 |
best | +0.043 |
fastest | +0.043 |
coolest | +0.042 |
Neg logit contributions | |
Lee | -0.133 |
riet | -0.133 |
ens | -0.131 |
lee | -0.127 |
Sophie | -0.126 |
Token | glass |
Token ID | 5405 |
Feature activation | +9.35e-02 |
Pos logit contributions | |
fastest | +0.460 |
best | +0.425 |
possible | +0.422 |
arest | +0.415 |
impossible | +0.407 |
Neg logit contributions | |
Sue | -0.360 |
Sophie | -0.337 |
Amy | -0.337 |
Tom | -0.331 |
Sally | -0.326 |
Token | . |
Token ID | 13 |
Feature activation | +1.90e-02 |
Pos logit contributions | |
biggest | +0.164 |
most | +0.163 |
fastest | +0.159 |
highest | +0.157 |
truly | +0.153 |
Neg logit contributions | |
ens | -0.155 |
Lee | -0.151 |
Sue | -0.146 |
Sophie | -0.145 |
handing | -0.144 |
Token |
|
Token ID | 198 |
Feature activation | +6.17e-02 |
Pos logit contributions | |
most | +0.038 |
biggest | +0.037 |
highest | +0.035 |
fastest | +0.035 |
best | +0.034 |
Neg logit contributions | |
ens | -0.027 |
pa | -0.027 |
kids | -0.026 |
riet | -0.025 |
olph | -0.025 |
Token |
|
Token ID | 198 |
Feature activation | +5.64e-02 |
Pos logit contributions | |
most | +0.079 |
biggest | +0.077 |
whole | +0.076 |
truly | +0.075 |
highest | +0.074 |
Neg logit contributions | |
ens | -0.130 |
Lee | -0.127 |
handing | -0.124 |
riet | -0.124 |
kids | -0.123 |
Token | As |
Token ID | 1722 |
Feature activation | +1.72e-02 |
Pos logit contributions | |
most | +0.092 |
biggest | +0.090 |
fastest | +0.089 |
whole | +0.088 |
highest | +0.088 |
Neg logit contributions | |
ens | -0.100 |
Lee | -0.098 |
kids | -0.094 |
riet | -0.094 |
handing | -0.093 |
Token | too |
Token ID | 1165 |
Feature activation | +9.73e-02 |
Pos logit contributions | |
ens | +0.081 |
Lee | +0.080 |
riet | +0.077 |
kids | +0.076 |
Sue | +0.076 |
Neg logit contributions | |
most | -0.078 |
biggest | -0.078 |
fastest | -0.075 |
highest | -0.075 |
best | -0.073 |
Token | close |
Token ID | 1969 |
Feature activation | -9.90e-03 |
Pos logit contributions | |
most | +0.146 |
biggest | +0.145 |
fastest | +0.138 |
highest | +0.137 |
truly | +0.135 |
Neg logit contributions | |
ens | -0.189 |
Lee | -0.183 |
Sue | -0.181 |
Sophie | -0.176 |
riet | -0.176 |
Token | . |
Token ID | 13 |
Feature activation | +3.59e-02 |
Pos logit contributions | |
ens | +0.015 |
Lee | +0.015 |
riet | +0.015 |
kids | +0.015 |
handing | +0.014 |
Neg logit contributions | |
most | -0.018 |
biggest | -0.018 |
fastest | -0.018 |
highest | -0.018 |
best | -0.017 |
Token | The |
Token ID | 383 |
Feature activation | +4.14e-02 |
Pos logit contributions | |
most | +0.054 |
biggest | +0.054 |
fastest | +0.052 |
highest | +0.052 |
best | +0.051 |
Neg logit contributions | |
ens | -0.068 |
Lee | -0.067 |
riet | -0.065 |
kids | -0.064 |
handing | -0.064 |
Token | ost |
Token ID | 23619 |
Feature activation | +2.15e-03 |
Pos logit contributions | |
ifying | +0.076 |
hest | +0.075 |
truly | +0.074 |
possible | +0.074 |
arest | +0.071 |
Neg logit contributions | |
Lee | -0.062 |
Sue | -0.061 |
ens | -0.059 |
Gig | -0.059 |
kids | -0.057 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.20e-01 |
Pos logit contributions | |
biggest | +0.003 |
most | +0.002 |
whole | +0.002 |
highest | +0.002 |
truly | +0.002 |
Neg logit contributions | |
Lee | -0.005 |
ens | -0.005 |
pa | -0.005 |
kids | -0.005 |
Sue | -0.005 |
Token | hes |
Token ID | 956 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
biggest | +0.383 |
most | +0.377 |
coolest | +0.373 |
fastest | +0.371 |
truly | +0.357 |
Neg logit contributions | |
Sue | -0.356 |
Tom | -0.345 |
ens | -0.345 |
Amy | -0.340 |
Lee | -0.333 |
Token | might |
Token ID | 1244 |
Feature activation | +9.52e-02 |
Pos logit contributions | |
biggest | +0.095 |
most | +0.090 |
highest | +0.089 |
fastest | +0.088 |
coolest | +0.087 |
Neg logit contributions | |
ens | -0.121 |
Sue | -0.113 |
Lee | -0.112 |
Tom | -0.108 |
hugged | -0.107 |
Token | pe |
Token ID | 613 |
Feature activation | -5.73e-03 |
Pos logit contributions | |
biggest | +0.186 |
most | +0.179 |
ifying | +0.178 |
coolest | +0.172 |
best | +0.169 |
Neg logit contributions | |
ens | -0.143 |
Sue | -0.139 |
her | -0.128 |
Tom | -0.126 |
begged | -0.126 |
Token | ck |
Token ID | 694 |
Feature activation | -9.08e-03 |
Pos logit contributions | |
ens | +0.006 |
ppa | +0.006 |
stro | +0.006 |
Lee | +0.006 |
pa | +0.006 |
Neg logit contributions | |
biggest | -0.013 |
most | -0.013 |
whole | -0.013 |
entire | -0.012 |
ifying | -0.012 |
Token | you |
Token ID | 345 |
Feature activation | +5.56e-02 |
Pos logit contributions | |
ens | +0.015 |
Lee | +0.014 |
riet | +0.013 |
Sue | +0.013 |
Sophie | +0.013 |
Neg logit contributions | |
biggest | -0.016 |
most | -0.016 |
fastest | -0.016 |
highest | -0.016 |
truly | -0.016 |
Token | and |
Token ID | 290 |
Feature activation | -9.73e-02 |
Pos logit contributions | |
riet | +0.091 |
kids | +0.089 |
ens | +0.085 |
annah | +0.081 |
Sophie | +0.080 |
Neg logit contributions | |
most | -0.093 |
best | -0.092 |
whole | -0.089 |
truly | -0.088 |
biggest | -0.088 |
Token | looked |
Token ID | 3114 |
Feature activation | +5.32e-02 |
Pos logit contributions | |
kids | +0.178 |
riet | +0.171 |
handing | +0.163 |
child | +0.162 |
ens | +0.161 |
Neg logit contributions | |
most | -0.155 |
truly | -0.149 |
biggest | -0.141 |
highest | -0.139 |
best | -0.139 |
Token | at |
Token ID | 379 |
Feature activation | -4.63e-02 |
Pos logit contributions | |
most | +0.087 |
biggest | +0.086 |
best | +0.083 |
highest | +0.082 |
fastest | +0.081 |
Neg logit contributions | |
Lee | -0.096 |
kids | -0.096 |
ens | -0.095 |
riet | -0.095 |
olph | -0.091 |
Token | the |
Token ID | 262 |
Feature activation | +2.78e-02 |
Pos logit contributions | |
ens | +0.061 |
Lee | +0.060 |
riet | +0.060 |
kids | +0.059 |
handing | +0.059 |
Neg logit contributions | |
most | -0.097 |
biggest | -0.096 |
best | -0.092 |
highest | -0.092 |
fastest | -0.092 |
Token | ost |
Token ID | 23619 |
Feature activation | -4.24e-03 |
Pos logit contributions | |
ifying | +0.057 |
hest | +0.054 |
possible | +0.054 |
fastest | +0.053 |
coolest | +0.053 |
Neg logit contributions | |
Lee | -0.034 |
kids | -0.034 |
Sue | -0.034 |
Gig | -0.033 |
begged | -0.033 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.19e-01 |
Pos logit contributions | |
Lee | +0.010 |
ens | +0.010 |
kids | +0.010 |
pa | +0.010 |
Sue | +0.010 |
Neg logit contributions | |
biggest | -0.004 |
most | -0.004 |
fastest | -0.004 |
highest | -0.004 |
whole | -0.004 |
Token | hes |
Token ID | 956 |
Feature activation | +6.08e-02 |
Pos logit contributions | |
biggest | +0.372 |
most | +0.372 |
fastest | +0.369 |
coolest | +0.365 |
highest | +0.349 |
Neg logit contributions | |
Sue | -0.361 |
ens | -0.353 |
Tom | -0.352 |
Amy | -0.346 |
Lee | -0.345 |
Token | . |
Token ID | 13 |
Feature activation | +6.67e-03 |
Pos logit contributions | |
biggest | +0.108 |
most | +0.107 |
fastest | +0.105 |
highest | +0.104 |
truly | +0.100 |
Neg logit contributions | |
ens | -0.101 |
Lee | -0.099 |
Sue | -0.094 |
handing | -0.093 |
Sophie | -0.093 |
Token | They |
Token ID | 1119 |
Feature activation | +2.48e-02 |
Pos logit contributions | |
most | +0.009 |
biggest | +0.009 |
highest | +0.009 |
best | +0.008 |
whole | +0.008 |
Neg logit contributions | |
ens | -0.013 |
pa | -0.013 |
riet | -0.013 |
ppa | -0.013 |
kids | -0.013 |
Token | saw |
Token ID | 2497 |
Feature activation | -3.91e-02 |
Pos logit contributions | |
most | +0.046 |
truly | +0.044 |
biggest | +0.043 |
best | +0.043 |
whole | +0.042 |
Neg logit contributions | |
riet | -0.039 |
kids | -0.039 |
Lee | -0.038 |
ens | -0.038 |
handing | -0.038 |
Token | one |
Token ID | 530 |
Feature activation | +8.81e-02 |
Pos logit contributions | |
ens | +0.075 |
riet | +0.074 |
Lee | +0.073 |
kids | +0.073 |
handing | +0.071 |
Neg logit contributions | |
most | -0.060 |
biggest | -0.060 |
best | -0.056 |
highest | -0.056 |
fastest | -0.055 |
Token | cried |
Token ID | 16896 |
Feature activation | -2.40e-02 |
Pos logit contributions | |
riet | +0.081 |
kids | +0.081 |
Lee | +0.077 |
purs | +0.074 |
ens | +0.074 |
Neg logit contributions | |
most | -0.192 |
biggest | -0.188 |
best | -0.186 |
highest | -0.184 |
fastest | -0.183 |
Token | . |
Token ID | 13 |
Feature activation | +6.53e-03 |
Pos logit contributions | |
ens | +0.044 |
Lee | +0.043 |
riet | +0.042 |
kids | +0.042 |
Sophie | +0.041 |
Neg logit contributions | |
most | -0.038 |
biggest | -0.038 |
fastest | -0.036 |
highest | -0.036 |
best | -0.036 |
Token | " |
Token ID | 366 |
Feature activation | +4.80e-02 |
Pos logit contributions | |
most | +0.011 |
biggest | +0.011 |
fastest | +0.011 |
highest | +0.011 |
best | +0.010 |
Neg logit contributions | |
ens | -0.011 |
Lee | -0.011 |
riet | -0.011 |
kids | -0.011 |
handing | -0.011 |
Token | The |
Token ID | 464 |
Feature activation | +5.67e-02 |
Pos logit contributions | |
biggest | +0.072 |
best | +0.070 |
fastest | +0.069 |
impossible | +0.068 |
possible | +0.068 |
Neg logit contributions | |
ens | -0.096 |
olph | -0.091 |
sob | -0.090 |
stro | -0.087 |
Tom | -0.087 |
Token | ost |
Token ID | 23619 |
Feature activation | +1.16e-02 |
Pos logit contributions | |
hest | +0.105 |
truly | +0.104 |
ifying | +0.104 |
fastest | +0.100 |
biggest | +0.099 |
Neg logit contributions | |
ens | -0.084 |
Lee | -0.083 |
Sue | -0.075 |
purs | -0.075 |
babys | -0.075 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.18e-01 |
Pos logit contributions | |
biggest | +0.013 |
most | +0.013 |
highest | +0.012 |
whole | +0.012 |
fastest | +0.011 |
Neg logit contributions | |
ens | -0.028 |
Lee | -0.027 |
pa | -0.026 |
olph | -0.026 |
Sue | -0.026 |
Token | hes |
Token ID | 956 |
Feature activation | +6.91e-02 |
Pos logit contributions | |
biggest | +0.416 |
fastest | +0.415 |
most | +0.408 |
coolest | +0.408 |
highest | +0.399 |
Neg logit contributions | |
Tom | -0.310 |
Sue | -0.307 |
ens | -0.304 |
hugged | -0.287 |
Amy | -0.286 |
Token | are |
Token ID | 389 |
Feature activation | +9.09e-02 |
Pos logit contributions | |
biggest | +0.102 |
most | +0.098 |
fastest | +0.098 |
highest | +0.096 |
truly | +0.096 |
Neg logit contributions | |
ens | -0.140 |
Sue | -0.133 |
Lee | -0.126 |
Tom | -0.126 |
her | -0.125 |
Token | mean |
Token ID | 1612 |
Feature activation | -6.18e-03 |
Pos logit contributions | |
biggest | +0.145 |
most | +0.140 |
fastest | +0.140 |
highest | +0.139 |
world | +0.137 |
Neg logit contributions | |
ens | -0.162 |
Sue | -0.158 |
handing | -0.154 |
hugged | -0.151 |
Tom | -0.151 |
Token | ! |
Token ID | 0 |
Feature activation | +4.48e-02 |
Pos logit contributions | |
ens | +0.011 |
Lee | +0.010 |
Sue | +0.010 |
riet | +0.010 |
kids | +0.010 |
Neg logit contributions | |
most | -0.011 |
biggest | -0.011 |
fastest | -0.010 |
highest | -0.010 |
truly | -0.010 |
Token | They |
Token ID | 1119 |
Feature activation | +4.79e-02 |
Pos logit contributions | |
most | +0.065 |
biggest | +0.065 |
fastest | +0.063 |
highest | +0.063 |
best | +0.061 |
Neg logit contributions | |
Lee | -0.089 |
ens | -0.088 |
Sue | -0.085 |
handing | -0.083 |
riet | -0.083 |
Token | and |
Token ID | 290 |
Feature activation | -8.59e-02 |
Pos logit contributions | |
most | +0.067 |
biggest | +0.066 |
fastest | +0.065 |
highest | +0.064 |
best | +0.063 |
Neg logit contributions | |
ens | -0.043 |
Lee | -0.042 |
riet | -0.040 |
kids | -0.039 |
handing | -0.039 |
Token | run |
Token ID | 1057 |
Feature activation | -5.99e-02 |
Pos logit contributions | |
ens | +0.132 |
kids | +0.128 |
Lee | +0.128 |
riet | +0.128 |
Sophie | +0.123 |
Neg logit contributions | |
most | -0.160 |
biggest | -0.159 |
highest | -0.157 |
fastest | -0.155 |
best | -0.150 |
Token | to |
Token ID | 284 |
Feature activation | -9.48e-02 |
Pos logit contributions | |
ens | +0.091 |
Lee | +0.089 |
Sue | +0.085 |
riet | +0.083 |
handing | +0.082 |
Neg logit contributions | |
most | -0.115 |
biggest | -0.115 |
fastest | -0.111 |
highest | -0.109 |
truly | -0.108 |
Token | the |
Token ID | 262 |
Feature activation | +1.21e-02 |
Pos logit contributions | |
riet | +0.151 |
ens | +0.150 |
kids | +0.146 |
ppa | +0.142 |
olph | +0.140 |
Neg logit contributions | |
biggest | -0.178 |
most | -0.178 |
highest | -0.173 |
whole | -0.170 |
best | -0.169 |
Token | ost |
Token ID | 23619 |
Feature activation | -4.39e-03 |
Pos logit contributions | |
most | +0.020 |
biggest | +0.020 |
fastest | +0.020 |
truly | +0.019 |
highest | +0.019 |
Neg logit contributions | |
ens | -0.021 |
Lee | -0.021 |
Sue | -0.019 |
kids | -0.019 |
purs | -0.019 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.17e-01 |
Pos logit contributions | |
Lee | +0.011 |
ens | +0.011 |
kids | +0.010 |
Sue | +0.010 |
pa | +0.010 |
Neg logit contributions | |
biggest | -0.004 |
most | -0.004 |
fastest | -0.004 |
highest | -0.004 |
whole | -0.004 |
Token | hes |
Token ID | 956 |
Feature activation | +5.83e-02 |
Pos logit contributions | |
most | +0.367 |
biggest | +0.365 |
fastest | +0.358 |
coolest | +0.355 |
truly | +0.340 |
Neg logit contributions | |
Sue | -0.371 |
Tom | -0.360 |
ens | -0.359 |
Amy | -0.351 |
Lee | -0.351 |
Token | . |
Token ID | 13 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
biggest | +0.104 |
most | +0.104 |
fastest | +0.102 |
highest | +0.099 |
truly | +0.099 |
Neg logit contributions | |
ens | -0.095 |
Lee | -0.095 |
Sue | -0.091 |
handing | -0.088 |
Sophie | -0.088 |
Token |
|
Token ID | 198 |
Feature activation | +4.94e-02 |
Pos logit contributions | |
most | +0.019 |
biggest | +0.018 |
highest | +0.018 |
fastest | +0.017 |
best | +0.017 |
Neg logit contributions | |
ens | -0.024 |
riet | -0.024 |
kids | -0.023 |
pa | -0.023 |
ppa | -0.023 |
Token |
|
Token ID | 198 |
Feature activation | +4.24e-02 |
Pos logit contributions | |
most | +0.069 |
biggest | +0.068 |
whole | +0.065 |
highest | +0.065 |
fastest | +0.065 |
Neg logit contributions | |
ens | -0.099 |
Lee | -0.097 |
riet | -0.096 |
kids | -0.094 |
handing | -0.094 |
Token | " |
Token ID | 1 |
Feature activation | -4.08e-03 |
Pos logit contributions | |
biggest | +0.069 |
most | +0.069 |
fastest | +0.066 |
highest | +0.066 |
best | +0.065 |
Neg logit contributions | |
ens | -0.076 |
Lee | -0.075 |
riet | -0.072 |
Sue | -0.071 |
kids | -0.071 |
Token | the |
Token ID | 262 |
Feature activation | +2.66e-02 |
Pos logit contributions | |
kids | +0.013 |
riet | +0.013 |
ior | +0.013 |
ens | +0.012 |
olph | +0.012 |
Neg logit contributions | |
most | -0.020 |
whole | -0.019 |
biggest | -0.019 |
best | -0.018 |
highest | -0.017 |
Token | necklace |
Token ID | 38987 |
Feature activation | +3.28e-02 |
Pos logit contributions | |
biggest | +0.046 |
most | +0.046 |
fastest | +0.046 |
highest | +0.045 |
truly | +0.045 |
Neg logit contributions | |
ens | -0.043 |
Lee | -0.042 |
Sue | -0.041 |
purs | -0.040 |
handing | -0.040 |
Token | and |
Token ID | 290 |
Feature activation | -8.30e-02 |
Pos logit contributions | |
most | +0.065 |
biggest | +0.064 |
highest | +0.063 |
best | +0.062 |
fastest | +0.062 |
Neg logit contributions | |
ens | -0.046 |
Lee | -0.045 |
kids | -0.045 |
riet | -0.045 |
olph | -0.043 |
Token | the |
Token ID | 262 |
Feature activation | +2.58e-02 |
Pos logit contributions | |
kids | +0.140 |
riet | +0.133 |
olph | +0.130 |
Lee | +0.129 |
child | +0.128 |
Neg logit contributions | |
most | -0.151 |
biggest | -0.139 |
best | -0.134 |
perfect | -0.133 |
highest | -0.133 |
Token | magn |
Token ID | 7842 |
Feature activation | +3.60e-02 |
Pos logit contributions | |
most | +0.046 |
biggest | +0.046 |
fastest | +0.045 |
highest | +0.044 |
best | +0.043 |
Neg logit contributions | |
ens | -0.041 |
Lee | -0.041 |
riet | -0.039 |
Sue | -0.039 |
kids | -0.038 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.15e-01 |
Pos logit contributions | |
most | +0.041 |
biggest | +0.039 |
fastest | +0.038 |
coolest | +0.037 |
best | +0.037 |
Neg logit contributions | |
riet | -0.083 |
Lee | -0.082 |
ens | -0.082 |
stro | -0.082 |
lee | -0.081 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.27e-02 |
Pos logit contributions | |
fastest | +0.429 |
best | +0.397 |
possible | +0.389 |
most | +0.388 |
biggest | +0.387 |
Neg logit contributions | |
Sue | -0.339 |
Sophie | -0.313 |
Mrs | -0.313 |
Tom | -0.310 |
Amy | -0.310 |
Token | and |
Token ID | 290 |
Feature activation | -7.18e-02 |
Pos logit contributions | |
most | +0.135 |
biggest | +0.135 |
fastest | +0.130 |
highest | +0.130 |
best | +0.127 |
Neg logit contributions | |
ens | -0.147 |
Lee | -0.144 |
riet | -0.140 |
kids | -0.138 |
handing | -0.137 |
Token | started |
Token ID | 2067 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
kids | +0.118 |
riet | +0.114 |
Lee | +0.114 |
ens | +0.113 |
olph | +0.112 |
Neg logit contributions | |
most | -0.134 |
biggest | -0.126 |
best | -0.120 |
highest | -0.119 |
whole | -0.119 |
Token | to |
Token ID | 284 |
Feature activation | -6.79e-02 |
Pos logit contributions | |
most | +0.020 |
biggest | +0.020 |
whole | +0.020 |
highest | +0.019 |
best | +0.019 |
Neg logit contributions | |
ens | -0.026 |
riet | -0.026 |
kids | -0.026 |
pa | -0.025 |
Lee | -0.024 |
Token | examine |
Token ID | 10716 |
Feature activation | -3.07e-02 |
Pos logit contributions | |
Lee | +0.119 |
riet | +0.118 |
kids | +0.117 |
ens | +0.117 |
olph | +0.111 |
Neg logit contributions | |
most | -0.115 |
biggest | -0.113 |
highest | -0.112 |
whole | -0.112 |
fastest | -0.108 |
Token | . |
Token ID | 13 |
Feature activation | +1.56e-02 |
Pos logit contributions | |
biggest | +0.083 |
most | +0.082 |
fastest | +0.079 |
highest | +0.079 |
coolest | +0.076 |
Neg logit contributions | |
ens | -0.069 |
Lee | -0.068 |
handing | -0.065 |
Sue | -0.065 |
stro | -0.065 |
Token | They |
Token ID | 1119 |
Feature activation | +3.12e-02 |
Pos logit contributions | |
most | +0.026 |
biggest | +0.024 |
highest | +0.023 |
ifying | +0.023 |
best | +0.023 |
Neg logit contributions | |
pa | -0.028 |
ens | -0.028 |
ppa | -0.027 |
kids | -0.027 |
riet | -0.026 |
Token | were |
Token ID | 547 |
Feature activation | +6.82e-02 |
Pos logit contributions | |
most | +0.055 |
truly | +0.053 |
biggest | +0.053 |
fastest | +0.053 |
highest | +0.052 |
Neg logit contributions | |
Lee | -0.051 |
kids | -0.050 |
ens | -0.050 |
riet | -0.050 |
handing | -0.050 |
Token | called |
Token ID | 1444 |
Feature activation | -5.18e-02 |
Pos logit contributions | |
most | +0.100 |
biggest | +0.099 |
fastest | +0.096 |
highest | +0.094 |
best | +0.094 |
Neg logit contributions | |
ens | -0.136 |
Lee | -0.132 |
riet | -0.130 |
kids | -0.128 |
handing | -0.126 |
Token | ost |
Token ID | 23619 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
ens | +0.065 |
riet | +0.062 |
kids | +0.061 |
Lee | +0.060 |
handing | +0.059 |
Neg logit contributions | |
most | -0.113 |
biggest | -0.113 |
highest | -0.109 |
best | -0.109 |
fastest | -0.108 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.15e-01 |
Pos logit contributions | |
ens | +0.030 |
Lee | +0.029 |
kids | +0.028 |
riet | +0.028 |
pa | +0.028 |
Neg logit contributions | |
biggest | -0.012 |
most | -0.012 |
fastest | -0.011 |
highest | -0.011 |
whole | -0.011 |
Token | hes |
Token ID | 956 |
Feature activation | +5.81e-02 |
Pos logit contributions | |
biggest | +0.356 |
most | +0.351 |
coolest | +0.344 |
fastest | +0.342 |
virtue | +0.334 |
Neg logit contributions | |
Sue | -0.356 |
Tom | -0.346 |
ens | -0.344 |
Amy | -0.343 |
Lee | -0.341 |
Token | . |
Token ID | 13 |
Feature activation | +2.00e-02 |
Pos logit contributions | |
biggest | +0.110 |
most | +0.109 |
fastest | +0.107 |
highest | +0.106 |
whole | +0.103 |
Neg logit contributions | |
ens | -0.089 |
Lee | -0.087 |
Sue | -0.082 |
riet | -0.081 |
kids | -0.081 |
Token | Lily |
Token ID | 20037 |
Feature activation | -6.80e-02 |
Pos logit contributions | |
most | +0.034 |
biggest | +0.033 |
highest | +0.031 |
fastest | +0.031 |
best | +0.031 |
Neg logit contributions | |
ens | -0.034 |
pa | -0.033 |
riet | -0.033 |
kids | -0.032 |
ppa | -0.032 |
Token | and |
Token ID | 290 |
Feature activation | -8.58e-02 |
Pos logit contributions | |
ens | +0.124 |
riet | +0.123 |
Lee | +0.122 |
handing | +0.119 |
kids | +0.118 |
Neg logit contributions | |
most | -0.107 |
truly | -0.101 |
fastest | -0.099 |
whole | -0.099 |
best | -0.098 |
Token | Ben |
Token ID | 3932 |
Feature activation | -9.11e-02 |
Pos logit contributions | |
handing | +0.073 |
kids | +0.070 |
riet | +0.069 |
ppa | +0.067 |
old | +0.063 |
Neg logit contributions | |
most | -0.216 |
best | -0.207 |
biggest | -0.204 |
fastest | -0.198 |
whole | -0.196 |
Token | a |
Token ID | 257 |
Feature activation | +4.74e-02 |
Pos logit contributions | |
biggest | +0.094 |
most | +0.093 |
fastest | +0.092 |
highest | +0.091 |
whole | +0.086 |
Neg logit contributions | |
ens | -0.151 |
Lee | -0.149 |
Sue | -0.146 |
kids | -0.144 |
Sophie | -0.143 |
Token | flashlight |
Token ID | 38371 |
Feature activation | +6.43e-02 |
Pos logit contributions | |
most | +0.059 |
biggest | +0.058 |
fastest | +0.057 |
highest | +0.056 |
truly | +0.053 |
Neg logit contributions | |
ens | -0.100 |
Lee | -0.100 |
Sue | -0.097 |
Sophie | -0.095 |
kids | -0.095 |
Token | and |
Token ID | 290 |
Feature activation | -7.69e-02 |
Pos logit contributions | |
biggest | +0.134 |
most | +0.133 |
fastest | +0.130 |
highest | +0.129 |
whole | +0.127 |
Neg logit contributions | |
Lee | -0.085 |
ens | -0.084 |
handing | -0.079 |
Sue | -0.078 |
riet | -0.076 |
Token | a |
Token ID | 257 |
Feature activation | +4.24e-02 |
Pos logit contributions | |
ens | +0.163 |
Lee | +0.161 |
riet | +0.158 |
kids | +0.155 |
handing | +0.155 |
Neg logit contributions | |
most | -0.101 |
biggest | -0.098 |
fastest | -0.092 |
highest | -0.092 |
best | -0.092 |
Token | magn |
Token ID | 7842 |
Feature activation | +4.84e-02 |
Pos logit contributions | |
most | +0.069 |
biggest | +0.069 |
fastest | +0.068 |
highest | +0.067 |
truly | +0.065 |
Neg logit contributions | |
ens | -0.073 |
Lee | -0.073 |
Sue | -0.069 |
handing | -0.068 |
riet | -0.068 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
most | +0.041 |
biggest | +0.038 |
best | +0.035 |
coolest | +0.035 |
fastest | +0.034 |
Neg logit contributions | |
Lee | -0.124 |
ens | -0.123 |
riet | -0.122 |
lee | -0.118 |
olph | -0.117 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.69e-02 |
Pos logit contributions | |
fastest | +0.431 |
possible | +0.420 |
arest | +0.417 |
best | +0.394 |
truly | +0.391 |
Neg logit contributions | |
Sue | -0.319 |
Amy | -0.299 |
Tom | -0.281 |
Mrs | -0.281 |
Mia | -0.280 |
Token | to |
Token ID | 284 |
Feature activation | -7.32e-02 |
Pos logit contributions | |
most | +0.154 |
biggest | +0.154 |
fastest | +0.150 |
highest | +0.148 |
truly | +0.146 |
Neg logit contributions | |
ens | -0.143 |
Lee | -0.140 |
Sue | -0.134 |
handing | -0.133 |
Sophie | -0.132 |
Token | examine |
Token ID | 10716 |
Feature activation | -3.08e-02 |
Pos logit contributions | |
Lee | +0.121 |
ens | +0.121 |
riet | +0.117 |
kids | +0.115 |
handing | +0.114 |
Neg logit contributions | |
most | -0.129 |
biggest | -0.129 |
highest | -0.124 |
fastest | -0.123 |
whole | -0.122 |
Token | their |
Token ID | 511 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
ens | +0.050 |
Lee | +0.049 |
riet | +0.047 |
Sue | +0.047 |
Sophie | +0.046 |
Neg logit contributions | |
biggest | -0.055 |
most | -0.055 |
fastest | -0.053 |
highest | -0.053 |
whole | -0.052 |
Token | finds |
Token ID | 7228 |
