Token | the |
Token ID | 262 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
Along | +0.093 |
nights | +0.091 |
On | +0.083 |
way | +0.083 |
whenever | +0.082 |
Neg logit contributions | |
paramed | -0.097 |
liar | -0.095 |
boys | -0.094 |
wasteful | -0.093 |
roc | -0.091 |
Token | ambulance |
Token ID | 22536 |
Feature activation | +1.41e-01 |
Pos logit contributions | |
paramed | +0.018 |
liar | +0.017 |
boys | +0.017 |
wasteful | +0.017 |
mixer | +0.017 |
Neg logit contributions | |
Along | -0.025 |
nights | -0.024 |
way | -0.023 |
On | -0.022 |
drifted | -0.022 |
Token | he |
Token ID | 339 |
Feature activation | -8.86e-02 |
Pos logit contributions | |
paramed | +0.231 |
liar | +0.229 |
wasteful | +0.223 |
mixer | +0.222 |
greedy | +0.221 |
Neg logit contributions | |
Along | -0.212 |
nights | -0.208 |
way | -0.194 |
On | -0.187 |
whenever | -0.185 |
Token | saw |
Token ID | 2497 |
Feature activation | +4.15e-02 |
Pos logit contributions | |
Along | +0.169 |
nights | +0.166 |
way | +0.157 |
On | +0.152 |
drifted | +0.151 |
Neg logit contributions | |
paramed | -0.110 |
mixer | -0.106 |
liar | -0.106 |
mean | -0.105 |
boys | -0.104 |
Token | a |
Token ID | 257 |
Feature activation | -2.19e-02 |
Pos logit contributions | |
paramed | +0.067 |
liar | +0.065 |
wasteful | +0.064 |
boys | +0.063 |
mixer | +0.063 |
Neg logit contributions | |
nights | -0.064 |
Along | -0.063 |
way | -0.059 |
whenever | -0.057 |
days | -0.057 |
Token | paramed |
Token ID | 34581 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
day | +0.028 |
nights | +0.028 |
night | +0.028 |
way | +0.027 |
journey | +0.027 |
Neg logit contributions | |
roc | -0.038 |
boys | -0.038 |
wasteful | -0.037 |
chefs | -0.035 |
bluff | -0.035 |
Token | ic |
Token ID | 291 |
Feature activation | +4.91e-02 |
Pos logit contributions | |
liar | +0.535 |
paramed | +0.525 |
mean | +0.524 |
mixer | +0.519 |
wasteful | +0.518 |
Neg logit contributions | |
nights | -0.384 |
Along | -0.374 |
way | -0.368 |
ingle | -0.333 |
more | -0.328 |
Token | wearing |
Token ID | 5762 |
Feature activation | -1.17e-01 |
Pos logit contributions | |
paramed | +0.097 |
liar | +0.095 |
mixer | +0.095 |
boys | +0.095 |
wasteful | +0.095 |
Neg logit contributions | |
Along | -0.056 |
nights | -0.055 |
way | -0.051 |
drifted | -0.047 |
On | -0.046 |
Token | a |
Token ID | 257 |
Feature activation | -2.24e-02 |
Pos logit contributions | |
nights | +0.175 |
Along | +0.174 |
way | +0.160 |
whenever | +0.159 |
days | +0.157 |
Neg logit contributions | |
paramed | -0.188 |
roc | -0.186 |
liar | -0.186 |
wasteful | -0.186 |
boys | -0.185 |
Token | big |
Token ID | 1263 |
Feature activation | -1.89e-04 |
Pos logit contributions | |
day | +0.028 |
Along | +0.028 |
journey | +0.027 |
night | +0.027 |
lifetime | +0.025 |
Neg logit contributions | |
roc | -0.042 |
boys | -0.042 |
wasteful | -0.039 |
chefs | -0.038 |
eper | -0.038 |
Token | coat |
Token ID | 13209 |
Feature activation | +1.17e-03 |
Pos logit contributions | |
day | +0.000 |
journey | +0.000 |
night | +0.000 |
morning | +0.000 |
Along | +0.000 |
Neg logit contributions | |
boys | -0.000 |
roc | -0.000 |
wasteful | -0.000 |
chefs | -0.000 |
bluff | -0.000 |
Token | and |
Token ID | 290 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
paramed | +0.277 |
mixer | +0.273 |
mean | +0.271 |
liar | +0.269 |
boys | +0.268 |
Neg logit contributions | |
Along | -0.123 |
nights | -0.122 |
way | -0.104 |
drifted | -0.101 |
On | -0.099 |
Token | wearing |
Token ID | 5762 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
Along | +0.031 |
nights | +0.030 |
way | +0.029 |
On | +0.027 |
drifted | +0.027 |
Neg logit contributions | |
paramed | -0.026 |
mixer | -0.025 |
liar | -0.025 |
mean | -0.025 |
boys | -0.024 |
Token | gloves |
Token ID | 18051 |
Feature activation | +3.67e-02 |
Pos logit contributions | |
Along | +0.176 |
nights | +0.175 |
whenever | +0.160 |
On | +0.156 |
days | +0.156 |
Neg logit contributions | |
liar | -0.196 |
paramed | -0.195 |
wasteful | -0.194 |
roc | -0.189 |
boys | -0.187 |
Token | . |
Token ID | 13 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
paramed | +0.081 |
liar | +0.080 |
roc | +0.079 |
wasteful | +0.079 |
mixer | +0.078 |
Neg logit contributions | |
Along | -0.035 |
nights | -0.033 |
whenever | -0.031 |
way | -0.031 |
On | -0.028 |
Token | The |
Token ID | 383 |
Feature activation | -7.29e-03 |
Pos logit contributions | |
nights | +0.029 |
Along | +0.026 |
way | +0.026 |
drifted | +0.025 |
days | +0.025 |
Neg logit contributions | |
mixer | -0.035 |
paramed | -0.035 |
greedy | -0.033 |
stern | -0.033 |
liar | -0.033 |
Token | paramed |
Token ID | 34581 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
nights | +0.015 |
Along | +0.015 |
way | +0.015 |
days | +0.014 |
day | +0.014 |
Neg logit contributions | |
boys | -0.007 |
bluff | -0.007 |
liar | -0.006 |
roc | -0.006 |
chefs | -0.006 |
Token | ic |
Token ID | 291 |
Feature activation | +4.77e-02 |
Pos logit contributions | |
liar | +0.532 |
mean | +0.524 |
paramed | +0.523 |
mixer | +0.519 |
boys | +0.519 |
Neg logit contributions | |
nights | -0.382 |
Along | -0.370 |
way | -0.364 |
more | -0.326 |
ingle | -0.325 |
Token | was |
Token ID | 373 |
Feature activation | +2.78e-02 |
Pos logit contributions | |
boys | +0.081 |
bluff | +0.080 |
liar | +0.079 |
wasteful | +0.078 |
greedy | +0.077 |
Neg logit contributions | |
nights | -0.068 |
way | -0.065 |
Along | -0.065 |
drifted | -0.061 |
days | -0.060 |
Token | very |
Token ID | 845 |
Feature activation | +8.00e-02 |
Pos logit contributions | |
roc | +0.044 |
liar | +0.044 |
boys | +0.043 |
paramed | +0.041 |
mixer | +0.041 |
Neg logit contributions | |
nights | -0.044 |
Along | -0.043 |
whenever | -0.040 |
drifted | -0.040 |
night | -0.040 |
Token | polite |
Token ID | 23507 |
Feature activation | +9.89e-02 |
Pos logit contributions | |
boys | +0.124 |
paramed | +0.122 |
liar | +0.121 |
mixer | +0.120 |
roc | +0.117 |
Neg logit contributions | |
Along | -0.127 |
nights | -0.125 |
way | -0.119 |
On | -0.115 |
drifted | -0.114 |
Token | and |
Token ID | 290 |
Feature activation | -1.90e-02 |
Pos logit contributions | |
paramed | +0.166 |
mixer | +0.163 |
liar | +0.162 |
mean | +0.161 |
boys | +0.159 |
Neg logit contributions | |
Along | -0.145 |
nights | -0.140 |
way | -0.130 |
drifted | -0.126 |
On | -0.126 |
Token | . |
Token ID | 13 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
Along | +0.018 |
nights | +0.017 |
way | +0.016 |
drifted | +0.015 |
whenever | +0.015 |
Neg logit contributions | |
paramed | -0.040 |
mixer | -0.040 |
liar | -0.039 |
wasteful | -0.039 |
roc | -0.039 |
Token | The |
Token ID | 383 |
Feature activation | -2.82e-04 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.042 |
way | +0.040 |
On | +0.040 |
drifted | +0.039 |
Neg logit contributions | |
paramed | -0.040 |
mixer | -0.039 |
liar | -0.038 |
boys | -0.038 |
wasteful | -0.037 |
Token | driver |
Token ID | 4639 |
Feature activation | +8.98e-02 |
Pos logit contributions | |
nights | +0.001 |
Along | +0.001 |
way | +0.000 |
days | +0.000 |
drifted | +0.000 |
Neg logit contributions | |
roc | -0.000 |
boys | -0.000 |
paramed | -0.000 |
liar | -0.000 |
wasteful | -0.000 |
Token | and |
Token ID | 290 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
paramed | +0.134 |
liar | +0.131 |
mixer | +0.131 |
wasteful | +0.128 |
boys | +0.127 |
Neg logit contributions | |
Along | -0.148 |
nights | -0.146 |
way | -0.137 |
drifted | -0.134 |
On | -0.131 |
Token | the |
Token ID | 262 |
Feature activation | +1.64e-02 |
Pos logit contributions | |
nights | +0.034 |
drifted | +0.034 |
Along | +0.034 |
days | +0.032 |
night | +0.031 |
Neg logit contributions | |
paramed | -0.030 |
roc | -0.030 |
mixer | -0.029 |
liar | -0.029 |
boys | -0.028 |
Token | paramed |
Token ID | 34581 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
roc | +0.021 |
boys | +0.021 |
wasteful | +0.021 |
bluff | +0.020 |
liar | +0.020 |
Neg logit contributions | |
nights | -0.030 |
Along | -0.029 |
way | -0.027 |
days | -0.027 |
night | -0.027 |
Token | ic |
Token ID | 291 |
Feature activation | +4.28e-02 |
Pos logit contributions | |
paramed | +0.482 |
liar | +0.475 |
mixer | +0.473 |
mean | +0.468 |
boys | +0.464 |
Neg logit contributions | |
Along | -0.428 |
nights | -0.426 |
way | -0.399 |
drifted | -0.375 |
On | -0.371 |
Token | jumped |
Token ID | 11687 |
Feature activation | -6.88e-02 |
Pos logit contributions | |
paramed | +0.063 |
mixer | +0.061 |
liar | +0.061 |
mean | +0.060 |
boys | +0.059 |
Neg logit contributions | |
Along | -0.073 |
nights | -0.071 |
way | -0.066 |
On | -0.065 |
drifted | -0.064 |
Token | out |
Token ID | 503 |
Feature activation | -3.00e-02 |
Pos logit contributions | |
Along | +0.091 |
nights | +0.089 |
way | +0.083 |
On | +0.079 |
drifted | +0.078 |
Neg logit contributions | |
paramed | -0.127 |
liar | -0.124 |
mixer | -0.123 |
wasteful | -0.122 |
boys | -0.121 |
Token | . |
Token ID | 13 |
Feature activation | -2.67e-02 |
Pos logit contributions | |
Along | +0.036 |
nights | +0.032 |
lifetime | +0.029 |
dark | +0.029 |
way | +0.028 |
Neg logit contributions | |
mean | -0.061 |
mad | -0.060 |
stern | -0.059 |
phone | -0.058 |
paramed | -0.058 |
Token | They |
Token ID | 1119 |
Feature activation | +4.09e-02 |
Pos logit contributions | |
Along | +0.040 |
nights | +0.040 |
way | +0.037 |
drifted | +0.035 |
more | +0.034 |
Neg logit contributions | |
paramed | -0.044 |
liar | -0.043 |
mixer | -0.043 |
mean | -0.042 |
wasteful | -0.042 |
Token |
|
Token ID | 198 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
nights | +0.042 |
Along | +0.039 |
drifted | +0.037 |
way | +0.035 |
days | +0.035 |
Neg logit contributions | |
paramed | -0.052 |
mixer | -0.051 |
mean | -0.050 |
stern | -0.049 |
greedy | -0.048 |
Token | L |
Token ID | 43 |
Feature activation | -3.39e-02 |
Pos logit contributions | |
Along | +0.047 |
nights | +0.046 |
way | +0.043 |
drifted | +0.042 |
On | +0.041 |
Neg logit contributions | |
paramed | -0.046 |
mixer | -0.045 |
liar | -0.044 |
mean | -0.044 |
wasteful | -0.044 |
Token | ily |
Token ID | 813 |
Feature activation | -5.08e-02 |
Pos logit contributions | |
Along | +0.069 |
nights | +0.068 |
drifted | +0.066 |
horizon | +0.065 |
lifetime | +0.064 |
Neg logit contributions | |
wasteful | -0.038 |
mean | -0.038 |
chefs | -0.038 |
liar | -0.038 |
mad | -0.037 |
Token | asked |
Token ID | 1965 |
Feature activation | +1.58e-02 |
Pos logit contributions | |
Along | +0.059 |
nights | +0.054 |
drifted | +0.054 |
way | +0.051 |
On | +0.049 |
Neg logit contributions | |
bluff | -0.104 |
liar | -0.103 |
wasteful | -0.100 |
mixer | -0.100 |
paramed | -0.099 |
Token | the |
Token ID | 262 |
Feature activation | +9.65e-03 |
Pos logit contributions | |
liar | +0.031 |
paramed | +0.028 |
bluff | +0.028 |
wasteful | +0.028 |
disrespectful | +0.028 |
Neg logit contributions | |
whenever | -0.022 |
nights | -0.020 |
Along | -0.019 |
days | -0.019 |
night | -0.018 |
Token | paramed |
Token ID | 34581 |
Feature activation | +2.91e-01 |
Pos logit contributions | |
roc | +0.010 |
boys | +0.010 |
liar | +0.009 |
wasteful | +0.009 |
bluff | +0.009 |
Neg logit contributions | |
Along | -0.020 |
nights | -0.019 |
way | -0.018 |
days | -0.018 |
day | -0.018 |
Token | ic |
Token ID | 291 |
Feature activation | +4.29e-02 |
Pos logit contributions | |
liar | +0.516 |
mean | +0.512 |
mixer | +0.507 |
boys | +0.502 |
paramed | +0.502 |
Neg logit contributions | |
nights | -0.378 |
way | -0.372 |
Along | -0.361 |
ingle | -0.339 |
more | -0.332 |
Token | , |
Token ID | 11 |
Feature activation | +7.24e-02 |
Pos logit contributions | |
boys | +0.078 |
bluff | +0.078 |
wasteful | +0.076 |
roc | +0.076 |
chefs | +0.075 |
Neg logit contributions | |
nights | -0.053 |
way | -0.052 |
Along | -0.051 |
days | -0.048 |
day | -0.047 |
Token | " |
Token ID | 366 |
Feature activation | +2.01e-01 |
Pos logit contributions | |
roc | +0.101 |
boys | +0.098 |
liar | +0.097 |
paramed | +0.090 |
mixer | +0.089 |
Neg logit contributions | |
Along | -0.136 |
nights | -0.124 |
On | -0.123 |
whenever | -0.123 |
As | -0.118 |
Token | Is |
Token ID | 3792 |
Feature activation | +2.39e-02 |
Pos logit contributions | |
liar | +0.307 |
roc | +0.295 |
wasteful | +0.287 |
boys | +0.286 |
mean | +0.286 |
Neg logit contributions | |
Along | -0.320 |
nights | -0.314 |
whenever | -0.294 |
more | -0.291 |
way | -0.290 |
Token | the |
Token ID | 262 |
Feature activation | +9.58e-03 |
Pos logit contributions | |
paramed | +0.041 |
liar | +0.040 |
roc | +0.039 |
mixer | +0.038 |
bluff | +0.038 |
Neg logit contributions | |
nights | -0.034 |
Along | -0.033 |
way | -0.031 |
whenever | -0.030 |
days | -0.030 |
Token | to |
Token ID | 284 |
Feature activation | +7.56e-02 |
Pos logit contributions | |
Along | +0.127 |
nights | +0.126 |
way | +0.118 |
drifted | +0.114 |
On | +0.113 |
Neg logit contributions | |
paramed | -0.125 |
mixer | -0.122 |
mean | -0.122 |
boys | -0.120 |
liar | -0.120 |
Token | be |
Token ID | 307 |
Feature activation | +3.56e-02 |
Pos logit contributions | |
paramed | +0.108 |
mixer | +0.103 |
mean | +0.103 |
liar | +0.103 |
boys | +0.102 |
Neg logit contributions | |
Along | -0.132 |
nights | -0.130 |
way | -0.122 |
drifted | -0.117 |
On | -0.117 |
Token | okay |
Token ID | 8788 |
Feature activation | +1.07e-01 |
Pos logit contributions | |
mixer | +0.048 |
paramed | +0.048 |
roc | +0.048 |
liar | +0.047 |
boys | +0.046 |
Neg logit contributions | |
nights | -0.063 |
Along | -0.060 |
whenever | -0.056 |
drifted | -0.056 |
days | -0.056 |
Token | ?" |
Token ID | 1701 |
Feature activation | +6.88e-02 |
Pos logit contributions | |
liar | +0.166 |
paramed | +0.162 |
wasteful | +0.161 |
roc | +0.160 |
bluff | +0.156 |
Neg logit contributions | |
Along | -0.166 |
nights | -0.162 |
whenever | -0.160 |
On | -0.155 |
way | -0.151 |
Token | The |
Token ID | 383 |
Feature activation | -1.17e-02 |
Pos logit contributions | |
liar | +0.128 |
paramed | +0.125 |