Feature activation | +8.74e-03 |
Pos logit contributions | |
ens | +0.036 |
Lee | +0.035 |
riet | +0.034 |
kids | +0.034 |
Sue | +0.034 |
Neg logit contributions | |
most | -0.034 |
biggest | -0.034 |
fastest | -0.033 |
highest | -0.033 |
whole | -0.032 |
Token | should |
Token ID | 815 |
Feature activation | -1.04e-03 |
Pos logit contributions | |
biggest | +0.149 |
highest | +0.146 |
fastest | +0.144 |
possible | +0.143 |
most | +0.141 |
Neg logit contributions | |
ens | -0.168 |
hugged | -0.163 |
begged | -0.156 |
stro | -0.155 |
Sue | -0.153 |
Token | not |
Token ID | 407 |
Feature activation | +2.98e-02 |
Pos logit contributions | |
ens | +0.002 |
Lee | +0.002 |
riet | +0.002 |
Sue | +0.002 |
kids | +0.002 |
Neg logit contributions | |
biggest | -0.002 |
most | -0.002 |
fastest | -0.002 |
highest | -0.002 |
whole | -0.002 |
Token | tease |
Token ID | 34691 |
Feature activation | -3.74e-02 |
Pos logit contributions | |
biggest | +0.062 |
most | +0.062 |
fastest | +0.059 |
highest | +0.059 |
best | +0.059 |
Neg logit contributions | |
ens | -0.041 |
Lee | -0.039 |
Sue | -0.038 |
handing | -0.037 |
riet | -0.037 |
Token | the |
Token ID | 262 |
Feature activation | +2.31e-02 |
Pos logit contributions | |
ens | +0.058 |
Lee | +0.057 |
kids | +0.057 |
riet | +0.057 |
handing | +0.055 |
Neg logit contributions | |
most | -0.069 |
biggest | -0.069 |
fastest | -0.067 |
highest | -0.066 |
best | -0.066 |
Token | ost |
Token ID | 23619 |
Feature activation | +5.96e-03 |
Pos logit contributions | |
ifying | +0.051 |
truly | +0.048 |
hest | +0.048 |
possible | +0.048 |
fastest | +0.047 |
Neg logit contributions | |
Lee | -0.027 |
Sue | -0.026 |
ens | -0.026 |
kids | -0.025 |
purs | -0.025 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
biggest | +0.006 |
most | +0.006 |
highest | +0.006 |
fastest | +0.005 |
whole | +0.005 |
Neg logit contributions | |
ens | -0.014 |
Lee | -0.014 |
kids | -0.014 |
pa | -0.014 |
Sue | -0.014 |
Token | hes |
Token ID | 956 |
Feature activation | +7.15e-02 |
Pos logit contributions | |
biggest | +0.350 |
most | +0.349 |
fastest | +0.336 |
coolest | +0.334 |
truly | +0.328 |
Neg logit contributions | |
ens | -0.376 |
Sue | -0.366 |
Tom | -0.357 |
Lee | -0.350 |
Amy | -0.349 |
Token | . |
Token ID | 13 |
Feature activation | +2.86e-02 |
Pos logit contributions | |
biggest | +0.129 |
most | +0.127 |
fastest | +0.124 |
highest | +0.124 |
truly | +0.122 |
Neg logit contributions | |
ens | -0.118 |
Lee | -0.114 |
Sue | -0.109 |
Sophie | -0.106 |
handing | -0.105 |
Token | They |
Token ID | 1119 |
Feature activation | +4.58e-02 |
Pos logit contributions | |
biggest | +0.048 |
most | +0.048 |
highest | +0.047 |
fastest | +0.047 |
whole | +0.047 |
Neg logit contributions | |
Lee | -0.050 |
ens | -0.049 |
Sue | -0.047 |
riet | -0.046 |
Sophie | -0.046 |
Token | are |
Token ID | 389 |
Feature activation | +8.20e-02 |
Pos logit contributions | |
biggest | +0.072 |
possible | +0.069 |
highest | +0.069 |
entire | +0.067 |
ifying | +0.066 |
Neg logit contributions | |
ens | -0.090 |
hugged | -0.083 |
begged | -0.080 |
stro | -0.079 |
hugs | -0.078 |
Token | not |
Token ID | 407 |
Feature activation | +4.99e-02 |
Pos logit contributions | |
biggest | +0.125 |
highest | +0.122 |
most | +0.121 |
fastest | +0.119 |
whole | +0.118 |
Neg logit contributions | |
ens | -0.151 |
Sue | -0.148 |
Lee | -0.145 |
handing | -0.144 |
Tom | -0.141 |
Token | that |
Token ID | 326 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
most | +0.048 |
biggest | +0.047 |
fastest | +0.046 |
highest | +0.046 |
best | +0.045 |
Neg logit contributions | |
ens | -0.052 |
Lee | -0.051 |
riet | -0.049 |
kids | -0.049 |
handing | -0.048 |
Token | said |
Token ID | 531 |
Feature activation | -3.72e-02 |
Pos logit contributions | |
biggest | +0.058 |
most | +0.058 |
fastest | +0.057 |
highest | +0.057 |
whole | +0.056 |
Neg logit contributions | |
ens | -0.018 |
Lee | -0.018 |
Sue | -0.016 |
riet | -0.016 |
Sophie | -0.016 |
Token | " |
Token ID | 366 |
Feature activation | +6.66e-02 |
Pos logit contributions | |
Lee | +0.050 |
ens | +0.048 |
Sue | +0.047 |
Sophie | +0.046 |
handing | +0.045 |
Neg logit contributions | |
most | -0.077 |
biggest | -0.077 |
fastest | -0.076 |
highest | -0.075 |
whole | -0.074 |
Token | O |
Token ID | 46 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
biggest | +0.119 |
most | +0.116 |
best | +0.113 |
fastest | +0.111 |
highest | +0.111 |
Neg logit contributions | |
ens | -0.111 |
Lee | -0.109 |
kids | -0.104 |
Sue | -0.103 |
riet | -0.102 |
Token | st |
Token ID | 301 |
Feature activation | +7.49e-02 |
Pos logit contributions | |
ens | +0.025 |
Lee | +0.023 |
kids | +0.022 |
Sue | +0.022 |
pa | +0.022 |
Neg logit contributions | |
most | -0.030 |
biggest | -0.030 |
coolest | -0.029 |
ifying | -0.029 |
highest | -0.028 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
most | +0.129 |
coolest | +0.124 |
biggest | +0.123 |
truly | +0.122 |
fastest | +0.121 |
Neg logit contributions | |
Lee | -0.128 |
Sue | -0.122 |
ens | -0.121 |
kids | -0.115 |
Tom | -0.114 |
Token | hes |
Token ID | 956 |
Feature activation | +7.50e-02 |
Pos logit contributions | |
most | +0.344 |
biggest | +0.342 |
coolest | +0.339 |
fastest | +0.337 |
highest | +0.322 |
Neg logit contributions | |
Sue | -0.357 |
Lee | -0.354 |
Tom | -0.351 |
ens | -0.347 |
Amy | -0.346 |
Token | ". |
Token ID | 1911 |
Feature activation | -1.79e-02 |
Pos logit contributions | |
biggest | +0.109 |
fastest | +0.108 |
highest | +0.107 |
most | +0.106 |
coolest | +0.105 |
Neg logit contributions | |
Lee | -0.144 |
ens | -0.144 |
Sue | -0.141 |
Tom | -0.136 |
Sophie | -0.133 |
Token | Sara |
Token ID | 24799 |
Feature activation | -1.12e-01 |
Pos logit contributions | |
ens | +0.015 |
riet | +0.013 |
pa | +0.013 |
ppa | +0.013 |
kids | +0.012 |
Neg logit contributions | |
most | -0.047 |
biggest | -0.047 |
highest | -0.045 |
best | -0.045 |
whole | -0.045 |
Token | and |
Token ID | 290 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
ens | +0.179 |
riet | +0.174 |
Lee | +0.170 |
Sophie | +0.168 |
kids | +0.166 |
Neg logit contributions | |
most | -0.205 |
biggest | -0.193 |
whole | -0.193 |
truly | -0.192 |
best | -0.191 |
Token | Tom |
Token ID | 4186 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
ppa | +0.114 |
stro | +0.105 |
riet | +0.105 |
handing | +0.103 |
olph | +0.101 |
Neg logit contributions | |
best | -0.235 |
most | -0.231 |
biggest | -0.224 |
whole | -0.213 |
perfect | -0.207 |
Token | wanted |
Token ID | 2227 |
Feature activation | +5.36e-03 |
Pos logit contributions | |
most | +0.031 |
truly | +0.030 |
biggest | +0.030 |
best | +0.029 |
whole | +0.029 |
Neg logit contributions | |
Lee | -0.030 |
riet | -0.029 |
kids | -0.028 |
handing | -0.028 |
ens | -0.028 |
Token | to |
Token ID | 284 |
Feature activation | -7.85e-02 |
Pos logit contributions | |
most | +0.008 |
best | +0.008 |
biggest | +0.008 |
whole | +0.007 |
highest | +0.007 |
Neg logit contributions | |
riet | -0.011 |
kids | -0.010 |
annah | -0.010 |
ens | -0.010 |
ppa | -0.010 |
Token | touch |
Token ID | 3638 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
ens | +0.123 |
Lee | +0.121 |
riet | +0.118 |
kids | +0.116 |
handing | +0.114 |
Neg logit contributions | |
most | -0.146 |
biggest | -0.146 |
fastest | -0.141 |
highest | -0.141 |
whole | -0.139 |
Token | the |
Token ID | 262 |
Feature activation | +2.51e-02 |
Pos logit contributions | |
ens | +0.024 |
Lee | +0.023 |
riet | +0.022 |
kids | +0.022 |
handing | +0.022 |
Neg logit contributions | |
most | -0.038 |
biggest | -0.038 |
fastest | -0.037 |
highest | -0.037 |
best | -0.036 |
Token | ost |
Token ID | 23619 |
Feature activation | -4.40e-03 |
Pos logit contributions | |
truly | +0.050 |
ifying | +0.050 |
possible | +0.050 |
fastest | +0.049 |
most | +0.049 |
Neg logit contributions | |
ens | -0.033 |
Lee | -0.031 |
kids | -0.031 |
purs | -0.030 |
Gig | -0.030 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
ens | +0.011 |
Lee | +0.011 |
kids | +0.010 |
Sue | +0.010 |
riet | +0.010 |
Neg logit contributions | |
biggest | -0.004 |
most | -0.004 |
fastest | -0.004 |
highest | -0.004 |
whole | -0.004 |
Token | hes |
Token ID | 956 |
Feature activation | +5.77e-02 |
Pos logit contributions | |
most | +0.324 |
biggest | +0.323 |
fastest | +0.314 |
highest | +0.304 |
coolest | +0.300 |
Neg logit contributions | |
ens | -0.406 |
Sue | -0.395 |
Lee | -0.389 |
Amy | -0.381 |
Tom | -0.378 |
Token | . |
Token ID | 13 |
Feature activation | +1.10e-02 |
Pos logit contributions | |
biggest | +0.103 |
most | +0.103 |
fastest | +0.100 |
highest | +0.099 |
whole | +0.097 |
Neg logit contributions | |
ens | -0.094 |
Lee | -0.092 |
Sue | -0.087 |
riet | -0.087 |
Sophie | -0.087 |
Token | They |
Token ID | 1119 |
Feature activation | +2.16e-02 |
Pos logit contributions | |
most | +0.018 |
biggest | +0.018 |
fastest | +0.018 |
highest | +0.018 |
whole | +0.017 |
Neg logit contributions | |
ens | -0.019 |
Lee | -0.019 |
riet | -0.018 |
Sue | -0.018 |
kids | -0.018 |
Token | reached |
Token ID | 4251 |
Feature activation | -4.35e-02 |
Pos logit contributions | |
most | +0.038 |
truly | +0.036 |
best | +0.036 |
biggest | +0.035 |
whole | +0.034 |
Neg logit contributions | |
Lee | -0.037 |
handing | -0.035 |
riet | -0.035 |
kids | -0.035 |
pa | -0.034 |
Token | out |
Token ID | 503 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
ens | +0.090 |
riet | +0.089 |
kids | +0.089 |
Lee | +0.088 |
handing | +0.086 |
Neg logit contributions | |
most | -0.059 |
biggest | -0.059 |
highest | -0.056 |
best | -0.056 |
fastest | -0.055 |
Token | When |
Token ID | 1649 |
Feature activation | +3.82e-02 |
Pos logit contributions | |
most | +0.048 |
biggest | +0.046 |
highest | +0.044 |
fastest | +0.044 |
best | +0.043 |
Neg logit contributions | |
ens | -0.056 |
kids | -0.054 |
ppa | -0.053 |
riet | -0.053 |
pa | -0.053 |
Token | Daisy |
Token ID | 40355 |
Feature activation | -8.94e-02 |
Pos logit contributions | |
most | +0.076 |
biggest | +0.074 |
best | +0.072 |
fastest | +0.072 |
whole | +0.072 |
Neg logit contributions | |
ens | -0.054 |
riet | -0.052 |
Lee | -0.052 |
kids | -0.052 |
handing | -0.050 |
Token | put |
Token ID | 1234 |
Feature activation | -3.86e-02 |
Pos logit contributions | |
kids | +0.159 |
riet | +0.154 |
Lee | +0.154 |
handing | +0.151 |
ens | +0.148 |
Neg logit contributions | |
truly | -0.135 |
most | -0.131 |
whole | -0.131 |
highest | -0.126 |
fastest | -0.125 |
Token | the |
Token ID | 262 |
Feature activation | +3.95e-02 |
Pos logit contributions | |
ens | +0.064 |
Lee | +0.063 |
riet | +0.061 |
kids | +0.061 |
handing | +0.060 |
Neg logit contributions | |
most | -0.068 |
biggest | -0.068 |
highest | -0.065 |
fastest | -0.065 |
best | -0.064 |
Token | magn |
Token ID | 7842 |
Feature activation | +3.19e-02 |
Pos logit contributions | |
fastest | +0.069 |
truly | +0.069 |
most | +0.068 |
biggest | +0.068 |
highest | +0.067 |
Neg logit contributions | |
ens | -0.060 |
Lee | -0.059 |
purs | -0.057 |
handing | -0.056 |
Sophie | -0.055 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
most | +0.027 |
biggest | +0.027 |
best | +0.026 |
fastest | +0.026 |
coolest | +0.025 |
Neg logit contributions | |
riet | -0.078 |
ens | -0.078 |
Lee | -0.077 |
Sophie | -0.076 |
handing | -0.075 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
fastest | +0.427 |
best | +0.395 |
impossible | +0.385 |
possible | +0.385 |
biggest | +0.384 |
Neg logit contributions | |
Sue | -0.339 |
Amy | -0.321 |
Sophie | -0.320 |
Tom | -0.309 |
Sally | -0.303 |
Token | down |
Token ID | 866 |
Feature activation | +2.05e-02 |
Pos logit contributions | |
biggest | +0.181 |
most | +0.180 |
fastest | +0.177 |
highest | +0.175 |
truly | +0.172 |
Neg logit contributions | |
ens | -0.109 |
Lee | -0.104 |
Sue | -0.099 |
Sophie | -0.099 |
riet | -0.097 |
Token | , |
Token ID | 11 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
most | +0.045 |
biggest | +0.045 |
highest | +0.043 |
fastest | +0.043 |
best | +0.043 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.025 |
kids | -0.024 |
riet | -0.024 |
handing | -0.023 |
Token | the |
Token ID | 262 |
Feature activation | +4.15e-02 |
Pos logit contributions | |
Lee | +0.121 |
ens | +0.120 |
kids | +0.120 |
riet | +0.115 |
handing | +0.115 |
Neg logit contributions | |
most | -0.241 |
biggest | -0.236 |
best | -0.229 |
highest | -0.228 |
fastest | -0.227 |
Token | cater |
Token ID | 20825 |
Feature activation | +4.68e-03 |
Pos logit contributions | |
biggest | +0.073 |
most | +0.073 |
fastest | +0.071 |
highest | +0.071 |
best | +0.069 |
Neg logit contributions | |
ens | -0.068 |
Lee | -0.066 |
riet | -0.064 |
kids | -0.063 |
Sue | -0.063 |
Token | , |
Token ID | 11 |
Feature activation | -1.15e-01 |
Pos logit contributions | |
kids | +0.124 |
ens | +0.123 |
riet | +0.120 |
Lee | +0.117 |
purs | +0.116 |
Neg logit contributions | |
best | -0.125 |
most | -0.123 |
biggest | -0.122 |
whole | -0.116 |
highest | -0.113 |
Token | handing |
Token ID | 22786 |
Feature activation | -7.16e-02 |
Pos logit contributions | |
kids | +0.236 |
pa | +0.233 |
ens | +0.231 |
riet | +0.230 |
purs | +0.225 |
Neg logit contributions | |
most | -0.165 |
whole | -0.154 |
biggest | -0.151 |
best | -0.151 |
highest | -0.149 |
Token | him |
Token ID | 683 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
kids | +0.086 |
handing | +0.084 |
lee | +0.082 |
blamed | +0.082 |
annah | +0.081 |
Neg logit contributions | |
most | -0.161 |
biggest | -0.157 |
best | -0.156 |
highest | -0.151 |
whole | -0.149 |
Token | the |
Token ID | 262 |
Feature activation | +3.99e-02 |
Pos logit contributions | |
kids | +0.036 |
riet | +0.035 |
handing | +0.035 |
olph | +0.034 |
Lee | +0.033 |
Neg logit contributions | |
most | -0.034 |
biggest | -0.033 |
best | -0.032 |
highest | -0.031 |
whole | -0.030 |
Token | magn |
Token ID | 7842 |
Feature activation | +4.09e-02 |
Pos logit contributions | |
most | +0.069 |
fastest | +0.068 |
biggest | +0.068 |
truly | +0.068 |
highest | +0.067 |
Neg logit contributions | |
Lee | -0.064 |
ens | -0.063 |
Sue | -0.060 |
handing | -0.059 |
purs | -0.058 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.13e-01 |
Pos logit contributions | |
most | +0.033 |
biggest | +0.032 |
best | +0.029 |
coolest | +0.028 |
fastest | +0.028 |
Neg logit contributions | |
Lee | -0.106 |
riet | -0.105 |
ens | -0.105 |
olph | -0.102 |
Sophie | -0.101 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.99e-02 |
Pos logit contributions | |
fastest | +0.404 |
biggest | +0.377 |
most | +0.377 |
best | +0.375 |
truly | +0.374 |
Neg logit contributions | |
Sue | -0.346 |
Lee | -0.331 |
Amy | -0.328 |
Sophie | -0.324 |
Mrs | -0.323 |
Token | . |
Token ID | 13 |
Feature activation | +5.71e-03 |
Pos logit contributions | |
most | +0.155 |
biggest | +0.155 |
fastest | +0.150 |
highest | +0.149 |
best | +0.146 |
Neg logit contributions | |
ens | -0.151 |
Lee | -0.149 |
riet | -0.143 |
Sue | -0.141 |
kids | -0.141 |
Token | She |
Token ID | 1375 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
most | +0.007 |
biggest | +0.006 |
highest | +0.006 |
best | +0.006 |
It | +0.006 |
Neg logit contributions | |
kids | -0.013 |
pa | -0.013 |
ppa | -0.013 |
riet | -0.013 |
child | -0.012 |
Token | looked |
Token ID | 3114 |
Feature activation | +5.78e-02 |
Pos logit contributions | |
most | +0.020 |
truly | +0.019 |
best | +0.018 |
biggest | +0.018 |
perfect | +0.017 |
Neg logit contributions | |
ens | -0.015 |
Lee | -0.015 |
riet | -0.015 |
kids | -0.014 |
Gig | -0.014 |
Token | around |
Token ID | 1088 |
Feature activation | +7.46e-03 |
Pos logit contributions | |
most | +0.086 |
biggest | +0.081 |
best | +0.080 |
whole | +0.076 |
highest | +0.076 |
Neg logit contributions | |
kids | -0.123 |
Lee | -0.122 |
ens | -0.119 |
riet | -0.116 |
child | -0.116 |
Token | the |
Token ID | 262 |
Feature activation | +4.30e-02 |
Pos logit contributions | |
kids | +0.036 |
riet | +0.034 |
ens | +0.033 |
olph | +0.032 |
ppa | +0.032 |
Neg logit contributions | |
most | -0.042 |
biggest | -0.042 |
best | -0.041 |
highest | -0.039 |
whole | -0.038 |
Token | necklace |
Token ID | 38987 |
Feature activation | +4.34e-02 |
Pos logit contributions | |
ifying | +0.085 |
ained | +0.080 |
hest | +0.079 |
fastest | +0.079 |
truly | +0.078 |
Neg logit contributions | |
purs | -0.062 |
ens | -0.054 |
handing | -0.053 |
her | -0.053 |
sob | -0.053 |
Token | and |
Token ID | 290 |
Feature activation | -8.10e-02 |
Pos logit contributions | |
biggest | +0.091 |
most | +0.091 |
fastest | +0.089 |
highest | +0.088 |
whole | +0.086 |
Neg logit contributions | |
ens | -0.057 |
Lee | -0.056 |
riet | -0.053 |
Sue | -0.053 |
handing | -0.052 |
Token | the |
Token ID | 262 |
Feature activation | +2.78e-02 |
Pos logit contributions | |
kids | +0.123 |
ens | +0.118 |
annah | +0.117 |
ior | +0.115 |
riet | +0.112 |
Neg logit contributions | |
most | -0.160 |
best | -0.148 |
biggest | -0.142 |
truly | -0.138 |
perfect | -0.136 |
Token | magn |
Token ID | 7842 |
Feature activation | +4.20e-02 |
Pos logit contributions | |
fastest | +0.049 |
biggest | +0.049 |
most | +0.048 |
highest | +0.048 |
truly | +0.048 |
Neg logit contributions | |
ens | -0.042 |
Lee | -0.042 |
purs | -0.041 |
handing | -0.040 |
Sue | -0.040 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.10e-01 |
Pos logit contributions | |
most | +0.047 |
biggest | +0.045 |
coolest | +0.044 |
best | +0.043 |
fastest | +0.043 |
Neg logit contributions | |
riet | -0.098 |
lee | -0.097 |
ens | -0.094 |
Lee | -0.093 |
purs | -0.093 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.71e-02 |
Pos logit contributions | |
fastest | +0.448 |
best | +0.409 |
coolest | +0.406 |
biggest | +0.402 |
possible | +0.401 |
Neg logit contributions | |
Sue | -0.308 |
Mrs | -0.288 |
Sara | -0.283 |
Sophie | -0.282 |
Tom | -0.278 |
Token | and |
Token ID | 290 |
Feature activation | -7.40e-02 |
Pos logit contributions | |
most | +0.154 |
biggest | +0.154 |
fastest | +0.152 |
highest | +0.148 |
truly | +0.145 |
Neg logit contributions | |
ens | -0.144 |
Lee | -0.141 |
Sue | -0.137 |
handing | -0.135 |
Sophie | -0.134 |
Token | told |
Token ID | 1297 |
Feature activation | -6.49e-02 |
Pos logit contributions | |
kids | +0.114 |
ens | +0.113 |
riet | +0.108 |
annah | +0.106 |
ior | +0.105 |
Neg logit contributions | |
most | -0.146 |
biggest | -0.135 |
best | -0.133 |
truly | -0.129 |
highest | -0.126 |
Token | him |
Token ID | 683 |
Feature activation | -2.55e-02 |
Pos logit contributions | |
stro | +0.086 |
kids | +0.086 |
olph | +0.085 |
ppa | +0.084 |
ens | +0.084 |
Neg logit contributions | |
biggest | -0.128 |
best | -0.127 |
most | -0.127 |
possible | -0.117 |
whole | -0.116 |
Token | that |
Token ID | 326 |
Feature activation | +2.01e-02 |
Pos logit contributions | |
ens | +0.041 |
kids | +0.040 |
Lee | +0.039 |
riet | +0.039 |
olph | +0.038 |
Neg logit contributions | |
most | -0.046 |
biggest | -0.046 |
best | -0.044 |
highest | -0.043 |
truly | -0.043 |
Token | did |
Token ID | 750 |
Feature activation | +4.52e-02 |
Pos logit contributions | |
ens | +0.187 |
Lee | +0.185 |
riet | +0.182 |
kids | +0.178 |
handing | +0.176 |
Neg logit contributions | |
most | -0.216 |
biggest | -0.212 |
fastest | -0.205 |
best | -0.204 |
highest | -0.204 |
Token | not |
Token ID | 407 |
Feature activation | +1.72e-02 |
Pos logit contributions | |
biggest | +0.076 |
fastest | +0.075 |
most | +0.074 |
highest | +0.073 |
ifying | +0.072 |
Neg logit contributions | |
Sue | -0.078 |
Lee | -0.077 |
ens | -0.074 |
Amy | -0.073 |
Tom | -0.073 |
Token | know |
Token ID | 760 |
Feature activation | -1.36e-02 |
Pos logit contributions | |
most | +0.032 |
biggest | +0.032 |
fastest | +0.032 |
highest | +0.031 |
whole | +0.031 |
Neg logit contributions | |
ens | -0.026 |
Lee | -0.026 |
riet | -0.025 |
kids | -0.025 |
handing | -0.024 |
Token | what |
Token ID | 644 |
Feature activation | +8.42e-02 |
Pos logit contributions | |
ens | +0.018 |
riet | +0.017 |
kids | +0.017 |
Lee | +0.017 |
handing | +0.016 |
Neg logit contributions | |
most | -0.029 |
biggest | -0.029 |
best | -0.028 |
fastest | -0.028 |
highest | -0.028 |
Token | ost |
Token ID | 23619 |
Feature activation | -3.09e-03 |
Pos logit contributions | |
biggest | +0.165 |
most | +0.164 |
fastest | +0.163 |
highest | +0.162 |
whole | +0.158 |
Neg logit contributions | |
Lee | -0.122 |
Sue | -0.118 |
ens | -0.118 |
Amy | -0.111 |
Sophie | -0.110 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.10e-01 |
Pos logit contributions | |
ens | +0.007 |
Lee | +0.007 |
pa | +0.007 |
kids | +0.007 |
Sue | +0.007 |
Neg logit contributions | |
biggest | -0.003 |
most | -0.003 |
highest | -0.003 |
fastest | -0.003 |
whole | -0.003 |
Token | hes |
Token ID | 956 |
Feature activation | +7.13e-02 |
Pos logit contributions | |
biggest | +0.348 |
most | +0.346 |
coolest | +0.338 |
fastest | +0.336 |
highest | +0.327 |
Neg logit contributions | |
ens | -0.352 |
Sue | -0.340 |
Lee | -0.338 |
Tom | -0.332 |
Amy | -0.327 |
Token | were |
Token ID | 547 |
Feature activation | +8.32e-02 |
Pos logit contributions | |
biggest | +0.142 |
highest | +0.140 |
most | +0.139 |
fastest | +0.139 |
entire | +0.135 |
Neg logit contributions | |
ens | -0.103 |
Lee | -0.102 |
Sue | -0.098 |
Sophie | -0.094 |
Tom | -0.092 |
Token | . |
Token ID | 13 |
Feature activation | +6.20e-03 |
Pos logit contributions | |
most | +0.154 |
biggest | +0.153 |
fastest | +0.149 |
highest | +0.148 |
best | +0.146 |
Neg logit contributions | |
ens | -0.129 |
Lee | -0.127 |
riet | -0.123 |
kids | -0.121 |
Sophie | -0.120 |
Token | They |
Token ID | 1119 |
Feature activation | +2.38e-02 |
Pos logit contributions | |
most | +0.010 |
biggest | +0.010 |
highest | +0.009 |
fastest | +0.009 |
best | +0.009 |
Neg logit contributions | |
ens | -0.011 |
riet | -0.011 |
pa | -0.011 |
kids | -0.011 |
ppa | -0.011 |
Token | looked |
Token ID | 3114 |
Feature activation | +5.43e-02 |
Pos logit contributions | |
most | +0.048 |
biggest | +0.046 |
fastest | +0.046 |
highest | +0.045 |
truly | +0.045 |
Neg logit contributions | |
Lee | -0.033 |
ens | -0.033 |
riet | -0.032 |
handing | -0.032 |
kids | -0.032 |
Token | you |
Token ID | 345 |
Feature activation | +1.29e-02 |
Pos logit contributions | |
ens | +0.070 |
Lee | +0.069 |
riet | +0.068 |
kids | +0.067 |
handing | +0.066 |
Neg logit contributions | |
most | -0.049 |
biggest | -0.049 |
fastest | -0.047 |
highest | -0.047 |
best | -0.046 |
Token | count |
Token ID | 954 |
Feature activation | -4.40e-02 |
Pos logit contributions | |
biggest | +0.026 |
most | +0.026 |
fastest | +0.025 |
highest | +0.025 |
best | +0.025 |
Neg logit contributions | |
ens | -0.018 |
Lee | -0.017 |
Sue | -0.016 |
riet | -0.016 |
handing | -0.016 |
Token | how |
Token ID | 703 |
Feature activation | +5.16e-02 |
Pos logit contributions | |
Lee | +0.064 |
ens | +0.064 |
Sue | +0.061 |
handing | +0.059 |
riet | +0.059 |
Neg logit contributions | |
biggest | -0.084 |
most | -0.083 |
fastest | -0.082 |
highest | -0.081 |
whole | -0.080 |
Token | many |
Token ID | 867 |
Feature activation | +2.61e-02 |
Pos logit contributions | |
fastest | +0.091 |
biggest | +0.091 |
most | +0.091 |
highest | +0.088 |
coolest | +0.087 |
Neg logit contributions | |
ens | -0.080 |
Sue | -0.079 |
Lee | -0.078 |
purs | -0.075 |
Gig | -0.075 |
Token | ost |
Token ID | 23619 |
Feature activation | -2.38e-03 |
Pos logit contributions | |