wasteful | +0.124 |
mean | +0.123 |
boys | +0.123 |
Neg logit contributions | |
Along | -0.093 |
nights | -0.086 |
On | -0.082 |
way | -0.081 |
whenever | -0.080 |
Token | paramed |
Token ID | 34581 |
Feature activation | +2.90e-01 |
Pos logit contributions | |
Along | +0.021 |
nights | +0.021 |
day | +0.020 |
days | +0.020 |
way | +0.020 |
Neg logit contributions | |
boys | -0.014 |
bluff | -0.013 |
chefs | -0.013 |
roc | -0.013 |
liar | -0.013 |
Token | ic |
Token ID | 291 |
Feature activation | +4.20e-02 |
Pos logit contributions | |
liar | +0.537 |
boys | +0.530 |
mean | +0.529 |
mixer | +0.520 |
paramed | +0.517 |
Neg logit contributions | |
nights | -0.352 |
way | -0.352 |
Along | -0.335 |
ingle | -0.321 |
more | -0.306 |
Token | smiled |
Token ID | 13541 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
bluff | +0.064 |
boys | +0.064 |
wasteful | +0.061 |
chefs | +0.061 |
liar | +0.060 |
Neg logit contributions | |
way | -0.066 |
nights | -0.065 |
Along | -0.062 |
days | -0.060 |
drifted | -0.060 |
Token | and |
Token ID | 290 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
paramed | +0.022 |
mixer | +0.021 |
liar | +0.021 |
mean | +0.021 |
boys | +0.021 |
Neg logit contributions | |
Along | -0.017 |
nights | -0.017 |
way | -0.015 |
On | -0.015 |
drifted | -0.015 |
Token | said |
Token ID | 531 |
Feature activation | +5.14e-02 |
Pos logit contributions | |
nights | +0.043 |
drifted | +0.042 |
Along | +0.042 |
way | +0.039 |
On | +0.039 |
Neg logit contributions | |
liar | -0.040 |
roc | -0.039 |
mixer | -0.038 |
paramed | -0.037 |
boys | -0.037 |
Token | , |
Token ID | 11 |
Feature activation | +7.19e-02 |
Pos logit contributions | |
roc | +0.115 |
boys | +0.111 |
app | +0.110 |
wasteful | +0.110 |
ram | +0.108 |
Neg logit contributions | |
whenever | -0.050 |
days | -0.041 |
nights | -0.040 |
way | -0.040 |
Along | -0.039 |
Token | Billy |
Token ID | 15890 |
Feature activation | +3.56e-02 |
Pos logit contributions | |
paramed | +0.013 |
mixer | +0.013 |
liar | +0.013 |
mean | +0.013 |
boys | +0.013 |
Neg logit contributions | |
Along | -0.019 |
nights | -0.019 |
way | -0.018 |
On | -0.017 |
drifted | -0.017 |
Token | and |
Token ID | 290 |
Feature activation | -9.36e-03 |
Pos logit contributions | |
paramed | +0.065 |
liar | +0.064 |
mixer | +0.064 |
wasteful | +0.063 |
mean | +0.062 |
Neg logit contributions | |
Along | -0.047 |
nights | -0.045 |
way | -0.042 |
On | -0.041 |
whenever | -0.040 |
Token | called |
Token ID | 1444 |
Feature activation | -5.95e-02 |
Pos logit contributions | |
drifted | +0.013 |
nights | +0.013 |
Along | +0.013 |
whenever | +0.013 |
way | +0.013 |
Neg logit contributions | |
roc | -0.015 |
mixer | -0.014 |
paramed | -0.014 |
chefs | -0.014 |
bluff | -0.014 |
Token | him |
Token ID | 683 |
Feature activation | +5.07e-02 |
Pos logit contributions | |
Along | +0.068 |
whenever | +0.062 |
way | +0.061 |
On | +0.060 |
nights | +0.059 |
Neg logit contributions | |
roc | -0.118 |
boys | -0.116 |
wasteful | -0.116 |
paramed | -0.115 |
chefs | -0.114 |
Token | a |
Token ID | 257 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
paramed | +0.110 |
mixer | +0.108 |
wasteful | +0.107 |
roc | +0.106 |
liar | +0.106 |
Neg logit contributions | |
Along | -0.051 |
whenever | -0.044 |
way | -0.044 |
nights | -0.043 |
On | -0.042 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.71e-01 |
Pos logit contributions | |
Along | +0.029 |
way | +0.028 |
day | +0.028 |
journey | +0.027 |
night | +0.027 |
Neg logit contributions | |
boys | -0.031 |
roc | -0.028 |
chefs | -0.028 |
wasteful | -0.027 |
bluff | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
paramed | +0.614 |
roc | +0.608 |
wasteful | +0.607 |
mixer | +0.605 |
bluff | +0.597 |
Neg logit contributions | |
Along | -0.223 |
nights | -0.211 |
way | -0.203 |
more | -0.184 |
whenever | -0.183 |
Token | He |
Token ID | 679 |
Feature activation | +1.01e-02 |
Pos logit contributions | |
Along | +0.021 |
On | +0.020 |
As | +0.018 |
way | +0.017 |
Everywhere | +0.017 |
Neg logit contributions | |
liar | -0.032 |
boys | -0.031 |
chefs | -0.031 |
wasteful | -0.031 |
paramed | -0.030 |
Token | said |
Token ID | 531 |
Feature activation | +6.77e-02 |
Pos logit contributions | |
bluff | +0.016 |
mixer | +0.016 |
paramed | +0.016 |
roc | +0.016 |
chefs | +0.016 |
Neg logit contributions | |
Along | -0.015 |
drifted | -0.014 |
way | -0.014 |
nights | -0.013 |
On | -0.013 |
Token | that |
Token ID | 326 |
Feature activation | -2.88e-02 |
Pos logit contributions | |
roc | +0.159 |
app | +0.158 |
boys | +0.157 |
bluff | +0.156 |
chefs | +0.151 |
Neg logit contributions | |
of | -0.048 |
whenever | -0.047 |
he | -0.041 |
way | -0.040 |
days | -0.039 |
Token | he |
Token ID | 339 |
Feature activation | -8.22e-02 |
Pos logit contributions | |
nights | +0.026 |
whenever | +0.024 |
Along | +0.023 |
day | +0.023 |
way | +0.023 |
Neg logit contributions | |
paramed | -0.064 |
roc | -0.062 |
boys | -0.061 |
liar | -0.061 |
mixer | -0.061 |
Token | was |
Token ID | 373 |
Feature activation | +4.18e-02 |
Pos logit contributions | |
paramed | +0.227 |
mixer | +0.220 |
liar | +0.214 |
bluff | +0.214 |
roc | +0.213 |
Neg logit contributions | |
Along | -0.143 |
nights | -0.140 |
way | -0.133 |
drifted | -0.130 |
whenever | -0.129 |
Token | a |
Token ID | 257 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
paramed | +0.066 |
mixer | +0.065 |
roc | +0.064 |
boys | +0.057 |
argue | +0.057 |
Neg logit contributions | |
before | -0.063 |
Along | -0.059 |
nights | -0.059 |
night | -0.058 |
day | -0.058 |
Token | thief |
Token ID | 25906 |
Feature activation | +1.22e-01 |
Pos logit contributions | |
day | +0.023 |
night | +0.023 |
Along | +0.022 |
nights | +0.022 |
days | +0.021 |
Neg logit contributions | |
boys | -0.022 |
roc | -0.022 |
ese | -0.020 |
oa | -0.019 |
wasteful | -0.019 |
Token | and |
Token ID | 290 |
Feature activation | -7.49e-03 |
Pos logit contributions | |
paramed | +0.244 |
mixer | +0.237 |
wasteful | +0.235 |
roc | +0.234 |
liar | +0.233 |
Neg logit contributions | |
Along | -0.137 |
nights | -0.137 |
way | -0.124 |
whenever | -0.120 |
days | -0.115 |
Token | a |
Token ID | 257 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
Along | +0.009 |
nights | +0.009 |
way | +0.008 |
drifted | +0.008 |
On | +0.008 |
Neg logit contributions | |
paramed | -0.014 |
mixer | -0.014 |
boys | -0.013 |
liar | -0.013 |
roc | -0.013 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.67e-01 |
Pos logit contributions | |
day | +0.027 |
nights | +0.027 |
Along | +0.027 |
night | +0.026 |
days | +0.025 |
Neg logit contributions | |
boys | -0.020 |
roc | -0.020 |
bluff | -0.017 |
ese | -0.017 |
oa | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
paramed | +0.553 |
mixer | +0.544 |
wasteful | +0.535 |
liar | +0.534 |
roc | +0.532 |
Neg logit contributions | |
Along | -0.281 |
nights | -0.280 |
way | -0.251 |
drifted | -0.237 |
whenever | -0.233 |
Token | He |
Token ID | 679 |
Feature activation | +1.54e-02 |
Pos logit contributions | |
Along | +0.010 |
On | +0.009 |
nights | +0.008 |
way | +0.008 |
whenever | +0.007 |
Neg logit contributions | |
paramed | -0.023 |
mixer | -0.023 |
boys | -0.023 |
liar | -0.022 |
wasteful | -0.022 |
Token | said |
Token ID | 531 |
Feature activation | +6.82e-02 |
Pos logit contributions | |
mixer | +0.024 |
bluff | +0.023 |
paramed | +0.023 |
roc | +0.023 |
liar | +0.022 |
Neg logit contributions | |
Along | -0.023 |
nights | -0.023 |
drifted | -0.022 |
way | -0.022 |
On | -0.021 |
Token | Tom |
Token ID | 4186 |
Feature activation | +1.17e-01 |
Pos logit contributions | |
app | +0.159 |
roc | +0.159 |
bluff | +0.159 |
boys | +0.155 |
intervene | +0.155 |
Neg logit contributions | |
of | -0.042 |
days | -0.042 |
whenever | -0.042 |
nights | -0.041 |
he | -0.035 |
Token | had |
Token ID | 550 |
Feature activation | -5.27e-02 |
Pos logit contributions | |
paramed | +0.222 |
mixer | +0.216 |
bluff | +0.210 |
liar | +0.209 |
wasteful | +0.208 |
Neg logit contributions | |
Along | -0.147 |
nights | -0.144 |
way | -0.135 |
whenever | -0.131 |
drifted | -0.131 |
Token | be |
Token ID | 307 |
Feature activation | +5.10e-02 |
Pos logit contributions | |
paramed | +0.144 |
mixer | +0.141 |
liar | +0.138 |
wasteful | +0.138 |
bluff | +0.136 |
Neg logit contributions | |
Along | -0.115 |
nights | -0.112 |
way | -0.110 |
whenever | -0.108 |
night | -0.104 |
Token | a |
Token ID | 257 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
mixer | +0.077 |
paramed | +0.076 |
tripod | +0.068 |
printer | +0.066 |
liar | +0.063 |
Neg logit contributions | |
nights | -0.080 |
whenever | -0.079 |
day | -0.076 |
as | -0.075 |
days | -0.075 |
Token | bully |
Token ID | 27410 |
Feature activation | +2.10e-01 |
Pos logit contributions | |
day | +0.022 |
night | +0.022 |
nights | +0.020 |
way | +0.020 |
Along | +0.020 |
Neg logit contributions | |
roc | -0.024 |
paramed | -0.023 |
boys | -0.023 |
wasteful | -0.023 |
mixer | -0.022 |
Token | and |
Token ID | 290 |
Feature activation | -5.83e-03 |
Pos logit contributions | |
paramed | +0.472 |
mixer | +0.452 |
roc | +0.446 |
bluff | +0.444 |
wasteful | +0.443 |
Neg logit contributions | |
nights | -0.187 |
whenever | -0.183 |
Along | -0.171 |
way | -0.159 |
night | -0.159 |
Token | a |
Token ID | 257 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
Along | +0.009 |
nights | +0.009 |
whenever | +0.009 |
night | +0.009 |
drifted | +0.008 |
Neg logit contributions | |
paramed | -0.009 |
mixer | -0.008 |
roc | -0.008 |
boys | -0.008 |
bluff | -0.008 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.67e-01 |
Pos logit contributions | |
day | +0.024 |
night | +0.023 |
nights | +0.023 |
way | +0.022 |
Along | +0.022 |
Neg logit contributions | |
roc | -0.023 |
boys | -0.022 |
paramed | -0.022 |
wasteful | -0.021 |
bluff | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
paramed | +0.553 |
mixer | +0.543 |
wasteful | +0.531 |
roc | +0.530 |
boys | +0.527 |
Neg logit contributions | |
nights | -0.279 |
Along | -0.275 |
way | -0.250 |
whenever | -0.243 |
more | -0.235 |
Token | I |
Token ID | 314 |
Feature activation | +3.39e-03 |
Pos logit contributions | |
Along | +0.019 |
On | +0.017 |
nights | +0.016 |
dark | +0.016 |
whenever | +0.016 |
Neg logit contributions | |
paramed | -0.024 |
mixer | -0.023 |
boys | -0.023 |
wasteful | -0.022 |
liar | -0.022 |
Token | want |
Token ID | 765 |
Feature activation | +4.17e-02 |
Pos logit contributions | |
mixer | +0.005 |
paramed | +0.005 |
liar | +0.005 |
boys | +0.005 |
wasteful | +0.005 |
Neg logit contributions | |
Along | -0.006 |
nights | -0.005 |
way | -0.005 |
On | -0.005 |
drifted | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | +8.47e-02 |
Pos logit contributions | |
paramed | +0.085 |
liar | +0.079 |
roc | +0.078 |
bluff | +0.078 |
mixer | +0.076 |
Neg logit contributions | |
nights | -0.047 |
Along | -0.044 |
whenever | -0.043 |
night | -0.043 |
days | -0.042 |
Token | be |
Token ID | 307 |
Feature activation | +4.84e-02 |
Pos logit contributions | |
paramed | +0.144 |
mixer | +0.142 |
wasteful | +0.140 |
bluff | +0.137 |
liar | +0.137 |
Neg logit contributions | |
Along | -0.116 |
nights | -0.109 |
whenever | -0.107 |
way | -0.107 |
On | -0.104 |
Token | nice |
Token ID | 3621 |
Feature activation | +7.07e-02 |
Pos logit contributions | |
paramed | +0.115 |
roc | +0.114 |
boys | +0.112 |
mixer | +0.111 |
liar | +0.107 |
Neg logit contributions | |
nights | -0.101 |
Along | -0.099 |
day | -0.095 |
night | -0.095 |
way | -0.093 |
Token | . |
Token ID | 13 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
paramed | +0.144 |
liar | +0.139 |
mixer | +0.139 |
wasteful | +0.137 |
boys | +0.137 |
Neg logit contributions | |
Along | -0.080 |
nights | -0.078 |
way | -0.071 |
whenever | -0.068 |
On | -0.067 |
Token | He |
Token ID | 679 |
Feature activation | +9.48e-03 |
Pos logit contributions | |
Along | +0.014 |
On | +0.013 |
way | +0.011 |
nights | +0.010 |
As | +0.010 |
Neg logit contributions | |
boys | -0.038 |
paramed | -0.037 |
mixer | -0.037 |
liar | -0.036 |
phone | -0.036 |
Token | was |
Token ID | 373 |
Feature activation | +3.54e-02 |
Pos logit contributions | |
paramed | +0.015 |
mixer | +0.015 |
liar | +0.014 |
roc | +0.014 |
chefs | +0.014 |
Neg logit contributions | |
Along | -0.015 |
nights | -0.014 |
drifted | -0.014 |
way | -0.013 |
On | -0.013 |
Token | a |
Token ID | 257 |
Feature activation | -1.76e-02 |
Pos logit contributions | |
roc | +0.054 |
boys | +0.051 |
paramed | +0.049 |
mixer | +0.048 |
liar | +0.046 |
Neg logit contributions | |
Along | -0.057 |
nights | -0.054 |
day | -0.053 |
before | -0.053 |
night | -0.052 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.65e-01 |
Pos logit contributions | |
day | +0.028 |
nights | +0.027 |
night | +0.027 |
Along | +0.027 |
way | +0.025 |
Neg logit contributions | |
boys | -0.028 |
roc | -0.026 |
wasteful | -0.024 |
paramed | -0.024 |
bluff | -0.024 |
Token | and |
Token ID | 290 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
paramed | +0.489 |
mixer | +0.477 |
liar | +0.474 |
wasteful | +0.468 |
boys | +0.467 |
Neg logit contributions | |
Along | -0.344 |
nights | -0.339 |
way | -0.309 |
drifted | -0.296 |
On | -0.294 |
Token | a |
Token ID | 257 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
Along | +0.016 |
nights | +0.016 |
way | +0.015 |
drifted | +0.014 |
On | +0.014 |
Neg logit contributions | |
paramed | -0.017 |
mixer | -0.017 |
liar | -0.016 |
boys | -0.016 |
wasteful | -0.016 |
Token | thief |
Token ID | 25906 |
Feature activation | +1.17e-01 |
Pos logit contributions | |
nights | +0.029 |
Along | +0.028 |
day | +0.028 |
night | +0.027 |
days | +0.027 |
Neg logit contributions | |
boys | -0.027 |
roc | -0.026 |
wasteful | -0.024 |
bluff | -0.024 |
paramed | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | -1.67e-02 |
Pos logit contributions | |
paramed | +0.202 |
liar | +0.199 |
mixer | +0.198 |
mean | +0.196 |
boys | +0.196 |