biggest | +0.047 |
most | +0.047 |
highest | +0.046 |
fastest | +0.046 |
whole | +0.046 |
Neg logit contributions | |
ens | -0.040 |
Lee | -0.039 |
Sue | -0.038 |
kids | -0.037 |
purs | -0.037 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.09e-01 |
Pos logit contributions | |
ens | +0.005 |
Lee | +0.005 |
pa | +0.005 |
kids | +0.005 |
lee | +0.005 |
Neg logit contributions | |
biggest | -0.003 |
most | -0.003 |
whole | -0.003 |
highest | -0.003 |
fastest | -0.003 |
Token | hes |
Token ID | 956 |
Feature activation | +6.50e-02 |
Pos logit contributions | |
biggest | +0.347 |
most | +0.343 |
fastest | +0.331 |
coolest | +0.329 |
ifying | +0.329 |
Neg logit contributions | |
ens | -0.353 |
Sue | -0.352 |
Amy | -0.335 |
Tom | -0.335 |
Lee | -0.333 |
Token | there |
Token ID | 612 |
Feature activation | +8.73e-02 |
Pos logit contributions | |
biggest | +0.114 |
entire | +0.111 |
fastest | +0.111 |
highest | +0.109 |
most | +0.109 |
Neg logit contributions | |
ens | -0.106 |
Sue | -0.101 |
Lee | -0.099 |
her | -0.096 |
Tom | -0.095 |
Token | are |
Token ID | 389 |
Feature activation | +8.89e-02 |
Pos logit contributions | |
ifying | +0.174 |
biggest | +0.173 |
entire | +0.169 |
most | +0.168 |
virtue | +0.168 |
Neg logit contributions | |
ens | -0.113 |
Lee | -0.107 |
Sue | -0.107 |
Tom | -0.101 |
Amy | -0.100 |
Token | ?" |
Token ID | 1701 |
Feature activation | -3.62e-02 |
Pos logit contributions | |
biggest | +0.145 |
most | +0.144 |
ifying | +0.142 |
fastest | +0.139 |
highest | +0.138 |
Neg logit contributions | |
ens | -0.147 |
Sue | -0.147 |
Lee | -0.146 |
handing | -0.139 |
Amy | -0.139 |
Token |
|
Token ID | 198 |
Feature activation | +4.30e-02 |
Pos logit contributions | |
ppa | +0.064 |
kids | +0.062 |
riet | +0.062 |
pa | +0.060 |
purs | +0.059 |
Neg logit contributions | |
most | -0.066 |
best | -0.063 |
biggest | -0.063 |
whole | -0.059 |
It | -0.059 |
Token | She |
Token ID | 1375 |
Feature activation | +8.67e-03 |
Pos logit contributions | |
ens | +0.005 |
kids | +0.004 |
riet | +0.004 |
pa | +0.004 |
purs | +0.004 |
Neg logit contributions | |
most | -0.004 |
biggest | -0.004 |
highest | -0.004 |
best | -0.004 |
fastest | -0.004 |
Token | takes |
Token ID | 2753 |
Feature activation | -3.33e-02 |
Pos logit contributions | |
most | +0.018 |
biggest | +0.018 |
fastest | +0.017 |
highest | +0.017 |
whole | +0.017 |
Neg logit contributions | |
ens | -0.012 |
Lee | -0.012 |
riet | -0.011 |
handing | -0.011 |
kids | -0.011 |
Token | out |
Token ID | 503 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
ens | +0.063 |
Lee | +0.062 |
riet | +0.060 |
kids | +0.059 |
handing | +0.059 |
Neg logit contributions | |
most | -0.051 |
biggest | -0.051 |
fastest | -0.049 |
highest | -0.049 |
best | -0.048 |
Token | a |
Token ID | 257 |
Feature activation | +3.52e-02 |
Pos logit contributions | |
ens | +0.023 |
riet | +0.023 |
kids | +0.022 |
Lee | +0.022 |
handing | +0.022 |
Neg logit contributions | |
most | -0.017 |
biggest | -0.016 |
whole | -0.016 |
best | -0.015 |
highest | -0.015 |
Token | magn |
Token ID | 7842 |
Feature activation | +2.95e-02 |
Pos logit contributions | |
most | +0.050 |
fastest | +0.049 |
biggest | +0.049 |
highest | +0.047 |
best | +0.046 |
Neg logit contributions | |
ens | -0.068 |
Lee | -0.067 |
Sue | -0.066 |
handing | -0.065 |
Amy | -0.065 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.08e-01 |
Pos logit contributions | |
most | +0.028 |
biggest | +0.026 |
best | +0.026 |
coolest | +0.025 |
fastest | +0.025 |
Neg logit contributions | |
Lee | -0.073 |
riet | -0.072 |
lee | -0.070 |
ens | -0.070 |
handing | -0.069 |
Token | glass |
Token ID | 5405 |
Feature activation | +8.45e-02 |
Pos logit contributions | |
fastest | +0.406 |
best | +0.376 |
biggest | +0.370 |
coolest | +0.366 |
truly | +0.364 |
Neg logit contributions | |
Sue | -0.331 |
Amy | -0.317 |
Sophie | -0.314 |
Mrs | -0.310 |
Lee | -0.307 |
Token | . |
Token ID | 13 |
Feature activation | +1.19e-02 |
Pos logit contributions | |
biggest | +0.162 |
most | +0.161 |
fastest | +0.161 |
truly | +0.155 |
highest | +0.155 |
Neg logit contributions | |
ens | -0.127 |
Lee | -0.125 |
handing | -0.120 |
Sue | -0.119 |
Sophie | -0.117 |
Token | She |
Token ID | 1375 |
Feature activation | +1.85e-02 |
Pos logit contributions | |
most | +0.019 |
biggest | +0.019 |
highest | +0.018 |
best | +0.018 |
fastest | +0.018 |
Neg logit contributions | |
ens | -0.021 |
kids | -0.021 |
riet | -0.021 |
pa | -0.020 |
Lee | -0.020 |
Token | looks |
Token ID | 3073 |
Feature activation | +6.38e-02 |
Pos logit contributions | |
most | +0.037 |
biggest | +0.037 |
fastest | +0.036 |
highest | +0.036 |
best | +0.035 |
Neg logit contributions | |
ens | -0.026 |
Lee | -0.026 |
riet | -0.025 |
kids | -0.024 |
Sophie | -0.024 |
Token | at |
Token ID | 379 |
Feature activation | -4.22e-02 |
Pos logit contributions | |
most | +0.103 |
biggest | +0.101 |
highest | +0.096 |
best | +0.095 |
whole | +0.094 |
Neg logit contributions | |
Lee | -0.122 |
ens | -0.121 |
kids | -0.118 |
riet | -0.116 |
olph | -0.114 |
Token | shouted |
Token ID | 17293 |
Feature activation | -7.31e-02 |
Pos logit contributions | |
riet | +0.229 |
kids | +0.220 |
ppa | +0.219 |
purs | +0.215 |
pa | +0.215 |
Neg logit contributions | |
most | -0.197 |
best | -0.180 |
highest | -0.178 |
fastest | -0.173 |
biggest | -0.172 |
Token | . |
Token ID | 13 |
Feature activation | +8.70e-04 |
Pos logit contributions | |
kids | +0.157 |
riet | +0.156 |
ens | +0.152 |
purs | +0.149 |
pa | +0.149 |
Neg logit contributions | |
best | -0.092 |
most | -0.092 |
biggest | -0.090 |
whole | -0.087 |
highest | -0.082 |
Token | " |
Token ID | 366 |
Feature activation | +3.83e-02 |
Pos logit contributions | |
most | +0.001 |
biggest | +0.001 |
whole | +0.001 |
best | +0.001 |
highest | +0.001 |
Neg logit contributions | |
ens | -0.002 |
riet | -0.002 |
kids | -0.002 |
pa | -0.001 |
Lee | -0.001 |
Token | The |
Token ID | 464 |
Feature activation | +4.40e-02 |
Pos logit contributions | |
biggest | +0.054 |
best | +0.053 |
most | +0.053 |
fastest | +0.052 |
truly | +0.050 |
Neg logit contributions | |
Lee | -0.077 |
ens | -0.076 |
Sue | -0.074 |
Tom | -0.073 |
stro | -0.072 |
Token | ost |
Token ID | 23619 |
Feature activation | +1.33e-03 |
Pos logit contributions | |
truly | +0.088 |
fastest | +0.088 |
most | +0.088 |
biggest | +0.088 |
ifying | +0.087 |
Neg logit contributions | |
ens | -0.060 |
Lee | -0.060 |
kids | -0.056 |
Sue | -0.054 |
Gig | -0.053 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.07e-01 |
Pos logit contributions | |
biggest | +0.001 |
most | +0.001 |
highest | +0.001 |
fastest | +0.001 |
whole | +0.001 |
Neg logit contributions | |
Lee | -0.003 |
ens | -0.003 |
pa | -0.003 |
kids | -0.003 |
Sue | -0.003 |
Token | hes |
Token ID | 956 |
Feature activation | +6.17e-02 |
Pos logit contributions | |
biggest | +0.357 |
most | +0.354 |
fastest | +0.350 |
coolest | +0.346 |
ifying | +0.335 |
Neg logit contributions | |
Sue | -0.345 |
Tom | -0.335 |
ens | -0.332 |
Amy | -0.331 |
Lee | -0.324 |
Token | are |
Token ID | 389 |
Feature activation | +8.48e-02 |
Pos logit contributions | |
biggest | +0.091 |
most | +0.087 |
fastest | +0.087 |
highest | +0.085 |
possible | +0.085 |
Neg logit contributions | |
ens | -0.119 |
Sue | -0.118 |
Tom | -0.112 |
Lee | -0.112 |
Amy | -0.110 |
Token | so |
Token ID | 523 |
Feature activation | +2.91e-02 |
Pos logit contributions | |
biggest | +0.127 |
most | +0.126 |
fastest | +0.121 |
highest | +0.121 |
whole | +0.117 |
Neg logit contributions | |
ens | -0.158 |
Lee | -0.156 |
Sue | -0.154 |
handing | -0.151 |
Sophie | -0.149 |
Token | funny |
Token ID | 8258 |
Feature activation | +2.34e-02 |
Pos logit contributions | |
biggest | +0.037 |
most | +0.037 |
fastest | +0.035 |
coolest | +0.035 |
highest | +0.034 |
Neg logit contributions | |
Sue | -0.062 |
ens | -0.061 |
Amy | -0.059 |
Tom | -0.058 |
Gig | -0.057 |
Token | !" |
Token ID | 2474 |
Feature activation | -3.76e-02 |
Pos logit contributions | |
biggest | +0.038 |
most | +0.038 |
fastest | +0.037 |
highest | +0.037 |
truly | +0.036 |
Neg logit contributions | |
Lee | -0.042 |
ens | -0.041 |
Sue | -0.040 |
handing | -0.039 |
riet | -0.039 |
Token | but |
Token ID | 475 |
Feature activation | +1.44e-02 |
Pos logit contributions | |
riet | +0.193 |
Lee | +0.192 |
kids | +0.192 |
stro | +0.190 |
olph | +0.189 |
Neg logit contributions | |
most | -0.159 |
biggest | -0.152 |
best | -0.149 |
fastest | -0.146 |
highest | -0.141 |
Token | be |
Token ID | 307 |
Feature activation | +4.77e-02 |
Pos logit contributions | |
biggest | +0.021 |
most | +0.021 |
fastest | +0.020 |
highest | +0.020 |
whole | +0.020 |
Neg logit contributions | |
ens | -0.028 |
Lee | -0.028 |
riet | -0.027 |
Sue | -0.027 |
Sophie | -0.027 |
Token | careful |
Token ID | 8161 |
Feature activation | -2.53e-02 |
Pos logit contributions | |
most | +0.085 |
biggest | +0.085 |
fastest | +0.082 |
highest | +0.082 |
whole | +0.080 |
Neg logit contributions | |
ens | -0.078 |
Lee | -0.077 |
riet | -0.074 |
kids | -0.073 |
Sue | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | +2.71e-02 |
Pos logit contributions | |
Lee | +0.038 |
ens | +0.038 |
Sue | +0.037 |
handing | +0.036 |
Amy | +0.035 |
Neg logit contributions | |
biggest | -0.048 |
most | -0.047 |
highest | -0.047 |
fastest | -0.046 |
truly | -0.046 |
Token | Ost |
Token ID | 38919 |
Feature activation | +3.51e-02 |
Pos logit contributions | |
biggest | +0.048 |
most | +0.048 |
fastest | +0.047 |
highest | +0.047 |
whole | +0.046 |
Neg logit contributions | |
ens | -0.044 |
Lee | -0.044 |
Sue | -0.042 |
riet | -0.041 |
handing | -0.041 |
Token | ric |
Token ID | 1173 |
Feature activation | +2.07e-01 |
Pos logit contributions | |
biggest | +0.043 |
most | +0.041 |
fastest | +0.041 |
highest | +0.040 |
best | +0.039 |
Neg logit contributions | |
Lee | -0.076 |
pa | -0.075 |
ens | -0.074 |
kids | -0.073 |
handing | -0.073 |
Token | hes |
Token ID | 956 |
Feature activation | +6.41e-02 |
Pos logit contributions | |
biggest | +0.363 |
fastest | +0.360 |
most | +0.357 |
coolest | +0.353 |
ifying | +0.351 |
Neg logit contributions | |
Tom | -0.328 |
Sue | -0.327 |
ens | -0.318 |
Amy | -0.307 |
Lee | -0.303 |
Token | can |
Token ID | 460 |
Feature activation | +3.60e-02 |
Pos logit contributions | |
biggest | +0.091 |
fastest | +0.088 |
highest | +0.088 |
most | +0.087 |
coolest | +0.085 |
Neg logit contributions | |
ens | -0.131 |
Sue | -0.122 |
Lee | -0.122 |
her | -0.117 |
Tom | -0.116 |
Token | be |
Token ID | 307 |
Feature activation | +5.24e-02 |
Pos logit contributions | |
biggest | +0.067 |
most | +0.066 |
fastest | +0.063 |
best | +0.063 |
highest | +0.062 |
Neg logit contributions | |
ens | -0.058 |
Lee | -0.054 |
Sue | -0.053 |
stro | -0.052 |
Gig | -0.052 |
Token | gr |
Token ID | 1036 |
Feature activation | +2.86e-02 |
Pos logit contributions | |
most | +0.080 |
biggest | +0.080 |
fastest | +0.077 |
highest | +0.076 |
whole | +0.076 |
Neg logit contributions | |
ens | -0.099 |
Lee | -0.098 |
riet | -0.096 |
kids | -0.094 |
Sophie | -0.093 |
Token | umpy |
Token ID | 32152 |
Feature activation | -2.16e-03 |
Pos logit contributions | |
biggest | +0.049 |
most | +0.049 |
best | +0.045 |
highest | +0.045 |
coolest | +0.044 |
Neg logit contributions | |
ens | -0.053 |
kids | -0.050 |
pa | -0.049 |
riet | -0.048 |
Lee | -0.048 |
Token | examine |
Token ID | 10716 |
Feature activation | -4.04e-02 |
Pos logit contributions | |
most | +0.029 |
biggest | +0.029 |
fastest | +0.029 |
highest | +0.028 |
best | +0.028 |
Neg logit contributions | |
ens | -0.019 |
Lee | -0.018 |
Sue | -0.018 |
Sophie | -0.017 |
riet | -0.017 |
Token | it |
Token ID | 340 |
Feature activation | +9.32e-02 |
Pos logit contributions | |
ens | +0.068 |
Lee | +0.067 |
riet | +0.065 |
kids | +0.064 |
Sue | +0.064 |
Neg logit contributions | |
most | -0.069 |
biggest | -0.069 |
fastest | -0.067 |
highest | -0.066 |
whole | -0.065 |
Token | with |
Token ID | 351 |
Feature activation | +8.73e-03 |
Pos logit contributions | |
biggest | +0.180 |
most | +0.179 |
fastest | +0.176 |
highest | +0.174 |
whole | +0.171 |
Neg logit contributions | |
ens | -0.134 |
Lee | -0.132 |
Sue | -0.124 |
handing | -0.124 |
Sophie | -0.124 |
Token | the |
Token ID | 262 |
Feature activation | +4.10e-02 |
Pos logit contributions | |
most | +0.016 |
biggest | +0.015 |
fastest | +0.015 |
highest | +0.015 |
best | +0.015 |
Neg logit contributions | |
ens | -0.014 |
Lee | -0.014 |
riet | -0.014 |
kids | -0.013 |
handing | -0.013 |
Token | magn |
Token ID | 7842 |
Feature activation | +4.09e-02 |
Pos logit contributions | |
biggest | +0.063 |
most | +0.063 |
fastest | +0.063 |
truly | +0.062 |
highest | +0.062 |
Neg logit contributions | |
ens | -0.073 |
Lee | -0.072 |
Sue | -0.069 |
handing | -0.068 |
purs | -0.068 |
Token | ifying |
Token ID | 4035 |
Feature activation | +2.07e-01 |
Pos logit contributions | |
most | +0.037 |
biggest | +0.035 |
best | +0.035 |
coolest | +0.034 |
fastest | +0.032 |
Neg logit contributions | |
Lee | -0.101 |
riet | -0.100 |
ens | -0.098 |
lee | -0.097 |
handing | -0.096 |
Token | glass |
Token ID | 5405 |
Feature activation | +9.29e-02 |
Pos logit contributions | |
fastest | +0.413 |
truly | +0.383 |
possible | +0.382 |
most | +0.379 |
best | +0.376 |
Neg logit contributions | |
Sue | -0.314 |
Lee | -0.294 |
Sara | -0.291 |
Sophie | -0.288 |
Mrs | -0.286 |
Token | ?" |
Token ID | 1701 |
Feature activation | -3.81e-02 |
Pos logit contributions | |
biggest | +0.180 |
most | +0.180 |
fastest | +0.178 |
highest | +0.174 |
truly | +0.172 |
Neg logit contributions | |
ens | -0.136 |
Lee | -0.132 |
Sue | -0.126 |
Sophie | -0.125 |
riet | -0.123 |
Token |
|
Token ID | 198 |
Feature activation | +3.93e-02 |
Pos logit contributions | |
riet | +0.077 |
kids | +0.075 |
pa | +0.075 |
ppa | +0.074 |
Lee | +0.073 |
Neg logit contributions | |
most | -0.056 |
biggest | -0.054 |
best | -0.052 |
whole | -0.051 |
truly | -0.048 |
Token |
|
Token ID | 198 |
Feature activation | +3.71e-02 |
Pos logit contributions | |
most | +0.059 |
biggest | +0.058 |
fastest | +0.056 |
highest | +0.055 |
whole | +0.055 |
Neg logit contributions | |
ens | -0.076 |
Lee | -0.075 |
riet | -0.073 |
kids | -0.072 |
handing | -0.071 |
Token | " |
Token ID | 1 |
Feature activation | -1.18e-02 |
Pos logit contributions | |
most | +0.069 |
biggest | +0.069 |
fastest | +0.067 |
highest | +0.066 |
best | +0.065 |
Neg logit contributions | |
ens | -0.058 |
Lee | -0.057 |
riet | -0.055 |
kids | -0.054 |
handing | -0.054 |
Token | counter |
Token ID | 3753 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
most | +0.087 |
truly | +0.086 |
fastest | +0.085 |
biggest | +0.085 |
possible | +0.084 |
Neg logit contributions | |
Lee | -0.060 |
ens | -0.060 |
kids | -0.056 |
handing | -0.055 |
riet | -0.055 |
Token | . |
Token ID | 13 |
Feature activation | +1.03e-02 |
Pos logit contributions | |
most | +0.028 |
biggest | +0.028 |
fastest | +0.027 |
highest | +0.026 |
best | +0.026 |
Neg logit contributions | |
ens | -0.030 |
Lee | -0.029 |
riet | -0.028 |
handing | -0.028 |
kids | -0.028 |
Token | |
Token ID | 220 |
Feature activation | +9.48e-02 |
Pos logit contributions | |
most | +0.017 |
biggest | +0.017 |
highest | +0.016 |
whole | +0.015 |
fastest | +0.015 |
Neg logit contributions | |
ens | -0.019 |
Lee | -0.019 |
kids | -0.018 |
riet | -0.018 |
purs | -0.018 |
Token |
|
Token ID | 198 |
Feature activation | +4.98e-02 |
Pos logit contributions | |
most | +0.162 |
biggest | +0.162 |
fastest | +0.160 |
highest | +0.158 |
whole | +0.151 |
Neg logit contributions | |
ens | -0.161 |
Lee | -0.157 |
Sue | -0.153 |
Sophie | -0.150 |
riet | -0.150 |
Token |
|
Token ID | 198 |
Feature activation | +4.47e-02 |
Pos logit contributions | |
most | +0.072 |
biggest | +0.071 |
whole | +0.068 |
highest | +0.068 |
fastest | +0.068 |
Neg logit contributions | |
ens | -0.098 |
Lee | -0.097 |
riet | -0.094 |
kids | -0.093 |
handing | -0.093 |
Token | Mic |
Token ID | 25437 |
Feature activation | -1.82e-01 |
Pos logit contributions | |
most | +0.107 |
biggest | +0.107 |
fastest | +0.105 |
highest | +0.104 |
best | +0.103 |
Neg logit contributions | |
ens | -0.045 |
Lee | -0.043 |
riet | -0.042 |
kids | -0.040 |
handing | -0.040 |
Token | key |
Token ID | 2539 |
Feature activation | -1.13e-01 |
Pos logit contributions | |
Lee | +0.250 |
Sue | +0.242 |
ens | +0.237 |
Sophie | +0.236 |
kids | +0.232 |
Neg logit contributions | |
most | -0.364 |
fastest | -0.354 |
whole | -0.350 |
best | -0.349 |
biggest | -0.348 |
Token | wanted |
Token ID | 2227 |
Feature activation | +9.72e-03 |
Pos logit contributions | |
Lee | +0.229 |
riet | +0.224 |
ens | +0.223 |
handing | +0.217 |
purs | +0.212 |
Neg logit contributions | |
most | -0.160 |
whole | -0.150 |
best | -0.146 |
truly | -0.146 |
biggest | -0.144 |
Token | some |
Token ID | 617 |
Feature activation | +6.08e-02 |
Pos logit contributions | |
most | +0.016 |
biggest | +0.016 |
best | +0.015 |
whole | +0.015 |
highest | +0.015 |
Neg logit contributions | |
riet | -0.018 |
ens | -0.017 |
Lee | -0.017 |
kids | -0.017 |
handing | -0.017 |
Token | popcorn |
Token ID | 38914 |
Feature activation | +4.46e-02 |
Pos logit contributions | |
fastest | +0.117 |
highest | +0.117 |
biggest | +0.116 |
most | +0.115 |
whole | +0.111 |
Neg logit contributions | |
ens | -0.088 |
Lee | -0.084 |
kids | -0.082 |
Sophie | -0.082 |
handing | -0.081 |
Token | , |
Token ID | 11 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
most | +0.085 |
biggest | +0.085 |
fastest | +0.082 |
highest | +0.082 |
whole | +0.080 |
Neg logit contributions | |
ens | -0.067 |
Lee | -0.066 |
riet | -0.063 |
kids | -0.062 |
Sophie | -0.062 |
Token | their |
Token ID | 511 |
Feature activation | -8.76e-03 |
Pos logit contributions | |
ens | +0.017 |
Lee | +0.017 |
riet | +0.016 |
kids | +0.016 |
handing | +0.016 |
Neg logit contributions | |
most | -0.026 |
biggest | -0.026 |
fastest | -0.025 |
highest | -0.025 |
best | -0.025 |
Token | impressive |
Token ID | 8036 |
Feature activation | +8.66e-03 |
Pos logit contributions | |
ens | +0.016 |
Lee | +0.015 |
riet | +0.015 |
handing | +0.015 |
Sophie | +0.015 |
Neg logit contributions | |
most | -0.015 |
biggest | -0.015 |
best | -0.014 |
whole | -0.014 |
fastest | -0.014 |
Token | strength |
Token ID | 4202 |
Feature activation | +5.63e-02 |
Pos logit contributions | |
most | +0.014 |
biggest | +0.014 |
best | +0.014 |
whole | +0.014 |
fastest | +0.014 |
Neg logit contributions | |
ens | -0.015 |
Lee | -0.015 |
riet | -0.015 |
Sophie | -0.014 |
Sue | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | +2.26e-02 |
Pos logit contributions | |
biggest | +0.112 |
most | +0.109 |
highest | +0.108 |
fastest | +0.108 |
coolest | +0.104 |
Neg logit contributions | |
ens | -0.076 |
Lee | -0.074 |
handing | -0.073 |
Sue | -0.073 |
Sophie | -0.072 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.00e-02 |
Pos logit contributions | |
biggest | +0.042 |
most | +0.041 |
fastest | +0.041 |
highest | +0.040 |
best | +0.039 |
Neg logit contributions | |
ens | -0.036 |
Lee | -0.036 |
Sue | -0.034 |
Sophie | -0.034 |
riet | -0.033 |
Token | Stan |
Token ID | 32140 |
Feature activation | -1.82e-01 |
Pos logit contributions | |
biggest | +0.017 |
fastest | +0.016 |
most | +0.016 |
highest | +0.016 |
coolest | +0.016 |
Neg logit contributions | |
Lee | -0.018 |
ens | -0.018 |
Sue | -0.017 |
Sophie | -0.017 |
Amy | -0.016 |
Token | ley |
Token ID | 1636 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
ens | +0.284 |
Lee | +0.279 |
handing | +0.274 |
Sue | +0.271 |
kids | +0.266 |
Neg logit contributions | |
most | -0.328 |
whole | -0.322 |
biggest | -0.317 |
truly | -0.313 |
highest | -0.312 |
Token | was |
Token ID | 373 |
Feature activation | +4.70e-02 |
Pos logit contributions | |
ens | +0.182 |
Lee | +0.179 |
riet | +0.178 |
handing | +0.176 |
kids | +0.172 |
Neg logit contributions | |
most | -0.167 |
best | -0.158 |
biggest | -0.157 |
whole | -0.157 |
truly | -0.156 |
Token | a |
Token ID | 257 |
Feature activation | +3.00e-02 |
Pos logit contributions | |
biggest | +0.068 |
most | +0.067 |
fastest | +0.066 |
highest | +0.065 |
whole | +0.062 |
Neg logit contributions | |
ens | -0.090 |
Lee | -0.090 |
Sophie | -0.087 |
Sue | -0.087 |
riet | -0.087 |
Token | curious |
Token ID | 11040 |
Feature activation | -1.77e-02 |
Pos logit contributions | |
biggest | +0.048 |
most | +0.048 |
fastest | +0.047 |
highest | +0.047 |
best | +0.045 |
Neg logit contributions | |
ens | -0.052 |
Lee | -0.052 |
Sophie | -0.050 |
riet | -0.049 |
Sue | -0.049 |
Token | 3 |
Token ID | 513 |
Feature activation | -8.09e-04 |
Pos logit contributions | |
ens | +0.033 |
Lee | +0.031 |
riet | +0.031 |
handing | +0.030 |
purs | +0.030 |
Neg logit contributions | |
most | -0.028 |
biggest | -0.028 |
whole | -0.027 |
best | -0.026 |
world | -0.026 |
Token | time |
Token ID | 640 |
Feature activation | +6.16e-02 |
Pos logit contributions | |
arest | +0.098 |
ifying | +0.093 |
highest | +0.092 |
hest | +0.092 |
fastest | +0.090 |
Neg logit contributions | |
Sue | -0.128 |
Sally | -0.126 |
ens | -0.123 |
her | -0.122 |
Sophie | -0.122 |
Token | to |
Token ID | 284 |
Feature activation | -6.54e-02 |
Pos logit contributions | |
most | +0.119 |
biggest | +0.119 |
fastest | +0.116 |
highest | +0.116 |
best | +0.112 |
Neg logit contributions | |
ens | -0.091 |
Lee | -0.089 |
riet | -0.084 |
handing | -0.083 |
kids | -0.083 |
Token | wash |
Token ID | 13502 |
Feature activation | -2.32e-02 |
Pos logit contributions | |
ens | +0.097 |
Lee | +0.091 |
Sophie | +0.090 |
Sue | +0.090 |
handing | +0.090 |
Neg logit contributions | |
biggest | -0.125 |
most | -0.124 |
highest | -0.122 |
fastest | -0.122 |
best | -0.116 |
Token | up |
Token ID | 510 |
Feature activation | -1.42e-02 |
Pos logit contributions | |
ens | +0.033 |
Lee | +0.032 |
riet | +0.030 |
Sophie | +0.030 |
Sue | +0.030 |
Neg logit contributions | |
biggest | -0.046 |
most | -0.046 |
fastest | -0.045 |
highest | -0.045 |
whole | -0.044 |
Token | ." |
Token ID | 526 |
Feature activation | -4.06e-02 |
Pos logit contributions | |
ens | +0.022 |
Lee | +0.022 |
riet | +0.021 |
kids | +0.021 |
handing | +0.020 |
Neg logit contributions | |
most | -0.026 |
biggest | -0.026 |
highest | -0.025 |
fastest | -0.025 |
whole | -0.025 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.81e-01 |
Pos logit contributions | |
riet | +0.061 |
kids | +0.060 |
ens | +0.060 |
pa | +0.060 |
purs | +0.058 |
Neg logit contributions | |
most | -0.078 |
biggest | -0.074 |
best | -0.073 |
whole | -0.070 |
world | -0.068 |
Token | i |
Token ID | 72 |
Feature activation | -7.87e-02 |
Pos logit contributions | |
Amy | +0.223 |
handing | +0.221 |
Lee | +0.213 |
Sue | +0.209 |
blamed | +0.200 |
Neg logit contributions | |
most | -0.387 |
best | -0.382 |
biggest | -0.375 |
whole | -0.373 |
truly | -0.372 |
Token | didn |
Token ID | 1422 |
Feature activation | +5.79e-02 |
Pos logit contributions | |
Lee | +0.130 |
riet | +0.128 |
ens | +0.127 |
handing | +0.124 |
Sophie | +0.122 |
Neg logit contributions | |
most | -0.139 |
biggest | -0.134 |
best | -0.133 |
whole | -0.131 |
truly | -0.127 |
Token | 't |
Token ID | 470 |
Feature activation | +1.60e-02 |
Pos logit contributions | |
biggest | +0.097 |
most | +0.096 |
fastest | +0.092 |
coolest | +0.092 |
highest | +0.088 |
Neg logit contributions | |
Lee | -0.101 |
ens | -0.101 |
handing | -0.098 |