Neg logit contributions | |
Along | -0.168 |
nights | -0.164 |
way | -0.151 |
On | -0.146 |
drifted | -0.144 |
Token | He |
Token ID | 679 |
Feature activation | +9.17e-03 |
Pos logit contributions | |
Along | +0.017 |
nights | +0.016 |
way | +0.014 |
On | +0.014 |
drifted | +0.013 |
Neg logit contributions | |
paramed | -0.036 |
mixer | -0.035 |
boys | -0.035 |
liar | -0.035 |
mean | -0.035 |
Token | He |
Token ID | 679 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
nights | +0.033 |
Along | +0.031 |
way | +0.029 |
drifted | +0.028 |
dark | +0.027 |
Neg logit contributions | |
paramed | -0.051 |
mean | -0.050 |
liar | -0.049 |
mixer | -0.049 |
greedy | -0.049 |
Token | said |
Token ID | 531 |
Feature activation | +6.09e-02 |
Pos logit contributions | |
paramed | +0.015 |
mixer | +0.015 |
liar | +0.014 |
bluff | +0.014 |
roc | +0.014 |
Neg logit contributions | |
Along | -0.016 |
nights | -0.016 |
drifted | -0.015 |
way | -0.015 |
On | -0.015 |
Token | Jen |
Token ID | 13364 |
Feature activation | -7.94e-03 |
Pos logit contributions | |
boys | +0.129 |
bluff | +0.125 |
roc | +0.123 |
paramed | +0.121 |
app | +0.120 |
Neg logit contributions | |
of | -0.056 |
whenever | -0.055 |
days | -0.051 |
way | -0.051 |
nights | -0.051 |
Token | was |
Token ID | 373 |
Feature activation | +3.27e-02 |
Pos logit contributions | |
Along | +0.011 |
nights | +0.011 |
way | +0.011 |
drifted | +0.011 |
whenever | +0.011 |
Neg logit contributions | |
paramed | -0.013 |
mixer | -0.013 |
chefs | -0.012 |
roc | -0.012 |
bluff | -0.012 |
Token | a |
Token ID | 257 |
Feature activation | -1.78e-02 |
Pos logit contributions | |
paramed | +0.039 |
mixer | +0.035 |
roc | +0.035 |
boys | +0.033 |
liar | +0.029 |
Neg logit contributions | |
before | -0.061 |
Along | -0.060 |
nights | -0.060 |
day | -0.060 |
night | -0.058 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
day | +0.032 |
night | +0.030 |
nights | +0.030 |
way | +0.030 |
days | +0.029 |
Neg logit contributions | |
boys | -0.024 |
roc | -0.022 |
paramed | -0.021 |
wasteful | -0.020 |
ese | -0.020 |
Token | . |
Token ID | 13 |
Feature activation | -2.65e-02 |
Pos logit contributions | |
paramed | +0.559 |
mixer | +0.548 |
roc | +0.543 |
wasteful | +0.537 |
liar | +0.536 |
Neg logit contributions | |
nights | -0.264 |
Along | -0.262 |
way | -0.237 |
whenever | -0.229 |
more | -0.224 |
Token | He |
Token ID | 679 |
Feature activation | +9.89e-03 |
Pos logit contributions | |
nights | +0.032 |
Along | +0.031 |
way | +0.028 |
drifted | +0.028 |
dark | +0.026 |
Neg logit contributions | |
paramed | -0.052 |
liar | -0.050 |
mean | -0.050 |
mixer | -0.050 |
greedy | -0.050 |
Token | said |
Token ID | 531 |
Feature activation | +6.08e-02 |
Pos logit contributions | |
paramed | +0.015 |
mixer | +0.014 |
bluff | +0.014 |
roc | +0.014 |
liar | +0.014 |
Neg logit contributions | |
Along | -0.016 |
nights | -0.015 |
way | -0.015 |
drifted | -0.015 |
On | -0.014 |
Token | Jen |
Token ID | 13364 |
Feature activation | -8.23e-03 |
Pos logit contributions | |
boys | +0.135 |
bluff | +0.134 |
roc | +0.131 |
app | +0.129 |
paramed | +0.129 |
Neg logit contributions | |
whenever | -0.046 |
of | -0.046 |
nights | -0.043 |
days | -0.043 |
Along | -0.041 |
Token | was |
Token ID | 373 |
Feature activation | +3.23e-02 |
Pos logit contributions | |
Along | +0.012 |
nights | +0.012 |
way | +0.011 |
drifted | +0.011 |
whenever | +0.011 |
Neg logit contributions | |
paramed | -0.014 |
mixer | -0.013 |
bluff | -0.013 |
roc | -0.013 |
liar | -0.013 |
Token | looked |
Token ID | 3114 |
Feature activation | -1.04e-02 |
Pos logit contributions | |
drifted | +0.123 |
Along | +0.121 |
nights | +0.116 |
way | +0.111 |
On | +0.106 |
Neg logit contributions | |
roc | -0.155 |
bluff | -0.150 |
liar | -0.150 |
wasteful | -0.146 |
paramed | -0.142 |
Token | at |
Token ID | 379 |
Feature activation | +9.48e-03 |
Pos logit contributions | |
Along | +0.012 |
nights | +0.011 |
On | +0.011 |
whenever | +0.011 |
drifted | +0.010 |
Neg logit contributions | |
paramed | -0.021 |
liar | -0.021 |
roc | -0.021 |
boys | -0.021 |
mixer | -0.020 |
Token | Lily |
Token ID | 20037 |
Feature activation | +3.79e-02 |
Pos logit contributions | |
mixer | +0.011 |
paramed | +0.011 |
mean | +0.011 |
wasteful | +0.011 |
mad | +0.010 |
Neg logit contributions | |
Along | -0.019 |
nights | -0.019 |
way | -0.018 |
drifted | -0.017 |
more | -0.017 |
Token | with |
Token ID | 351 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
paramed | +0.065 |
liar | +0.064 |
mixer | +0.063 |
wasteful | +0.062 |
mean | +0.062 |
Neg logit contributions | |
Along | -0.055 |
nights | -0.053 |
way | -0.050 |
On | -0.048 |
drifted | -0.048 |
Token | a |
Token ID | 257 |
Feature activation | -1.79e-02 |
Pos logit contributions | |
drifted | +0.030 |
Along | +0.030 |
nights | +0.029 |
On | +0.029 |
night | +0.028 |
Neg logit contributions | |
roc | -0.040 |
boys | -0.040 |
paramed | -0.040 |
chefs | -0.037 |
mixer | -0.037 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
drifted | +0.020 |
moment | +0.019 |
particular | +0.019 |
drift | +0.018 |
day | +0.018 |
Neg logit contributions | |
boys | -0.038 |
roc | -0.036 |
chefs | -0.035 |
wasteful | -0.034 |
oa | -0.033 |
Token | face |
Token ID | 1986 |
Feature activation | +6.52e-02 |
Pos logit contributions | |
boys | +0.406 |
wasteful | +0.403 |
paramed | +0.395 |
mixer | +0.393 |
liar | +0.386 |
Neg logit contributions | |
Along | -0.406 |
nights | -0.397 |
night | -0.380 |
drifted | -0.378 |
On | -0.371 |
Token | . |
Token ID | 13 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
paramed | +0.136 |
liar | +0.134 |
roc | +0.132 |
wasteful | +0.131 |
mixer | +0.131 |
Neg logit contributions | |
Along | -0.069 |
nights | -0.067 |
way | -0.062 |
whenever | -0.059 |
On | -0.059 |
Token | Lily |
Token ID | 20037 |
Feature activation | +4.00e-02 |
Pos logit contributions | |
Along | +0.021 |
nights | +0.020 |
On | +0.019 |
way | +0.019 |
drifted | +0.018 |
Neg logit contributions | |
paramed | -0.033 |
liar | -0.032 |
boys | -0.032 |
mixer | -0.032 |
mean | -0.032 |
Token | felt |
Token ID | 2936 |
Feature activation | -3.84e-04 |
Pos logit contributions | |
liar | +0.083 |
roc | +0.082 |
paramed | +0.081 |
bluff | +0.080 |
wasteful | +0.079 |
Neg logit contributions | |
drifted | -0.046 |
Along | -0.039 |
whenever | -0.037 |
nights | -0.036 |
way | -0.036 |
Token | sorry |
Token ID | 7926 |
Feature activation | +2.00e-01 |
Pos logit contributions | |
whenever | +0.001 |
nights | +0.001 |
drifted | +0.001 |
night | +0.001 |
On | +0.001 |
Neg logit contributions | |
boys | -0.000 |
roc | -0.000 |
mixer | -0.000 |
paramed | -0.000 |
printer | -0.000 |
Token | But |
Token ID | 887 |
Feature activation | -3.03e-02 |
Pos logit contributions | |
nights | +0.025 |
Along | +0.023 |
way | +0.022 |
ingle | +0.022 |
outer | +0.022 |
Neg logit contributions | |
paramed | -0.029 |
mixer | -0.028 |
stern | -0.027 |
greedy | -0.027 |
Lee | -0.027 |
Token | her |
Token ID | 607 |
Feature activation | -5.91e-02 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.042 |
way | +0.039 |
On | +0.038 |
drifted | +0.038 |
Neg logit contributions | |
paramed | -0.052 |
mixer | -0.051 |
mean | -0.050 |
liar | -0.050 |
boys | -0.050 |
Token | mother |
Token ID | 2802 |
Feature activation | -9.76e-03 |
Pos logit contributions | |
Along | +0.111 |
nights | +0.109 |
way | +0.103 |
On | +0.100 |
drifted | +0.100 |
Neg logit contributions | |
paramed | -0.076 |
mixer | -0.073 |
liar | -0.073 |
mean | -0.071 |
boys | -0.071 |
Token | told |
Token ID | 1297 |
Feature activation | +5.78e-02 |
Pos logit contributions | |
nights | +0.015 |
Along | +0.014 |
way | +0.014 |
drifted | +0.013 |
whenever | +0.013 |
Neg logit contributions | |
liar | -0.016 |
chefs | -0.015 |
bluff | -0.015 |
paramed | -0.015 |
roc | -0.015 |
Token | her |
Token ID | 607 |
Feature activation | -5.87e-02 |
Pos logit contributions | |
paramed | +0.074 |
mean | +0.072 |
mixer | +0.071 |
mad | +0.069 |
wasteful | +0.069 |
Neg logit contributions | |
Along | -0.111 |
nights | -0.106 |
way | -0.101 |
On | -0.098 |
drifted | -0.097 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
Along | +0.094 |
nights | +0.092 |
way | +0.085 |
On | +0.083 |
drifted | +0.083 |
Neg logit contributions | |
paramed | -0.091 |
mixer | -0.089 |
liar | -0.089 |
mean | -0.088 |
wasteful | -0.087 |
Token | ly |
Token ID | 306 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
paramed | +0.741 |
liar | +0.739 |
mixer | +0.733 |
boys | +0.732 |
chefs | +0.730 |
Neg logit contributions | |
nights | -0.077 |
Along | -0.075 |
way | -0.059 |
whenever | -0.043 |
night | -0.041 |
Token | , |
Token ID | 11 |
Feature activation | +8.54e-02 |
Pos logit contributions | |
nights | +0.162 |
Along | +0.159 |
way | +0.147 |
whenever | +0.147 |
drifted | +0.143 |
Neg logit contributions | |
paramed | -0.208 |
liar | -0.204 |
bluff | -0.202 |
roc | -0.201 |
boys | -0.201 |
Token | " |
Token ID | 366 |
Feature activation | +2.11e-01 |
Pos logit contributions | |
roc | +0.130 |
liar | +0.128 |
boys | +0.125 |
paramed | +0.122 |
chefs | +0.121 |
Neg logit contributions | |
Along | -0.141 |
nights | -0.139 |
way | -0.130 |
whenever | -0.129 |
On | -0.128 |
Token | No |
Token ID | 2949 |
Feature activation | +8.32e-02 |
Pos logit contributions | |
paramed | +0.293 |
mixer | +0.281 |
mean | +0.278 |
wasteful | +0.276 |
greedy | +0.273 |
Neg logit contributions | |
Along | -0.375 |
nights | -0.367 |
way | -0.337 |
On | -0.334 |
drifted | -0.333 |
Token | more |
Token ID | 517 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
paramed | +0.144 |
boys | +0.140 |
liar | +0.139 |
chefs | +0.136 |
bluff | +0.136 |
Neg logit contributions | |
nights | -0.121 |
Along | -0.114 |
way | -0.112 |
days | -0.109 |
night | -0.106 |
Token | ?" |
Token ID | 1701 |
Feature activation | +7.96e-02 |
Pos logit contributions | |
paramed | +0.141 |
liar | +0.133 |
boys | +0.125 |
stern | +0.123 |
bluff | +0.123 |
Neg logit contributions | |
nights | -0.105 |
whenever | -0.105 |
night | -0.102 |
before | -0.100 |
day | -0.099 |
Token | Mom |
Token ID | 11254 |
Feature activation | +6.93e-02 |
Pos logit contributions | |
paramed | +0.170 |
liar | +0.166 |
mixer | +0.166 |
boys | +0.165 |
wasteful | +0.164 |
Neg logit contributions | |
Along | -0.081 |
nights | -0.077 |
way | -0.069 |
On | -0.068 |
drifted | -0.066 |
Token | said |
Token ID | 531 |
Feature activation | +6.29e-02 |
Pos logit contributions | |
liar | +0.121 |
paramed | +0.114 |
roc | +0.110 |
chefs | +0.109 |
wasteful | +0.109 |
Neg logit contributions | |
Along | -0.095 |
On | -0.093 |
drifted | -0.090 |
way | -0.089 |
night | -0.089 |
Token | in |
Token ID | 287 |
Feature activation | -5.84e-02 |
Pos logit contributions | |
roc | +0.147 |
paramed | +0.147 |
boys | +0.144 |
liar | +0.144 |
wasteful | +0.142 |
Neg logit contributions | |
Along | -0.048 |
nights | -0.045 |
whenever | -0.044 |
way | -0.044 |
more | -0.040 |
Token | a |
Token ID | 257 |
Feature activation | -1.62e-02 |
Pos logit contributions | |
Along | +0.092 |
nights | +0.088 |
On | +0.083 |
drifted | +0.081 |
whenever | +0.080 |
Neg logit contributions | |
paramed | -0.092 |
boys | -0.091 |
mixer | -0.090 |
liar | -0.089 |
wasteful | -0.089 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
day | +0.025 |
Along | +0.025 |
nights | +0.024 |
particular | +0.024 |
night | +0.024 |
Neg logit contributions | |
boys | -0.028 |
roc | -0.024 |
chefs | -0.024 |
oa | -0.024 |
ese | -0.023 |
Token | but |
Token ID | 475 |
Feature activation | +2.08e-02 |
Pos logit contributions | |
boys | +0.425 |
wasteful | +0.424 |
mixer | +0.420 |
liar | +0.419 |
paramed | +0.418 |
Neg logit contributions | |
Along | -0.380 |
nights | -0.379 |
night | -0.360 |
way | -0.356 |
ly | -0.348 |
Token | gentle |
Token ID | 10296 |
Feature activation | +6.81e-02 |
Pos logit contributions | |
roc | +0.055 |
liar | +0.054 |
paramed | +0.053 |
boys | +0.052 |
chefs | +0.051 |
Neg logit contributions | |
Along | -0.010 |
nights | -0.010 |
night | -0.010 |
way | -0.009 |
drifted | -0.009 |
Token | voice |
Token ID | 3809 |
Feature activation | +4.43e-02 |
Pos logit contributions | |
boys | +0.101 |
wasteful | +0.100 |
liar | +0.098 |
paramed | +0.098 |
mixer | +0.098 |
Neg logit contributions | |
nights | -0.111 |
Along | -0.108 |
way | -0.103 |
drifted | -0.101 |
night | -0.100 |
Token | . |
Token ID | 13 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
roc | +0.098 |
chefs | +0.096 |
liar | +0.096 |
boys | +0.096 |
paramed | +0.092 |
Neg logit contributions | |
way | -0.039 |
ly | -0.034 |
nights | -0.034 |
dark | -0.033 |
On | -0.033 |
Token |
|
Token ID | 198 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
Along | +0.024 |
nights | +0.023 |
way | +0.020 |
On | +0.020 |
drifted | +0.020 |
Neg logit contributions | |
paramed | -0.041 |
liar | -0.040 |
mixer | -0.040 |
boys | -0.040 |
mean | -0.040 |
Token | holes |
Token ID | 10421 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
Along | +0.048 |
nights | +0.047 |
way | +0.043 |
drifted | +0.042 |
On | +0.041 |
Neg logit contributions | |
paramed | -0.057 |
mixer | -0.055 |
liar | -0.055 |
boys | -0.054 |
mean | -0.054 |
Token | !" |
Token ID | 2474 |
Feature activation | +1.13e-01 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.040 |
On | +0.038 |
way | +0.037 |
drifted | +0.037 |
Neg logit contributions | |
mean | -0.042 |
mixer | -0.042 |
wasteful | -0.041 |
greedy | -0.040 |
liar | -0.040 |
Token | Mom |
Token ID | 11254 |
Feature activation | +7.03e-02 |
Pos logit contributions | |
paramed | +0.220 |
liar | +0.217 |
wasteful | +0.215 |
mixer | +0.213 |
roc | +0.212 |
Neg logit contributions | |
Along | -0.133 |
nights | -0.124 |
On | -0.121 |
whenever | -0.116 |
drifted | -0.115 |
Token | said |
Token ID | 531 |
Feature activation | +6.49e-02 |
Pos logit contributions | |
liar | +0.128 |
paramed | +0.116 |
roc | +0.116 |
wasteful | +0.114 |
bluff | +0.114 |
Neg logit contributions | |
Along | -0.093 |
way | -0.090 |
On | -0.090 |
of | -0.089 |
nights | -0.088 |
Token | , |
Token ID | 11 |
Feature activation | +8.94e-02 |
Pos logit contributions | |
roc | +0.144 |
paramed | +0.143 |
liar | +0.143 |
wasteful | +0.141 |
boys | +0.140 |
Neg logit contributions | |
Along | -0.057 |
nights | -0.055 |
way | -0.053 |
whenever | -0.051 |
drifted | -0.047 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.63e-01 |
Pos logit contributions | |
roc | +0.146 |
liar | +0.126 |
wasteful | +0.126 |
boys | +0.123 |
mixer | +0.118 |
Neg logit contributions | |
Along | -0.147 |
drifted | -0.145 |
nights | -0.136 |
whenever | -0.135 |
before | -0.133 |
Token | ly |
Token ID | 306 |
Feature activation | -1.17e-01 |
Pos logit contributions | |
liar | +0.727 |
paramed | +0.726 |
mixer | +0.719 |
wasteful | +0.714 |
boys | +0.711 |
Neg logit contributions | |
Along | -0.094 |
nights | -0.087 |
way | -0.066 |
On | -0.051 |
whenever | -0.049 |
Token | . |
Token ID | 13 |
Feature activation | -1.75e-02 |
Pos logit contributions | |
Along | +0.091 |
nights | +0.090 |
way | +0.087 |
drifted | +0.085 |
before | +0.084 |
Neg logit contributions | |
roc | -0.275 |
liar | -0.264 |
wasteful | -0.262 |
bluff | -0.257 |
paramed | -0.254 |
Token |
|
Token ID | 198 |
Feature activation | -1.22e-02 |
Pos logit contributions | |
Along | +0.017 |
nights | +0.016 |
On | +0.014 |
way | +0.014 |
drifted | +0.014 |
Neg logit contributions | |
paramed | -0.038 |
liar | -0.038 |
mixer | -0.038 |
boys | -0.037 |
wasteful | -0.037 |
Token |
|
Token ID | 198 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
Along | +0.014 |
nights | +0.014 |
way | +0.013 |
drifted | +0.012 |
On | +0.011 |
Neg logit contributions | |
paramed | -0.024 |
mixer | -0.024 |
liar | -0.023 |
mean | -0.023 |
boys | -0.023 |
Token | Max |
Token ID | 11518 |
Feature activation | +3.81e-02 |
Pos logit contributions | |
Along | +0.017 |
nights | +0.016 |
way | +0.015 |
On | +0.015 |
more | +0.015 |
Neg logit contributions | |
paramed | -0.017 |
liar | -0.016 |
mixer | -0.016 |
wasteful | -0.016 |
greedy | -0.016 |
Token | walked |
Token ID | 6807 |
Feature activation | -1.53e-01 |
Pos logit contributions | |
mixer | +0.028 |
bluff | +0.028 |
paramed | +0.028 |
liar | +0.027 |
roc | +0.027 |
Neg logit contributions | |
Along | -0.021 |
drifted | -0.021 |
nights | -0.020 |
way | -0.019 |
On | -0.019 |
Token | towards |
Token ID | 3371 |
Feature activation | -6.40e-02 |
Pos logit contributions | |
Along | +0.195 |
nights | +0.165 |
way | +0.164 |
On | +0.159 |
whenever | +0.157 |
Neg logit contributions | |
liar | -0.317 |
mixer | -0.309 |
paramed | -0.305 |
wasteful | -0.297 |
boys | -0.296 |
Token | them |
Token ID | 606 |
Feature activation | +9.54e-02 |
Pos logit contributions | |
Along | +0.127 |
nights | +0.127 |
way | +0.118 |
whenever | +0.117 |
On | +0.116 |
Neg logit contributions | |
paramed | -0.074 |
liar | -0.072 |
wasteful | -0.070 |
mixer | -0.069 |
boys | -0.069 |
Token | with |
Token ID | 351 |
Feature activation | -2.04e-02 |
Pos logit contributions | |
paramed | +0.167 |
mixer | +0.164 |
boys | +0.163 |
mean | +0.163 |
liar | +0.162 |
Neg logit contributions | |
Along | -0.131 |
nights | -0.131 |
way | -0.117 |
drifted | -0.115 |
dark | -0.113 |
Token | a |
Token ID | 257 |
Feature activation | -1.59e-02 |
Pos logit contributions | |
nights | +0.031 |
Along | +0.030 |
whenever | +0.028 |
days | +0.028 |
drifted | +0.028 |
Neg logit contributions | |
paramed | -0.033 |
roc | -0.032 |
wasteful | -0.031 |
liar | -0.031 |
boys | -0.031 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
Along | +0.021 |
day | +0.020 |
drifted | +0.019 |
journey | +0.019 |
On | +0.019 |
Neg logit contributions | |
boys | -0.029 |
roc | -0.028 |
wasteful | -0.026 |
chefs | -0.026 |
oa | -0.026 |
Token | face |
Token ID | 1986 |
Feature activation | +6.15e-02 |
Pos logit contributions | |
boys | +0.409 |
wasteful | +0.404 |
mixer | +0.398 |
paramed | +0.394 |
chefs | +0.388 |
Neg logit contributions | |
Along | -0.401 |
nights | -0.395 |
night | -0.373 |
way | -0.371 |
drifted | -0.365 |
Token | . |
Token ID | 13 |
Feature activation | -1.59e-02 |
Pos logit contributions | |
paramed | +0.128 |
liar | +0.126 |
mixer | +0.125 |
wasteful | +0.124 |
boys | +0.123 |
Neg logit contributions | |
Along | -0.065 |
nights | -0.063 |
way | -0.058 |
On | -0.054 |
drifted | -0.054 |
Token |
|
Token ID | 198 |
Feature activation | -5.94e-03 |
Pos logit contributions | |
Along | +0.017 |
nights | +0.016 |
way | +0.014 |
On | +0.014 |
drifted | +0.013 |
Neg logit contributions | |
paramed | -0.034 |
mixer | -0.033 |
liar | -0.033 |
boys | -0.033 |
mean | -0.032 |
Token |
|
Token ID | 198 |
Feature activation | -5.53e-03 |
Pos logit contributions | |
nights | +0.008 |
Along | +0.008 |
drifted | +0.007 |
way | +0.007 |
days | +0.007 |
Neg logit contributions | |
paramed | -0.011 |
mean | -0.010 |
mixer | -0.010 |
liar | -0.010 |
stern | -0.010 |
Token | " |
Token ID | 1 |
Feature activation | +1.75e-01 |
Pos logit contributions | |
Along | +0.009 |
nights | +0.009 |
way | +0.008 |
On | +0.008 |
more | +0.008 |
Neg logit contributions | |
paramed | -0.008 |
mixer | -0.008 |
liar | -0.008 |
boys | -0.008 |
wasteful | -0.008 |
Token | Dad |
Token ID | 17415 |
Feature activation | +9.28e-02 |
Pos logit contributions | |
boys | +0.124 |
paramed | +0.122 |
liar | +0.120 |
roc | +0.119 |
mixer | +0.116 |
Neg logit contributions | |
Along | -0.128 |
nights | -0.123 |
On | -0.120 |
drifted | -0.116 |
night | -0.115 |
Token | asked |
Token ID | 1965 |
Feature activation | +2.21e-02 |
Pos logit contributions | |
liar | +0.190 |
roc | +0.181 |
wasteful | +0.175 |
paramed | +0.169 |
chefs | +0.168 |
Neg logit contributions | |
drifted | -0.109 |
Along | -0.109 |
On | -0.107 |
nights | -0.100 |
way | -0.100 |
Token | . |
Token ID | 13 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
paramed | +0.050 |
liar | +0.049 |
mixer | +0.049 |
wasteful | +0.049 |
boys | +0.048 |
Neg logit contributions | |
Along | -0.019 |
nights | -0.018 |
way | -0.016 |
whenever | -0.016 |
On | -0.015 |
Token | He |
Token ID | 679 |
Feature activation | +1.06e-02 |
Pos logit contributions | |
Along | +0.022 |
On | +0.019 |
nights | +0.019 |
way | +0.018 |
drifted | +0.017 |
Neg logit contributions | |
boys | -0.051 |
paramed | -0.050 |
roc | -0.049 |
liar | -0.049 |
mixer | -0.049 |
Token | was |
Token ID | 373 |
Feature activation | +3.46e-02 |
Pos logit contributions | |
roc | +0.019 |
liar | +0.019 |
wasteful | +0.018 |
boys | +0.018 |
paramed | +0.018 |
Neg logit contributions | |
drifted | -0.014 |
Along | -0.014 |
On | -0.013 |
nights | -0.013 |
way | -0.012 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
roc | +0.057 |
boys | +0.051 |
liar | +0.048 |
Cubs | +0.048 |
mixer | +0.048 |
Neg logit contributions | |
of | -0.049 |
before | -0.047 |
as | -0.047 |
drifted | -0.046 |
Along | -0.046 |
Token | and |
Token ID | 290 |
Feature activation | -1.58e-02 |
Pos logit contributions | |
paramed | +0.501 |
mixer | +0.488 |
mean | +0.487 |
liar | +0.485 |
greedy | +0.480 |
Neg logit contributions | |
Along | -0.326 |
nights | -0.320 |
way | -0.287 |
drifted | -0.277 |
On | -0.274 |
Token | sad |
Token ID | 6507 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
drifted | +0.029 |
nights | +0.027 |
Along | +0.027 |
day | +0.026 |
days | +0.026 |
Neg logit contributions | |
roc | -0.020 |
boys | -0.020 |
mixer | -0.018 |
chefs | -0.018 |
paramed | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
mean | +0.185 |
paramed | +0.181 |
mixer | +0.180 |
mad | +0.179 |
phone | +0.178 |
Neg logit contributions | |
Along | -0.147 |
nights | -0.139 |
way | -0.124 |
dark | -0.120 |
drifted | -0.120 |
Token |
|
Token ID | 198 |
Feature activation | -1.21e-02 |
Pos logit contributions | |
Along | +0.022 |
nights | +0.020 |
On | +0.019 |
way | +0.018 |
drifted | +0.017 |
Neg logit contributions | |
paramed | -0.049 |
boys | -0.048 |
mixer | -0.048 |
liar | -0.048 |
wasteful | -0.047 |
Token |
|
Token ID | 198 |
Feature activation | -1.20e-02 |
Pos logit contributions | |
nights | +0.017 |
Along | +0.017 |
drifted | +0.015 |
way | +0.015 |
filled | +0.014 |
Neg logit contributions | |
paramed | -0.021 |
mean | -0.020 |
mixer | -0.020 |
stern | -0.020 |
mad | -0.020 |
Token | would |
Token ID | 561 |
Feature activation | +2.05e-02 |
Pos logit contributions | |
whenever | +0.034 |
drifted | +0.033 |
nights | +0.032 |
Along | +0.031 |
way | +0.030 |
Neg logit contributions | |
roc | -0.046 |
liar | -0.046 |
paramed | -0.046 |
bluff | -0.044 |
mixer | -0.043 |
Token | call |
Token ID | 869 |
Feature activation | +8.43e-04 |
Pos logit contributions | |
paramed | +0.031 |
roc | +0.030 |
chefs | +0.029 |
mixer | +0.029 |
liar | +0.029 |
Neg logit contributions | |
Along | -0.034 |
nights | -0.032 |
whenever | -0.032 |
drifted | -0.031 |
way | -0.031 |
Token | him |
Token ID | 683 |
Feature activation | +4.03e-02 |
Pos logit contributions | |
paramed | +0.002 |
greedy | +0.001 |
liar | +0.001 |
wasteful | +0.001 |
roc | +0.001 |
Neg logit contributions | |
nights | -0.001 |
Along | -0.001 |
whenever | -0.001 |
night | -0.001 |
days | -0.001 |
Token | a |
Token ID | 257 |
Feature activation | -2.57e-02 |
Pos logit contributions | |
paramed | +0.074 |
mixer | +0.071 |
liar | +0.071 |
wasteful | +0.070 |
greedy | +0.068 |
Neg logit contributions | |
Along | -0.056 |
nights | -0.052 |
whenever | -0.050 |
On | -0.049 |
way | -0.048 |
Token | stupid |
Token ID | 8531 |
Feature activation | +1.25e-01 |
Pos logit contributions | |
day | +0.037 |
Along | +0.036 |
night | +0.036 |
way | +0.035 |
nights | +0.035 |
Neg logit contributions | |
boys | -0.044 |
roc | -0.041 |
wasteful | -0.040 |
chefs | -0.039 |
eper | -0.039 |
Token | liar |
Token ID | 31866 |
Feature activation | +2.61e-01 |
Pos logit contributions | |
boys | +0.192 |
roc | +0.182 |
wasteful | +0.181 |
paramed | +0.180 |
bluff | +0.174 |
Neg logit contributions | |
day | -0.194 |
night | -0.193 |
nights | -0.189 |
Along | -0.186 |
way | -0.184 |
Token | . |
Token ID | 13 |
Feature activation | -2.77e-02 |
Pos logit contributions | |
paramed | +0.569 |
mixer | +0.557 |
roc | +0.552 |
wasteful | +0.552 |
liar | +0.551 |
Neg logit contributions | |
Along | -0.246 |
nights | -0.243 |
way | -0.218 |
whenever | -0.207 |
drifted | -0.202 |
Token |
|
Token ID | 198 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
nights | +0.044 |
Along | +0.043 |
way | +0.040 |
drifted | +0.039 |
more | +0.038 |
Neg logit contributions | |
paramed | -0.042 |
mixer | -0.042 |
boys | -0.041 |
mean | -0.041 |
liar | -0.041 |
Token |
|
Token ID | 198 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
nights | +0.036 |
Along | +0.034 |
drifted | +0.032 |
way | +0.031 |
more | +0.030 |
Neg logit contributions | |
mixer | -0.048 |
mean | -0.048 |
paramed | -0.048 |
phone | -0.047 |
liar | -0.047 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
Along | +0.033 |
nights | +0.032 |
way | +0.029 |
On | +0.028 |
drifted | +0.027 |
Neg logit contributions | |
paramed | -0.052 |
liar | -0.051 |
mixer | -0.051 |
wasteful | -0.050 |
boys | -0.050 |
Token | day |
Token ID | 1110 |
Feature activation | -2.17e-01 |
Pos logit contributions | |
Along | +0.129 |
nights | +0.115 |
drifted | +0.106 |
growth | +0.103 |
horizon | +0.102 |
Neg logit contributions | |
mean | -0.401 |
mad | -0.371 |
greedy | -0.371 |
mixer | -0.365 |
liar | -0.363 |
Token | ?" |
Token ID | 1701 |
Feature activation | +8.66e-02 |
Pos logit contributions | |
roc | +0.065 |
liar | +0.065 |
wasteful | +0.065 |
paramed | +0.065 |
greedy | +0.063 |
Neg logit contributions | |
day | -0.043 |
night | -0.041 |
whenever | -0.041 |
nights | -0.040 |
way | -0.040 |
Token | she |
Token ID | 673 |
Feature activation | -1.45e-01 |
Pos logit contributions | |
paramed | +0.180 |
liar | +0.179 |
boys | +0.178 |
mixer | +0.176 |
roc | +0.175 |
Neg logit contributions | |
Along | -0.090 |
nights | -0.086 |
On | -0.079 |
way | -0.076 |
drifted | -0.076 |
Token | asked |
Token ID | 1965 |
Feature activation | +2.27e-02 |
Pos logit contributions | |
drifted | +0.173 |
nights | +0.172 |
Along | +0.166 |
On | +0.155 |
night | +0.148 |
Neg logit contributions | |
liar | -0.274 |
paramed | -0.272 |
roc | -0.269 |
mixer | -0.266 |
wasteful | -0.264 |
Token | in |
Token ID | 287 |
Feature activation | -6.16e-02 |
Pos logit contributions | |
paramed | +0.051 |
liar | +0.050 |
wasteful | +0.050 |
roc | +0.049 |
mixer | +0.048 |
Neg logit contributions | |
Along | -0.021 |
nights | -0.020 |
whenever | -0.019 |
way | -0.018 |
On | -0.017 |
Token | a |
Token ID | 257 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
Along | +0.108 |
nights | +0.106 |
way | +0.099 |
drifted | +0.096 |
On | +0.096 |
Neg logit contributions | |
paramed | -0.087 |
mixer | -0.085 |
liar | -0.084 |
mean | -0.083 |
boys | -0.083 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.61e-01 |
Pos logit contributions | |
Along | +0.023 |
nights | +0.022 |
particular | +0.022 |
drifted | +0.022 |
day | +0.022 |
Neg logit contributions | |
boys | -0.037 |
oa | -0.033 |
roc | -0.032 |
chefs | -0.032 |
ese | -0.031 |
Token | voice |
Token ID | 3809 |
Feature activation | +4.20e-02 |
Pos logit contributions | |
wasteful | +0.406 |
mixer | +0.401 |
liar | +0.399 |
paramed | +0.396 |
boys | +0.392 |
Neg logit contributions | |
Along | -0.402 |
nights | -0.393 |
way | -0.371 |
On | -0.366 |
night | -0.363 |
Token | . |
Token ID | 13 |
Feature activation | -1.88e-02 |
Pos logit contributions | |
liar | +0.092 |
paramed | +0.092 |
mixer | +0.090 |
wasteful | +0.090 |
roc | +0.090 |
Neg logit contributions | |
Along | -0.038 |
nights | -0.036 |
way | -0.034 |
On | -0.032 |
whenever | -0.031 |
Token | " |
Token ID | 366 |
Feature activation | +2.11e-01 |
Pos logit contributions | |
Along | +0.015 |
On | +0.014 |
way | +0.011 |
As | +0.011 |
drifted | +0.011 |
Neg logit contributions | |
wasteful | -0.045 |
roc | -0.045 |