Tom | -0.098 |
Sue | -0.097 |
Token | want |
Token ID | 765 |
Feature activation | +1.38e-02 |
Pos logit contributions | |
biggest | +0.030 |
most | +0.030 |
highest | +0.029 |
fastest | +0.029 |
best | +0.028 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.024 |
riet | -0.023 |
Sue | -0.023 |
kids | -0.023 |
Token | to |
Token ID | 284 |
Feature activation | -7.02e-02 |
Pos logit contributions | |
most | +0.024 |
biggest | +0.023 |
best | +0.022 |
highest | +0.022 |
fastest | +0.021 |
Neg logit contributions | |
Lee | -0.024 |
riet | -0.024 |
kids | -0.024 |
ens | -0.023 |
handing | -0.023 |
Token | was |
Token ID | 373 |
Feature activation | +6.31e-02 |
Pos logit contributions | |
ens | +0.104 |
riet | +0.102 |
Lee | +0.101 |
kids | +0.098 |
pa | +0.095 |
Neg logit contributions | |
most | -0.052 |
truly | -0.049 |
best | -0.046 |
biggest | -0.045 |
whole | -0.044 |
Token | so |
Token ID | 523 |
Feature activation | +1.54e-02 |
Pos logit contributions | |
biggest | +0.088 |
most | +0.088 |
fastest | +0.085 |
highest | +0.085 |
whole | +0.081 |
Neg logit contributions | |
ens | -0.126 |
Lee | -0.125 |
riet | -0.121 |
Sue | -0.121 |
handing | -0.120 |
Token | happy |
Token ID | 3772 |
Feature activation | -2.19e-02 |
Pos logit contributions | |
most | +0.026 |
biggest | +0.026 |
whole | +0.025 |
best | +0.025 |
truly | +0.025 |
Neg logit contributions | |
ens | -0.027 |
Lee | -0.027 |
kids | -0.026 |
riet | -0.026 |
pa | -0.025 |
Token | and |
Token ID | 290 |
Feature activation | -7.67e-02 |
Pos logit contributions | |
ens | +0.040 |
kids | +0.040 |
riet | +0.040 |
Lee | +0.038 |
purs | +0.038 |
Neg logit contributions | |
most | -0.034 |
best | -0.032 |
truly | -0.032 |
biggest | -0.032 |
whole | -0.032 |
Token | thanked |
Token ID | 26280 |
Feature activation | -1.35e-01 |
Pos logit contributions | |
Lee | +0.099 |
kids | +0.097 |
pa | +0.096 |
ens | +0.094 |
olph | +0.092 |
Neg logit contributions | |
most | -0.170 |
truly | -0.164 |
biggest | -0.163 |
best | -0.160 |
perfect | -0.157 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.79e-01 |
Pos logit contributions | |
kids | +0.135 |
handing | +0.124 |
blamed | +0.123 |
riet | +0.122 |
purs | +0.122 |
Neg logit contributions | |
most | -0.328 |
best | -0.317 |
biggest | -0.314 |
truly | -0.308 |
whole | -0.302 |
Token | i |
Token ID | 72 |
Feature activation | -7.53e-02 |
Pos logit contributions | |
handing | +0.214 |
Amy | +0.212 |
blamed | +0.210 |
Lee | +0.209 |
blame | +0.204 |
Neg logit contributions | |
most | -0.386 |
best | -0.385 |
biggest | -0.380 |
whole | -0.372 |
truly | -0.368 |
Token | and |
Token ID | 290 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
kids | +0.142 |
ens | +0.141 |
Lee | +0.141 |
riet | +0.140 |
Samantha | +0.138 |
Neg logit contributions | |
most | -0.113 |
best | -0.112 |
biggest | -0.111 |
whole | -0.105 |
highest | -0.105 |
Token | Duck |
Token ID | 21867 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
kids | +0.159 |
riet | +0.149 |
pa | +0.148 |
Lee | +0.147 |
olph | +0.145 |
Neg logit contributions | |
most | -0.207 |
best | -0.204 |
truly | -0.196 |
biggest | -0.191 |
perfect | -0.190 |
Token | ie |
Token ID | 494 |
Feature activation | -5.88e-02 |
Pos logit contributions | |
Lee | +0.122 |
handing | +0.115 |
riet | +0.112 |
Samantha | +0.107 |
kids | +0.106 |
Neg logit contributions | |
best | -0.243 |
most | -0.239 |
whole | -0.236 |
biggest | -0.232 |
truly | -0.224 |
Token | . |
Token ID | 13 |
Feature activation | -4.80e-03 |
Pos logit contributions | |
ens | +0.111 |
Lee | +0.110 |
riet | +0.107 |
kids | +0.107 |
Sophie | +0.105 |
Neg logit contributions | |
most | -0.091 |
biggest | -0.089 |
best | -0.087 |
whole | -0.086 |
highest | -0.084 |
Token | !" |
Token ID | 2474 |
Feature activation | -4.25e-02 |
Pos logit contributions | |
most | +0.196 |
biggest | +0.195 |
fastest | +0.190 |
coolest | +0.188 |
highest | +0.188 |
Neg logit contributions | |
ens | -0.175 |
Sue | -0.171 |
Lee | -0.169 |
Tom | -0.168 |
Amy | -0.167 |
Token | |
Token ID | 220 |
Feature activation | +7.87e-02 |
Pos logit contributions | |
riet | +0.064 |
ens | +0.064 |
Lee | +0.063 |
kids | +0.063 |
handing | +0.061 |
Neg logit contributions | |
most | -0.081 |
biggest | -0.078 |
best | -0.077 |
whole | -0.074 |
highest | -0.073 |
Token |
|
Token ID | 198 |
Feature activation | +3.96e-02 |
Pos logit contributions | |
most | +0.120 |
biggest | +0.119 |
fastest | +0.114 |
highest | +0.114 |
best | +0.112 |
Neg logit contributions | |
ens | -0.148 |
Lee | -0.145 |
riet | -0.142 |
kids | -0.140 |
handing | -0.139 |
Token |
|
Token ID | 198 |
Feature activation | +3.77e-02 |
Pos logit contributions | |
most | +0.050 |
biggest | +0.048 |
whole | +0.047 |
best | +0.046 |
truly | +0.046 |
Neg logit contributions | |
Lee | -0.085 |
ens | -0.085 |
riet | -0.083 |
handing | -0.083 |
kids | -0.081 |
Token | So |
Token ID | 2396 |
Feature activation | -3.98e-02 |
Pos logit contributions | |
most | +0.065 |
biggest | +0.063 |
whole | +0.062 |
highest | +0.061 |
best | +0.060 |
Neg logit contributions | |
Lee | -0.067 |
ens | -0.065 |
riet | -0.063 |
handing | -0.062 |
kids | -0.062 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
kids | +0.039 |
Lee | +0.038 |
handing | +0.037 |
ens | +0.036 |
riet | +0.035 |
Neg logit contributions | |
most | -0.099 |
best | -0.096 |
biggest | -0.096 |
whole | -0.093 |
highest | -0.092 |
Token | i |
Token ID | 72 |
Feature activation | -8.09e-02 |
Pos logit contributions | |
Amy | +0.222 |
handing | +0.216 |
Sue | +0.210 |
Lee | +0.209 |
blamed | +0.202 |
Neg logit contributions | |
most | -0.382 |
best | -0.381 |
whole | -0.374 |
biggest | -0.370 |
truly | -0.365 |
Token | went |
Token ID | 1816 |
Feature activation | -9.26e-02 |
Pos logit contributions | |
Lee | +0.128 |
riet | +0.125 |
ens | +0.121 |
Sophie | +0.120 |
handing | +0.120 |
Neg logit contributions | |
most | -0.149 |
biggest | -0.145 |
best | -0.143 |
whole | -0.142 |
fastest | -0.140 |
Token | to |
Token ID | 284 |
Feature activation | -1.09e-01 |
Pos logit contributions | |
riet | +0.182 |
kids | +0.181 |
Lee | +0.177 |
ens | +0.177 |
pa | +0.173 |
Neg logit contributions | |
whole | -0.136 |
biggest | -0.133 |
most | -0.132 |
highest | -0.131 |
best | -0.128 |
Token | the |
Token ID | 262 |
Feature activation | -6.68e-03 |
Pos logit contributions | |
riet | +0.189 |
kids | +0.188 |
ens | +0.181 |
Lee | +0.178 |
ppa | +0.178 |
Neg logit contributions | |
most | -0.188 |
whole | -0.188 |
biggest | -0.185 |
highest | -0.175 |
perfect | -0.174 |
Token | river |
Token ID | 7850 |
Feature activation | -6.14e-03 |
Pos logit contributions | |
ens | +0.010 |
Lee | +0.010 |
handing | +0.009 |
Sue | +0.009 |
Sophie | +0.009 |
Neg logit contributions | |
most | -0.012 |
fastest | -0.012 |
biggest | -0.012 |
truly | -0.012 |
possible | -0.012 |
Token | to |
Token ID | 284 |
Feature activation | -7.86e-02 |
Pos logit contributions | |
most | +0.176 |
biggest | +0.176 |
highest | +0.170 |
fastest | +0.169 |
whole | +0.167 |
Neg logit contributions | |
ens | -0.152 |
Lee | -0.150 |
riet | -0.144 |
kids | -0.142 |
Sophie | -0.140 |
Token | Mickey |
Token ID | 27232 |
Feature activation | -1.12e-01 |
Pos logit contributions | |
ens | +0.077 |
Lee | +0.076 |
kids | +0.074 |
riet | +0.072 |
purs | +0.070 |
Neg logit contributions | |
most | -0.192 |
biggest | -0.190 |
whole | -0.184 |
best | -0.184 |
highest | -0.183 |
Token | . |
Token ID | 13 |
Feature activation | +7.25e-03 |
Pos logit contributions | |
ens | +0.211 |
Lee | +0.209 |
kids | +0.209 |
riet | +0.206 |
Samantha | +0.201 |
Neg logit contributions | |
most | -0.167 |
biggest | -0.162 |
best | -0.160 |
whole | -0.159 |
truly | -0.152 |
Token |
|
Token ID | 198 |
Feature activation | +5.43e-02 |
Pos logit contributions | |
most | +0.012 |
biggest | +0.012 |
highest | +0.011 |
whole | +0.011 |
best | +0.011 |
Neg logit contributions | |
ens | -0.013 |
kids | -0.012 |
Lee | -0.012 |
pa | -0.012 |
riet | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | +4.92e-02 |
Pos logit contributions | |
most | +0.073 |
biggest | +0.073 |
whole | +0.071 |
highest | +0.069 |
fastest | +0.068 |
Neg logit contributions | |
ens | -0.111 |
Lee | -0.111 |
riet | -0.107 |
kids | -0.107 |
purs | -0.105 |
Token | Mic |
Token ID | 25437 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
biggest | +0.086 |
most | +0.086 |
fastest | +0.083 |
highest | +0.083 |
best | +0.082 |
Neg logit contributions | |
ens | -0.082 |
Lee | -0.079 |
riet | -0.078 |
handing | -0.077 |
Sophie | -0.076 |
Token | key |
Token ID | 2539 |
Feature activation | -1.05e-01 |
Pos logit contributions | |
Lee | +0.247 |
Sue | +0.243 |
Sophie | +0.237 |
ens | +0.235 |
kids | +0.234 |
Neg logit contributions | |
most | -0.342 |
fastest | -0.334 |
truly | -0.328 |
best | -0.327 |
whole | -0.325 |
Token | was |
Token ID | 373 |
Feature activation | +6.92e-02 |
Pos logit contributions | |
ens | +0.202 |
Lee | +0.200 |
riet | +0.198 |
kids | +0.190 |
Samantha | +0.188 |
Neg logit contributions | |
most | -0.158 |
truly | -0.148 |
whole | -0.146 |
best | -0.144 |
biggest | -0.144 |
Token | very |
Token ID | 845 |
Feature activation | +4.03e-02 |
Pos logit contributions | |
most | +0.090 |
biggest | +0.089 |
fastest | +0.085 |
highest | +0.084 |
best | +0.083 |
Neg logit contributions | |
ens | -0.151 |
Lee | -0.147 |
riet | -0.145 |
kids | -0.143 |
Sophie | -0.139 |
Token | happy |
Token ID | 3772 |
Feature activation | -8.02e-03 |
Pos logit contributions | |
most | +0.064 |
biggest | +0.063 |
best | +0.063 |
whole | +0.061 |
truly | +0.061 |
Neg logit contributions | |
ens | -0.080 |
Lee | -0.076 |
kids | -0.076 |
riet | -0.074 |
pa | -0.073 |
Token | and |
Token ID | 290 |
Feature activation | -6.38e-02 |
Pos logit contributions | |
ens | +0.013 |
riet | +0.012 |
Lee | +0.012 |
kids | +0.012 |
purs | +0.012 |
Neg logit contributions | |
most | -0.015 |
biggest | -0.014 |
best | -0.014 |
whole | -0.014 |
fastest | -0.014 |
Token | 't |
Token ID | 470 |
Feature activation | +2.99e-02 |
Pos logit contributions | |
most | +0.031 |
biggest | +0.031 |
fastest | +0.029 |
highest | +0.029 |
best | +0.028 |
Neg logit contributions | |
ens | -0.051 |
Lee | -0.051 |
riet | -0.049 |
kids | -0.049 |
handing | -0.049 |
Token | find |
Token ID | 1064 |
Feature activation | +1.86e-03 |
Pos logit contributions | |
most | +0.061 |
biggest | +0.060 |
fastest | +0.059 |
highest | +0.059 |
best | +0.057 |
Neg logit contributions | |
ens | -0.042 |
Lee | -0.041 |
riet | -0.039 |
kids | -0.038 |
handing | -0.038 |
Token | it |
Token ID | 340 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
most | +0.003 |
biggest | +0.003 |
fastest | +0.003 |
highest | +0.003 |
best | +0.003 |
Neg logit contributions | |
ens | -0.003 |
Lee | -0.003 |
riet | -0.003 |
kids | -0.003 |
handing | -0.003 |
Token | anywhere |
Token ID | 6609 |
Feature activation | +6.72e-02 |
Pos logit contributions | |
most | +0.172 |
biggest | +0.171 |
fastest | +0.165 |
highest | +0.164 |
best | +0.160 |
Neg logit contributions | |
ens | -0.218 |
Lee | -0.215 |
riet | -0.209 |
kids | -0.207 |
handing | -0.205 |
Token | ." |
Token ID | 526 |
Feature activation | -3.04e-02 |
Pos logit contributions | |
most | +0.093 |
biggest | +0.092 |
fastest | +0.089 |
highest | +0.087 |
best | +0.087 |
Neg logit contributions | |
ens | -0.137 |
Lee | -0.137 |
riet | -0.133 |
kids | -0.132 |
Sophie | -0.129 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
kids | +0.036 |
riet | +0.034 |
pa | +0.033 |
purs | +0.033 |
ens | +0.031 |
Neg logit contributions | |
most | -0.068 |
biggest | -0.065 |
best | -0.064 |
highest | -0.063 |
truly | -0.062 |
Token | i |
Token ID | 72 |
Feature activation | -7.60e-02 |
Pos logit contributions | |
Lee | +0.207 |
handing | +0.204 |
Amy | +0.203 |
kids | +0.192 |
Sue | +0.190 |
Neg logit contributions | |
most | -0.387 |
best | -0.379 |
biggest | -0.377 |
whole | -0.374 |
truly | -0.371 |
Token | said |
Token ID | 531 |
Feature activation | -7.67e-02 |
Pos logit contributions | |
Lee | +0.121 |
riet | +0.117 |
ens | +0.117 |
kids | +0.116 |
handing | +0.114 |
Neg logit contributions | |
most | -0.139 |
best | -0.130 |
truly | -0.129 |
whole | -0.126 |
biggest | -0.124 |
Token | , |
Token ID | 11 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
kids | +0.125 |
ens | +0.120 |
riet | +0.118 |
Lee | +0.117 |
pa | +0.116 |
Neg logit contributions | |
most | -0.145 |
best | -0.144 |
biggest | -0.141 |
whole | -0.134 |
truly | -0.133 |
Token | " |
Token ID | 366 |
Feature activation | +3.92e-02 |
Pos logit contributions | |
pa | +0.250 |
riet | +0.246 |
ens | +0.244 |
purs | +0.243 |
kids | +0.242 |
Neg logit contributions | |
most | -0.166 |
whole | -0.149 |
best | -0.149 |
biggest | -0.149 |
truly | -0.142 |
Token | Don |
Token ID | 3987 |
Feature activation | +3.77e-02 |
Pos logit contributions | |
biggest | +0.067 |
best | +0.065 |
ifying | +0.063 |
fastest | +0.062 |
highest | +0.062 |
Neg logit contributions | |
ens | -0.064 |
stro | -0.063 |
Sue | -0.063 |
riet | -0.062 |
Lee | -0.061 |
Token | vanished |
Token ID | 23717 |
Feature activation | -8.34e-03 |
Pos logit contributions | |
ens | +0.180 |
riet | +0.176 |
Lee | +0.176 |
kids | +0.170 |
Sophie | +0.165 |
Neg logit contributions | |
most | -0.187 |
whole | -0.185 |
biggest | -0.183 |
truly | -0.178 |
best | -0.178 |
Token | ! |
Token ID | 0 |
Feature activation | +1.90e-02 |
Pos logit contributions | |
Lee | +0.012 |
ens | +0.012 |
Sophie | +0.012 |
Sue | +0.012 |
handing | +0.011 |
Neg logit contributions | |
fastest | -0.016 |
biggest | -0.015 |
most | -0.015 |
arest | -0.015 |
highest | -0.015 |
Token | |
Token ID | 220 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
most | +0.039 |
biggest | +0.039 |
highest | +0.038 |
whole | +0.037 |
fastest | +0.037 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.025 |
kids | -0.024 |
riet | -0.024 |
purs | -0.024 |
Token |
|
Token ID | 198 |
Feature activation | +5.47e-02 |
Pos logit contributions | |
fastest | +0.198 |
most | +0.197 |
biggest | +0.197 |
highest | +0.195 |
ifying | +0.188 |
Neg logit contributions | |
ens | -0.150 |
Sue | -0.147 |
Lee | -0.145 |
Tom | -0.145 |
Amy | -0.143 |
Token |
|
Token ID | 198 |
Feature activation | +4.94e-02 |
Pos logit contributions | |
most | +0.076 |
biggest | +0.075 |
whole | +0.073 |
highest | +0.072 |
fastest | +0.071 |
Neg logit contributions | |
ens | -0.110 |
Lee | -0.110 |
kids | -0.106 |
riet | -0.106 |
handing | -0.104 |
Token | Mic |
Token ID | 25437 |
Feature activation | -1.76e-01 |
Pos logit contributions | |
most | +0.114 |
biggest | +0.114 |
fastest | +0.111 |
highest | +0.111 |
best | +0.109 |
Neg logit contributions | |
ens | -0.055 |
Lee | -0.053 |
riet | -0.051 |
kids | -0.049 |
handing | -0.049 |
Token | key |
Token ID | 2539 |
Feature activation | -1.08e-01 |
Pos logit contributions | |
Lee | +0.229 |
ens | +0.224 |
Sue | +0.222 |
Sophie | +0.219 |
kids | +0.217 |
Neg logit contributions | |
most | -0.362 |
fastest | -0.352 |
biggest | -0.349 |
whole | -0.348 |
best | -0.346 |
Token | was |
Token ID | 373 |
Feature activation | +6.63e-02 |
Pos logit contributions | |
ens | +0.224 |
Lee | +0.220 |
riet | +0.214 |
kids | +0.211 |
handing | +0.209 |
Neg logit contributions | |
most | -0.149 |
biggest | -0.142 |
whole | -0.138 |
truly | -0.137 |
best | -0.136 |
Token | scared |
Token ID | 12008 |
Feature activation | -1.60e-02 |
Pos logit contributions | |
biggest | +0.089 |
fastest | +0.089 |
highest | +0.088 |
most | +0.086 |
possible | +0.082 |
Neg logit contributions | |
Sue | -0.132 |
sob | -0.130 |
Lee | -0.130 |
handing | -0.130 |
hugged | -0.128 |
Token | , |
Token ID | 11 |
Feature activation | -9.93e-02 |
Pos logit contributions | |
kids | +0.026 |
Lee | +0.025 |
riet | +0.025 |
purs | +0.025 |
ens | +0.024 |
Neg logit contributions | |
most | -0.029 |
best | -0.028 |
biggest | -0.028 |
whole | -0.027 |
fastest | -0.027 |
Token | but |
Token ID | 475 |
Feature activation | -4.67e-03 |
Pos logit contributions | |
ens | +0.153 |
riet | +0.150 |
Lee | +0.150 |
kids | +0.150 |
purs | +0.149 |
Neg logit contributions | |
most | -0.189 |
biggest | -0.179 |
best | -0.174 |
fastest | -0.173 |
highest | -0.173 |
Token | play |
Token ID | 711 |
Feature activation | -2.85e-02 |
Pos logit contributions | |
most | +0.033 |
biggest | +0.033 |
fastest | +0.032 |
highest | +0.032 |
best | +0.032 |
Neg logit contributions | |
ens | -0.026 |
Lee | -0.025 |
riet | -0.025 |
kids | -0.024 |
handing | -0.024 |
Token | with |
Token ID | 351 |
Feature activation | +5.96e-03 |
Pos logit contributions | |
Lee | +0.043 |
ens | +0.042 |
Sue | +0.042 |
Amy | +0.041 |
handing | +0.041 |
Neg logit contributions | |
biggest | -0.053 |
fastest | -0.052 |
most | -0.052 |
highest | -0.051 |
ifying | -0.050 |
Token | you |
Token ID | 345 |
Feature activation | +5.43e-02 |
Pos logit contributions | |
most | +0.010 |
biggest | +0.010 |
best | +0.009 |
highest | +0.009 |
fastest | +0.009 |
Neg logit contributions | |
ens | -0.011 |
riet | -0.010 |
Lee | -0.010 |
kids | -0.010 |
handing | -0.010 |
Token | too |
Token ID | 1165 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
most | +0.088 |
biggest | +0.087 |
fastest | +0.085 |
highest | +0.084 |
best | +0.083 |
Neg logit contributions | |
ens | -0.098 |
Lee | -0.096 |
riet | -0.094 |
kids | -0.093 |
Sophie | -0.092 |
Token | ." |
Token ID | 526 |
Feature activation | -3.34e-02 |
Pos logit contributions | |
most | +0.176 |
biggest | +0.175 |
fastest | +0.169 |
highest | +0.168 |
best | +0.163 |
Neg logit contributions | |
ens | -0.186 |
Lee | -0.184 |
riet | -0.176 |
Sue | -0.176 |
kids | -0.174 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.72e-01 |
Pos logit contributions | |
pa | +0.051 |
kids | +0.051 |
riet | +0.050 |
purs | +0.049 |
ens | +0.049 |
Neg logit contributions | |
most | -0.062 |
biggest | -0.061 |
best | -0.058 |
highest | -0.058 |
whole | -0.055 |
Token | i |
Token ID | 72 |
Feature activation | -7.41e-02 |
Pos logit contributions | |
handing | +0.179 |
Lee | +0.177 |
Amy | +0.175 |
kids | +0.172 |
Sue | +0.169 |
Neg logit contributions | |
most | -0.393 |
biggest | -0.382 |
best | -0.382 |
whole | -0.377 |
truly | -0.376 |
Token | , |
Token ID | 11 |
Feature activation | -1.25e-01 |
Pos logit contributions | |
riet | +0.122 |
ens | +0.119 |
Lee | +0.117 |
kids | +0.115 |
Samantha | +0.114 |
Neg logit contributions | |
most | -0.134 |
best | -0.127 |
biggest | -0.123 |
whole | -0.122 |
truly | -0.121 |
Token | Bree |
Token ID | 40110 |
Feature activation | -1.09e-01 |
Pos logit contributions | |
riet | +0.179 |
kids | +0.172 |
annah | +0.170 |
pa | +0.166 |
lee | +0.163 |
Neg logit contributions | |
most | -0.253 |
best | -0.238 |
truly | -0.229 |
whole | -0.221 |
perfect | -0.219 |
Token | , |
Token ID | 11 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
riet | +0.164 |
Lee | +0.163 |
kids | +0.158 |
handing | +0.154 |
ens | +0.152 |
Neg logit contributions | |
most | -0.203 |
best | -0.198 |
fastest | -0.193 |
whole | -0.193 |
biggest | -0.193 |
Token | and |
Token ID | 290 |
Feature activation | -9.96e-02 |
Pos logit contributions | |
riet | +0.219 |
annah | +0.217 |
kids | +0.210 |
lee | +0.209 |
blame | +0.204 |
Neg logit contributions | |
most | -0.196 |
best | -0.181 |
whole | -0.173 |
truly | -0.170 |
perfect | -0.166 |
Token | I |
Token ID | 314 |
Feature activation | +1.63e-02 |
Pos logit contributions | |
ens | +0.056 |
kids | +0.055 |
Lee | +0.055 |
riet | +0.055 |
pa | +0.053 |
Neg logit contributions | |
most | -0.031 |
biggest | -0.029 |
fastest | -0.028 |
highest | -0.028 |
best | -0.027 |
Token | play |
Token ID | 711 |
Feature activation | -2.62e-02 |
Pos logit contributions | |
biggest | +0.032 |
most | +0.032 |
fastest | +0.031 |
highest | +0.031 |
best | +0.031 |
Neg logit contributions | |
ens | -0.023 |
Lee | -0.022 |
Sue | -0.022 |
handing | -0.021 |
Sophie | -0.021 |
Token | with |
Token ID | 351 |
Feature activation | +2.82e-03 |
Pos logit contributions | |
ens | +0.045 |
Lee | +0.044 |
riet | +0.042 |
kids | +0.042 |
handing | +0.042 |
Neg logit contributions | |
most | -0.045 |
biggest | -0.044 |
fastest | -0.043 |
highest | -0.043 |
best | -0.042 |
Token | you |
Token ID | 345 |
Feature activation | +4.97e-02 |
Pos logit contributions | |
most | +0.005 |
biggest | +0.004 |
highest | +0.004 |
fastest | +0.004 |
best | +0.004 |
Neg logit contributions | |
Lee | -0.005 |
ens | -0.005 |
riet | -0.005 |
kids | -0.005 |
sob | -0.005 |
Token | ?" |
Token ID | 1701 |
Feature activation | -3.46e-02 |
Pos logit contributions | |
most | +0.088 |
biggest | +0.086 |
fastest | +0.083 |
highest | +0.083 |
best | +0.083 |
Neg logit contributions | |
ens | -0.084 |
Lee | -0.082 |
riet | -0.082 |
kids | -0.079 |
Sue | -0.078 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.71e-01 |
Pos logit contributions | |
kids | +0.057 |
riet | +0.056 |
pa | +0.053 |
purs | +0.051 |
ens | +0.051 |
Neg logit contributions | |
most | -0.062 |
biggest | -0.057 |
best | -0.057 |
truly | -0.055 |
highest | -0.054 |
Token | i |
Token ID | 72 |
Feature activation | -7.18e-02 |
Pos logit contributions | |
handing | +0.182 |
Lee | +0.180 |
Amy | +0.175 |
kids | +0.172 |
Sue | +0.171 |
Neg logit contributions | |
most | -0.387 |
biggest | -0.377 |
truly | -0.375 |
best | -0.374 |
fastest | -0.372 |
Token | was |
Token ID | 373 |
Feature activation | +5.73e-02 |
Pos logit contributions | |
ens | +0.103 |
riet | +0.102 |
kids | +0.101 |
Lee | +0.101 |
purs | +0.097 |
Neg logit contributions | |
most | -0.143 |
best | -0.136 |
truly | -0.134 |
biggest | -0.133 |
whole | -0.131 |
Token | happy |
Token ID | 3772 |
Feature activation | -1.49e-02 |
Pos logit contributions | |
most | +0.082 |
biggest | +0.081 |
fastest | +0.077 |
highest | +0.077 |
best | +0.076 |
Neg logit contributions | |
ens | -0.117 |
Lee | -0.114 |
riet | -0.112 |
kids | -0.111 |
handing | -0.108 |
Token | and |
Token ID | 290 |
Feature activation | -7.04e-02 |
Pos logit contributions | |
riet | +0.032 |
kids | +0.032 |
purs | +0.030 |
ppa | +0.030 |
ens | +0.030 |
Neg logit contributions | |
most | -0.020 |
best | -0.019 |
truly | -0.018 |
world | -0.018 |
whole | -0.017 |
Token | said |
Token ID | 531 |
Feature activation | -6.49e-02 |
Pos logit contributions | |
Lee | +0.093 |
pa | +0.089 |
kids | +0.089 |
ens | +0.087 |
olph | +0.086 |
Neg logit contributions | |
most | -0.157 |
biggest | -0.153 |
truly | -0.149 |
best | -0.149 |
highest | -0.145 |
Token |
|
Token ID | 198 |
Feature activation | +5.41e-02 |
Pos logit contributions | |
most | +0.107 |
biggest | +0.106 |
fastest | +0.103 |
highest | +0.103 |
best | +0.101 |
Neg logit contributions | |
ens | -0.096 |
Lee | -0.094 |
riet | -0.091 |
kids | -0.090 |
handing | -0.089 |
Token | The |
Token ID | 464 |
Feature activation | +2.95e-02 |
Pos logit contributions | |
biggest | +0.100 |
fastest | +0.095 |
highest | +0.094 |
most | +0.092 |
coolest | +0.089 |
Neg logit contributions | |
Amy | -0.083 |
kids | -0.083 |
pa | -0.081 |
child | -0.080 |
old | -0.080 |
Token | end |
Token ID | 886 |
Feature activation | +6.80e-02 |
Pos logit contributions | |
biggest | +0.043 |
ifying | +0.043 |
highest | +0.043 |
most | +0.042 |
fastest | +0.042 |
Neg logit contributions | |
ens | -0.055 |
Lee | -0.053 |
Sue | -0.052 |
kids | -0.052 |
Sophie | -0.051 |
Token | . |
Token ID | 13 |
Feature activation | +3.49e-02 |
Pos logit contributions | |
biggest | +0.143 |
fastest | +0.141 |
highest | +0.139 |
most | +0.138 |
coolest | +0.132 |
Neg logit contributions | |
Amy | -0.095 |
Sue | -0.094 |
Lee | -0.093 |
ens | -0.093 |
Mary | -0.089 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +2.01e-02 |
Pos logit contributions | |
fastest | +0.082 |
biggest | +0.081 |
highest | +0.080 |
tallest | +0.076 |
most | +0.075 |
Neg logit contributions | |
Tom | -0.044 |
Mrs | -0.044 |
Lee | -0.043 |
Amy | -0.043 |
Sophie | -0.041 |
Token | Mon |
Token ID | 9069 |
Feature activation | -1.71e-01 |
Pos logit contributions | |
biggest | +0.035 |
fastest | +0.035 |
highest | +0.034 |
most | +0.034 |
coolest | +0.033 |
Neg logit contributions | |
Lee | -0.034 |
ens | -0.034 |
Sue | -0.033 |
Sophie | -0.033 |
Amy | -0.032 |
Token | key |
Token ID | 2539 |
Feature activation | -1.29e-01 |
Pos logit contributions | |
ens | +0.240 |
Lee | +0.231 |
riet | +0.228 |
handing | +0.220 |
Sophie | +0.215 |
Neg logit contributions | |
most | -0.357 |
biggest | -0.342 |
best | -0.340 |
fastest | -0.338 |
truly | -0.338 |
Token | was |
Token ID | 373 |
Feature activation | +4.73e-02 |
Pos logit contributions | |
Lee | +0.234 |
handing | +0.226 |
ens | +0.223 |
riet | +0.221 |
kids | +0.219 |
Neg logit contributions | |
most | -0.203 |
whole | -0.192 |
truly | -0.192 |
biggest | -0.187 |
best | -0.186 |
Token | very |
Token ID | 845 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
biggest | +0.076 |
most | +0.076 |
fastest | +0.074 |
highest | +0.074 |
whole | +0.071 |
Neg logit contributions | |
ens | -0.084 |
Lee | -0.083 |
riet | -0.080 |
kids | -0.079 |
Sophie | -0.079 |
Token | excited |
Token ID | 6568 |
Feature activation | -4.47e-02 |
Pos logit contributions | |
most | +0.038 |
biggest | +0.038 |
fastest | +0.037 |
highest | +0.037 |
whole | +0.036 |
Neg logit contributions | |
ens | -0.028 |
Lee | -0.028 |
riet | -0.027 |
kids | -0.026 |
Sophie | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | +2.49e-02 |
Pos logit contributions | |
ens | +0.072 |
Lee | +0.071 |
riet | +0.068 |
kids | +0.067 |
Sophie | +0.067 |
Neg logit contributions | |
biggest | -0.081 |
most | -0.080 |
fastest | -0.079 |
highest | -0.078 |
best | -0.076 |
Token | a |
Token ID | 257 |
Feature activation | +9.97e-03 |
Pos logit contributions | |
Sue | +0.261 |
ens | +0.261 |
Lee | +0.253 |
Amy | +0.252 |
Sophie | +0.250 |
Neg logit contributions | |
biggest | -0.196 |
fastest | -0.195 |
highest | -0.192 |
most | -0.189 |
truly | -0.178 |
Token | tall |
Token ID | 7331 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
most | +0.017 |
biggest | +0.017 |
fastest | +0.017 |
highest | +0.017 |
truly | +0.016 |
Neg logit contributions | |
ens | -0.017 |
Lee | -0.016 |
Sue | -0.015 |
riet | -0.015 |
Sophie | -0.015 |
Token | gir |
Token ID | 37370 |
Feature activation | -5.36e-02 |
Pos logit contributions | |
riet | +0.029 |
ens | +0.028 |
handing | +0.028 |
stro | +0.028 |
Lee | +0.028 |
Neg logit contributions | |
most | -0.028 |
biggest | -0.027 |
whole | -0.027 |
best | -0.027 |
world | -0.027 |
Token | affe |
Token ID | 21223 |
Feature activation | -6.57e-02 |
Pos logit contributions | |
ens | +0.080 |
Lee | +0.079 |
riet | +0.076 |
kids | +0.075 |
handing | +0.074 |
Neg logit contributions | |
most | -0.102 |
biggest | -0.102 |
fastest | -0.099 |
highest | -0.098 |
best | -0.096 |
Token | named |
Token ID | 3706 |
Feature activation | -1.32e-01 |
Pos logit contributions | |
Lee | +0.066 |
riet | +0.065 |
kids | +0.063 |
ens | +0.063 |
Sophie | +0.062 |
Neg logit contributions | |
most | -0.157 |
whole | -0.154 |
biggest | -0.152 |
world | -0.151 |
best | -0.150 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
stro | +0.132 |
sob | +0.126 |
kids | +0.120 |
handing | +0.120 |
pa | +0.117 |
Neg logit contributions | |
biggest | -0.333 |
most | -0.328 |
highest | -0.326 |
whole | -0.319 |
fastest | -0.316 |
Token | i |
Token ID | 72 |
Feature activation | -7.03e-02 |
Pos logit contributions | |
handing | +0.213 |
Lee | +0.195 |
Amy | +0.184 |
blame | +0.182 |
kids | +0.178 |
Neg logit contributions | |
whole | -0.363 |
most | -0.361 |
best | -0.354 |
truly | -0.352 |
world | -0.351 |
Token | went |
Token ID | 1816 |
Feature activation | -9.01e-02 |
Pos logit contributions | |
ens | +0.125 |
Lee | +0.124 |
riet | +0.120 |
kids | +0.118 |
handing | +0.118 |
Neg logit contributions | |
most | -0.115 |
biggest | -0.114 |
fastest | -0.109 |
whole | -0.109 |
highest | -0.109 |
Token | to |
Token ID | 284 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
Lee | +0.158 |
ens | +0.158 |
riet | +0.155 |
kids | +0.154 |
Sophie | +0.151 |
Neg logit contributions | |
most | -0.150 |
biggest | -0.150 |
highest | -0.146 |
fastest | -0.144 |
best | -0.142 |
Token | the |
Token ID | 262 |
Feature activation | +2.49e-03 |
Pos logit contributions | |
ens | +0.162 |
riet | +0.162 |
Lee | +0.160 |
kids | +0.157 |
handing | +0.155 |
Neg logit contributions | |
most | -0.189 |
biggest | -0.186 |
whole | -0.185 |
highest | -0.180 |
fastest | -0.175 |
Token | stage |
Token ID | 3800 |
Feature activation | +2.18e-02 |
Pos logit contributions | |
most | +0.004 |
biggest | +0.004 |
fastest | +0.004 |
ifying | +0.004 |
highest | +0.004 |
Neg logit contributions | |
Lee | -0.004 |
ens | -0.004 |
Sue | -0.004 |
kids | -0.004 |
Sophie | -0.004 |
Token | was |
Token ID | 373 |
Feature activation | +7.89e-02 |
Pos logit contributions | |
biggest | +0.019 |
most | +0.018 |
fastest | +0.018 |
highest | +0.018 |
whole | +0.017 |
Neg logit contributions | |
ens | -0.024 |
Lee | -0.024 |
Sue | -0.023 |
Sophie | -0.022 |
pa | -0.022 |
Token | Steve |
Token ID | 6542 |
Feature activation | -4.98e-02 |
Pos logit contributions | |
most | +0.197 |
biggest | +0.197 |
fastest | +0.193 |
highest | +0.193 |
whole | +0.190 |
Neg logit contributions | |
ens | -0.071 |
Lee | -0.069 |
riet | -0.064 |
Sue | -0.062 |
Sophie | -0.062 |
Token | . |
Token ID | 13 |
Feature activation | +3.06e-03 |
Pos logit contributions | |
ens | +0.091 |
Lee | +0.089 |
riet | +0.087 |
kids | +0.086 |
handing | +0.085 |
Neg logit contributions | |
most | -0.079 |
biggest | -0.078 |
fastest | -0.075 |
highest | -0.075 |
best | -0.074 |
Token |
|
Token ID | 198 |
Feature activation | +4.63e-02 |
Pos logit contributions | |
most | +0.006 |
biggest | +0.006 |
highest | +0.006 |
fastest | +0.006 |
whole | +0.006 |
Neg logit contributions | |
ens | -0.004 |
Lee | -0.004 |
riet | -0.004 |
kids | -0.004 |
handing | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +4.13e-02 |
Pos logit contributions | |
most | +0.071 |
biggest | +0.071 |
fastest | +0.068 |
highest | +0.068 |
whole | +0.067 |
Neg logit contributions | |
ens | -0.087 |
Lee | -0.085 |
riet | -0.083 |
kids | -0.082 |
handing | -0.081 |
Token | Stan |
Token ID | 32140 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
biggest | +0.082 |
most | +0.082 |
fastest | +0.079 |
highest | +0.078 |
best | +0.078 |
Neg logit contributions | |
ens | -0.058 |
Lee | -0.056 |
riet | -0.056 |
Sophie | -0.054 |
kids | -0.054 |
Token | ley |
Token ID | 1636 |
Feature activation | -8.96e-02 |
Pos logit contributions | |
Lee | +0.265 |
ens | +0.263 |
handing | +0.255 |
purs | +0.249 |
kids | +0.248 |
Neg logit contributions | |
most | -0.307 |
whole | -0.298 |
biggest | -0.294 |
truly | -0.291 |
fastest | -0.290 |
Token | and |
Token ID | 290 |
Feature activation | -1.08e-01 |
Pos logit contributions | |
ens | +0.168 |
Lee | +0.166 |
riet | +0.163 |
handing | +0.159 |
kids | +0.157 |
Neg logit contributions | |
most | -0.140 |
truly | -0.132 |
best | -0.130 |
biggest | -0.125 |
whole | -0.124 |
Token | Steve |
Token ID | 6542 |
Feature activation | -9.84e-02 |
Pos logit contributions | |
ens | +0.153 |
annah | +0.149 |
kids | +0.148 |
ppa | +0.146 |
Lee | +0.142 |
Neg logit contributions | |
most | -0.232 |
biggest | -0.214 |
whole | -0.212 |
best | -0.212 |
truly | -0.205 |
Token | got |
Token ID | 1392 |
Feature activation | +6.82e-03 |
Pos logit contributions | |
Lee | +0.167 |
ens | +0.160 |
riet | +0.156 |
handing | +0.156 |
kids | +0.156 |
Neg logit contributions | |
most | -0.175 |
truly | -0.169 |
biggest | -0.166 |
whole | -0.165 |
best | -0.165 |
Token | talking |
Token ID | 3375 |
Feature activation | -3.05e-02 |
Pos logit contributions | |
most | +0.011 |
biggest | +0.011 |
best | +0.010 |
highest | +0.010 |
fastest | +0.010 |
Neg logit contributions | |
ens | -0.013 |
Lee | -0.013 |
kids | -0.013 |
riet | -0.013 |
handing | -0.012 |
Token | the |
Token ID | 262 |
Feature activation | +8.15e-03 |
Pos logit contributions | |
Lee | +0.105 |
kids | +0.100 |
handing | +0.097 |
pa | +0.096 |
child | +0.094 |
Neg logit contributions | |
most | -0.260 |
whole | -0.252 |
highest | -0.242 |
biggest | -0.241 |
best | -0.240 |
Token | fox |
Token ID | 21831 |
Feature activation | -3.37e-02 |
Pos logit contributions | |
ifying | +0.016 |
most | +0.016 |
truly | +0.016 |
biggest | +0.016 |
highest | +0.016 |
Neg logit contributions | |
ens | -0.011 |
Lee | -0.010 |
kids | -0.010 |
Sue | -0.010 |
begged | -0.010 |
Token | away |
Token ID | 1497 |
Feature activation | -2.15e-02 |
Pos logit contributions | |
ens | +0.054 |
Lee | +0.054 |
riet | +0.052 |
kids | +0.051 |
Sophie | +0.051 |
Neg logit contributions | |
most | -0.060 |
biggest | -0.060 |
fastest | -0.058 |
highest | -0.058 |
best | -0.057 |
Token | and |
Token ID | 290 |
Feature activation | -8.46e-02 |
Pos logit contributions | |
ens | +0.024 |
Lee | +0.024 |
Sue | +0.023 |
Amy | +0.022 |
Sophie | +0.022 |
Neg logit contributions | |
biggest | -0.049 |
most | -0.048 |
fastest | -0.047 |
highest | -0.047 |
coolest | -0.046 |
Token | helped |
Token ID | 4193 |
Feature activation | -9.62e-02 |
Pos logit contributions | |
Lee | +0.120 |
kids | +0.118 |
riet | +0.116 |
ens | +0.115 |
pa | +0.112 |
Neg logit contributions | |
most | -0.171 |
biggest | -0.161 |
truly | -0.161 |
best | -0.159 |
whole | -0.159 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
Lee | +0.096 |
kids | +0.093 |
riet | +0.089 |
ens | +0.088 |
handing | +0.087 |
Neg logit contributions | |
most | -0.234 |
biggest | -0.229 |
whole | -0.225 |
best | -0.224 |
highest | -0.221 |
Token | i |
Token ID | 72 |
Feature activation | -6.34e-02 |
Pos logit contributions | |
Lee | +0.165 |
handing | +0.158 |
Amy | +0.153 |
ens | +0.153 |
kids | +0.152 |
Neg logit contributions | |
most | -0.407 |
biggest | -0.402 |
best | -0.395 |
truly | -0.394 |
fastest | -0.393 |
Token | get |
Token ID | 651 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
ens | +0.104 |
Lee | +0.102 |
riet | +0.098 |
kids | +0.097 |
handing | +0.096 |
Neg logit contributions | |
most | -0.112 |
biggest | -0.112 |
fastest | -0.109 |
highest | -0.108 |
best | -0.106 |
Token | safely |
Token ID | 11512 |
Feature activation | -2.56e-02 |
Pos logit contributions | |
most | +0.032 |
biggest | +0.031 |
best | +0.030 |
highest | +0.030 |
whole | +0.030 |
Neg logit contributions | |
Lee | -0.035 |
kids | -0.034 |
ens | -0.034 |
riet | -0.034 |
pa | -0.033 |
Token | back |
Token ID | 736 |
Feature activation | -6.78e-02 |
Pos logit contributions | |
ens | +0.041 |
Lee | +0.040 |
Sue | +0.039 |
handing | +0.039 |
Sophie | +0.038 |
Neg logit contributions | |
biggest | -0.046 |
most | -0.046 |
fastest | -0.044 |
highest | -0.044 |
best | -0.042 |
Token | to |
Token ID | 284 |
Feature activation | -7.35e-02 |
Pos logit contributions | |
ens | +0.112 |
Lee | +0.110 |
riet | +0.108 |
kids | +0.107 |
handing | +0.105 |
Neg logit contributions | |
most | -0.120 |
biggest | -0.119 |
highest | -0.116 |
fastest | -0.115 |
whole | -0.114 |
Token | to |
Token ID | 284 |
Feature activation | -6.26e-02 |
Pos logit contributions | |
most | +0.075 |
biggest | +0.075 |
fastest | +0.073 |
highest | +0.072 |
best | +0.071 |
Neg logit contributions | |
ens | -0.054 |
Lee | -0.053 |
riet | -0.050 |
Sue | -0.050 |
kids | -0.049 |
Token | come |
Token ID | 1282 |
Feature activation | -3.53e-02 |
Pos logit contributions | |
ens | +0.103 |
Lee | +0.100 |
Sue | +0.098 |
Sophie | +0.095 |
kids | +0.093 |
Neg logit contributions | |
most | -0.110 |
biggest | -0.110 |
fastest | -0.105 |
highest | -0.104 |
best | -0.102 |
Token | with |
Token ID | 351 |
Feature activation | -2.32e-03 |
Pos logit contributions | |
Lee | +0.056 |
ens | +0.056 |
Sue | +0.054 |
handing | +0.052 |
Sophie | +0.052 |
Neg logit contributions | |
biggest | -0.063 |
most | -0.062 |
fastest | -0.060 |
highest | -0.059 |
best | -0.059 |
Token | me |
Token ID | 502 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
ens | +0.004 |
Lee | +0.004 |
riet | +0.004 |
stro | +0.004 |
handing | +0.004 |
Neg logit contributions | |
most | -0.004 |
biggest | -0.004 |
highest | -0.004 |
whole | -0.004 |
fastest | -0.004 |
Token | ," |
Token ID | 553 |
Feature activation | -9.19e-02 |
Pos logit contributions | |
ens | +0.029 |
riet | +0.028 |
Lee | +0.028 |
handing | +0.027 |
kids | +0.027 |
Neg logit contributions | |
most | -0.037 |
biggest | -0.036 |
highest | -0.035 |
fastest | -0.035 |
best | -0.035 |
Token | Cic |
Token ID | 37212 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
kids | +0.114 |
riet | +0.107 |
ens | +0.104 |
Lee | +0.104 |
handing | +0.104 |
Neg logit contributions | |
most | -0.208 |
biggest | -0.199 |
best | -0.196 |
highest | -0.191 |
truly | -0.188 |
Token | i |
Token ID | 72 |
Feature activation | -8.08e-02 |
Pos logit contributions | |
handing | +0.219 |
Amy | +0.219 |
kids | +0.216 |
Lee | +0.216 |
ens | +0.215 |
Neg logit contributions | |
best | -0.317 |
most | -0.311 |
biggest | -0.307 |
ble | -0.297 |
truly | -0.297 |
Token | said |
Token ID | 531 |
Feature activation | -8.55e-02 |
Pos logit contributions | |
kids | +0.098 |
ens | +0.098 |
riet | +0.097 |
Lee | +0.095 |
purs | +0.094 |
Neg logit contributions | |
most | -0.172 |
best | -0.168 |
biggest | -0.166 |
highest | -0.163 |
fastest | -0.163 |
Token | to |
Token ID | 284 |
Feature activation | -9.61e-02 |
Pos logit contributions | |
kids | +0.173 |
ens | +0.172 |
riet | +0.168 |
purs | +0.165 |
pa | +0.164 |
Neg logit contributions | |
most | -0.120 |
best | -0.118 |
biggest | -0.117 |
whole | -0.113 |
highest | -0.110 |
Token | her |
Token ID | 607 |
Feature activation | -5.69e-02 |
Pos logit contributions | |
kids | +0.131 |
ens | +0.130 |
riet | +0.125 |
olph | +0.121 |
Lee | +0.121 |
Neg logit contributions | |
most | -0.204 |
biggest | -0.200 |
highest | -0.193 |
best | -0.191 |
whole | -0.191 |
Token | friends |
Token ID | 2460 |
Feature activation | -7.32e-02 |
Pos logit contributions | |
riet | +0.106 |
ens | +0.106 |
kids | +0.105 |
olph | +0.101 |
Lee | +0.100 |
Neg logit contributions | |
most | -0.094 |
best | -0.094 |
biggest | -0.094 |
highest | -0.087 |
fastest | -0.087 |
Token | trees |
Token ID | 7150 |
Feature activation | +3.32e-02 |
Pos logit contributions | |
possible | +0.045 |
arest | +0.044 |
ute | +0.042 |
virtue | +0.041 |
truly | +0.040 |
Neg logit contributions | |
ens | -0.099 |
Sue | -0.094 |
Mrs | -0.092 |
Gig | -0.092 |
Sophie | -0.091 |
Token | . |
Token ID | 13 |
Feature activation | +1.57e-02 |
Pos logit contributions | |
biggest | +0.063 |
most | +0.063 |
fastest | +0.061 |
coolest | +0.059 |
highest | +0.059 |
Neg logit contributions | |
ens | -0.049 |
Lee | -0.047 |
Sue | -0.046 |
handing | -0.046 |
Amy | -0.044 |
Token | One |
Token ID | 1881 |
Feature activation | +4.80e-02 |
Pos logit contributions | |
biggest | +0.023 |
most | +0.022 |
world | +0.020 |
best | +0.020 |
highest | +0.020 |
Neg logit contributions | |
riet | -0.031 |
ens | -0.031 |
Lee | -0.031 |
handing | -0.031 |
kids | -0.031 |
Token | day |
Token ID | 1110 |
Feature activation | +5.51e-03 |
Pos logit contributions | |
biggest | +0.052 |
most | +0.051 |
fastest | +0.050 |
highest | +0.050 |
best | +0.046 |
Neg logit contributions | |
ens | -0.113 |
Lee | -0.109 |
Sue | -0.108 |
kids | -0.106 |
riet | -0.106 |
Token | , |
Token ID | 11 |
Feature activation | -1.17e-01 |
Pos logit contributions | |
most | +0.010 |
whole | +0.010 |
best | +0.010 |
biggest | +0.010 |
br | +0.009 |
Neg logit contributions | |
kids | -0.009 |
pa | -0.008 |
riet | -0.008 |
Lee | -0.008 |
child | -0.008 |
Token | Gig |
Token ID | 19525 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
kids | +0.087 |
pa | +0.081 |
Lee | +0.081 |
riet | +0.081 |
lee | +0.079 |
Neg logit contributions | |
most | -0.327 |
biggest | -0.317 |
whole | -0.311 |
best | -0.309 |
highest | -0.303 |
Token | i |
Token ID | 72 |
Feature activation | -7.24e-02 |
Pos logit contributions | |
Lee | +0.167 |
handing | +0.161 |
Amy | +0.159 |
Sue | +0.156 |
ens | +0.152 |
Neg logit contributions | |
most | -0.408 |
biggest | -0.403 |
best | -0.396 |
whole | -0.395 |
fastest | -0.392 |
Token | 's |
Token ID | 338 |
Feature activation | +1.66e-02 |
Pos logit contributions | |
Lee | +0.118 |
riet | +0.115 |
ens | +0.114 |
handing | +0.112 |
Sophie | +0.112 |
Neg logit contributions | |
most | -0.133 |
biggest | -0.129 |
whole | -0.127 |
fastest | -0.125 |
best | -0.125 |
Token | neck |
Token ID | 7393 |
Feature activation | +1.73e-02 |
Pos logit contributions | |
most | +0.028 |
biggest | +0.028 |
fastest | +0.027 |
highest | +0.027 |
best | +0.027 |
Neg logit contributions | |
ens | -0.028 |
Lee | -0.028 |
riet | -0.027 |
kids | -0.027 |
handing | -0.026 |
Token | started |
Token ID | 2067 |
Feature activation | +8.90e-03 |
Pos logit contributions | |
biggest | +0.032 |
most | +0.032 |
fastest | +0.032 |
possible | +0.031 |
highest | +0.031 |
Neg logit contributions | |
ens | -0.026 |
begged | -0.023 |
Sue | -0.023 |
hugged | -0.023 |
stro | -0.023 |
Token | to |
Token ID | 284 |
Feature activation | -6.68e-02 |
Pos logit contributions | |
most | +0.017 |
biggest | +0.017 |
fastest | +0.017 |
highest | +0.017 |
best | +0.016 |
Neg logit contributions | |
ens | -0.013 |
Lee | -0.013 |
riet | -0.012 |
kids | -0.012 |
Sophie | -0.012 |
Token | animals |
Token ID | 4695 |
Feature activation | +1.84e-02 |
Pos logit contributions | |
ifying | +0.064 |
biggest | +0.063 |
highest | +0.061 |
most | +0.061 |
fastest | +0.060 |
Neg logit contributions | |
ens | -0.070 |
Lee | -0.068 |
purs | -0.067 |
kids | -0.067 |
pa | -0.067 |
Token | cheered |
Token ID | 34767 |
Feature activation | -5.17e-02 |
Pos logit contributions | |
biggest | +0.034 |
most | +0.033 |
fastest | +0.033 |
highest | +0.033 |
best | +0.031 |
Neg logit contributions | |
ens | -0.028 |
hugged | -0.027 |
Lee | -0.027 |
stro | -0.027 |
Sue | -0.026 |
Token | happily |
Token ID | 18177 |
Feature activation | -6.37e-02 |
Pos logit contributions | |
Lee | +0.071 |
ens | +0.070 |
Sophie | +0.070 |
handing | +0.069 |
her | +0.068 |
Neg logit contributions | |
biggest | -0.104 |
fastest | -0.104 |
highest | -0.103 |
most | -0.099 |
tallest | -0.098 |
Token | ! |
Token ID | 0 |
Feature activation | +3.29e-02 |
Pos logit contributions | |
Lee | +0.085 |
handing | +0.083 |
Sue | +0.082 |
ens | +0.082 |
Amy | +0.079 |
Neg logit contributions | |
biggest | -0.127 |
coolest | -0.123 |
fastest | -0.122 |
most | -0.121 |
highest | -0.120 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.77e-02 |
Pos logit contributions | |
biggest | +0.060 |
most | +0.060 |
fastest | +0.059 |
highest | +0.059 |
best | +0.056 |
Neg logit contributions | |
Lee | -0.053 |
ens | -0.052 |
Sue | -0.051 |
Sophie | -0.051 |
riet | -0.050 |
Token | Mon |
Token ID | 9069 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
biggest | +0.031 |
most | +0.030 |
fastest | +0.030 |
highest | +0.030 |
coolest | +0.029 |
Neg logit contributions | |
Lee | -0.031 |
ens | -0.031 |
Sue | -0.029 |
Sophie | -0.029 |
Amy | -0.029 |
Token | ica |
Token ID | 3970 |
Feature activation | -1.13e-01 |
Pos logit contributions | |
ens | +0.247 |
Lee | +0.240 |
riet | +0.237 |
handing | +0.226 |
Sophie | +0.225 |
Neg logit contributions | |
most | -0.337 |
biggest | -0.325 |
fastest | -0.322 |
highest | -0.319 |
best | -0.319 |
Token | had |
Token ID | 550 |
Feature activation | +5.23e-02 |
Pos logit contributions | |
Lee | +0.199 |
ens | +0.193 |
handing | +0.190 |
kids | +0.188 |
riet | +0.185 |
Neg logit contributions | |
most | -0.180 |
whole | -0.171 |
truly | -0.170 |
biggest | -0.166 |
entire | -0.166 |
Token | a |
Token ID | 257 |
Feature activation | +3.14e-02 |
Pos logit contributions | |
fastest | +0.072 |
biggest | +0.072 |
highest | +0.071 |
most | +0.070 |
best | +0.065 |
Neg logit contributions | |
ens | -0.102 |
Sophie | -0.102 |
Sue | -0.100 |
Lee | -0.100 |
kids | -0.098 |
Token | puzzle |
Token ID | 15027 |
Feature activation | +6.16e-02 |
Pos logit contributions | |
fastest | +0.035 |
biggest | +0.034 |
most | +0.034 |
highest | +0.034 |
hest | +0.032 |
Neg logit contributions | |
ens | -0.070 |
Lee | -0.069 |
Sophie | -0.068 |
Sue | -0.067 |
Amy | -0.066 |
Token | . |
Token ID | 13 |
Feature activation | +2.30e-02 |
Pos logit contributions | |
biggest | +0.112 |
fastest | +0.112 |
highest | +0.110 |
most | +0.108 |
ifying | +0.107 |
Neg logit contributions | |
Sue | -0.090 |
handing | -0.089 |
ens | -0.088 |
Amy | -0.087 |
Lee | -0.087 |
Token | and |
Token ID | 290 |
Feature activation | -7.78e-02 |
Pos logit contributions | |
ens | +0.086 |
Lee | +0.085 |
riet | +0.082 |
kids | +0.081 |
Sue | +0.081 |
Neg logit contributions | |
biggest | -0.082 |
most | -0.082 |
fastest | -0.080 |
highest | -0.079 |
whole | -0.077 |
Token | play |
Token ID | 711 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
ens | +0.119 |
Lee | +0.116 |
riet | +0.112 |
kids | +0.110 |
handing | +0.109 |
Neg logit contributions | |
most | -0.147 |
biggest | -0.146 |
fastest | -0.142 |
highest | -0.141 |
best | -0.139 |
Token | together |
Token ID | 1978 |
Feature activation | +2.88e-02 |
Pos logit contributions | |
Amy | +0.038 |
Sue | +0.037 |
ens | +0.037 |
Lee | +0.037 |
handing | +0.036 |
Neg logit contributions | |
biggest | -0.037 |
fastest | -0.037 |
highest | -0.036 |
ifying | -0.036 |
most | -0.036 |
Token | again |
Token ID | 757 |
Feature activation | +6.52e-03 |
Pos logit contributions | |
biggest | +0.046 |
most | +0.046 |
fastest | +0.044 |
highest | +0.044 |
truly | +0.043 |
Neg logit contributions | |
ens | -0.051 |
Lee | -0.050 |
Sue | -0.048 |
Sophie | -0.048 |
kids | -0.048 |
Token | !" |
Token ID | 2474 |
Feature activation | -4.04e-02 |
Pos logit contributions | |
biggest | +0.010 |
most | +0.010 |
fastest | +0.010 |
highest | +0.010 |
best | +0.010 |
Neg logit contributions | |
ens | -0.012 |
Lee | -0.011 |
Sue | -0.011 |
riet | -0.011 |
kids | -0.011 |
Token | Fif |
Token ID | 12830 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
riet | +0.045 |
pa | +0.044 |
kids | +0.043 |
ppa | +0.041 |
handing | +0.039 |
Neg logit contributions | |
most | -0.095 |
biggest | -0.092 |
best | -0.092 |
highest | -0.090 |
whole | -0.089 |
Token | i |
Token ID | 72 |
Feature activation | -7.36e-02 |
Pos logit contributions | |
handing | +0.244 |
Lee | +0.239 |
kids | +0.237 |
Amy | +0.237 |
Mrs | +0.227 |
Neg logit contributions | |
best | -0.299 |
most | -0.295 |
truly | -0.294 |
fastest | -0.287 |
ute | -0.281 |
Token | smiled |
Token ID | 13541 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
riet | +0.098 |
Lee | +0.096 |
ens | +0.096 |
kids | +0.094 |
pa | +0.092 |
Neg logit contributions | |
most | -0.154 |
best | -0.148 |
biggest | -0.144 |
truly | -0.143 |
whole | -0.142 |
Token | and |
Token ID | 290 |