boys | -0.045 |
liar | -0.045 |
paramed | -0.044 |
Token | Why |
Token ID | 5195 |
Feature activation | +2.44e-02 |
Pos logit contributions | |
paramed | +0.303 |
mixer | +0.295 |
liar | +0.294 |
mean | +0.288 |
boys | +0.288 |
Neg logit contributions | |
Along | -0.363 |
nights | -0.356 |
way | -0.334 |
On | -0.324 |
drifted | -0.323 |
Token | are |
Token ID | 389 |
Feature activation | +1.39e-01 |
Pos logit contributions | |
paramed | +0.040 |
roc | +0.035 |
mixer | +0.035 |
wasteful | +0.035 |
liar | +0.034 |
Neg logit contributions | |
whenever | -0.035 |
night | -0.035 |
Along | -0.034 |
On | -0.034 |
day | -0.034 |
Token | , |
Token ID | 11 |
Feature activation | +9.04e-02 |
Pos logit contributions | |
wasteful | +0.125 |
roc | +0.124 |
mixer | +0.121 |
chefs | +0.121 |
paramed | +0.121 |
Neg logit contributions | |
Along | -0.092 |
nights | -0.089 |
whenever | -0.087 |
way | -0.085 |
drifted | -0.082 |
Token | " |
Token ID | 366 |
Feature activation | +2.08e-01 |
Pos logit contributions | |
liar | +0.100 |
paramed | +0.099 |
wasteful | +0.098 |
roc | +0.097 |
mixer | +0.096 |
Neg logit contributions | |
Along | -0.187 |
nights | -0.178 |
way | -0.171 |
On | -0.170 |
drifted | -0.169 |
Token | Follow |
Token ID | 7155 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
paramed | +0.243 |
mean | +0.221 |
greedy | +0.220 |
Fighting | +0.218 |
boys | +0.217 |
Neg logit contributions | |
nights | -0.387 |
Along | -0.382 |
way | -0.365 |
On | -0.359 |
lights | -0.358 |
Token | your |
Token ID | 534 |
Feature activation | +7.59e-02 |
Pos logit contributions | |
Along | +0.217 |
nights | +0.214 |
way | +0.201 |
On | +0.195 |
drifted | +0.194 |
Neg logit contributions | |
paramed | -0.160 |
liar | -0.155 |
mixer | -0.154 |
wasteful | -0.150 |
mean | -0.150 |
Token | dreams |
Token ID | 10625 |
Feature activation | -1.63e-01 |
Pos logit contributions | |
mean | +0.119 |
phone | +0.119 |
liar | +0.117 |
boys | +0.116 |
mixer | +0.116 |
Neg logit contributions | |
Along | -0.121 |
nights | -0.113 |
way | -0.113 |
On | -0.109 |
dark | -0.109 |
Token | ' |
Token ID | 6 |
Feature activation | +2.61e-01 |
Pos logit contributions | |
Along | +0.268 |
nights | +0.263 |
drifted | +0.243 |
way | +0.242 |
dark | +0.240 |
Neg logit contributions | |
mean | -0.234 |
phone | -0.228 |
boys | -0.227 |
mixer | -0.227 |
liar | -0.220 |
Token |
|
Token ID | 198 |
Feature activation | -5.77e-03 |
Pos logit contributions | |
mean | +0.421 |
boys | +0.403 |
greedy | +0.394 |
phone | +0.389 |
mad | +0.385 |
Neg logit contributions | |
Along | -0.403 |
nights | -0.374 |
drifted | -0.363 |
ingle | -0.360 |
way | -0.359 |
Token |
|
Token ID | 198 |
Feature activation | -4.20e-03 |
Pos logit contributions | |
nights | +0.009 |
Along | +0.009 |
way | +0.008 |
drifted | +0.008 |
ingle | +0.008 |
Neg logit contributions | |
paramed | -0.009 |
liar | -0.009 |
mean | -0.008 |
mixer | -0.008 |
greedy | -0.008 |
Token | Lu |
Token ID | 25596 |
Feature activation | -3.63e-02 |
Pos logit contributions | |
nights | +0.006 |
Along | +0.006 |
way | +0.006 |
drifted | +0.006 |
On | +0.006 |
Neg logit contributions | |
paramed | -0.007 |
liar | -0.007 |
mixer | -0.007 |
boys | -0.007 |
mean | -0.007 |
Token | cy |
Token ID | 948 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
nights | +0.051 |
Along | +0.050 |
way | +0.046 |
On | +0.044 |
drifted | +0.043 |
Neg logit contributions | |
paramed | -0.063 |
mixer | -0.062 |
boys | -0.061 |
greedy | -0.059 |
roc | -0.059 |
Token | smiled |
Token ID | 13541 |
Feature activation | +2.85e-02 |
Pos logit contributions | |
Along | +0.038 |
nights | +0.037 |
way | +0.036 |
dark | +0.035 |
On | +0.035 |
Neg logit contributions | |
boys | -0.027 |
mean | -0.026 |
phone | -0.025 |
stern | -0.025 |
paramed | -0.025 |
Token | couch |
Token ID | 18507 |
Feature activation | +4.29e-02 |
Pos logit contributions | |
paramed | +0.033 |
liar | +0.032 |
mixer | +0.031 |
mean | +0.031 |
boys | +0.031 |
Neg logit contributions | |
Along | -0.033 |
nights | -0.033 |
way | -0.030 |
drifted | -0.029 |
On | -0.029 |
Token | and |
Token ID | 290 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
paramed | +0.084 |
liar | +0.083 |
wasteful | +0.082 |
roc | +0.081 |
mixer | +0.081 |
Neg logit contributions | |
Along | -0.052 |
nights | -0.050 |
way | -0.047 |
whenever | -0.044 |
On | -0.044 |
Token | gave |
Token ID | 2921 |
Feature activation | +3.37e-02 |
Pos logit contributions | |
nights | +0.024 |
Along | +0.024 |
drifted | +0.023 |
whenever | +0.022 |
way | +0.022 |
Neg logit contributions | |
paramed | -0.024 |
liar | -0.024 |
roc | -0.023 |
mixer | -0.023 |
wasteful | -0.023 |
Token | them |
Token ID | 606 |
Feature activation | +8.90e-02 |
Pos logit contributions | |
paramed | +0.056 |
liar | +0.055 |
wasteful | +0.052 |
bluff | +0.051 |
dishonest | +0.050 |
Neg logit contributions | |
nights | -0.052 |
whenever | -0.051 |
night | -0.050 |
Along | -0.048 |
days | -0.047 |
Token | a |
Token ID | 257 |
Feature activation | -1.88e-02 |
Pos logit contributions | |
liar | +0.138 |
paramed | +0.136 |
wasteful | +0.131 |
mixer | +0.128 |
bluff | +0.127 |
Neg logit contributions | |
Along | -0.144 |
nights | -0.139 |
whenever | -0.131 |
night | -0.130 |
way | -0.128 |
Token | stern |
Token ID | 26370 |
Feature activation | +2.60e-01 |
Pos logit contributions | |
Along | +0.030 |
nights | +0.027 |
way | +0.027 |
On | +0.027 |
dark | +0.027 |
Neg logit contributions | |
boys | -0.030 |
roc | -0.029 |
paramed | -0.028 |
wasteful | -0.028 |
liar | -0.027 |
Token | look |
Token ID | 804 |
Feature activation | +3.81e-02 |
Pos logit contributions | |
boys | +0.392 |
paramed | +0.382 |
wasteful | +0.382 |
mixer | +0.375 |
liar | +0.371 |
Neg logit contributions | |
Along | -0.411 |
nights | -0.405 |
way | -0.386 |
night | -0.385 |
drifted | -0.376 |
Token | . |
Token ID | 13 |
Feature activation | -2.21e-02 |
Pos logit contributions | |
paramed | +0.072 |
mixer | +0.071 |
mean | +0.071 |
liar | +0.070 |
greedy | +0.069 |
Neg logit contributions | |
Along | -0.050 |
nights | -0.048 |
way | -0.043 |
On | -0.042 |
drifted | -0.041 |
Token |
|
Token ID | 198 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
Along | +0.027 |
nights | +0.024 |
On | +0.024 |
way | +0.023 |
whenever | +0.022 |
Neg logit contributions | |
paramed | -0.043 |
liar | -0.043 |
boys | -0.043 |
wasteful | -0.042 |
mixer | -0.042 |
Token |
|
Token ID | 198 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
nights | +0.018 |
Along | +0.017 |
drifted | +0.016 |
way | +0.016 |
days | +0.015 |
Neg logit contributions | |
paramed | -0.022 |
mixer | -0.022 |
mean | -0.022 |
stern | -0.021 |
mad | -0.021 |
Token | " |
Token ID | 1 |
Feature activation | +1.69e-01 |
Pos logit contributions | |
Along | +0.020 |
nights | +0.019 |
way | +0.018 |
On | +0.017 |
more | +0.017 |
Neg logit contributions | |
paramed | -0.019 |
mixer | -0.019 |
liar | -0.019 |
boys | -0.019 |
mean | -0.018 |
Token | , |
Token ID | 11 |
Feature activation | +7.54e-02 |
Pos logit contributions | |
Along | +0.067 |
nights | +0.065 |
way | +0.060 |
drifted | +0.057 |
On | +0.057 |
Neg logit contributions | |
paramed | -0.088 |
liar | -0.086 |
mixer | -0.086 |
wasteful | -0.084 |
mean | -0.084 |
Token | and |
Token ID | 290 |
Feature activation | -2.36e-02 |
Pos logit contributions | |
paramed | +0.117 |
boys | +0.117 |
mean | +0.116 |
mixer | +0.115 |
liar | +0.114 |
Neg logit contributions | |
Along | -0.118 |
nights | -0.112 |
way | -0.111 |
On | -0.105 |
ingle | -0.103 |
Token | even |
Token ID | 772 |
Feature activation | -4.54e-02 |
Pos logit contributions | |
Along | +0.041 |
way | +0.038 |
nights | +0.036 |
On | +0.036 |
ingle | +0.036 |
Neg logit contributions | |
mean | -0.034 |
paramed | -0.034 |
boys | -0.033 |
liar | -0.033 |
greedy | -0.033 |
Token | went |
Token ID | 1816 |
Feature activation | -1.25e-01 |
Pos logit contributions | |
Along | +0.084 |
way | +0.080 |
ingle | +0.076 |
nights | +0.074 |
On | +0.074 |
Neg logit contributions | |
mean | -0.064 |
chefs | -0.060 |
greedy | -0.059 |
boys | -0.059 |
mad | -0.058 |
Token | to |
Token ID | 284 |
Feature activation | +7.69e-02 |
Pos logit contributions | |
Along | +0.196 |
nights | +0.193 |
way | +0.182 |
On | +0.174 |
drifted | +0.172 |
Neg logit contributions | |
paramed | -0.195 |
mean | -0.190 |
mixer | -0.188 |
boys | -0.188 |
liar | -0.187 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.32e-01 |
Pos logit contributions | |
mean | +0.138 |
paramed | +0.138 |
liar | +0.137 |
boys | +0.137 |
mixer | +0.137 |
Neg logit contributions | |
Along | -0.100 |
nights | -0.097 |
way | -0.094 |
drifted | -0.086 |
On | -0.085 |
Token | space |
Token ID | 2272 |
Feature activation | -1.53e-01 |
Pos logit contributions | |
Along | +0.396 |
way | +0.361 |
nights | +0.350 |
drifted | +0.332 |
ingle | +0.323 |
Neg logit contributions | |
mean | -0.628 |
liar | -0.625 |
chefs | -0.620 |
boys | -0.618 |
paramed | -0.617 |
Token | ! |
Token ID | 0 |
Feature activation | -2.81e-02 |
Pos logit contributions | |
Along | +0.202 |
nights | +0.196 |
way | +0.180 |
On | +0.173 |
drifted | +0.172 |
Neg logit contributions | |
paramed | -0.279 |
mixer | -0.274 |
liar | -0.273 |
boys | -0.270 |
mean | -0.270 |
Token | It |
Token ID | 632 |
Feature activation | +6.89e-02 |
Pos logit contributions | |
nights | +0.050 |
ge | +0.046 |
drifted | +0.046 |
journey | +0.046 |
ingle | +0.045 |
Neg logit contributions | |
paramed | -0.039 |
mixer | -0.035 |
mean | -0.033 |
wasteful | -0.032 |
greedy | -0.032 |
Token | was |
Token ID | 373 |
Feature activation | +2.34e-02 |
Pos logit contributions | |
mean | +0.101 |
boys | +0.099 |
paramed | +0.097 |
mixer | +0.094 |
greedy | +0.094 |
Neg logit contributions | |
Along | -0.118 |
nights | -0.109 |
way | -0.106 |
On | -0.104 |
more | -0.100 |
Token | the |
Token ID | 262 |
Feature activation | +7.76e-03 |
Pos logit contributions | |
paramed | +0.033 |
liar | +0.032 |
mixer | +0.032 |
boys | +0.032 |
wasteful | +0.031 |
Neg logit contributions | |
Along | -0.040 |
nights | -0.040 |
way | -0.037 |
On | -0.036 |
drifted | -0.036 |
Token | to |
Token ID | 284 |
Feature activation | +7.84e-02 |
Pos logit contributions | |
Along | +0.059 |
nights | +0.058 |
way | +0.053 |
drifted | +0.053 |
On | +0.052 |
Neg logit contributions | |
paramed | -0.049 |
mean | -0.049 |
mixer | -0.048 |
liar | -0.048 |
phone | -0.047 |
Token | take |
Token ID | 1011 |
Feature activation | -8.78e-02 |
Pos logit contributions | |
paramed | +0.132 |
mean | +0.132 |
mixer | +0.132 |
boys | +0.130 |
phone | +0.128 |
Neg logit contributions | |
nights | -0.112 |
Along | -0.109 |
way | -0.101 |
drifted | -0.098 |
dark | -0.097 |
Token | a |
Token ID | 257 |
Feature activation | -2.73e-02 |
Pos logit contributions | |
Along | +0.169 |
nights | +0.166 |
way | +0.165 |
drifted | +0.155 |
more | +0.155 |
Neg logit contributions | |
paramed | -0.099 |
mean | -0.098 |
mixer | -0.097 |
phone | -0.095 |
mad | -0.092 |
Token | journey |
Token ID | 7002 |
Feature activation | -2.09e-01 |
Pos logit contributions | |
nights | +0.036 |
Along | +0.033 |
drifted | +0.031 |
more | +0.030 |
way | +0.029 |
Neg logit contributions | |
mixer | -0.053 |
mean | -0.053 |
phone | -0.052 |
paramed | -0.051 |
stern | -0.050 |
Token | to |
Token ID | 284 |
Feature activation | +7.83e-02 |
Pos logit contributions | |
Along | +0.320 |
nights | +0.318 |
drifted | +0.304 |
dark | +0.292 |
On | +0.286 |
Neg logit contributions | |
phone | -0.315 |
mean | -0.313 |
mixer | -0.311 |
boys | -0.310 |
paramed | -0.309 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.28e-01 |
Pos logit contributions | |
mean | +0.102 |
phone | +0.095 |
mad | +0.092 |
paramed | +0.092 |
boys | +0.090 |
Neg logit contributions | |
nights | -0.143 |
Along | -0.139 |
way | -0.137 |
drifted | -0.130 |
lights | -0.130 |
Token | space |
Token ID | 2272 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
Along | +0.371 |
nights | +0.350 |
way | +0.341 |
drifted | +0.329 |
dark | +0.312 |
Neg logit contributions | |
paramed | -0.636 |
mean | -0.630 |
mixer | -0.627 |
liar | -0.623 |
greedy | -0.616 |
Token | . |
Token ID | 13 |
Feature activation | -2.98e-02 |
Pos logit contributions | |
Along | +0.191 |
nights | +0.184 |
way | +0.170 |
drifted | +0.164 |
On | +0.162 |
Neg logit contributions | |
paramed | -0.277 |
mixer | -0.275 |
boys | -0.272 |
mean | -0.272 |
liar | -0.271 |
Token | He |
Token ID | 679 |
Feature activation | -3.11e-03 |
Pos logit contributions | |
nights | +0.046 |
way | +0.043 |
days | +0.041 |
dark | +0.041 |
more | +0.040 |
Neg logit contributions | |
paramed | -0.048 |
mixer | -0.045 |
Lee | -0.044 |
stern | -0.041 |
chefs | -0.040 |
Token | hopped |
Token ID | 45230 |
Feature activation | -1.39e-01 |
Pos logit contributions | |
Along | +0.005 |
way | +0.005 |
more | +0.005 |
nights | +0.005 |
ingle | +0.005 |
Neg logit contributions | |
mean | -0.004 |
boys | -0.004 |
phone | -0.004 |
mixer | -0.004 |
paramed | -0.004 |
Token | into |
Token ID | 656 |
Feature activation | -7.35e-02 |
Pos logit contributions | |
nights | +0.223 |
more | +0.210 |
way | +0.207 |
dark | +0.206 |
ingle | +0.205 |
Neg logit contributions | |
stern | -0.192 |
mean | -0.190 |
phone | -0.188 |
mixer | -0.187 |
paramed | -0.184 |
Token | izz |
Token ID | 6457 |
Feature activation | -7.19e-02 |
Pos logit contributions | |
paramed | +0.120 |
liar | +0.118 |
boys | +0.117 |
mixer | +0.115 |
mean | +0.115 |
Neg logit contributions | |
Along | -0.102 |
nights | -0.101 |
way | -0.095 |
drifted | -0.089 |
On | -0.088 |
Token | ed |
Token ID | 276 |
Feature activation | -3.72e-02 |
Pos logit contributions | |
Along | +0.097 |
On | +0.089 |
nights | +0.088 |
way | +0.083 |
drifted | +0.080 |
Neg logit contributions | |
boys | -0.129 |
paramed | -0.129 |
mixer | -0.125 |
phone | -0.124 |
mean | -0.121 |
Token | off |
Token ID | 572 |
Feature activation | -9.41e-02 |
Pos logit contributions | |
Along | +0.054 |
nights | +0.054 |