Feature activation | -8.94e-02 |
Pos logit contributions | |
ens | +0.029 |
Lee | +0.029 |
Sophie | +0.028 |
Sue | +0.027 |
handing | +0.026 |
Neg logit contributions | |
fastest | -0.071 |
biggest | -0.069 |
highest | -0.069 |
most | -0.068 |
ifying | -0.067 |
Token | said |
Token ID | 531 |
Feature activation | -7.01e-02 |
Pos logit contributions | |
Lee | +0.105 |
kids | +0.105 |
pa | +0.103 |
riet | +0.101 |
ens | +0.100 |
Neg logit contributions | |
most | -0.201 |
biggest | -0.195 |
best | -0.195 |
truly | -0.190 |
whole | -0.187 |
Token | , |
Token ID | 11 |
Feature activation | -1.12e-01 |
Pos logit contributions | |
ens | +0.088 |
Lee | +0.086 |
riet | +0.083 |
kids | +0.083 |
handing | +0.079 |
Neg logit contributions | |
most | -0.152 |
biggest | -0.152 |
best | -0.147 |
fastest | -0.146 |
highest | -0.146 |
Token | stop |
Token ID | 2245 |
Feature activation | -2.51e-02 |
Pos logit contributions | |
biggest | +0.046 |
most | +0.045 |
fastest | +0.044 |
highest | +0.044 |
best | +0.043 |
Neg logit contributions | |
ens | -0.036 |
Lee | -0.034 |
riet | -0.034 |
Sophie | -0.034 |
handing | -0.034 |
Token | smiling |
Token ID | 16755 |
Feature activation | -3.34e-02 |
Pos logit contributions | |
ens | +0.033 |
Lee | +0.031 |
riet | +0.031 |
kids | +0.030 |
Sue | +0.029 |
Neg logit contributions | |
most | -0.054 |
biggest | -0.053 |
best | -0.052 |
whole | -0.052 |
fastest | -0.052 |
Token | . |
Token ID | 13 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
ens | +0.055 |
Lee | +0.054 |
riet | +0.052 |
Sophie | +0.052 |
handing | +0.052 |
Neg logit contributions | |
biggest | -0.058 |
most | -0.058 |
fastest | -0.057 |
highest | -0.056 |
whole | -0.054 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.63e-02 |
Pos logit contributions | |
most | +0.034 |
biggest | +0.034 |
fastest | +0.033 |
highest | +0.033 |
whole | +0.032 |
Neg logit contributions | |
ens | -0.033 |
Lee | -0.032 |
riet | -0.031 |
kids | -0.030 |
handing | -0.030 |
Token | He |
Token ID | 1544 |
Feature activation | -9.70e-03 |
Pos logit contributions | |
biggest | +0.027 |
most | +0.026 |
fastest | +0.026 |
highest | +0.026 |
coolest | +0.025 |
Neg logit contributions | |
ens | -0.030 |
Lee | -0.029 |
Sue | -0.028 |
Sophie | -0.028 |
Amy | -0.027 |
Token | idi |
Token ID | 19830 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
Lee | +0.016 |
ens | +0.016 |
riet | +0.016 |
kids | +0.016 |
Sue | +0.015 |
Neg logit contributions | |
most | -0.017 |
biggest | -0.016 |
whole | -0.016 |
fastest | -0.016 |
truly | -0.016 |
Token | had |
Token ID | 550 |
Feature activation | +6.12e-02 |
Pos logit contributions | |
Lee | +0.345 |
kids | +0.331 |
handing | +0.330 |
ens | +0.329 |
ppa | +0.327 |
Neg logit contributions | |
most | -0.233 |
whole | -0.217 |
biggest | -0.213 |
best | -0.209 |
truly | -0.206 |
Token | a |
Token ID | 257 |
Feature activation | +3.72e-02 |
Pos logit contributions | |
biggest | +0.082 |
fastest | +0.082 |
most | +0.081 |
highest | +0.080 |
best | +0.075 |
Neg logit contributions | |
ens | -0.123 |
Lee | -0.121 |
Sophie | -0.120 |
Sue | -0.119 |
kids | -0.118 |
Token | big |
Token ID | 1263 |
Feature activation | -1.92e-02 |
Pos logit contributions | |
most | +0.045 |
fastest | +0.045 |
biggest | +0.045 |
highest | +0.043 |
ifying | +0.040 |
Neg logit contributions | |
ens | -0.080 |
Lee | -0.079 |
Sophie | -0.076 |
Sue | -0.076 |
kids | -0.075 |
Token | problem |
Token ID | 1917 |
Feature activation | +5.92e-02 |
Pos logit contributions | |
ens | +0.030 |
riet | +0.029 |
Lee | +0.029 |
handing | +0.028 |
ppa | +0.028 |
Neg logit contributions | |
most | -0.036 |
biggest | -0.036 |
best | -0.035 |
whole | -0.034 |
highest | -0.034 |
Token | , |
Token ID | 11 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
biggest | +0.104 |
most | +0.102 |
highest | +0.100 |
fastest | +0.100 |
whole | +0.097 |
Neg logit contributions | |
Lee | -0.098 |
ens | -0.098 |
handing | -0.094 |
Sue | -0.093 |
Sophie | -0.092 |
Token | home |
Token ID | 1363 |
Feature activation | -4.71e-02 |
Pos logit contributions | |
ens | +0.112 |
Lee | +0.111 |
riet | +0.106 |
Sue | +0.105 |
kids | +0.104 |
Neg logit contributions | |
most | -0.127 |
biggest | -0.127 |
fastest | -0.123 |
highest | -0.123 |
best | -0.120 |
Token | safely |
Token ID | 11512 |
Feature activation | -5.18e-03 |
Pos logit contributions | |
ens | +0.084 |
Lee | +0.082 |
riet | +0.080 |
kids | +0.079 |
handing | +0.078 |
Neg logit contributions | |
most | -0.077 |
biggest | -0.077 |
fastest | -0.074 |
highest | -0.074 |
best | -0.072 |
Token | . |
Token ID | 13 |
Feature activation | +2.11e-02 |
Pos logit contributions | |
ens | +0.007 |
Lee | +0.007 |
Sue | +0.007 |
handing | +0.007 |
Sophie | +0.007 |
Neg logit contributions | |
biggest | -0.010 |
most | -0.010 |
highest | -0.010 |
fastest | -0.010 |
coolest | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | +6.13e-02 |
Pos logit contributions | |
most | +0.035 |
biggest | +0.035 |
fastest | +0.034 |
highest | +0.034 |
best | +0.033 |
Neg logit contributions | |
ens | -0.037 |
Lee | -0.036 |
riet | -0.035 |
kids | -0.034 |
Sue | -0.034 |
Token |
|
Token ID | 198 |
Feature activation | +5.65e-02 |
Pos logit contributions | |
most | +0.106 |
biggest | +0.105 |
whole | +0.102 |
highest | +0.101 |
fastest | +0.101 |
Neg logit contributions | |
ens | -0.103 |
Lee | -0.103 |
riet | -0.098 |
handing | -0.097 |
kids | -0.096 |
Token | Mic |
Token ID | 25437 |
Feature activation | -1.67e-01 |
Pos logit contributions | |
biggest | +0.100 |
most | +0.097 |
fastest | +0.096 |
highest | +0.095 |
best | +0.092 |
Neg logit contributions | |
ens | -0.095 |
riet | -0.091 |
kids | -0.088 |
Sophie | -0.088 |
Lee | -0.088 |
Token | key |
Token ID | 2539 |
Feature activation | -9.73e-02 |
Pos logit contributions | |
Lee | +0.210 |
ens | +0.207 |
Sue | +0.201 |
Sophie | +0.199 |
handing | +0.195 |
Neg logit contributions | |
most | -0.357 |
fastest | -0.348 |
biggest | -0.347 |
whole | -0.342 |
best | -0.342 |
Token | learned |
Token ID | 4499 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
ens | +0.186 |
Lee | +0.186 |
riet | +0.179 |
handing | +0.176 |
Sue | +0.175 |
Neg logit contributions | |
most | -0.149 |
biggest | -0.142 |
fastest | -0.139 |
whole | -0.138 |
best | -0.138 |
Token | that |
Token ID | 326 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
ens | +0.032 |
Lee | +0.031 |
Sue | +0.030 |
riet | +0.030 |
Sophie | +0.030 |
Neg logit contributions | |
biggest | -0.037 |
fastest | -0.037 |
most | -0.037 |
highest | -0.036 |
whole | -0.035 |
Token | it |
Token ID | 340 |
Feature activation | +1.15e-01 |
Pos logit contributions | |
ifying | +0.022 |
highest | +0.021 |
biggest | +0.021 |
fastest | +0.021 |
entire | +0.020 |
Neg logit contributions | |
ens | -0.018 |
Lee | -0.018 |
Sue | -0.017 |
handing | -0.017 |
Sophie | -0.017 |
Token | 's |
Token ID | 338 |
Feature activation | +5.92e-02 |
Pos logit contributions | |
ifying | +0.242 |
arest | +0.212 |
highest | +0.206 |
ute | +0.205 |
biggest | +0.205 |
Neg logit contributions | |
ens | -0.169 |
hugs | -0.166 |
hugged | -0.165 |
Amy | -0.153 |
her | -0.152 |
Token | when |
Token ID | 618 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
ens | +0.033 |
Lee | +0.032 |
riet | +0.030 |
Sue | +0.030 |
Sophie | +0.030 |
Neg logit contributions | |
biggest | -0.057 |
most | -0.057 |
highest | -0.056 |
fastest | -0.056 |
whole | -0.055 |
Token | you |
Token ID | 345 |
Feature activation | +2.39e-02 |
Pos logit contributions | |
highest | +0.030 |
biggest | +0.030 |
hest | +0.029 |
fastest | +0.029 |
ifying | +0.029 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.025 |
Sue | -0.025 |
Sophie | -0.024 |
Tom | -0.024 |
Token | want |
Token ID | 765 |
Feature activation | +2.27e-02 |
Pos logit contributions | |
biggest | +0.042 |
most | +0.040 |
highest | +0.038 |
possible | +0.038 |
fastest | +0.038 |
Neg logit contributions | |
ens | -0.042 |
Lee | -0.039 |
stro | -0.039 |
begged | -0.038 |
hugged | -0.038 |
Token | to |
Token ID | 284 |
Feature activation | -6.54e-02 |
Pos logit contributions | |
biggest | +0.044 |
fastest | +0.043 |
most | +0.043 |
highest | +0.043 |
ifying | +0.042 |
Neg logit contributions | |
ens | -0.032 |
Lee | -0.031 |
Sophie | -0.031 |
Sue | -0.030 |
handing | -0.030 |
Token | do |
Token ID | 466 |
Feature activation | +3.68e-02 |
Pos logit contributions | |
ens | +0.102 |
Lee | +0.101 |
Sophie | +0.096 |
Sue | +0.095 |
handing | +0.094 |
Neg logit contributions | |
biggest | -0.121 |
most | -0.120 |
highest | -0.116 |
fastest | -0.115 |
whole | -0.112 |
Token | something |
Token ID | 1223 |
Feature activation | +1.33e-01 |
Pos logit contributions | |
biggest | +0.060 |
most | +0.059 |
fastest | +0.059 |
highest | +0.059 |
whole | +0.056 |
Neg logit contributions | |
ens | -0.063 |
Lee | -0.062 |
Sophie | -0.060 |
Sue | -0.060 |
handing | -0.060 |
Token | really |
Token ID | 1107 |
Feature activation | +4.73e-02 |
Pos logit contributions | |
ifying | +0.209 |
highest | +0.204 |
biggest | +0.201 |
entire | +0.201 |
most | +0.200 |
Neg logit contributions | |
ens | -0.220 |
handing | -0.218 |
Sophie | -0.217 |
Sue | -0.212 |
Lee | -0.209 |
Token | fun |
Token ID | 1257 |
Feature activation | +1.57e-02 |
Pos logit contributions | |
ifying | +0.067 |
biggest | +0.064 |
highest | +0.064 |
most | +0.063 |
fastest | +0.061 |
Neg logit contributions | |
ens | -0.091 |
Sue | -0.088 |
Sophie | -0.086 |
begged | -0.086 |
Lee | -0.085 |
Token | ." |
Token ID | 526 |
Feature activation | -3.09e-02 |
Pos logit contributions | |
biggest | +0.026 |
most | +0.025 |
highest | +0.025 |
fastest | +0.025 |
whole | +0.024 |
Neg logit contributions | |
ens | -0.028 |
Lee | -0.028 |
handing | -0.026 |
Sue | -0.026 |
riet | -0.026 |
Token | |
Token ID | 220 |
Feature activation | +7.99e-02 |
Pos logit contributions | |
riet | +0.062 |
kids | +0.062 |
ens | +0.061 |
ppa | +0.061 |
purs | +0.060 |
Neg logit contributions | |
most | -0.046 |
biggest | -0.043 |
best | -0.041 |
whole | -0.040 |
world | -0.038 |
Token |
|
Token ID | 198 |
Feature activation | +4.59e-02 |
Pos logit contributions | |
most | +0.122 |
biggest | +0.121 |
fastest | +0.116 |
highest | +0.116 |
best | +0.114 |
Neg logit contributions | |
ens | -0.151 |
Lee | -0.148 |
riet | -0.145 |
kids | -0.143 |
handing | -0.141 |
Token | and |
Token ID | 290 |
Feature activation | -8.74e-02 |
Pos logit contributions | |
ens | +0.005 |
Lee | +0.005 |
riet | +0.005 |
kids | +0.005 |
Sophie | +0.005 |
Neg logit contributions | |
most | -0.008 |
biggest | -0.007 |
fastest | -0.007 |
highest | -0.007 |
best | -0.007 |
Token | went |
Token ID | 1816 |
Feature activation | -7.60e-02 |
Pos logit contributions | |
Lee | +0.135 |
riet | +0.134 |
ens | +0.134 |
kids | +0.130 |
handing | +0.128 |
Neg logit contributions | |
most | -0.167 |
biggest | -0.158 |
whole | -0.154 |
best | -0.153 |
truly | -0.147 |
Token | outside |
Token ID | 2354 |
Feature activation | -5.56e-02 |
Pos logit contributions | |
ens | +0.123 |
Lee | +0.121 |
riet | +0.115 |
handing | +0.114 |
Sue | +0.114 |
Neg logit contributions | |
most | -0.136 |
biggest | -0.135 |
fastest | -0.132 |
highest | -0.130 |
best | -0.127 |
Token | . |
Token ID | 13 |
Feature activation | +2.58e-02 |
Pos logit contributions | |
ens | +0.091 |
Lee | +0.090 |
handing | +0.086 |
Sue | +0.085 |
riet | +0.085 |
Neg logit contributions | |
biggest | -0.097 |
most | -0.097 |
fastest | -0.095 |
highest | -0.094 |
best | -0.091 |
Token | Everything |
Token ID | 11391 |
Feature activation | +1.26e-01 |
Pos logit contributions | |
most | +0.049 |
biggest | +0.048 |
fastest | +0.048 |
highest | +0.047 |
best | +0.046 |
Neg logit contributions | |
Lee | -0.039 |
ens | -0.038 |
Sue | -0.037 |
kids | -0.037 |
Sophie | -0.036 |
Token | seemed |
Token ID | 3947 |
Feature activation | +1.40e-01 |
Pos logit contributions | |
ifying | +0.214 |
biggest | +0.211 |
highest | +0.207 |
fastest | +0.205 |
most | +0.202 |
Neg logit contributions | |
ens | -0.215 |
Sue | -0.206 |
Amy | -0.200 |
hugged | -0.199 |
Lucy | -0.197 |
Token | peaceful |
Token ID | 12309 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
arest | +0.250 |
ute | +0.247 |
ifying | +0.243 |
hest | +0.241 |
uch | +0.239 |
Neg logit contributions | |
Sue | -0.220 |
ens | -0.211 |
Sally | -0.211 |
Sophie | -0.210 |
her | -0.207 |
Token | and |
Token ID | 290 |
Feature activation | -4.23e-02 |
Pos logit contributions | |
biggest | +0.036 |
fastest | +0.035 |
most | +0.035 |
highest | +0.035 |
truly | +0.034 |
Neg logit contributions | |
Lee | -0.023 |
ens | -0.023 |
Sophie | -0.022 |
handing | -0.022 |
Sue | -0.022 |
Token | calm |
Token ID | 9480 |
Feature activation | -2.20e-02 |
Pos logit contributions | |
ens | +0.076 |
Sue | +0.076 |
Amy | +0.074 |
Sophie | +0.074 |
Mrs | +0.073 |
Neg logit contributions | |
ute | -0.065 |
biggest | -0.063 |
highest | -0.062 |
ifying | -0.062 |
fastest | -0.061 |
Token | . |
Token ID | 13 |
Feature activation | +2.73e-02 |
Pos logit contributions | |
ens | +0.032 |
Lee | +0.031 |
Sophie | +0.030 |
Sue | +0.029 |
Tom | +0.028 |
Neg logit contributions | |
biggest | -0.043 |
fastest | -0.043 |
highest | -0.043 |
most | -0.042 |
truly | -0.041 |
Token | But |
Token ID | 887 |
Feature activation | +7.10e-03 |
Pos logit contributions | |
biggest | +0.056 |
most | +0.056 |
fastest | +0.056 |
highest | +0.055 |
best | +0.054 |
Neg logit contributions | |
Lee | -0.035 |
ens | -0.035 |
Sue | -0.034 |
kids | -0.034 |
Sophie | -0.033 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.11e-01 |
Pos logit contributions | |
ens | +0.077 |
riet | +0.075 |
kids | +0.070 |
olph | +0.069 |
Gig | +0.068 |
Neg logit contributions | |
best | -0.081 |
most | -0.076 |
biggest | -0.076 |
whole | -0.069 |
perfect | -0.068 |
Token | said |
Token ID | 531 |
Feature activation | -5.73e-02 |
Pos logit contributions | |
ens | +0.156 |
riet | +0.153 |
Lee | +0.149 |
kids | +0.144 |
Gig | +0.142 |
Neg logit contributions | |
truly | -0.219 |
most | -0.216 |
certainly | -0.209 |
best | -0.205 |
whole | -0.195 |
Token | , |
Token ID | 11 |
Feature activation | -1.10e-01 |
Pos logit contributions | |
kids | +0.097 |
ens | +0.096 |
riet | +0.093 |
ior | +0.088 |
pa | +0.087 |
Neg logit contributions | |
most | -0.103 |
best | -0.101 |
biggest | -0.099 |
whole | -0.094 |
truly | -0.094 |
Token | " |
Token ID | 366 |
Feature activation | +5.52e-02 |
Pos logit contributions | |
ens | +0.244 |
kids | +0.235 |
riet | +0.235 |
pa | +0.227 |
purs | +0.226 |
Neg logit contributions | |
most | -0.141 |
whole | -0.128 |
truly | -0.126 |
biggest | -0.126 |
best | -0.123 |
Token | That |
Token ID | 2504 |
Feature activation | +1.59e-01 |
Pos logit contributions | |
best | +0.079 |
biggest | +0.078 |
fastest | +0.072 |
most | +0.070 |
impossible | +0.069 |
Neg logit contributions | |
ens | -0.110 |
Sue | -0.108 |
stro | -0.108 |
Lee | -0.106 |
riet | -0.105 |
Token | is |
Token ID | 318 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
fastest | +0.289 |
ifying | +0.286 |
arest | +0.281 |
possible | +0.280 |
hest | +0.274 |
Neg logit contributions | |
ens | -0.241 |
Lee | -0.240 |
pa | -0.233 |
Sue | -0.228 |
Amy | -0.227 |
Token | a |
Token ID | 257 |
Feature activation | +8.53e-02 |
Pos logit contributions | |
fastest | +0.192 |
highest | +0.189 |
biggest | +0.188 |
hest | +0.186 |
arest | +0.182 |
Neg logit contributions | |
Sue | -0.244 |
Mrs | -0.238 |
Lee | -0.238 |
Sophie | -0.236 |
ens | -0.235 |
Token | long |
Token ID | 890 |
Feature activation | +6.19e-02 |
Pos logit contributions | |
fastest | +0.128 |
most | +0.126 |
biggest | +0.125 |
ained | +0.125 |
hest | +0.125 |
Neg logit contributions | |
Lee | -0.151 |
handing | -0.147 |
ens | -0.145 |
Sophie | -0.143 |
Sue | -0.142 |
Token | plane |
Token ID | 6614 |
Feature activation | +7.61e-02 |
Pos logit contributions | |
most | +0.089 |
biggest | +0.088 |
whole | +0.084 |
fastest | +0.084 |
highest | +0.083 |
Neg logit contributions | |
ens | -0.122 |
Lee | -0.120 |
riet | -0.118 |
kids | -0.115 |
Sophie | -0.114 |
Token | , |
Token ID | 11 |
Feature activation | -7.76e-02 |
Pos logit contributions | |
biggest | +0.149 |
most | +0.145 |
highest | +0.143 |
fastest | +0.140 |
coolest | +0.138 |
Neg logit contributions | |
Lee | -0.110 |
ens | -0.109 |
handing | -0.106 |
Sue | -0.103 |
pa | -0.101 |
Token | dear |
Token ID | 13674 |
Feature activation | -1.40e-02 |
Pos logit contributions | |
ens | +0.091 |
riet | +0.087 |
Lee | +0.085 |
olph | +0.084 |
kids | +0.084 |
Neg logit contributions | |
most | -0.178 |
biggest | -0.172 |
best | -0.168 |
fastest | -0.166 |
highest | -0.165 |
Token | � |
Token ID | 6184 |
Feature activation | -2.13e-02 |
Pos logit contributions | |
ens | +0.160 |
riet | +0.151 |
Lee | +0.150 |
kids | +0.149 |
purs | +0.145 |
Neg logit contributions | |
most | -0.184 |
biggest | -0.181 |
highest | -0.175 |
fastest | -0.174 |
whole | -0.173 |
Token | � |
Token ID | 95 |
Feature activation | -6.16e-04 |
Pos logit contributions | |
Lee | +0.037 |
ens | +0.037 |
riet | +0.035 |
handing | +0.035 |
Sophie | +0.035 |
Neg logit contributions | |
most | -0.036 |
biggest | -0.036 |
fastest | -0.034 |
best | -0.034 |
highest | -0.034 |
Token | € |
Token ID | 26391 |
Feature activation | -1.25e-05 |
Pos logit contributions | |
Lee | +0.001 |
handing | +0.001 |
riet | +0.001 |
kids | +0.001 |
ens | +0.001 |
Neg logit contributions | |
biggest | -0.001 |
best | -0.001 |
most | -0.001 |
highest | -0.001 |
fastest | -0.001 |
Token | � |
Token ID | 129 |
Feature activation | -3.06e-03 |
Pos logit contributions | |
Lee | +0.000 |
ens | +0.000 |
Sue | +0.000 |
kids | +0.000 |
Tom | +0.000 |
Neg logit contributions | |
biggest | -0.000 |
most | -0.000 |
fastest | -0.000 |
highest | -0.000 |
best | -0.000 |
Token | � |
Token ID | 241 |
Feature activation | +6.03e-02 |
Pos logit contributions | |
ens | +0.006 |
Lee | +0.005 |
riet | +0.005 |
Sophie | +0.005 |
Sue | +0.005 |
Neg logit contributions | |
most | -0.005 |
biggest | -0.005 |
fastest | -0.005 |
highest | -0.005 |
whole | -0.005 |
Token | It |
Token ID | 1026 |
Feature activation | +1.54e-01 |
Pos logit contributions | |
biggest | +0.110 |
highest | +0.106 |
fastest | +0.105 |
most | +0.105 |
best | +0.104 |
Neg logit contributions | |
Lee | -0.099 |
Sue | -0.098 |
Sophie | -0.097 |
Samantha | -0.096 |
ens | -0.095 |
Token | looks |
Token ID | 3073 |
Feature activation | +9.70e-02 |
Pos logit contributions | |
ifying | +0.252 |
highest | +0.248 |
biggest | +0.248 |
ute | +0.244 |
whole | +0.240 |
Neg logit contributions | |
hugged | -0.261 |
ens | -0.257 |
Amy | -0.257 |
begged | -0.252 |
Sue | -0.251 |
Token | like |
Token ID | 588 |
Feature activation | +4.73e-02 |
Pos logit contributions | |
fastest | +0.126 |
biggest | +0.125 |
highest | +0.125 |
ute | +0.124 |
most | +0.121 |
Neg logit contributions | |
Sophie | -0.188 |
Sue | -0.185 |
ens | -0.185 |
handing | -0.183 |
Amy | -0.181 |
Token | someone |
Token ID | 2130 |
Feature activation | +8.58e-02 |
Pos logit contributions | |
biggest | +0.071 |
highest | +0.071 |
fastest | +0.071 |
most | +0.069 |
whole | +0.068 |
Neg logit contributions | |
ens | -0.086 |
Lee | -0.086 |
Sue | -0.085 |
Sophie | -0.085 |
Tom | -0.082 |
Token | is |
Token ID | 318 |
Feature activation | +1.04e-01 |
Pos logit contributions | |
biggest | +0.146 |
most | +0.144 |
highest | +0.143 |
fastest | +0.140 |
whole | +0.138 |
Neg logit contributions | |
ens | -0.146 |
Lee | -0.142 |
riet | -0.137 |
Sue | -0.136 |
Sophie | -0.135 |
Token | hope |
Token ID | 2911 |
Feature activation | -8.09e-03 |
Pos logit contributions | |
biggest | +0.211 |
most | +0.210 |
highest | +0.206 |
fastest | +0.205 |
whole | +0.200 |
Neg logit contributions | |
Lee | -0.141 |
ens | -0.140 |
handing | -0.133 |
riet | -0.133 |
Sue | -0.132 |
Token | he |
Token ID | 339 |
Feature activation | -1.67e-03 |
Pos logit contributions | |
ens | +0.143 |
Lee | +0.141 |
Sue | +0.132 |
riet | +0.130 |
Sophie | +0.130 |
Neg logit contributions | |
biggest | -0.225 |
most | -0.223 |
fastest | -0.219 |
highest | -0.219 |
best | -0.212 |
Token | heard |
Token ID | 2982 |
Feature activation | -7.11e-03 |
Pos logit contributions | |
ens | +0.002 |
Lee | +0.002 |
riet | +0.002 |
kids | +0.002 |
Sophie | +0.002 |
Neg logit contributions | |
biggest | -0.004 |
most | -0.004 |
fastest | -0.003 |
highest | -0.003 |
best | -0.003 |
Token | someone |
Token ID | 2130 |
Feature activation | +5.60e-02 |
Pos logit contributions | |
ens | +0.015 |
Lee | +0.015 |
riet | +0.015 |
kids | +0.014 |
Sue | +0.014 |
Neg logit contributions | |
most | -0.009 |
biggest | -0.009 |
fastest | -0.009 |
highest | -0.009 |
best | -0.008 |
Token | shout |
Token ID | 9191 |
Feature activation | -2.64e-02 |
Pos logit contributions | |
most | +0.122 |
biggest | +0.121 |
fastest | +0.119 |
truly | +0.118 |
highest | +0.118 |
Neg logit contributions | |
ens | -0.069 |
Lee | -0.067 |
riet | -0.067 |
kids | -0.065 |
Sophie | -0.062 |
Token | . |
Token ID | 13 |
Feature activation | +1.64e-02 |
Pos logit contributions | |
ens | +0.047 |
Lee | +0.046 |
riet | +0.044 |
kids | +0.044 |
handing | +0.043 |
Neg logit contributions | |
most | -0.043 |
biggest | -0.043 |
fastest | -0.042 |
highest | -0.042 |
best | -0.041 |
Token | It |
Token ID | 632 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
most | +0.031 |
biggest | +0.031 |
fastest | +0.030 |
highest | +0.030 |
whole | +0.030 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.024 |
riet | -0.023 |
kids | -0.023 |
handing | -0.023 |
Token | was |
Token ID | 373 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
biggest | +0.242 |
highest | +0.235 |
ifying | +0.234 |
fastest | +0.230 |
most | +0.228 |
Neg logit contributions | |
Sue | -0.233 |
Amy | -0.227 |
ens | -0.225 |
Lee | -0.223 |
begged | -0.223 |
Token | so |
Token ID | 523 |
Feature activation | +4.70e-02 |
Pos logit contributions | |
most | +0.217 |
biggest | +0.215 |
fastest | +0.209 |
highest | +0.208 |
best | +0.206 |
Neg logit contributions | |
ens | -0.146 |
Lee | -0.141 |
riet | -0.138 |
kids | -0.135 |
handing | -0.133 |
Token | loud |
Token ID | 7812 |
Feature activation | +1.28e-02 |
Pos logit contributions | |
biggest | +0.067 |
most | +0.067 |
fastest | +0.065 |
highest | +0.064 |
ifying | +0.062 |
Neg logit contributions | |
ens | -0.094 |
Sue | -0.094 |
Lee | -0.091 |
Amy | -0.090 |
Tom | -0.089 |
Token | ! |
Token ID | 0 |
Feature activation | +2.87e-02 |
Pos logit contributions | |
biggest | +0.026 |
most | +0.025 |
fastest | +0.025 |
highest | +0.025 |
best | +0.024 |
Neg logit contributions | |
ens | -0.018 |
Lee | -0.018 |
riet | -0.017 |
handing | -0.017 |
Sue | -0.017 |
Token | Max |
Token ID | 5436 |
Feature activation | -9.57e-02 |
Pos logit contributions | |
most | +0.066 |
biggest | +0.066 |
fastest | +0.064 |
highest | +0.064 |
whole | +0.063 |
Neg logit contributions | |
ens | -0.032 |
Lee | -0.031 |
riet | -0.029 |
kids | -0.029 |
handing | -0.028 |
Token | . |
Token ID | 13 |
Feature activation | +8.51e-03 |
Pos logit contributions | |
ens | +0.032 |
Lee | +0.031 |
Sue | +0.030 |
riet | +0.029 |
Sophie | +0.029 |
Neg logit contributions | |
biggest | -0.040 |
most | -0.040 |
fastest | -0.039 |
highest | -0.039 |
truly | -0.038 |
Token | He |
Token ID | 679 |
Feature activation | +9.01e-03 |
Pos logit contributions | |
most | +0.013 |
biggest | +0.013 |
whole | +0.012 |
best | +0.012 |
highest | +0.012 |