way | +0.050 |
drifted | +0.048 |
On | +0.047 |
Neg logit contributions | |
paramed | -0.062 |
mixer | -0.061 |
liar | -0.060 |
mean | -0.060 |
boys | -0.060 |
Token | back |
Token ID | 736 |
Feature activation | +5.30e-02 |
Pos logit contributions | |
Along | +0.142 |
nights | +0.142 |
way | +0.130 |
On | +0.125 |
drifted | +0.124 |
Neg logit contributions | |
paramed | -0.152 |
mixer | -0.148 |
liar | -0.146 |
mean | -0.146 |
boys | -0.145 |
Token | into |
Token ID | 656 |
Feature activation | -7.05e-02 |
Pos logit contributions | |
phone | +0.075 |
boys | +0.071 |
mad | +0.069 |
mixer | +0.069 |
Lee | +0.068 |
Neg logit contributions | |
nights | -0.087 |
way | -0.083 |
Along | -0.083 |
dark | -0.081 |
drifted | -0.081 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.26e-01 |
Pos logit contributions | |
Along | +0.132 |
nights | +0.130 |
way | +0.122 |
On | +0.119 |
drifted | +0.119 |
Neg logit contributions | |
paramed | -0.091 |
mixer | -0.088 |
liar | -0.087 |
mean | -0.086 |
boys | -0.085 |
Token | space |
Token ID | 2272 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
Along | +0.380 |
way | +0.351 |
nights | +0.345 |
drifted | +0.330 |
dark | +0.329 |
Neg logit contributions | |
liar | -0.612 |
paramed | -0.610 |
mean | -0.606 |
greedy | -0.604 |
mixer | -0.597 |
Token | . |
Token ID | 13 |
Feature activation | -2.04e-02 |
Pos logit contributions | |
Along | +0.180 |
nights | +0.173 |
way | +0.158 |
On | +0.153 |
drifted | +0.151 |
Neg logit contributions | |
paramed | -0.291 |
mixer | -0.287 |
liar | -0.285 |
boys | -0.283 |
mean | -0.283 |
Token | Earth |
Token ID | 3668 |
Feature activation | -1.44e-01 |
Pos logit contributions | |
nights | +0.038 |
ingle | +0.033 |
way | +0.033 |
drifted | +0.032 |
days | +0.032 |
Neg logit contributions | |
paramed | -0.028 |
Lee | -0.027 |
mixer | -0.026 |
sorry | -0.024 |
stern | -0.024 |
Token | was |
Token ID | 373 |
Feature activation | +2.89e-02 |
Pos logit contributions | |
Along | +0.221 |
nights | +0.214 |
way | +0.199 |
On | +0.195 |
drifted | +0.190 |
Neg logit contributions | |
paramed | -0.231 |
mixer | -0.226 |
boys | -0.224 |
liar | -0.224 |
mean | -0.223 |
Token | an |
Token ID | 281 |
Feature activation | -4.36e-03 |
Pos logit contributions | |
paramed | +0.042 |
liar | +0.041 |
mixer | +0.041 |
boys | +0.040 |
wasteful | +0.040 |
Neg logit contributions | |
Along | -0.049 |
nights | -0.049 |
way | -0.045 |
drifted | -0.044 |
whenever | -0.044 |
Token | looked |
Token ID | 3114 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
paramed | +0.080 |
wasteful | +0.079 |
liar | +0.079 |
dishonest | +0.077 |
chefs | +0.077 |
Neg logit contributions | |
Along | -0.027 |
way | -0.026 |
On | -0.025 |
nights | -0.024 |
drifted | -0.024 |
Token | like |
Token ID | 588 |
Feature activation | -2.57e-03 |
Pos logit contributions | |
nights | +0.024 |
Along | +0.024 |
whenever | +0.023 |
On | +0.023 |
way | +0.022 |
Neg logit contributions | |
paramed | -0.028 |
mixer | -0.026 |
liar | -0.026 |
boys | -0.026 |
Lee | -0.025 |
Token | it |
Token ID | 340 |
Feature activation | +3.44e-02 |
Pos logit contributions | |
Along | +0.004 |
nights | +0.004 |
way | +0.004 |
drifted | +0.004 |
On | +0.004 |
Neg logit contributions | |
paramed | -0.004 |
mixer | -0.004 |
liar | -0.004 |
mean | -0.003 |
boys | -0.003 |
Token | came |
Token ID | 1625 |
Feature activation | +4.73e-02 |
Pos logit contributions | |
mean | +0.053 |
boys | +0.049 |
mixer | +0.049 |
greedy | +0.048 |
phone | +0.048 |
Neg logit contributions | |
Along | -0.060 |
nights | -0.057 |
way | -0.053 |
On | -0.051 |
drifted | -0.051 |
Token | from |
Token ID | 422 |
Feature activation | -2.81e-02 |
Pos logit contributions | |
mean | +0.075 |
mad | +0.069 |
boys | +0.068 |
greedy | +0.066 |
phone | +0.065 |
Neg logit contributions | |
nights | -0.082 |
Along | -0.082 |
drifted | -0.076 |
days | -0.072 |
ingle | -0.072 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.26e-01 |
Pos logit contributions | |
Along | +0.051 |
nights | +0.049 |
way | +0.047 |
drifted | +0.045 |
ingle | +0.045 |
Neg logit contributions | |
mean | -0.038 |
wasteful | -0.037 |
mad | -0.037 |
mixer | -0.037 |
greedy | -0.037 |
Token | space |
Token ID | 2272 |
Feature activation | -1.49e-01 |
Pos logit contributions | |
Along | +0.399 |
nights | +0.370 |
way | +0.359 |
drifted | +0.337 |
dark | +0.332 |
Neg logit contributions | |
mean | -0.614 |
liar | -0.612 |
paramed | -0.610 |
greedy | -0.605 |
wasteful | -0.600 |
Token | . |
Token ID | 13 |
Feature activation | -2.36e-02 |
Pos logit contributions | |
Along | +0.170 |
nights | +0.166 |
way | +0.149 |
On | +0.141 |
drifted | +0.141 |
Neg logit contributions | |
paramed | -0.300 |
mixer | -0.293 |
liar | -0.292 |
mean | -0.288 |
wasteful | -0.288 |
Token |
|
Token ID | 198 |
Feature activation | -2.76e-02 |
Pos logit contributions | |
Along | +0.027 |
nights | +0.025 |
way | +0.023 |
On | +0.023 |
drifted | +0.022 |
Neg logit contributions | |
paramed | -0.048 |
liar | -0.048 |
mixer | -0.047 |
wasteful | -0.047 |
mean | -0.047 |
Token |
|
Token ID | 198 |
Feature activation | -2.67e-02 |
Pos logit contributions | |
nights | +0.044 |
Along | +0.041 |
drifted | +0.039 |
horizon | +0.037 |
more | +0.037 |
Neg logit contributions | |
mixer | -0.042 |
mean | -0.042 |
paramed | -0.041 |
stern | -0.040 |
mad | -0.040 |
Token | L |
Token ID | 43 |
Feature activation | -3.06e-02 |
Pos logit contributions | |
Along | +0.042 |
nights | +0.041 |
way | +0.037 |
drifted | +0.037 |
On | +0.036 |
Neg logit contributions | |
paramed | -0.043 |
mixer | -0.042 |
mean | -0.041 |
liar | -0.041 |
wasteful | -0.041 |
Token | special |
Token ID | 2041 |
Feature activation | -4.52e-02 |
Pos logit contributions | |
nights | +0.031 |
Along | +0.030 |
way | +0.028 |
night | +0.028 |
day | +0.027 |
Neg logit contributions | |
paramed | -0.038 |
boys | -0.037 |
roc | -0.037 |
wasteful | -0.037 |
liar | -0.036 |
Token | star |
Token ID | 3491 |
Feature activation | -2.59e-02 |
Pos logit contributions | |
Along | +0.072 |
nights | +0.068 |
On | +0.063 |
drifted | +0.062 |
way | +0.062 |
Neg logit contributions | |
mixer | -0.073 |
mean | -0.070 |
phone | -0.070 |
paramed | -0.069 |
liar | -0.069 |
Token | that |
Token ID | 326 |
Feature activation | -3.43e-02 |
Pos logit contributions | |
Along | +0.040 |
nights | +0.037 |
On | +0.035 |
way | +0.035 |
drifted | +0.034 |
Neg logit contributions | |
mixer | -0.043 |
mean | -0.042 |
phone | -0.041 |
boys | -0.041 |
liar | -0.041 |
Token | comes |
Token ID | 2058 |
Feature activation | +7.16e-02 |
Pos logit contributions | |
Along | +0.063 |
nights | +0.059 |
way | +0.057 |
On | +0.055 |
more | +0.054 |
Neg logit contributions | |
mean | -0.046 |
mixer | -0.045 |
greedy | -0.043 |
boys | -0.043 |
liar | -0.043 |
Token | from |
Token ID | 422 |
Feature activation | -3.03e-02 |
Pos logit contributions | |
mean | +0.111 |
mad | +0.106 |
boys | +0.104 |
greedy | +0.103 |
phone | +0.103 |
Neg logit contributions | |
Along | -0.118 |
nights | -0.117 |
drifted | -0.106 |
On | -0.103 |
days | -0.103 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.25e-01 |
Pos logit contributions | |
Along | +0.050 |
nights | +0.048 |
way | +0.046 |
drifted | +0.045 |
On | +0.044 |
Neg logit contributions | |
mean | -0.045 |
mixer | -0.044 |
paramed | -0.043 |
mad | -0.043 |
greedy | -0.043 |
Token | space |
Token ID | 2272 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
Along | +0.380 |
nights | +0.345 |
way | +0.334 |
drifted | +0.318 |
ingle | +0.314 |
Neg logit contributions | |
liar | -0.622 |
mixer | -0.621 |
paramed | -0.621 |
mean | -0.620 |
greedy | -0.609 |
Token | ," |
Token ID | 553 |
Feature activation | +1.27e-01 |
Pos logit contributions | |
Along | +0.205 |
nights | +0.197 |
way | +0.181 |
drifted | +0.177 |
On | +0.176 |
Neg logit contributions | |
paramed | -0.268 |
mixer | -0.266 |
liar | -0.264 |
mean | -0.263 |
boys | -0.261 |
Token | she |
Token ID | 673 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
paramed | +0.236 |
mixer | +0.231 |
mean | +0.225 |
boys | +0.223 |
phone | +0.223 |
Neg logit contributions | |
Along | -0.166 |
nights | -0.163 |
way | -0.151 |
dark | -0.138 |
drifted | -0.138 |
Token | explained |
Token ID | 4893 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
drifted | +0.210 |
Along | +0.207 |
way | +0.199 |
nights | +0.197 |
On | +0.197 |
Neg logit contributions | |
liar | -0.260 |
wasteful | -0.259 |
chefs | -0.258 |
mixer | -0.251 |
roc | -0.247 |
Token | . |
Token ID | 13 |
Feature activation | -2.64e-02 |
Pos logit contributions | |
liar | +0.251 |
wasteful | +0.248 |
paramed | +0.245 |
roc | +0.245 |
chefs | +0.243 |
Neg logit contributions | |
Along | -0.082 |
nights | -0.080 |
way | -0.078 |
whenever | -0.078 |
On | -0.069 |
Token | . |
Token ID | 13 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
nights | +0.146 |
Along | +0.146 |
way | +0.134 |
dark | +0.134 |
ingle | +0.131 |
Neg logit contributions | |
mixer | -0.129 |
boys | -0.127 |
paramed | -0.127 |
phone | -0.126 |
stern | -0.125 |
Token | Soon |
Token ID | 15894 |
Feature activation | -1.75e-01 |
Pos logit contributions | |
nights | +0.027 |
ingle | +0.025 |
drifted | +0.025 |
way | +0.024 |
more | +0.024 |
Neg logit contributions | |
paramed | -0.025 |
mixer | -0.024 |
boys | -0.024 |
stern | -0.023 |
Lee | -0.023 |
Token | they |
Token ID | 484 |
Feature activation | -1.08e-02 |
Pos logit contributions | |
Along | +0.305 |
nights | +0.284 |
On | +0.273 |
way | +0.273 |
dark | +0.272 |
Neg logit contributions | |
paramed | -0.245 |
boys | -0.240 |
mean | -0.239 |
mixer | -0.237 |
liar | -0.233 |
Token | were |
Token ID | 547 |
Feature activation | +4.13e-02 |
Pos logit contributions | |
Along | +0.021 |
nights | +0.020 |
way | +0.020 |
On | +0.019 |
dark | +0.019 |
Neg logit contributions | |
mean | -0.013 |
boys | -0.012 |
paramed | -0.012 |
mad | -0.011 |
liar | -0.011 |
Token | in |
Token ID | 287 |
Feature activation | -6.06e-02 |
Pos logit contributions | |
mean | +0.062 |
paramed | +0.062 |
mad | +0.061 |
boys | +0.061 |
mixer | +0.060 |
Neg logit contributions | |
Along | -0.066 |
nights | -0.064 |
way | -0.061 |
On | -0.060 |
dark | -0.059 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.24e-01 |
Pos logit contributions | |
Along | +0.108 |
nights | +0.106 |
way | +0.101 |
dark | +0.097 |
drifted | +0.096 |
Neg logit contributions | |
paramed | -0.082 |
mixer | -0.080 |
liar | -0.079 |
mean | -0.079 |
boys | -0.077 |
Token | space |
Token ID | 2272 |
Feature activation | -1.50e-01 |
Pos logit contributions | |
Along | +0.373 |
way | +0.351 |
nights | +0.332 |
ingle | +0.325 |
dark | +0.321 |
Neg logit contributions | |
mean | -0.603 |
paramed | -0.596 |
mixer | -0.596 |
liar | -0.593 |
chefs | -0.587 |
Token | and |
Token ID | 290 |
Feature activation | -1.59e-02 |
Pos logit contributions | |
Along | +0.227 |
nights | +0.214 |
way | +0.201 |
On | +0.200 |
dark | +0.197 |
Neg logit contributions | |
boys | -0.239 |
mixer | -0.239 |
paramed | -0.237 |
mean | -0.236 |
liar | -0.235 |
Token | having |
Token ID | 1719 |
Feature activation | -5.81e-02 |
Pos logit contributions | |
Along | +0.026 |
way | +0.023 |
On | +0.023 |
nights | +0.023 |
ingle | +0.022 |
Neg logit contributions | |
mean | -0.024 |
boys | -0.024 |
paramed | -0.023 |
mixer | -0.023 |
phone | -0.023 |
Token | fun |
Token ID | 1257 |
Feature activation | -2.71e-03 |
Pos logit contributions | |
Along | +0.092 |
nights | +0.088 |
way | +0.085 |
On | +0.082 |
dark | +0.081 |
Neg logit contributions | |
paramed | -0.088 |
mean | -0.087 |
boys | -0.087 |
mixer | -0.086 |
liar | -0.086 |
Token | . |
Token ID | 13 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
Along | +0.005 |
dark | +0.004 |
ingle | +0.004 |
On | +0.004 |
way | +0.004 |
Neg logit contributions | |
boys | -0.004 |
fighting | -0.004 |
chefs | -0.004 |
arguing | -0.004 |
phone | -0.004 |
Token | the |
Token ID | 262 |
Feature activation | +1.09e-02 |
Pos logit contributions | |
paramed | +0.010 |
mixer | +0.010 |
mean | +0.010 |
liar | +0.010 |
boys | +0.010 |
Neg logit contributions | |
Along | -0.013 |
nights | -0.013 |
way | -0.012 |
drifted | -0.012 |
On | -0.012 |
Token | stars |
Token ID | 5788 |
Feature activation | -1.71e-01 |
Pos logit contributions | |
mean | +0.020 |
rude | +0.017 |
phone | +0.017 |
greedy | +0.016 |
bluff | +0.016 |
Neg logit contributions | |
Along | -0.017 |
ingle | -0.016 |
drifted | -0.015 |
ment | -0.015 |
way | -0.015 |
Token | and |
Token ID | 290 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
Along | +0.255 |
nights | +0.244 |
way | +0.228 |
On | +0.221 |
drifted | +0.216 |
Neg logit contributions | |
mixer | -0.283 |
mean | -0.281 |
paramed | -0.278 |
boys | -0.277 |
liar | -0.273 |
Token | planets |
Token ID | 14705 |
Feature activation | -2.00e-01 |
Pos logit contributions | |
Along | +0.032 |
nights | +0.031 |
way | +0.030 |
On | +0.029 |
drifted | +0.029 |
Neg logit contributions | |
paramed | -0.028 |
mixer | -0.027 |
liar | -0.027 |
mean | -0.027 |
boys | -0.027 |
Token | in |
Token ID | 287 |
Feature activation | -5.92e-02 |
Pos logit contributions | |
Along | +0.263 |
nights | +0.253 |
way | +0.232 |
On | +0.224 |
drifted | +0.219 |
Neg logit contributions | |
mean | -0.363 |
mixer | -0.361 |
boys | -0.359 |
liar | -0.357 |
paramed | -0.357 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.24e-01 |
Pos logit contributions | |
Along | +0.118 |
nights | +0.114 |
way | +0.113 |
dark | +0.107 |
drifted | +0.106 |
Neg logit contributions | |
mean | -0.069 |
mixer | -0.067 |
mad | -0.066 |
paramed | -0.065 |
liar | -0.064 |
Token | space |
Token ID | 2272 |
Feature activation | -1.49e-01 |
Pos logit contributions | |
Along | +0.375 |
way | +0.358 |
nights | +0.344 |
ingle | +0.330 |
drifted | +0.321 |
Neg logit contributions | |
mean | -0.600 |
liar | -0.585 |
mixer | -0.583 |
greedy | -0.581 |
roc | -0.580 |
Token | . |
Token ID | 13 |
Feature activation | -2.41e-02 |
Pos logit contributions | |
Along | +0.172 |
nights | +0.165 |
way | +0.151 |
On | +0.144 |
drifted | +0.143 |
Neg logit contributions | |
paramed | -0.294 |
mixer | -0.291 |
liar | -0.290 |
mean | -0.289 |
boys | -0.287 |
Token | So |
Token ID | 1406 |
Feature activation | -2.76e-03 |
Pos logit contributions | |
nights | +0.040 |
drifted | +0.037 |
outer | +0.034 |
ingle | +0.034 |
Along | +0.033 |
Neg logit contributions | |
mixer | -0.038 |
paramed | -0.034 |
mean | -0.033 |
Lee | -0.033 |
stern | -0.033 |
Token | one |
Token ID | 530 |
Feature activation | -6.68e-02 |
Pos logit contributions | |
Along | +0.005 |
ingle | +0.005 |
powers | +0.005 |
dark | +0.005 |
nights | +0.005 |
Neg logit contributions | |
mixer | -0.003 |
mean | -0.003 |
mad | -0.003 |
stern | -0.003 |
sorry | -0.003 |
Token | day |
Token ID | 1110 |
Feature activation | -2.14e-01 |
Pos logit contributions | |
Along | +0.076 |
growth | +0.067 |
lights | +0.067 |
horizon | +0.067 |
way | +0.065 |
Neg logit contributions | |
mean | -0.148 |
mixer | -0.135 |
mad | -0.134 |
angry | -0.131 |
phone | -0.130 |
Token | that |
Token ID | 326 |
Feature activation | -3.21e-02 |
Pos logit contributions | |
Along | +0.028 |
nights | +0.027 |
way | +0.025 |
On | +0.024 |
drifted | +0.024 |
Neg logit contributions | |
paramed | -0.026 |
mixer | -0.025 |
liar | -0.025 |
mean | -0.025 |
boys | -0.025 |
Token | they |
Token ID | 484 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
nights | +0.050 |
Along | +0.049 |
drifted | +0.045 |
way | +0.044 |
whenever | +0.044 |
Neg logit contributions | |
paramed | -0.053 |
liar | -0.052 |
mixer | -0.050 |
wasteful | -0.050 |
greedy | -0.050 |
Token | were |
Token ID | 547 |
Feature activation | +4.11e-02 |
Pos logit contributions | |
Along | +0.022 |
nights | +0.021 |
way | +0.020 |
On | +0.019 |
more | +0.018 |
Neg logit contributions | |
paramed | -0.020 |
mean | -0.020 |
boys | -0.020 |
liar | -0.020 |
mixer | -0.020 |
Token | going |
Token ID | 1016 |
Feature activation | -7.42e-02 |
Pos logit contributions | |
paramed | +0.066 |
liar | +0.063 |
mixer | +0.062 |
roc | +0.062 |
wasteful | +0.061 |
Neg logit contributions | |
nights | -0.066 |
Along | -0.066 |
way | -0.059 |
whenever | -0.059 |
drifted | -0.059 |
Token | to |
Token ID | 284 |
Feature activation | +7.89e-02 |
Pos logit contributions | |
Along | +0.116 |
nights | +0.114 |
way | +0.106 |
drifted | +0.103 |
On | +0.102 |
Neg logit contributions | |
paramed | -0.117 |
mixer | -0.115 |
liar | -0.114 |
mean | -0.113 |
boys | -0.113 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.23e-01 |
Pos logit contributions | |
paramed | +0.129 |
mixer | +0.125 |
liar | +0.125 |
mean | +0.124 |
boys | +0.123 |
Neg logit contributions | |
Along | -0.120 |
nights | -0.117 |
way | -0.109 |
drifted | -0.105 |
On | -0.105 |
Token | space |
Token ID | 2272 |
Feature activation | -1.50e-01 |
Pos logit contributions | |
Along | +0.364 |
nights | +0.328 |
way | +0.325 |
drifted | +0.323 |
ingle | +0.305 |
Neg logit contributions | |
mean | -0.620 |
chefs | -0.615 |
liar | -0.612 |
mixer | -0.611 |
boys | -0.607 |
Token | . |
Token ID | 13 |
Feature activation | -2.13e-02 |
Pos logit contributions | |
Along | +0.205 |
nights | +0.193 |
way | +0.180 |
On | +0.178 |
drifted | +0.177 |
Neg logit contributions | |
boys | -0.261 |
mixer | -0.259 |
mean | -0.258 |
liar | -0.255 |
phone | -0.254 |
Token | They |
Token ID | 1119 |
Feature activation | +4.21e-02 |
Pos logit contributions | |
nights | +0.030 |
Along | +0.029 |
way | +0.028 |
drifted | +0.027 |
more | +0.026 |
Neg logit contributions | |
mixer | -0.036 |
paramed | -0.036 |
wasteful | -0.034 |
boys | -0.034 |
stern | -0.034 |
Token | watched |
Token ID | 7342 |
Feature activation | -8.16e-02 |
Pos logit contributions | |
paramed | +0.060 |
boys | +0.060 |
mean | +0.060 |
mixer | +0.060 |
liar | +0.059 |
Neg logit contributions | |
Along | -0.070 |
nights | -0.068 |
way | -0.065 |
On | -0.063 |
more | -0.061 |
Token | out |
Token ID | 503 |
Feature activation | -2.75e-02 |
Pos logit contributions | |
Along | +0.146 |
nights | +0.140 |
way | +0.135 |
more | +0.130 |
drifted | +0.130 |
Neg logit contributions | |
mixer | -0.108 |
paramed | -0.106 |
boys | -0.106 |
mean | -0.104 |
phone | -0.103 |
Token | what |
Token ID | 644 |
Feature activation | -5.76e-02 |
Pos logit contributions | |
nights | +0.162 |
Along | +0.157 |
way | +0.144 |
days | +0.143 |
whenever | +0.142 |
Neg logit contributions | |
paramed | -0.163 |
liar | -0.153 |
mixer | -0.153 |
wasteful | -0.150 |
roc | -0.149 |
Token | the |
Token ID | 262 |
Feature activation | +1.38e-02 |
Pos logit contributions | |
nights | +0.080 |
Along | +0.078 |
way | +0.072 |
drifted | +0.069 |
whenever | +0.069 |
Neg logit contributions | |
paramed | -0.103 |
liar | -0.099 |
mixer | -0.098 |
wasteful | -0.098 |
roc | -0.097 |
Token | rocket |
Token ID | 10701 |
Feature activation | -6.54e-02 |
Pos logit contributions | |
mean | +0.021 |
mixer | +0.020 |
liar | +0.020 |
greedy | +0.020 |
phone | +0.020 |
Neg logit contributions | |
Along | -0.024 |
On | -0.022 |
nights | -0.022 |
ingle | -0.022 |
way | -0.021 |
Token | discovered |
Token ID | 5071 |
Feature activation | -4.57e-02 |
Pos logit contributions | |
Along | +0.108 |
nights | +0.102 |
On | +0.096 |
way | +0.095 |
dark | +0.091 |
Neg logit contributions | |
boys | -0.097 |
mean | -0.097 |
mixer | -0.097 |
paramed | -0.096 |
phone | -0.095 |
Token | in |
Token ID | 287 |
Feature activation | -5.99e-02 |
Pos logit contributions | |
Along | +0.070 |
nights | +0.064 |
On | +0.059 |
way | +0.059 |
drifted | +0.058 |
Neg logit contributions | |
mean | -0.074 |
boys | -0.073 |
mixer | -0.072 |
phone | -0.071 |
mad | -0.070 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.23e-01 |
Pos logit contributions | |
Along | +0.102 |
nights | +0.101 |
way | +0.094 |
On | +0.091 |
drifted | +0.091 |
Neg logit contributions | |
paramed | -0.087 |
liar | -0.084 |
mixer | -0.084 |
boys | -0.082 |
wasteful | -0.082 |
Token | space |
Token ID | 2272 |
Feature activation | -1.48e-01 |
Pos logit contributions | |
Along | +0.355 |
nights | +0.322 |
way | +0.318 |
drifted | +0.297 |
ingle | +0.294 |
Neg logit contributions | |
liar | -0.631 |
paramed | -0.629 |
mean | -0.629 |
mixer | -0.626 |
greedy | -0.620 |
Token | . |
Token ID | 13 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
Along | +0.198 |
nights | +0.188 |
way | +0.173 |
On | +0.170 |
drifted | +0.167 |
Neg logit contributions | |
paramed | -0.265 |
mixer | -0.264 |
liar | -0.262 |
mean | -0.262 |
boys | -0.261 |
Token | It |
Token ID | 632 |
Feature activation | +7.31e-02 |
Pos logit contributions | |
nights | +0.029 |
drifted | +0.027 |
ingle | +0.026 |
way | +0.026 |
days | +0.025 |
Neg logit contributions | |
paramed | -0.028 |
mixer | -0.027 |
stern | -0.027 |
Lee | -0.027 |
greedy | -0.026 |
Token | was |
Token ID | 373 |
Feature activation | +3.19e-02 |
Pos logit contributions | |
mean | +0.100 |
boys | +0.095 |
paramed | +0.091 |
mad | +0.091 |
stern | +0.090 |
Neg logit contributions | |
Along | -0.132 |
nights | -0.121 |
On | -0.117 |
way | -0.117 |
ingle | -0.114 |
Token | a |
Token ID | 257 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
paramed | +0.045 |
mixer | +0.044 |
liar | +0.043 |
mean | +0.043 |
boys | +0.043 |
Neg logit contributions | |
Along | -0.055 |
nights | -0.054 |
way | -0.051 |
On | -0.049 |
drifted | -0.049 |
Token | Emma |
Token ID | 18966 |
Feature activation | +4.72e-02 |
Pos logit contributions | |
nights | +0.074 |
drifted | +0.072 |
Along | +0.071 |
way | +0.067 |
ingle | +0.066 |
Neg logit contributions | |
mean | -0.036 |
mixer | -0.035 |
boys | -0.034 |
paramed | -0.033 |
greedy | -0.033 |
Token | who |
Token ID | 508 |
Feature activation | -2.32e-02 |
Pos logit contributions | |
paramed | +0.087 |
mixer | +0.086 |
mean | +0.086 |
boys | +0.086 |
liar | +0.085 |
Neg logit contributions | |
Along | -0.061 |
nights | -0.060 |
way | -0.055 |
drifted | -0.052 |
On | -0.051 |
Token | loved |
Token ID | 6151 |
Feature activation | -3.76e-02 |
Pos logit contributions | |
Along | +0.037 |
nights | +0.037 |
way | +0.034 |
On | +0.033 |
drifted | +0.033 |
Neg logit contributions | |
paramed | -0.036 |
mixer | -0.035 |
liar | -0.035 |
wasteful | -0.034 |
boys | -0.034 |
Token | learning |
Token ID | 4673 |
Feature activation | -8.99e-03 |
Pos logit contributions | |
nights | +0.053 |
Along | +0.053 |
whenever | +0.046 |
night | +0.046 |
days | +0.046 |
Neg logit contributions | |
paramed | -0.065 |
liar | -0.065 |
wasteful | -0.064 |
roc | -0.063 |
boys | -0.062 |
Token | about |
Token ID | 546 |
Feature activation | +6.49e-03 |
Pos logit contributions | |
Along | +0.013 |
nights | +0.012 |
way | +0.011 |
drifted | +0.011 |
On | +0.011 |
Neg logit contributions | |
paramed | -0.016 |
mixer | -0.015 |
liar | -0.015 |
mean | -0.015 |
boys | -0.015 |
Token | outer |
Token ID | 12076 |
Feature activation | -3.21e-01 |
Pos logit contributions | |
paramed | +0.012 |
liar | +0.011 |
wasteful | +0.011 |
mixer | +0.011 |
stern | +0.011 |
Neg logit contributions | |
nights | -0.009 |
Along | -0.009 |
whenever | -0.008 |
days | -0.008 |
On | -0.008 |
Token | space |
Token ID | 2272 |
Feature activation | -1.46e-01 |
Pos logit contributions | |
Along | +0.409 |
way | +0.386 |
nights | +0.382 |
drifted | +0.378 |
ingle | +0.352 |
Neg logit contributions | |
mean | -0.580 |
liar | -0.573 |
mixer | -0.571 |
greedy | -0.569 |
roc | -0.567 |
Token | . |
Token ID | 13 |
Feature activation | -2.04e-02 |
Pos logit contributions | |
Along | +0.198 |
nights | +0.187 |
way | +0.178 |
drifted | +0.176 |
On | +0.169 |
Neg logit contributions | |
mean | -0.259 |
boys | -0.258 |
mixer | -0.257 |
phone | -0.255 |
liar | -0.254 |
Token | Everyday |
Token ID | 48154 |
Feature activation | -1.61e-01 |
Pos logit contributions | |
nights | +0.030 |
way | +0.028 |
Along | +0.028 |
drifted | +0.028 |
ingle | +0.026 |
Neg logit contributions | |
mixer | -0.034 |
paramed | -0.034 |
mean | -0.032 |
bluff | -0.031 |
liar | -0.031 |
Token | , |
Token ID | 11 |
Feature activation | +8.19e-02 |
Pos logit contributions | |
nights | +0.225 |
Along | +0.220 |
days | +0.200 |
whenever | +0.198 |
night | +0.198 |
Neg logit contributions | |
paramed | -0.290 |
liar | -0.282 |
wasteful | -0.281 |
roc | -0.280 |
boys | -0.276 |
Token | she |
Token ID | 673 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
mixer | +0.147 |
mean | +0.144 |
paramed | +0.142 |
boys | +0.139 |
phone | +0.138 |
Neg logit contributions | |
Along | -0.112 |
way | -0.107 |
nights | -0.107 |
ingle | -0.099 |
drifted | -0.099 |
Token | of |
Token ID | 286 |
Feature activation | -1.10e-01 |
Pos logit contributions | |
Along | +0.267 |
way | +0.233 |
more | +0.231 |
dark | +0.230 |
drifted | +0.226 |
Neg logit contributions | |
stern | -0.254 |
mean | -0.252 |
mixer | -0.244 |
sorry | -0.241 |
phone | -0.234 |
Token | something |
Token ID | 1223 |
Feature activation | +1.75e-03 |
Pos logit contributions | |
nights | +0.167 |
days | +0.152 |
Along | +0.151 |
night | +0.150 |
whenever | +0.150 |
Neg logit contributions | |
paramed | -0.188 |
roc | -0.181 |
liar | -0.180 |
boys | -0.175 |
wasteful | -0.172 |
Token | to |
Token ID | 284 |
Feature activation | +7.23e-02 |
Pos logit contributions | |
mixer | +0.003 |
mean | +0.003 |
paramed | +0.003 |
stern | +0.003 |
greedy | +0.003 |
Neg logit contributions | |
Along | -0.003 |
nights | -0.003 |
way | -0.002 |
drifted | -0.002 |
On | -0.002 |
Token | eat |
Token ID | 4483 |
Feature activation | -3.48e-02 |
Pos logit contributions | |
paramed | +0.139 |
liar | +0.133 |
boys | +0.132 |
mixer | +0.131 |
greedy | +0.128 |
Neg logit contributions | |
Along | -0.089 |
nights | -0.089 |
On | -0.081 |
whenever | -0.080 |
night | -0.079 |
Token | . |
Token ID | 13 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
Along | +0.039 |
nights | +0.038 |
way | +0.034 |
whenever | +0.033 |
On | +0.033 |
Neg logit contributions | |
paramed | -0.071 |
liar | -0.069 |
mixer | -0.068 |
boys | -0.068 |
wasteful | -0.068 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.97e-01 |
Pos logit contributions | |
nights | +0.022 |
drifted | +0.020 |
way | +0.019 |
ingle | +0.019 |
lights | +0.019 |
Neg logit contributions | |
paramed | -0.020 |
mixer | -0.020 |
Lee | -0.020 |
stern | -0.020 |
greedy | -0.019 |
Token | the |
Token ID | 262 |
Feature activation | +1.27e-02 |
Pos logit contributions | |
Along | +0.615 |
nights | +0.602 |
way | +0.578 |
dark | +0.577 |
more | +0.570 |
Neg logit contributions | |
mixer | -0.294 |
boys | -0.283 |
mean | -0.272 |
phone | -0.270 |
paramed | -0.265 |
Token | way |
Token ID | 835 |
Feature activation | -7.36e-02 |
Pos logit contributions | |
mean | +0.020 |
phone | +0.019 |
mixer | +0.019 |
quarrel | +0.018 |
mad | +0.017 |
Neg logit contributions | |
ingle | -0.021 |
drifted | -0.019 |
ment | -0.019 |
dark | -0.019 |
Along | -0.019 |
Token | , |
Token ID | 11 |
Feature activation | +8.43e-02 |
Pos logit contributions | |
Along | +0.135 |
nights | +0.133 |
drifted | +0.126 |
dark | +0.126 |
way | +0.123 |
Neg logit contributions | |
mixer | -0.090 |
mean | -0.086 |
boys | -0.085 |
phone | -0.085 |
paramed | -0.083 |
Token | she |
Token ID | 673 |
Feature activation | -1.53e-01 |
Pos logit contributions | |
boys | +0.149 |
paramed | +0.148 |
mixer | +0.148 |
mean | +0.146 |
phone | +0.144 |
Neg logit contributions | |
Along | -0.110 |
nights | -0.103 |
way | -0.100 |
dark | -0.099 |
ingle | -0.098 |
Token | spotted |
Token ID | 13489 |
Feature activation | -4.55e-03 |
Pos logit contributions | |
Along | +0.270 |
nights | +0.267 |
way | +0.247 |
drifted | +0.244 |
On | +0.240 |
Neg logit contributions | |
paramed | -0.216 |
mixer | -0.209 |
liar | -0.208 |
wasteful | -0.204 |
mean | -0.201 |
Token | on |
Token ID | 319 |
Feature activation | -3.73e-02 |
Pos logit contributions | |
Along | +0.109 |
nights | +0.106 |
way | +0.096 |
On | +0.092 |
drifted | +0.092 |