Neg logit contributions | |
kids | -0.016 |
ens | -0.016 |
riet | -0.016 |
pa | -0.016 |
purs | -0.015 |
Token | does |
Token ID | 857 |
Feature activation | +5.85e-02 |
Pos logit contributions | |
most | +0.016 |
biggest | +0.016 |
fastest | +0.015 |
highest | +0.015 |
whole | +0.015 |
Neg logit contributions | |
ens | -0.015 |
Lee | -0.015 |
riet | -0.014 |
kids | -0.014 |
handing | -0.014 |
Token | not |
Token ID | 407 |
Feature activation | +2.20e-02 |
Pos logit contributions | |
biggest | +0.103 |
most | +0.101 |
fastest | +0.101 |
highest | +0.099 |
ifying | +0.097 |
Neg logit contributions | |
ens | -0.096 |
handing | -0.095 |
Sue | -0.095 |
Lee | -0.093 |
Tom | -0.092 |
Token | like |
Token ID | 588 |
Feature activation | +6.79e-03 |
Pos logit contributions | |
most | +0.040 |
biggest | +0.040 |
fastest | +0.038 |
highest | +0.038 |
best | +0.037 |
Neg logit contributions | |
ens | -0.037 |
Lee | -0.035 |
riet | -0.034 |
Sue | -0.034 |
handing | -0.033 |
Token | it |
Token ID | 340 |
Feature activation | +9.22e-02 |
Pos logit contributions | |
most | +0.014 |
biggest | +0.014 |
highest | +0.013 |
best | +0.013 |
fastest | +0.013 |
Neg logit contributions | |
riet | -0.009 |
ens | -0.009 |
kids | -0.009 |
Lee | -0.009 |
stro | -0.009 |
Token | when |
Token ID | 618 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
biggest | +0.161 |
most | +0.161 |
highest | +0.154 |
fastest | +0.154 |
truly | +0.153 |
Neg logit contributions | |
ens | -0.156 |
Lee | -0.150 |
Sue | -0.146 |
handing | -0.142 |
Amy | -0.140 |
Token | Anna |
Token ID | 11735 |
Feature activation | -7.67e-02 |
Pos logit contributions | |
most | +0.036 |
biggest | +0.036 |
fastest | +0.035 |
highest | +0.035 |
best | +0.035 |
Neg logit contributions | |
ens | -0.014 |
Lee | -0.013 |
riet | -0.013 |
kids | -0.013 |
handing | -0.012 |
Token | throws |
Token ID | 12542 |
Feature activation | -6.40e-02 |
Pos logit contributions | |
ens | +0.145 |
Lee | +0.143 |
riet | +0.140 |
kids | +0.138 |
Sophie | +0.136 |
Neg logit contributions | |
most | -0.116 |
biggest | -0.115 |
fastest | -0.112 |
highest | -0.111 |
best | -0.109 |
Token | things |
Token ID | 1243 |
Feature activation | +8.75e-02 |
Pos logit contributions | |
ens | +0.077 |
Lee | +0.075 |
riet | +0.072 |
kids | +0.071 |
handing | +0.070 |
Neg logit contributions | |
most | -0.141 |
biggest | -0.140 |
fastest | -0.137 |
highest | -0.137 |
best | -0.134 |
Token | at |
Token ID | 379 |
Feature activation | -3.93e-02 |
Pos logit contributions | |
biggest | +0.153 |
most | +0.153 |
fastest | +0.145 |
highest | +0.144 |
truly | +0.143 |
Neg logit contributions | |
ens | -0.151 |
Lee | -0.143 |
Sue | -0.138 |
handing | -0.136 |
Sophie | -0.136 |
Token | might |
Token ID | 1244 |
Feature activation | +9.01e-02 |
Pos logit contributions | |
biggest | +0.024 |
most | +0.024 |
highest | +0.023 |
fastest | +0.023 |
whole | +0.022 |
Neg logit contributions | |
ens | -0.043 |
Lee | -0.042 |
riet | -0.041 |
kids | -0.041 |
handing | -0.041 |
Token | get |
Token ID | 651 |
Feature activation | +2.58e-02 |
Pos logit contributions | |
biggest | +0.162 |
most | +0.158 |
highest | +0.154 |
fastest | +0.154 |
best | +0.149 |
Neg logit contributions | |
ens | -0.149 |
Lee | -0.143 |
Sue | -0.143 |
Sophie | -0.140 |
handing | -0.139 |
Token | angry |
Token ID | 7954 |
Feature activation | -3.84e-02 |
Pos logit contributions | |
biggest | +0.042 |
most | +0.042 |
fastest | +0.041 |
highest | +0.041 |
whole | +0.040 |
Neg logit contributions | |
ens | -0.045 |
Lee | -0.044 |
riet | -0.043 |
Sophie | -0.042 |
Sue | -0.042 |
Token | ." |
Token ID | 526 |
Feature activation | -3.76e-02 |
Pos logit contributions | |
ens | +0.061 |
Lee | +0.059 |
riet | +0.059 |
kids | +0.058 |
Sophie | +0.056 |
Neg logit contributions | |
most | -0.070 |
biggest | -0.069 |
fastest | -0.068 |
best | -0.067 |
highest | -0.066 |
Token |
|
Token ID | 198 |
Feature activation | +3.94e-02 |
Pos logit contributions | |
kids | +0.093 |
purs | +0.090 |
riet | +0.089 |
ens | +0.088 |
pa | +0.087 |
Neg logit contributions | |
most | -0.037 |
biggest | -0.034 |
best | -0.034 |
whole | -0.031 |
highest | -0.030 |
Token |
|
Token ID | 198 |
Feature activation | +3.63e-02 |
Pos logit contributions | |
most | +0.047 |
biggest | +0.047 |
whole | +0.046 |
best | +0.044 |
highest | +0.043 |
Neg logit contributions | |
Lee | -0.085 |
ens | -0.084 |
riet | -0.084 |
kids | -0.083 |
handing | -0.082 |
Token | L |
Token ID | 43 |
Feature activation | -5.56e-02 |
Pos logit contributions | |
most | +0.070 |
biggest | +0.069 |
fastest | +0.067 |
highest | +0.067 |
whole | +0.067 |
Neg logit contributions | |
ens | -0.055 |
Lee | -0.054 |
kids | -0.053 |
riet | -0.052 |
handing | -0.051 |
Token | ily |
Token ID | 813 |
Feature activation | -9.40e-02 |
Pos logit contributions | |
ens | +0.067 |
Lee | +0.066 |
kids | +0.063 |
riet | +0.062 |
handing | +0.062 |
Neg logit contributions | |
most | -0.121 |
biggest | -0.121 |
fastest | -0.118 |
highest | -0.117 |
best | -0.116 |
Token | felt |
Token ID | 2936 |
Feature activation | +5.18e-02 |
Pos logit contributions | |
Gig | +0.166 |
riet | +0.163 |
kids | +0.162 |
ens | +0.160 |
Samantha | +0.157 |
Neg logit contributions | |
most | -0.157 |
best | -0.144 |
truly | -0.139 |
biggest | -0.138 |
whole | -0.136 |
Token | sad |
Token ID | 6507 |
Feature activation | -1.51e-03 |
Pos logit contributions | |
most | +0.098 |
biggest | +0.098 |
fastest | +0.097 |
highest | +0.096 |
whole | +0.094 |
Neg logit contributions | |
ens | -0.078 |
Lee | -0.076 |
riet | -0.073 |
Sophie | -0.072 |
Sue | -0.072 |
Token | . |
Token ID | 13 |
Feature activation | -4.77e-03 |
Pos logit contributions | |
ens | +0.002 |
Lee | +0.002 |
handing | +0.002 |
Sue | +0.002 |
riet | +0.002 |
Neg logit contributions | |
biggest | -0.003 |
most | -0.003 |
fastest | -0.003 |
highest | -0.003 |
whole | -0.003 |
Token | still |
Token ID | 991 |
Feature activation | +3.44e-02 |
Pos logit contributions | |
biggest | +0.133 |
most | +0.131 |
highest | +0.129 |
fastest | +0.129 |
whole | +0.123 |
Neg logit contributions | |
Lee | -0.131 |
ens | -0.130 |
handing | -0.128 |
Sophie | -0.127 |
Sue | -0.126 |
Token | very |
Token ID | 845 |
Feature activation | +4.39e-02 |
Pos logit contributions | |
biggest | +0.064 |
highest | +0.063 |
most | +0.063 |
fastest | +0.063 |
whole | +0.059 |
Neg logit contributions | |
Lee | -0.051 |
ens | -0.051 |
handing | -0.050 |
Sophie | -0.049 |
Sue | -0.049 |
Token | sad |
Token ID | 6507 |
Feature activation | +2.39e-02 |
Pos logit contributions | |
most | +0.071 |
biggest | +0.071 |
fastest | +0.069 |
highest | +0.068 |
best | +0.067 |
Neg logit contributions | |
ens | -0.080 |
Lee | -0.078 |
riet | -0.076 |
kids | -0.075 |
handing | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
biggest | +0.043 |
most | +0.042 |
highest | +0.042 |
fastest | +0.041 |
whole | +0.040 |
Neg logit contributions | |
ens | -0.039 |
Lee | -0.039 |
Sophie | -0.037 |
handing | -0.037 |
riet | -0.036 |
Token | |
Token ID | 220 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
most | +0.027 |
biggest | +0.026 |
truly | +0.025 |
highest | +0.025 |
fastest | +0.025 |
Neg logit contributions | |
ens | -0.030 |
kids | -0.029 |
pa | -0.029 |
Lee | -0.029 |
riet | -0.029 |
Token |
|
Token ID | 198 |
Feature activation | +5.41e-02 |
Pos logit contributions | |
most | +0.159 |
biggest | +0.159 |
fastest | +0.154 |
highest | +0.153 |
best | +0.149 |
Neg logit contributions | |
ens | -0.181 |
Lee | -0.179 |
riet | -0.172 |
kids | -0.170 |
Sue | -0.170 |
Token |
|
Token ID | 198 |
Feature activation | +4.98e-02 |
Pos logit contributions | |
most | +0.063 |
whole | +0.063 |
biggest | +0.061 |
truly | +0.060 |
fastest | +0.059 |
Neg logit contributions | |
ens | -0.120 |
Lee | -0.118 |
pa | -0.115 |
riet | -0.114 |
kids | -0.114 |
Token | The |
Token ID | 464 |
Feature activation | +2.60e-02 |
Pos logit contributions | |
most | +0.068 |
biggest | +0.067 |
fastest | +0.065 |
highest | +0.065 |
whole | +0.064 |
Neg logit contributions | |
ens | -0.102 |
Lee | -0.101 |
riet | -0.098 |
kids | -0.097 |
Sue | -0.096 |
Token | old |
Token ID | 1468 |
Feature activation | -2.74e-02 |
Pos logit contributions | |
most | +0.043 |
biggest | +0.043 |
fastest | +0.041 |
whole | +0.040 |
best | +0.040 |
Neg logit contributions | |
ens | -0.047 |
Lee | -0.047 |
riet | -0.045 |
Sue | -0.045 |
Sophie | -0.045 |
Token | woman |
Token ID | 2415 |
Feature activation | -7.60e-02 |
Pos logit contributions | |
ens | +0.043 |
Sue | +0.042 |
Sophie | +0.041 |
Amy | +0.041 |
lee | +0.041 |
Neg logit contributions | |
most | -0.050 |
whole | -0.049 |
biggest | -0.048 |
best | -0.048 |
world | -0.048 |
Token | then |
Token ID | 788 |
Feature activation | -6.29e-02 |
Pos logit contributions | |
ens | +0.121 |
Lee | +0.119 |
riet | +0.116 |
kids | +0.110 |
Sophie | +0.110 |
Neg logit contributions | |
most | -0.142 |
truly | -0.136 |
best | -0.133 |
biggest | -0.130 |
whole | -0.130 |
Token | autumn |
Token ID | 23608 |
Feature activation | +4.52e-02 |
Pos logit contributions | |
biggest | +0.017 |
most | +0.017 |
highest | +0.016 |
fastest | +0.016 |
whole | +0.016 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.025 |
riet | -0.024 |
kids | -0.024 |
handing | -0.024 |
Token | sun |
Token ID | 4252 |
Feature activation | +4.87e-02 |
Pos logit contributions | |
biggest | +0.086 |
most | +0.086 |
fastest | +0.084 |
truly | +0.083 |
best | +0.082 |
Neg logit contributions | |
ens | -0.068 |
Lee | -0.063 |
Sue | -0.062 |
Sophie | -0.061 |
pa | -0.060 |
Token | and |
Token ID | 290 |
Feature activation | -6.67e-02 |
Pos logit contributions | |
biggest | +0.094 |
most | +0.094 |
fastest | +0.092 |
highest | +0.091 |
truly | +0.090 |
Neg logit contributions | |
ens | -0.072 |
Lee | -0.070 |
Sue | -0.066 |
Sophie | -0.065 |
riet | -0.065 |
Token | have |
Token ID | 423 |
Feature activation | +6.64e-02 |
Pos logit contributions | |
ens | +0.097 |
Lee | +0.094 |
Sue | +0.093 |
Sophie | +0.092 |
handing | +0.090 |
Neg logit contributions | |
biggest | -0.126 |
most | -0.124 |
fastest | -0.124 |
highest | -0.121 |
best | -0.121 |
Token | a |
Token ID | 257 |
Feature activation | +5.41e-02 |
Pos logit contributions | |
fastest | +0.087 |
biggest | +0.086 |
most | +0.085 |
highest | +0.084 |
best | +0.079 |
Neg logit contributions | |
ens | -0.135 |
Sue | -0.133 |
Sophie | -0.131 |
Amy | -0.130 |
Lee | -0.130 |
Token | pumpkin |
Token ID | 30089 |
Feature activation | +2.78e-02 |
Pos logit contributions | |
most | +0.076 |
fastest | +0.075 |
biggest | +0.075 |
highest | +0.073 |
entire | +0.069 |
Neg logit contributions | |
ens | -0.104 |
Lee | -0.103 |
Sophie | -0.101 |
Sue | -0.101 |
Amy | -0.099 |
Token | party |
Token ID | 2151 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
biggest | +0.054 |
most | +0.054 |
fastest | +0.053 |
highest | +0.053 |
whole | +0.051 |
Neg logit contributions | |
ens | -0.040 |
Lee | -0.039 |
riet | -0.037 |
kids | -0.037 |
Sue | -0.037 |
Token | !", |
Token ID | 40754 |
Feature activation | -8.72e-02 |
Pos logit contributions | |
ens | +0.020 |
Lee | +0.019 |
riet | +0.019 |
kids | +0.018 |
handing | +0.018 |
Neg logit contributions | |
most | -0.019 |
biggest | -0.019 |
fastest | -0.018 |
highest | -0.018 |
best | -0.018 |
Token | said |
Token ID | 531 |
Feature activation | -7.80e-02 |
Pos logit contributions | |
kids | +0.112 |
riet | +0.108 |
ppa | +0.106 |
purs | +0.101 |
pa | +0.101 |
Neg logit contributions | |
most | -0.188 |
best | -0.170 |
truly | -0.167 |
biggest | -0.167 |
highest | -0.162 |
Token | the |
Token ID | 262 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
kids | +0.085 |
ppa | +0.077 |
riet | +0.076 |
ens | +0.072 |
ior | +0.071 |
Neg logit contributions | |
best | -0.186 |
most | -0.185 |
biggest | -0.180 |
highest | -0.174 |
whole | -0.172 |
Token | pump |
Token ID | 8901 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
most | +0.020 |
biggest | +0.019 |
fastest | +0.019 |
highest | +0.019 |
truly | +0.019 |
Neg logit contributions | |
ens | -0.015 |
Lee | -0.015 |
riet | -0.015 |
Sophie | -0.014 |
Sue | -0.014 |
Token | sk |
Token ID | 1341 |
Feature activation | +6.42e-03 |
Pos logit contributions | |
most | +0.013 |
biggest | +0.013 |
fastest | +0.013 |
highest | +0.012 |
whole | +0.012 |
Neg logit contributions | |
ens | -0.013 |
Lee | -0.012 |
riet | -0.012 |
kids | -0.012 |
handing | -0.012 |
Token | unk |
Token ID | 2954 |
Feature activation | -3.31e-03 |
Pos logit contributions | |
biggest | +0.011 |
most | +0.011 |
highest | +0.010 |
coolest | +0.010 |
fastest | +0.010 |
Neg logit contributions | |
Lee | -0.011 |
ens | -0.011 |
olph | -0.010 |
kids | -0.010 |
handing | -0.010 |
Token | ! |
Token ID | 0 |
Feature activation | +3.46e-02 |
Pos logit contributions | |
ens | +0.006 |
Lee | +0.006 |
riet | +0.006 |
kids | +0.006 |
handing | +0.005 |
Neg logit contributions | |
most | -0.005 |
biggest | -0.005 |
fastest | -0.005 |
highest | -0.005 |
best | -0.005 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +2.00e-02 |
Pos logit contributions | |
most | +0.055 |
biggest | +0.054 |
highest | +0.052 |
whole | +0.051 |
fastest | +0.051 |
Neg logit contributions | |
ens | -0.062 |
kids | -0.060 |
Lee | -0.060 |
riet | -0.059 |
pa | -0.059 |
Token | Once |
Token ID | 7454 |
Feature activation | -4.49e-02 |
Pos logit contributions | |
biggest | +0.033 |
most | +0.032 |
fastest | +0.032 |
highest | +0.032 |
coolest | +0.031 |
Neg logit contributions | |
ens | -0.036 |
Lee | -0.036 |
Sue | -0.034 |
Sophie | -0.034 |
Amy | -0.033 |
Token | upon |
Token ID | 2402 |
Feature activation | +8.69e-02 |
Pos logit contributions | |
Lee | +0.089 |
ens | +0.089 |
kids | +0.087 |
handing | +0.087 |
riet | +0.087 |
Neg logit contributions | |
most | -0.063 |
biggest | -0.063 |
best | -0.060 |
whole | -0.060 |
fastest | -0.059 |
Token | a |
Token ID | 257 |
Feature activation | -1.15e-03 |
Pos logit contributions | |
fastest | +0.155 |
highest | +0.147 |
biggest | +0.145 |
most | +0.139 |
virtue | +0.139 |
Neg logit contributions | |
Sue | -0.150 |
Amy | -0.147 |
her | -0.146 |
Tom | -0.145 |
ens | -0.140 |
Token | time |
Token ID | 640 |
Feature activation | +2.15e-02 |
Pos logit contributions | |
ens | +0.003 |
Lee | +0.003 |
riet | +0.003 |
kids | +0.003 |
Sue | +0.003 |
Neg logit contributions | |
biggest | -0.001 |
most | -0.001 |
fastest | -0.001 |
highest | -0.001 |
best | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -1.31e-01 |
Pos logit contributions | |
most | +0.046 |
biggest | +0.045 |
best | +0.044 |
fastest | +0.044 |
whole | +0.044 |
Neg logit contributions | |
ens | -0.027 |
riet | -0.026 |
kids | -0.026 |
Lee | -0.026 |
handing | -0.025 |
Token | there |
Token ID | 612 |
Feature activation | +5.29e-02 |
Pos logit contributions | |
Sue | +0.194 |
ens | +0.189 |
Amy | +0.189 |
Lee | +0.183 |
Tom | +0.180 |
Neg logit contributions | |
biggest | -0.252 |
highest | -0.252 |
fastest | -0.249 |
most | -0.243 |
truly | -0.236 |
Token | was |
Token ID | 373 |
Feature activation | +6.42e-02 |
Pos logit contributions | |
biggest | +0.079 |
highest | +0.078 |
fastest | +0.078 |
most | +0.077 |
ifying | +0.072 |
Neg logit contributions | |
ens | -0.102 |
Lee | -0.100 |
Sue | -0.100 |
Sophie | -0.097 |
Amy | -0.096 |
Token | nap |
Token ID | 25422 |
Feature activation | +1.83e-02 |
Pos logit contributions | |
biggest | +0.028 |
most | +0.028 |
fastest | +0.027 |
highest | +0.027 |
best | +0.026 |
Neg logit contributions | |
ens | -0.044 |
Lee | -0.043 |
kids | -0.042 |
riet | -0.042 |
Sophie | -0.042 |
Token | . |
Token ID | 13 |
Feature activation | +1.38e-02 |
Pos logit contributions | |
biggest | +0.032 |
highest | +0.032 |
fastest | +0.031 |
most | +0.031 |
truly | +0.029 |
Neg logit contributions | |
ens | -0.029 |
Lee | -0.029 |
handing | -0.028 |
Sophie | -0.028 |
Sue | -0.028 |
Token | When |
Token ID | 1649 |
Feature activation | +2.89e-02 |
Pos logit contributions | |
fastest | +0.027 |
biggest | +0.027 |
most | +0.027 |
highest | +0.027 |
best | +0.026 |
Neg logit contributions | |
Lee | -0.019 |
ens | -0.019 |
Sophie | -0.019 |
Sue | -0.019 |
kids | -0.018 |
Token | he |
Token ID | 339 |
Feature activation | -9.89e-03 |
Pos logit contributions | |
biggest | +0.065 |
fastest | +0.065 |
highest | +0.065 |
most | +0.064 |
ifying | +0.061 |
Neg logit contributions | |
Lee | -0.032 |
ens | -0.032 |
Sue | -0.032 |
Sophie | -0.032 |
pa | -0.030 |
Token | woke |
Token ID | 19092 |
Feature activation | -5.93e-02 |
Pos logit contributions | |
Lee | +0.015 |
ens | +0.015 |
riet | +0.014 |
handing | +0.014 |
kids | +0.014 |
Neg logit contributions | |
most | -0.019 |
biggest | -0.018 |
whole | -0.018 |
truly | -0.018 |
fastest | -0.018 |
Token | up |
Token ID | 510 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
ens | +0.137 |
kids | +0.130 |
Lee | +0.130 |
riet | +0.130 |
stro | +0.127 |
Neg logit contributions | |
most | -0.069 |
biggest | -0.068 |
best | -0.066 |
whole | -0.064 |
truly | -0.062 |
Token | , |
Token ID | 11 |
Feature activation | -1.07e-01 |
Pos logit contributions | |
ens | +0.031 |
riet | +0.030 |
kids | +0.030 |
Lee | +0.030 |
purs | +0.029 |
Neg logit contributions | |
most | -0.042 |
biggest | -0.040 |
best | -0.040 |
whole | -0.040 |
truly | -0.039 |
Token | he |
Token ID | 339 |
Feature activation | -5.82e-03 |
Pos logit contributions | |
Lee | +0.116 |
ens | +0.114 |
Sophie | +0.110 |
Sue | +0.110 |
pa | +0.107 |
Neg logit contributions | |
biggest | -0.244 |
fastest | -0.242 |
highest | -0.241 |
most | -0.239 |
whole | -0.231 |
Token | felt |
Token ID | 2936 |
Feature activation | +6.84e-02 |
Pos logit contributions | |
ens | +0.009 |
Lee | +0.009 |
riet | +0.009 |
kids | +0.009 |
Sue | +0.009 |
Neg logit contributions | |
most | -0.011 |
biggest | -0.010 |
truly | -0.010 |
whole | -0.010 |
fastest | -0.010 |
Token | rested |
Token ID | 29263 |
Feature activation | -3.32e-02 |
Pos logit contributions | |
fastest | +0.106 |
highest | +0.105 |
biggest | +0.103 |
most | +0.101 |
ute | +0.101 |
Neg logit contributions | |
Sophie | -0.123 |
stro | -0.119 |
hugged | -0.118 |
her | -0.117 |
ens | -0.117 |
Token | and |
Token ID | 290 |
Feature activation | -7.07e-02 |
Pos logit contributions | |
ens | +0.042 |
Lee | +0.042 |
Sophie | +0.041 |
handing | +0.040 |
her | +0.039 |
Neg logit contributions | |
biggest | -0.069 |
highest | -0.068 |
fastest | -0.068 |
most | -0.065 |
coolest | -0.064 |
Token | each |
Token ID | 1123 |
Feature activation | +5.30e-03 |
Pos logit contributions | |
most | +0.018 |
biggest | +0.017 |
best | +0.016 |
whole | +0.016 |
highest | +0.016 |
Neg logit contributions | |
handing | -0.022 |
riet | -0.022 |
Lee | -0.022 |
kids | -0.022 |
ens | -0.022 |
Token | a |
Token ID | 257 |
Feature activation | +4.45e-02 |
Pos logit contributions | |
most | +0.006 |
biggest | +0.006 |
highest | +0.006 |
whole | +0.006 |
fastest | +0.006 |
Neg logit contributions | |
ens | -0.012 |
Lee | -0.012 |
riet | -0.011 |
handing | -0.011 |
kids | -0.011 |
Token | hug |
Token ID | 16225 |
Feature activation | -5.28e-02 |
Pos logit contributions | |
fastest | +0.096 |
most | +0.095 |
biggest | +0.093 |
highest | +0.092 |
truly | +0.091 |
Neg logit contributions | |
Lee | -0.055 |
ens | -0.055 |
Sue | -0.052 |
handing | -0.051 |
stro | -0.050 |
Token | and |
Token ID | 290 |
Feature activation | -8.60e-02 |
Pos logit contributions | |
ens | +0.072 |
Lee | +0.072 |
Sue | +0.066 |
handing | +0.065 |
riet | +0.065 |
Neg logit contributions | |
biggest | -0.110 |
most | -0.109 |
fastest | -0.107 |
highest | -0.107 |
truly | -0.103 |
Token | a |
Token ID | 257 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
ens | +0.156 |
Lee | +0.152 |
riet | +0.151 |
kids | +0.149 |
handing | +0.145 |
Neg logit contributions | |
most | -0.140 |
biggest | -0.137 |
best | -0.132 |
highest | -0.128 |
whole | -0.127 |
Token | kiss |
Token ID | 9245 |
Feature activation | -4.35e-02 |
Pos logit contributions | |
most | +0.075 |
fastest | +0.075 |
biggest | +0.075 |
highest | +0.073 |
truly | +0.072 |
Neg logit contributions | |
ens | -0.029 |
Lee | -0.029 |
Sue | -0.027 |
handing | -0.027 |
stro | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | +2.85e-03 |
Pos logit contributions | |
Lee | +0.074 |
ens | +0.074 |
Sue | +0.069 |
handing | +0.068 |
Sophie | +0.068 |
Neg logit contributions | |
biggest | -0.076 |
most | -0.075 |
highest | -0.074 |
fastest | -0.074 |
truly | -0.071 |
Token | He |
Token ID | 679 |
Feature activation | +7.85e-03 |
Pos logit contributions | |
most | +0.004 |
biggest | +0.004 |
best | +0.004 |
whole | +0.004 |
highest | +0.004 |
Neg logit contributions | |
pa | -0.005 |
ens | -0.005 |
ppa | -0.005 |
riet | -0.005 |
kids | -0.005 |
Token | gave |
Token ID | 2921 |
Feature activation | -7.77e-02 |
Pos logit contributions | |
most | +0.017 |
best | +0.016 |
biggest | +0.016 |
truly | +0.016 |
fastest | +0.016 |
Neg logit contributions | |
ens | -0.010 |
riet | -0.010 |
Lee | -0.010 |
kids | -0.009 |
handing | -0.009 |
Token | them |
Token ID | 606 |
Feature activation | +1.43e-02 |
Pos logit contributions | |
handing | +0.115 |
ens | +0.112 |
kids | +0.111 |
Lee | +0.111 |
riet | +0.110 |
Neg logit contributions | |
most | -0.152 |
biggest | -0.149 |
best | -0.146 |
fastest | -0.140 |
highest | -0.140 |
Token | a |
Token ID | 257 |
Feature activation | +3.77e-02 |
Pos logit contributions | |
most | +0.020 |
biggest | +0.020 |
best | +0.019 |
whole | +0.019 |
highest | +0.018 |
Neg logit contributions | |
Lee | -0.029 |
riet | -0.029 |
ens | -0.029 |
handing | -0.029 |
kids | -0.028 |
Token | Jim |
Token ID | 5395 |
Feature activation | -9.01e-02 |
Pos logit contributions | |
most | +0.040 |
biggest | +0.039 |
highest | +0.038 |
fastest | +0.037 |
whole | +0.037 |
Neg logit contributions | |
ens | -0.038 |
riet | -0.036 |
pa | -0.036 |
kids | -0.035 |
Lee | -0.035 |
Token | and |
Token ID | 290 |
Feature activation | -8.78e-02 |
Pos logit contributions | |
ens | +0.173 |
Lee | +0.171 |
riet | +0.167 |
kids | +0.164 |
Sophie | +0.161 |
Neg logit contributions | |
most | -0.136 |
biggest | -0.129 |
whole | -0.128 |
fastest | -0.127 |
truly | -0.127 |
Token | his |
Token ID | 465 |
Feature activation | -1.59e-02 |
Pos logit contributions | |
ppa | +0.122 |
ens | +0.117 |
kids | +0.114 |
annah | +0.112 |
riet | +0.111 |
Neg logit contributions | |
most | -0.175 |
whole | -0.168 |
best | -0.165 |
biggest | -0.161 |
power | -0.158 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
ens | +0.018 |
riet | +0.018 |
Lee | +0.018 |
handing | +0.017 |
Sophie | +0.017 |
Neg logit contributions | |
most | -0.037 |
biggest | -0.037 |
best | -0.036 |
whole | -0.035 |
highest | -0.035 |
Token | cheered |
Token ID | 34767 |
Feature activation | -5.59e-02 |
Pos logit contributions | |
Lee | +0.156 |
riet | +0.155 |
ens | +0.152 |
kids | +0.147 |
Sophie | +0.147 |
Neg logit contributions | |
most | -0.196 |
biggest | -0.190 |
fastest | -0.189 |
whole | -0.188 |
best | -0.188 |
Token | . |
Token ID | 13 |
Feature activation | +9.47e-03 |
Pos logit contributions | |
ens | +0.077 |
Lee | +0.076 |
riet | +0.072 |
handing | +0.071 |
Sophie | +0.071 |
Neg logit contributions | |
biggest | -0.115 |
most | -0.114 |
fastest | -0.112 |
highest | -0.112 |
best | -0.108 |
Token | They |
Token ID | 1119 |
Feature activation | +3.00e-02 |
Pos logit contributions | |
most | +0.016 |
biggest | +0.016 |
highest | +0.015 |
fastest | +0.015 |
whole | +0.015 |
Neg logit contributions | |
ens | -0.017 |
Lee | -0.016 |
riet | -0.016 |
kids | -0.016 |
purs | -0.016 |
Token | had |
Token ID | 550 |
Feature activation | +6.78e-02 |
Pos logit contributions | |
most | +0.054 |
biggest | +0.053 |
truly | +0.053 |
whole | +0.052 |
fastest | +0.052 |
Neg logit contributions | |
Lee | -0.048 |
ens | -0.047 |
riet | -0.046 |
kids | -0.045 |
handing | -0.045 |
Token | done |
Token ID | 1760 |
Feature activation | +3.51e-02 |
Pos logit contributions | |
ute | +0.114 |
biggest | +0.113 |
fastest | +0.111 |
highest | +0.110 |
most | +0.109 |
Neg logit contributions | |
Sue | -0.113 |
Tom | -0.111 |
begged | -0.111 |
hugged | -0.110 |
Sophie | -0.110 |
Token | the |
Token ID | 262 |
Feature activation | +3.47e-02 |
Pos logit contributions | |
fastest | +0.070 |
biggest | +0.069 |
highest | +0.068 |
most | +0.066 |
coolest | +0.065 |
Neg logit contributions | |
Lee | -0.052 |
Tom | -0.050 |
Sophie | -0.049 |
Sue | -0.049 |
handing | -0.049 |
Token | bush |
Token ID | 24484 |
Feature activation | +4.40e-02 |
Pos logit contributions | |
ifying | +0.046 |
possible | +0.040 |
ute | +0.039 |
hest | +0.039 |
arest | +0.039 |
Neg logit contributions | |
ens | -0.072 |
Sue | -0.070 |
Amy | -0.069 |
Lee | -0.067 |
Sophie | -0.067 |
Token | They |
Token ID | 1119 |
Feature activation | +2.51e-02 |
Pos logit contributions | |
most | +0.028 |
biggest | +0.028 |
fastest | +0.028 |
highest | +0.028 |
whole | +0.027 |
Neg logit contributions | |
ens | -0.025 |
Lee | -0.024 |
riet | -0.023 |
Sophie | -0.023 |
Sue | -0.023 |
Token | were |
Token ID | 547 |
Feature activation | +6.47e-02 |
Pos logit contributions | |
most | +0.039 |
biggest | +0.037 |
truly | +0.037 |
best | +0.037 |
whole | +0.036 |
Neg logit contributions | |
Lee | -0.048 |
ens | -0.046 |
handing | -0.046 |
riet | -0.046 |
kids | -0.045 |
Token | best |
Token ID | 1266 |
Feature activation | -3.55e-03 |
Pos logit contributions | |
biggest | +0.109 |
highest | +0.108 |
fastest | +0.107 |
most | +0.106 |
ifying | +0.104 |
Neg logit contributions | |
Sue | -0.105 |
ens | -0.104 |
Sophie | -0.103 |
Lee | -0.102 |
Amy | -0.100 |
Token | friends |
Token ID | 2460 |
Feature activation | -3.41e-02 |
Pos logit contributions | |
ens | +0.006 |
Lee | +0.006 |
riet | +0.006 |
Sophie | +0.005 |
handing | +0.005 |
Neg logit contributions | |
most | -0.006 |
biggest | -0.006 |
best | -0.006 |
fastest | -0.006 |
highest | -0.006 |
Token | forever |
Token ID | 8097 |
Feature activation | +4.51e-03 |
Pos logit contributions | |
ens | +0.055 |
Lee | +0.055 |
Sophie | +0.053 |
Sue | +0.053 |
Amy | +0.053 |
Neg logit contributions | |
biggest | -0.062 |
fastest | -0.061 |
highest | -0.060 |
most | -0.059 |
coolest | -0.056 |
Token | . |
Token ID | 13 |
Feature activation | +1.75e-02 |
Pos logit contributions | |
biggest | +0.010 |
fastest | +0.009 |
highest | +0.009 |
most | +0.009 |
coolest | +0.009 |
Neg logit contributions | |
ens | -0.006 |
Sue | -0.006 |
stro | -0.006 |
Lee | -0.006 |
Sophie | -0.006 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.79e-02 |
Pos logit contributions | |
fastest | +0.035 |
highest | +0.034 |
biggest | +0.034 |
most | +0.033 |
possible | +0.032 |
Neg logit contributions | |
ens | -0.027 |
Sophie | -0.025 |
Lee | -0.024 |
Sue | -0.024 |
Amy | -0.024 |
Token | Once |
Token ID | 7454 |
Feature activation | -4.58e-02 |
Pos logit contributions | |
biggest | +0.032 |
fastest | +0.031 |
most | +0.031 |
highest | +0.031 |
coolest | +0.029 |
Neg logit contributions | |
Lee | -0.030 |
ens | -0.030 |
Sue | -0.029 |
Sophie | -0.028 |
Amy | -0.028 |
Token | upon |
Token ID | 2402 |
Feature activation | +8.00e-02 |
Pos logit contributions | |
Lee | +0.091 |
ens | +0.090 |
kids | +0.089 |
handing | +0.089 |
riet | +0.088 |
Neg logit contributions | |
most | -0.065 |
biggest | -0.064 |
best | -0.062 |
whole | -0.061 |
fastest | -0.060 |
Token | a |
Token ID | 257 |
Feature activation | -8.96e-03 |
Pos logit contributions | |
fastest | +0.144 |
highest | +0.137 |
biggest | +0.135 |
most | +0.129 |
ifying | +0.127 |
Neg logit contributions | |
Sue | -0.137 |
Amy | -0.135 |
her | -0.134 |
Tom | -0.132 |
ens | -0.129 |
Token | time |
Token ID | 640 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
ens | +0.022 |
Sue | +0.021 |
Lee | +0.020 |
Amy | +0.020 |
Sophie | +0.020 |
Neg logit contributions | |
biggest | -0.009 |
most | -0.009 |
fastest | -0.009 |
highest | -0.009 |
ifying | -0.008 |
Token | it |
Token ID | 340 |
Feature activation | +8.47e-02 |
Pos logit contributions | |
ens | +0.046 |
Lee | +0.044 |
Sue | +0.041 |
Sophie | +0.041 |
riet | +0.041 |
Neg logit contributions | |
biggest | -0.068 |
most | -0.068 |
fastest | -0.067 |
highest | -0.066 |
truly | -0.065 |
Token | to |
Token ID | 284 |
Feature activation | -8.98e-02 |
Pos logit contributions | |
most | +0.149 |
biggest | +0.148 |
fastest | +0.144 |
highest | +0.142 |
best | +0.139 |
Neg logit contributions | |
ens | -0.140 |
Lee | -0.137 |
riet | -0.131 |
Sue | -0.131 |
handing | -0.131 |
Token | the |
Token ID | 262 |
Feature activation | +8.89e-03 |
Pos logit contributions | |
ens | +0.120 |
riet | +0.117 |
Lee | +0.116 |
kids | +0.115 |
handing | +0.110 |
Neg logit contributions | |
most | -0.188 |
biggest | -0.187 |
highest | -0.182 |
whole | -0.180 |
fastest | -0.179 |
Token | ducks |
Token ID | 39694 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
truly | +0.016 |
possible | +0.016 |
most | +0.015 |
arest | +0.015 |
fastest | +0.015 |
Neg logit contributions | |
ens | -0.013 |
Lee | -0.013 |
handing | -0.013 |
purs | -0.013 |
Sue | -0.013 |
Token | . |
Token ID | 13 |
Feature activation | +4.42e-03 |
Pos logit contributions | |
ens | +0.039 |
Lee | +0.039 |
riet | +0.038 |
kids | +0.038 |
Sophie | +0.037 |
Neg logit contributions | |
most | -0.032 |
biggest | -0.032 |
fastest | -0.031 |
highest | -0.031 |
whole | -0.030 |
Token | The |
Token ID | 383 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
most | +0.006 |
biggest | +0.006 |
highest | +0.006 |
best | +0.006 |
whole | +0.006 |
Neg logit contributions | |
pa | -0.009 |
riet | -0.009 |
kids | -0.009 |
ens | -0.009 |
ppa | -0.009 |
Token | ducks |
Token ID | 39694 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
ifying | +0.037 |
truly | +0.036 |
hest | +0.035 |
possible | +0.035 |
highest | +0.035 |
Neg logit contributions | |
ens | -0.019 |
Sue | -0.017 |
Tom | -0.017 |
Lee | -0.017 |
dads | -0.017 |
Token | splash |
Token ID | 22870 |
Feature activation | +4.18e-03 |
Pos logit contributions | |
ens | +0.039 |
Lee | +0.038 |
riet | +0.036 |
kids | +0.036 |
handing | +0.036 |
Neg logit contributions | |
most | -0.045 |
biggest | -0.045 |
fastest | -0.044 |
highest | -0.043 |
best | -0.043 |
Token | in |
Token ID | 287 |
Feature activation | -2.15e-02 |
Pos logit contributions | |
biggest | +0.009 |
most | +0.009 |
fastest | +0.008 |
highest | +0.008 |
truly | +0.008 |
Neg logit contributions | |
ens | -0.006 |
Lee | -0.006 |
Sue | -0.005 |
riet | -0.005 |
Sophie | -0.005 |
Token | the |
Token ID | 262 |
Feature activation | +1.12e-02 |
Pos logit contributions | |
riet | +0.032 |
ens | +0.032 |
kids | +0.031 |
Lee | +0.030 |
pa | +0.029 |
Neg logit contributions | |
most | -0.043 |
biggest | -0.042 |
whole | -0.042 |
highest | -0.040 |
best | -0.040 |
Token | water |
Token ID | 1660 |
Feature activation | +5.22e-02 |
Pos logit contributions | |
possible | +0.021 |
truly | +0.020 |
most | +0.019 |
fastest | +0.019 |
impossible | +0.019 |
Neg logit contributions | |
ens | -0.018 |
Sue | -0.017 |
purs | -0.016 |
hugged | -0.016 |
hugs | -0.016 |
Token | !" |
Token ID | 2474 |
Feature activation | -3.25e-02 |
Pos logit contributions | |
biggest | +0.065 |
most | +0.065 |
highest | +0.062 |
fastest | +0.062 |
truly | +0.061 |
Neg logit contributions | |
Lee | -0.067 |
ens | -0.066 |
Sue | -0.063 |
handing | -0.062 |
kids | -0.061 |
Token | He |
Token ID | 679 |
Feature activation | -4.07e-04 |
Pos logit contributions | |
ens | +0.053 |
Lee | +0.051 |
riet | +0.051 |
kids | +0.049 |
handing | +0.048 |
Neg logit contributions | |
most | -0.059 |
biggest | -0.058 |
highest | -0.056 |
whole | -0.055 |
fastest | -0.055 |
Token | waited |
Token ID | 13488 |
Feature activation | -1.81e-02 |
Pos logit contributions | |
ens | +0.001 |
riet | +0.001 |
Lee | +0.001 |
handing | +0.001 |
kids | +0.001 |
Neg logit contributions | |
most | -0.001 |
biggest | -0.001 |
best | -0.001 |
fastest | -0.001 |
whole | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | -8.68e-02 |
Pos logit contributions | |
ens | +0.029 |
Lee | +0.028 |
riet | +0.028 |
kids | +0.028 |
purs | +0.027 |
Neg logit contributions | |
most | -0.033 |
biggest | -0.033 |
highest | -0.031 |
best | -0.031 |
fastest | -0.031 |
Token | waited |
Token ID | 13488 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
riet | +0.140 |
ens | +0.140 |
kids | +0.139 |
Lee | +0.137 |
purs | +0.136 |
Neg logit contributions | |
most | -0.152 |
whole | -0.143 |
truly | -0.143 |
biggest | -0.140 |
best | -0.140 |
Token | , |
Token ID | 11 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
ens | +0.024 |
Lee | +0.024 |
riet | +0.023 |
kids | +0.023 |
Sue | +0.022 |
Neg logit contributions | |
most | -0.026 |
biggest | -0.026 |
fastest | -0.025 |
highest | -0.025 |
best | -0.025 |
Token | but |
Token ID | 475 |
Feature activation | +5.39e-03 |
Pos logit contributions | |
riet | +0.179 |
kids | +0.175 |
ens | +0.174 |
pa | +0.170 |
purs | +0.169 |
Neg logit contributions | |
most | -0.192 |
biggest | -0.177 |
fastest | -0.171 |
highest | -0.171 |
best | -0.171 |
Token | nothing |
Token ID | 2147 |
Feature activation | +9.90e-02 |
Pos logit contributions | |
biggest | +0.010 |
most | +0.009 |
highest | +0.009 |
fastest | +0.009 |
whole | +0.009 |
Neg logit contributions | |
ens | -0.009 |
Lee | -0.009 |
Sue | -0.008 |
Sophie | -0.008 |
riet | -0.008 |
Token | happened |
Token ID | 3022 |
Feature activation | -1.68e-03 |
Pos logit contributions | |
biggest | +0.183 |
ifying | +0.181 |
highest | +0.181 |
most | +0.177 |
fastest | +0.173 |
Neg logit contributions | |
ens | -0.170 |
Sue | -0.155 |
begged | -0.150 |
hugged | -0.147 |
Amy | -0.147 |
Token | . |
Token ID | 13 |
Feature activation | +2.23e-02 |
Pos logit contributions | |
Lee | +0.003 |
handing | +0.003 |
ens | +0.002 |
Sue | +0.002 |
Sophie | +0.002 |
Neg logit contributions | |
biggest | -0.003 |
most | -0.003 |
highest | -0.003 |
fastest | -0.003 |
coolest | -0.003 |
Token | Then |
Token ID | 3244 |
Feature activation | -3.18e-02 |
Pos logit contributions | |
most | +0.040 |
biggest | +0.040 |
fastest | +0.039 |
highest | +0.039 |
best | +0.038 |
Neg logit contributions | |
ens | -0.036 |
Lee | -0.035 |
riet | -0.034 |
Sue | -0.034 |
Sophie | -0.033 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +2.48e-02 |
Pos logit contributions | |
biggest | +0.043 |
most | +0.043 |
fastest | +0.042 |
highest | +0.041 |
best | +0.040 |
Neg logit contributions | |
ens | -0.046 |
Lee | -0.044 |
riet | -0.043 |
Sue | -0.043 |
Sophie | -0.043 |
Token | Once |
Token ID | 7454 |
Feature activation | -4.07e-02 |
Pos logit contributions | |
biggest | +0.042 |
most | +0.042 |
fastest | +0.041 |
highest | +0.041 |
best | +0.039 |
Neg logit contributions | |
ens | -0.043 |
Lee | -0.042 |
Sue | -0.041 |
Sophie | -0.040 |
Amy | -0.039 |
Token | upon |
Token ID | 2402 |
Feature activation | +9.16e-02 |
Pos logit contributions | |
Lee | +0.080 |
ens | +0.080 |
kids | +0.078 |
handing | +0.078 |
riet | +0.078 |
Neg logit contributions | |
most | -0.058 |
biggest | -0.057 |
best | -0.055 |
whole | -0.054 |
fastest | -0.054 |
Token | a |
Token ID | 257 |
Feature activation | -9.12e-05 |
Pos logit contributions | |
fastest | +0.154 |
highest | +0.147 |
biggest | +0.147 |
most | +0.141 |
ifying | +0.141 |
Neg logit contributions | |
Sue | -0.162 |
Amy | -0.158 |
ens | -0.157 |
Tom | -0.156 |
her | -0.155 |
Token | time |
Token ID | 640 |
Feature activation | +2.23e-02 |
Pos logit contributions | |
ens | +0.000 |
Lee | +0.000 |
riet | +0.000 |
handing | +0.000 |
kids | +0.000 |
Neg logit contributions | |
most | -0.000 |
biggest | -0.000 |
fastest | -0.000 |
highest | -0.000 |
whole | -0.000 |
Token | , |
Token ID | 11 |
Feature activation | -1.35e-01 |
Pos logit contributions | |
most | +0.046 |
biggest | +0.046 |
best | +0.045 |
whole | +0.044 |
fastest | +0.044 |
Neg logit contributions | |
ens | -0.030 |
riet | -0.029 |
kids | -0.029 |
Lee | -0.029 |
handing | -0.028 |
Token | there |
Token ID | 612 |
Feature activation | +5.37e-02 |
Pos logit contributions | |
ens | +0.201 |
Sue | +0.201 |
Amy | +0.196 |
Lee | +0.195 |
Tom | +0.188 |
Neg logit contributions | |
biggest | -0.253 |
highest | -0.252 |
fastest | -0.249 |
most | -0.246 |
truly | -0.240 |
Token | was |
Token ID | 373 |
Feature activation | +6.20e-02 |
Pos logit contributions | |
biggest | +0.079 |
highest | +0.079 |
fastest | +0.078 |
most | +0.077 |
ifying | +0.073 |
Neg logit contributions | |
ens | -0.104 |
Lee | -0.103 |
Sue | -0.102 |
Sophie | -0.099 |
Amy | -0.098 |
Token | a |
Token ID | 257 |
Feature activation | +3.42e-02 |
Pos logit contributions | |
most | +0.076 |
biggest | +0.075 |
fastest | +0.072 |
highest | +0.071 |
best | +0.069 |
Neg logit contributions | |
ens | -0.136 |
Lee | -0.134 |
riet | -0.131 |
kids | -0.130 |
handing | -0.129 |
Token | little |
Token ID | 1310 |
Feature activation | -2.86e-02 |
Pos logit contributions | |
most | +0.052 |
biggest | +0.052 |
whole | +0.051 |
best | +0.050 |
fastest | +0.049 |
Neg logit contributions | |
ens | -0.066 |
riet | -0.064 |
Lee | -0.064 |
kids | -0.063 |
handing | -0.062 |
Token | girl |
Token ID | 2576 |
Feature activation | -2.99e-02 |
Pos logit contributions | |
ens | +0.047 |
Lee | +0.046 |
riet | +0.045 |
kids | +0.045 |
handing | +0.045 |
Neg logit contributions | |
most | -0.050 |
biggest | -0.050 |
fastest | -0.048 |
highest | -0.048 |
best | -0.048 |
Token | His |
Token ID | 6653 |
Feature activation | -2.22e-02 |
Pos logit contributions | |
most | +0.084 |
whole | +0.081 |
fastest | +0.081 |
biggest | +0.080 |
highest | +0.079 |
Neg logit contributions | |
ens | -0.093 |
Lee | -0.091 |
kids | -0.087 |
handing | -0.087 |
riet | -0.087 |
Token | friends |
Token ID | 2460 |
Feature activation | -5.28e-02 |
Pos logit contributions | |
ens | +0.026 |
Lee | +0.025 |
riet | +0.024 |
handing | +0.023 |
kids | +0.023 |
Neg logit contributions | |
most | -0.051 |
biggest | -0.051 |
best | -0.050 |
highest | -0.049 |
fastest | -0.049 |
Token | nodded |
Token ID | 14464 |
Feature activation | -2.85e-02 |
Pos logit contributions | |
ens | +0.078 |
Lee | +0.077 |
riet | +0.075 |
kids | +0.073 |
handing | +0.071 |
Neg logit contributions | |
most | -0.103 |
best | -0.098 |
truly | -0.098 |
biggest | -0.097 |
fastest | -0.096 |
Token | and |
Token ID | 290 |
Feature activation | -8.86e-02 |
Pos logit contributions | |
ens | +0.049 |
riet | +0.046 |
Lee | +0.045 |
kids | +0.045 |
purs | +0.045 |
Neg logit contributions | |
most | -0.051 |
whole | -0.049 |
best | -0.049 |
biggest | -0.048 |
truly | -0.047 |
Token | said |
Token ID | 531 |
Feature activation | -5.25e-02 |
Pos logit contributions | |
kids | +0.114 |
ens | +0.111 |
riet | +0.109 |
Lee | +0.107 |
olph | +0.100 |
Neg logit contributions | |
most | -0.192 |
best | -0.187 |
truly | -0.182 |
biggest | -0.180 |
whole | -0.173 |
Token | , |
Token ID | 11 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
ens | +0.071 |
kids | +0.067 |
Lee | +0.067 |
riet | +0.067 |
pa | +0.063 |
Neg logit contributions | |
most | -0.110 |
biggest | -0.108 |
best | -0.106 |
fastest | -0.104 |
highest | -0.103 |
Token | " |
Token ID | 366 |
Feature activation | +5.88e-02 |
Pos logit contributions | |
ens | +0.173 |
riet | +0.163 |
kids | +0.162 |
pa | +0.159 |
purs | +0.158 |
Neg logit contributions | |
most | -0.177 |
biggest | -0.172 |
best | -0.166 |
whole | -0.166 |
highest | -0.165 |
Token | By |
Token ID | 3886 |
Feature activation | -1.23e-03 |
Pos logit contributions | |
biggest | +0.107 |
best | +0.103 |
most | +0.102 |
highest | +0.100 |
fastest | +0.100 |
Neg logit contributions | |
ens | -0.096 |
riet | -0.093 |
Sophie | -0.091 |
Lee | -0.090 |
Sue | -0.090 |
Token | e |
Token ID | 68 |
Feature activation | +5.18e-03 |
Pos logit contributions | |
Lee | +0.002 |
ens | +0.002 |
Sue | +0.002 |
Amy | +0.002 |
pa | +0.002 |
Neg logit contributions | |
biggest | -0.002 |
highest | -0.002 |
most | -0.002 |
fastest | -0.002 |
coolest | -0.002 |
Token | ! |
Token ID | 0 |
Feature activation | +4.67e-02 |
Pos logit contributions | |
most | +0.009 |
whole | +0.009 |
biggest | +0.009 |
best | +0.009 |
fastest | +0.009 |
Neg logit contributions | |
riet | -0.008 |
ens | -0.008 |
Lee | -0.008 |
purs | -0.008 |
kids | -0.008 |
Token | We |
Token ID | 775 |
Feature activation | +4.39e-02 |
Pos logit contributions | |
biggest | +0.081 |
most | +0.080 |
highest | +0.079 |
fastest | +0.078 |
whole | +0.078 |
Neg logit contributions | |
Lee | -0.081 |
ens | -0.077 |
Sue | -0.077 |
Tom | -0.075 |
Sophie | -0.074 |
Token | cob |
Token ID | 22843 |
Feature activation | +3.61e-02 |
Pos logit contributions | |
biggest | +0.065 |
most | +0.064 |
highest | +0.064 |
fastest | +0.064 |
truly | +0.064 |
Neg logit contributions | |
ens | -0.067 |
Lee | -0.064 |
purs | -0.063 |
kids | -0.062 |
pa | -0.061 |
Token | we |
Token ID | 732 |
Feature activation | +5.91e-02 |
Pos logit contributions | |
biggest | +0.061 |
most | +0.059 |
fastest | +0.059 |
highest | +0.059 |
whole | +0.059 |
Neg logit contributions | |
ens | -0.061 |
Lee | -0.061 |
pa | -0.059 |
Sue | -0.057 |
olph | -0.056 |
Token | bs |
Token ID | 1443 |
Feature activation | +2.94e-02 |
Pos logit contributions | |
biggest | +0.124 |
most | +0.124 |
fastest | +0.120 |
highest | +0.118 |
best | +0.117 |
Neg logit contributions | |
ens | -0.081 |
hugged | -0.073 |
Lee | -0.071 |
riet | -0.070 |
stro | -0.070 |
Token | ," |
Token ID | 553 |
Feature activation | -9.50e-02 |
Pos logit contributions | |
biggest | +0.057 |
most | +0.056 |
fastest | +0.054 |
highest | +0.054 |
truly | +0.053 |
Neg logit contributions | |
ens | -0.043 |
Lee | -0.042 |
Sue | -0.039 |
Sophie | -0.039 |
Amy | -0.039 |
Token | Lily |
Token ID | 20037 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
kids | +0.109 |
ppa | +0.104 |
pa | +0.099 |
handing | +0.097 |
Lee | +0.094 |
Neg logit contributions | |
most | -0.231 |
best | -0.216 |
biggest | -0.215 |
perfect | -0.204 |
whole | -0.202 |
Token | said |
Token ID | 531 |
Feature activation | -8.62e-02 |
Pos logit contributions | |
ens | +0.159 |
riet | +0.153 |
kids | +0.150 |
Lee | +0.148 |
olph | +0.147 |
Neg logit contributions | |
most | -0.189 |
truly | -0.175 |
best | -0.168 |
biggest | -0.164 |
highest | -0.163 |
Token | . |
Token ID | 13 |
Feature activation | -7.97e-03 |
Pos logit contributions | |
kids | +0.191 |
ens | +0.187 |
riet | +0.181 |
purs | +0.180 |
Lee | +0.176 |
Neg logit contributions | |
best | -0.106 |
most | -0.104 |
biggest | -0.102 |
whole | -0.094 |
highest | -0.094 |
Token | She |
Token ID | 1375 |
Feature activation | -6.33e-03 |
Pos logit contributions | |
kids | +0.018 |
pa | +0.017 |
ens | +0.017 |
purs | +0.016 |
riet | +0.016 |
Neg logit contributions | |
most | -0.010 |
biggest | -0.010 |
highest | -0.010 |
best | -0.009 |
whole | -0.009 |
Token | touched |
Token ID | 12615 |
Feature activation | -4.59e-02 |
Pos logit contributions | |
ens | +0.011 |
Lee | +0.010 |
riet | +0.010 |
kids | +0.010 |
Gig | +0.010 |
Neg logit contributions | |
most | -0.012 |
biggest | -0.011 |
truly | -0.011 |
whole | -0.011 |
highest | -0.011 |
Token | one |
Token ID | 530 |
Feature activation | +7.37e-02 |
Pos logit contributions | |
ens | +0.056 |
Lee | +0.054 |
Sue | +0.050 |
Sophie | +0.050 |
riet | +0.050 |
Neg logit contributions | |
biggest | -0.101 |
most | -0.101 |
fastest | -0.100 |
highest | -0.099 |
whole | -0.098 |
Token | with |
Token ID | 351 |
Feature activation | -9.23e-03 |
Pos logit contributions | |
fastest | +0.132 |
biggest | +0.129 |
most | +0.128 |
truly | +0.125 |
whole | +0.124 |
Neg logit contributions | |
ens | -0.112 |
Sue | -0.109 |
Lee | -0.108 |
Amy | -0.105 |
handing | -0.104 |
Token | with |
Token ID | 351 |
Feature activation | +2.08e-03 |
Pos logit contributions | |
biggest | +0.083 |
most | +0.083 |
fastest | +0.080 |
highest | +0.080 |
whole | +0.078 |
Neg logit contributions | |
ens | -0.076 |
Lee | -0.075 |
riet | -0.072 |
Sue | -0.071 |
kids | -0.071 |
Token | animals |
Token ID | 4695 |
Feature activation | +1.60e-02 |
Pos logit contributions | |
most | +0.004 |
biggest | +0.004 |
best | +0.003 |
whole | +0.003 |
highest | +0.003 |
Neg logit contributions | |
ens | -0.004 |
riet | -0.004 |
Lee | -0.004 |
olph | -0.003 |
handing | -0.003 |
Token | , |
Token ID | 11 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
most | +0.033 |
biggest | +0.033 |
fastest | +0.032 |
highest | +0.032 |
whole | +0.031 |
Neg logit contributions | |
ens | -0.022 |
Lee | -0.021 |
riet | -0.020 |
kids | -0.020 |
handing | -0.020 |
Token | cars |
Token ID | 5006 |
Feature activation | +6.00e-03 |
Pos logit contributions | |
ens | +0.164 |
riet | +0.164 |
Lee | +0.159 |
kids | +0.153 |
olph | +0.152 |
Neg logit contributions | |
most | -0.195 |
biggest | -0.187 |
fastest | -0.181 |
best | -0.179 |
highest | -0.178 |
Token | , |
Token ID | 11 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
most | +0.012 |
biggest | +0.012 |
fastest | +0.012 |
best | +0.012 |
whole | +0.012 |
Neg logit contributions | |
riet | -0.008 |
ens | -0.008 |
Lee | -0.007 |
Sophie | -0.007 |
kids | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | -7.98e-02 |
Pos logit contributions | |
riet | +0.162 |
ens | +0.157 |
Lee | +0.153 |
Sophie | +0.152 |
Samantha | +0.151 |
Neg logit contributions | |
most | -0.203 |
biggest | -0.195 |
best | -0.189 |
fastest | -0.189 |
highest | -0.187 |
Token | princess |
Token ID | 21752 |
Feature activation | -4.43e-02 |
Pos logit contributions | |
ens | +0.132 |
Lee | +0.129 |
riet | +0.128 |
kids | +0.122 |
Sophie | +0.122 |
Neg logit contributions | |
most | -0.145 |
biggest | -0.143 |
fastest | -0.136 |
best | -0.136 |
highest | -0.135 |
Token | es |
Token ID | 274 |
Feature activation | -3.68e-02 |
Pos logit contributions | |
ens | +0.066 |
riet | +0.065 |
Sophie | +0.063 |
Lee | +0.062 |
handing | +0.062 |
Neg logit contributions | |
most | -0.084 |
biggest | -0.083 |
best | -0.081 |
fastest | -0.081 |
whole | -0.080 |
Token | . |
Token ID | 13 |
Feature activation | +1.08e-02 |
Pos logit contributions | |
ens | +0.067 |
riet | +0.067 |
Lee | +0.065 |
kids | +0.064 |
handing | +0.063 |
Neg logit contributions | |
most | -0.058 |
biggest | -0.056 |
fastest | -0.054 |
highest | -0.054 |
best | -0.054 |
Token | Sometimes |
Token ID | 8975 |
Feature activation | +3.60e-02 |
Pos logit contributions | |
most | +0.020 |
biggest | +0.019 |
whole | +0.018 |
best | +0.018 |
highest | +0.018 |
Neg logit contributions | |
ppa | -0.017 |
ens | -0.017 |
riet | -0.017 |
pa | -0.017 |
olph | -0.016 |
Token | they |
Token ID | 484 |
Feature activation | +2.02e-02 |
Pos logit contributions | |
most | +0.069 |
biggest | +0.067 |
best | +0.066 |
fastest | +0.065 |
highest | +0.064 |
Neg logit contributions | |
ens | -0.055 |
riet | -0.053 |
Lee | -0.052 |
handing | -0.051 |
kids | -0.051 |