Neg logit contributions | |
paramed | -0.180 |
mixer | -0.176 |
liar | -0.176 |
mean | -0.173 |
boys | -0.173 |
Token | her |
Token ID | 607 |
Feature activation | -7.40e-02 |
Pos logit contributions | |
Along | +0.072 |
nights | +0.070 |
way | +0.066 |
drifted | +0.065 |
dark | +0.064 |
Neg logit contributions | |
mixer | -0.045 |
paramed | -0.045 |
mean | -0.045 |
mad | -0.043 |
boys | -0.043 |
Token | sc |
Token ID | 629 |
Feature activation | +3.15e-02 |
Pos logit contributions | |
Along | +0.128 |
nights | +0.123 |
way | +0.118 |
On | +0.117 |
drifted | +0.115 |
Neg logit contributions | |
phone | -0.108 |
mixer | -0.108 |
boys | -0.105 |
mean | -0.103 |
greedy | -0.100 |
Token | ooter |
Token ID | 25141 |
Feature activation | -4.75e-02 |
Pos logit contributions | |
paramed | +0.041 |
roc | +0.040 |
liar | +0.040 |
greedy | +0.039 |
mean | +0.038 |
Neg logit contributions | |
nights | -0.060 |
drifted | -0.057 |
night | -0.053 |
Along | -0.053 |
days | -0.053 |
Token | . |
Token ID | 13 |
Feature activation | -2.97e-02 |
Pos logit contributions | |
nights | +0.063 |
Along | +0.063 |
drifted | +0.059 |
dark | +0.056 |
way | +0.054 |
Neg logit contributions | |
mean | -0.084 |
phone | -0.083 |
mixer | -0.083 |
boys | -0.083 |
paramed | -0.082 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.97e-01 |
Pos logit contributions | |
nights | +0.058 |
lights | +0.054 |
drifted | +0.054 |
lifetime | +0.054 |
ge | +0.053 |
Neg logit contributions | |
paramed | -0.037 |
Lee | -0.036 |
mixer | -0.034 |
phone | -0.031 |
stern | -0.030 |
Token | the |
Token ID | 262 |
Feature activation | +6.33e-03 |
Pos logit contributions | |
Along | +0.648 |
nights | +0.641 |
way | +0.607 |
dark | +0.604 |
more | +0.603 |
Neg logit contributions | |
mixer | -0.268 |
boys | -0.254 |
mean | -0.253 |
phone | -0.242 |
paramed | -0.240 |
Token | way |
Token ID | 835 |
Feature activation | -7.53e-02 |
Pos logit contributions | |
mean | +0.012 |
mixer | +0.011 |
mad | +0.010 |
phone | +0.010 |
paramed | +0.010 |
Neg logit contributions | |
dark | -0.009 |
Along | -0.009 |
nights | -0.009 |
ingle | -0.008 |
outer | -0.008 |
Token | , |
Token ID | 11 |
Feature activation | +7.06e-02 |
Pos logit contributions | |
Along | +0.145 |
nights | +0.143 |
drifted | +0.137 |
dark | +0.135 |
more | +0.133 |
Neg logit contributions | |
mixer | -0.086 |
mean | -0.083 |
phone | -0.080 |
boys | -0.078 |
paramed | -0.078 |
Token | she |
Token ID | 673 |
Feature activation | -1.60e-01 |
Pos logit contributions | |
mixer | +0.117 |
mean | +0.116 |
boys | +0.114 |
greedy | +0.112 |
chefs | +0.111 |
Neg logit contributions | |
Along | -0.101 |
way | -0.095 |
ingle | -0.095 |
nights | -0.094 |
dark | -0.092 |
Token | saw |
Token ID | 2497 |
Feature activation | +3.48e-02 |
Pos logit contributions | |
Along | +0.286 |
nights | +0.285 |
way | +0.265 |
drifted | +0.263 |
whenever | +0.257 |
Neg logit contributions | |
paramed | -0.220 |
liar | -0.213 |
mixer | -0.212 |
wasteful | -0.210 |
roc | -0.204 |
Token | this |
Token ID | 428 |
Feature activation | -5.23e-03 |
Pos logit contributions | |
Along | +0.093 |
nights | +0.091 |
way | +0.083 |
On | +0.081 |
drifted | +0.080 |
Neg logit contributions | |
paramed | -0.109 |
liar | -0.108 |
mixer | -0.105 |
wasteful | -0.105 |
boys | -0.104 |
Token | way |
Token ID | 835 |
Feature activation | -7.57e-02 |
Pos logit contributions | |
Along | +0.009 |
nights | +0.009 |
drifted | +0.008 |
growth | +0.008 |
dark | +0.008 |
Neg logit contributions | |
mean | -0.008 |
phone | -0.007 |
paramed | -0.007 |
mixer | -0.007 |
mad | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | -2.75e-02 |
Pos logit contributions | |
nights | +0.129 |
Along | +0.127 |
growth | +0.121 |
dark | +0.120 |
drifted | +0.118 |
Neg logit contributions | |
mean | -0.103 |
mixer | -0.102 |
phone | -0.102 |
mad | -0.102 |
fighting | -0.096 |
Token | that |
Token ID | 326 |
Feature activation | -3.84e-02 |
Pos logit contributions | |
Along | +0.043 |
nights | +0.043 |
way | +0.040 |
drifted | +0.038 |
On | +0.038 |
Neg logit contributions | |
paramed | -0.043 |
liar | -0.042 |
mixer | -0.042 |
wasteful | -0.041 |
boys | -0.041 |
Token | . |
Token ID | 13 |
Feature activation | -3.16e-02 |
Pos logit contributions | |
nights | +0.057 |
Along | +0.054 |
whenever | +0.052 |
days | +0.050 |
night | +0.049 |
Neg logit contributions | |
paramed | -0.067 |
liar | -0.065 |
mixer | -0.064 |
roc | -0.063 |
wasteful | -0.062 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.96e-01 |
Pos logit contributions | |
nights | +0.046 |
way | +0.040 |
drifted | +0.040 |
Along | +0.040 |
ingle | +0.039 |
Neg logit contributions | |
paramed | -0.056 |
mixer | -0.053 |
liar | -0.052 |
greedy | -0.051 |
wasteful | -0.051 |
Token | the |
Token ID | 262 |
Feature activation | +7.27e-03 |
Pos logit contributions | |
Along | +0.563 |
nights | +0.551 |
way | +0.521 |
On | +0.505 |
dark | +0.504 |
Neg logit contributions | |
paramed | -0.358 |
mixer | -0.356 |
boys | -0.345 |
mean | -0.341 |
liar | -0.334 |
Token | way |
Token ID | 835 |
Feature activation | -7.58e-02 |
Pos logit contributions | |
mean | +0.013 |
phone | +0.012 |
paramed | +0.012 |
mad | +0.012 |
fighting | +0.012 |
Neg logit contributions | |
ingle | -0.010 |
dark | -0.010 |
Along | -0.009 |
drifted | -0.009 |
ment | -0.009 |
Token | , |
Token ID | 11 |
Feature activation | +7.10e-02 |
Pos logit contributions | |
Along | +0.132 |
nights | +0.130 |
drifted | +0.120 |
dark | +0.119 |
way | +0.119 |
Neg logit contributions | |
mixer | -0.102 |
paramed | -0.100 |
mean | -0.099 |
boys | -0.097 |
phone | -0.096 |
Token | the |
Token ID | 262 |
Feature activation | +7.37e-03 |
Pos logit contributions | |
paramed | +0.105 |
mixer | +0.103 |
mean | +0.101 |
liar | +0.101 |
boys | +0.101 |
Neg logit contributions | |
Along | -0.117 |
nights | -0.115 |
way | -0.108 |
drifted | -0.105 |
On | -0.104 |
Token | girl |
Token ID | 2576 |
Feature activation | +6.85e-03 |
Pos logit contributions | |
paramed | +0.010 |
mean | +0.009 |
mixer | +0.009 |
phone | +0.009 |
greedy | +0.008 |
Neg logit contributions | |
Along | -0.014 |
nights | -0.014 |
drifted | -0.014 |
ingle | -0.014 |
dark | -0.013 |
Token | a |
Token ID | 257 |
Feature activation | -2.77e-02 |
Pos logit contributions | |
paramed | +0.106 |
mean | +0.103 |
liar | +0.103 |
mixer | +0.102 |
boys | +0.101 |
Neg logit contributions | |
Along | -0.137 |
nights | -0.135 |
way | -0.128 |
drifted | -0.124 |
On | -0.122 |
Token | far |
Token ID | 1290 |
Feature activation | -1.20e-01 |
Pos logit contributions | |
Along | +0.036 |
nights | +0.034 |
way | +0.033 |
night | +0.033 |
day | +0.031 |
Neg logit contributions | |
paramed | -0.051 |
liar | -0.049 |
wasteful | -0.049 |
mixer | -0.049 |
boys | -0.048 |
Token | off |
Token ID | 572 |
Feature activation | -9.44e-02 |
Pos logit contributions | |
Along | +0.137 |
nights | +0.135 |
drifted | +0.127 |
dark | +0.113 |
On | +0.113 |
Neg logit contributions | |
mean | -0.243 |
paramed | -0.236 |
mad | -0.231 |
mixer | -0.231 |
greedy | -0.231 |
Token | place |
Token ID | 1295 |
Feature activation | -5.91e-02 |
Pos logit contributions | |
Along | +0.200 |
more | +0.193 |
way | +0.191 |
nights | +0.191 |
drifted | +0.191 |
Neg logit contributions | |
mean | -0.096 |
angry | -0.078 |
mad | -0.077 |
greedy | -0.076 |
wasteful | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | -2.88e-02 |
Pos logit contributions | |
Along | +0.088 |
nights | +0.087 |
drifted | +0.082 |
dark | +0.081 |
way | +0.077 |
Neg logit contributions | |
mean | -0.093 |
stern | -0.092 |
mixer | -0.091 |
phone | -0.091 |
mad | -0.090 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
nights | +0.050 |
days | +0.045 |
lights | +0.044 |
drifted | +0.043 |
way | +0.042 |
Neg logit contributions | |
paramed | -0.044 |
Lee | -0.042 |
mixer | -0.041 |
mean | -0.037 |
greedy | -0.037 |
Token | the |
Token ID | 262 |
Feature activation | +9.00e-03 |
Pos logit contributions | |
nights | +0.629 |
Along | +0.623 |
way | +0.606 |
more | +0.599 |
dark | +0.592 |
Neg logit contributions | |
mixer | -0.256 |
paramed | -0.248 |
boys | -0.246 |
Lee | -0.240 |
phone | -0.236 |
Token | way |
Token ID | 835 |
Feature activation | -7.41e-02 |
Pos logit contributions | |
mean | +0.016 |
quarrel | +0.014 |
boys | +0.014 |
phone | +0.014 |
rude | +0.014 |
Neg logit contributions | |
ingle | -0.013 |
ment | -0.012 |
dark | -0.012 |
edly | -0.012 |
drifted | -0.012 |
Token | , |
Token ID | 11 |
Feature activation | +7.13e-02 |
Pos logit contributions | |
Along | +0.142 |
nights | +0.142 |
drifted | +0.137 |
dark | +0.137 |
more | +0.135 |
Neg logit contributions | |
mixer | -0.077 |
mean | -0.074 |
phone | -0.073 |
Lee | -0.072 |
boys | -0.069 |
Token | she |
Token ID | 673 |
Feature activation | -1.60e-01 |
Pos logit contributions | |
paramed | +0.111 |
mixer | +0.111 |
Lee | +0.107 |
greedy | +0.107 |
mean | +0.107 |
Neg logit contributions | |
nights | -0.105 |
Along | -0.104 |
way | -0.103 |
ingle | -0.101 |
dark | -0.098 |
Token | found |
Token ID | 1043 |
Feature activation | -9.38e-03 |
Pos logit contributions | |
Along | +0.311 |
nights | +0.305 |
way | +0.289 |
On | +0.283 |
dark | +0.279 |
Neg logit contributions | |
paramed | -0.188 |
mean | -0.183 |
boys | -0.181 |
liar | -0.179 |
mixer | -0.178 |
Token | the |
Token ID | 262 |
Feature activation | +7.42e-03 |
Pos logit contributions | |
Along | +0.094 |
nights | +0.094 |
way | +0.086 |
On | +0.084 |
drifted | +0.084 |
Neg logit contributions | |
paramed | -0.084 |
liar | -0.082 |
mixer | -0.082 |
wasteful | -0.080 |
mean | -0.079 |
Token | mouse |
Token ID | 10211 |
Feature activation | +9.12e-02 |
Pos logit contributions | |
paramed | +0.010 |
mean | +0.009 |
mixer | +0.009 |
liar | +0.009 |
boys | +0.009 |
Neg logit contributions | |
Along | -0.014 |
nights | -0.013 |
way | -0.013 |
drifted | -0.013 |
On | -0.013 |
Token | 's |
Token ID | 338 |
Feature activation | +1.49e-01 |
Pos logit contributions | |
boys | +0.143 |
phone | +0.139 |
mean | +0.138 |
greedy | +0.136 |
stern | +0.135 |
Neg logit contributions | |
Along | -0.141 |
nights | -0.136 |
way | -0.131 |
ingle | -0.129 |
drifted | -0.127 |
Token | home |
Token ID | 1363 |
Feature activation | -5.35e-02 |
Pos logit contributions | |
phone | +0.214 |
boys | +0.209 |
mean | +0.205 |
paramed | +0.199 |
greedy | +0.197 |
Neg logit contributions | |
Along | -0.258 |
dark | -0.249 |
ingle | -0.248 |
way | -0.239 |
nights | -0.238 |
Token | . |
Token ID | 13 |
Feature activation | -3.15e-02 |
Pos logit contributions | |
Along | +0.078 |
nights | +0.073 |
On | +0.070 |
way | +0.070 |
drifted | +0.069 |
Neg logit contributions | |
boys | -0.088 |
phone | -0.086 |
mean | -0.086 |
mixer | -0.085 |
greedy | -0.085 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
nights | +0.049 |
way | +0.044 |
outer | +0.043 |
dark | +0.043 |
ingle | +0.042 |
Neg logit contributions | |
paramed | -0.051 |
boys | -0.048 |
stern | -0.047 |
greedy | -0.047 |
mixer | -0.046 |
Token | the |
Token ID | 262 |
Feature activation | +6.92e-03 |
Pos logit contributions | |
Along | +0.595 |
nights | +0.588 |
way | +0.569 |
dark | +0.562 |
more | +0.550 |
Neg logit contributions | |
boys | -0.304 |
mixer | -0.294 |
paramed | -0.286 |
mean | -0.279 |
phone | -0.274 |
Token | way |
Token ID | 835 |
Feature activation | -7.63e-02 |
Pos logit contributions | |
mean | +0.012 |
boys | +0.011 |
phone | +0.011 |
quarrel | +0.010 |
mad | +0.010 |
Neg logit contributions | |
ingle | -0.010 |
dark | -0.009 |
Along | -0.009 |
ment | -0.009 |
powers | -0.009 |
Token | , |
Token ID | 11 |
Feature activation | +7.24e-02 |
Pos logit contributions | |
Along | +0.136 |
nights | +0.135 |
drifted | +0.127 |
dark | +0.127 |
way | +0.126 |
Neg logit contributions | |
mixer | -0.095 |
boys | -0.092 |
paramed | -0.092 |
mean | -0.092 |
phone | -0.089 |
Token | they |
Token ID | 484 |
Feature activation | -2.21e-02 |
Pos logit contributions | |
paramed | +0.105 |
boys | +0.104 |
mixer | +0.103 |
mean | +0.102 |
greedy | +0.101 |
Neg logit contributions | |
Along | -0.120 |
nights | -0.117 |
way | -0.113 |
dark | -0.109 |
drifted | -0.107 |
Token | met |
Token ID | 1138 |
Feature activation | -4.33e-02 |
Pos logit contributions | |
Along | +0.037 |
nights | +0.037 |
drifted | +0.034 |
way | +0.033 |
whenever | +0.033 |
Neg logit contributions | |
paramed | -0.033 |
mixer | -0.033 |
liar | -0.033 |
wasteful | -0.032 |
bluff | -0.031 |
Token | to |
Token ID | 284 |
Feature activation | +7.64e-02 |
Pos logit contributions | |
nights | +0.211 |
Along | +0.195 |
dark | +0.194 |
drifted | +0.191 |
way | +0.190 |
Neg logit contributions | |
boys | -0.181 |
mean | -0.176 |
phone | -0.175 |
paramed | -0.171 |
mixer | -0.170 |
Token | explore |
Token ID | 7301 |
Feature activation | -2.24e-01 |
Pos logit contributions | |
paramed | +0.100 |
mixer | +0.099 |
boys | +0.099 |
mean | +0.098 |
liar | +0.098 |
Neg logit contributions | |
nights | -0.137 |
Along | -0.135 |
way | -0.129 |
drifted | -0.124 |
dark | -0.122 |
Token | the |
Token ID | 262 |
Feature activation | +9.00e-03 |
Pos logit contributions | |
Along | +0.403 |
nights | +0.388 |
way | +0.377 |
drifted | +0.369 |
ingle | +0.363 |
Neg logit contributions | |
boys | -0.286 |
mean | -0.284 |
mixer | -0.281 |
stern | -0.271 |
greedy | -0.269 |
Token | forest |
Token ID | 8222 |
Feature activation | -1.31e-01 |
Pos logit contributions | |
mean | +0.016 |
boys | +0.015 |
phone | +0.014 |
greedy | +0.014 |
roc | +0.014 |
Neg logit contributions | |
ingle | -0.013 |
drifted | -0.013 |
Along | -0.013 |
lights | -0.012 |
ment | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | -2.93e-02 |
Pos logit contributions | |
Along | +0.184 |
nights | +0.181 |
drifted | +0.173 |
dark | +0.172 |
On | +0.163 |
Neg logit contributions | |
boys | -0.223 |
mean | -0.210 |
phone | -0.209 |
mixer | -0.207 |
greedy | -0.206 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
nights | +0.043 |
ingle | +0.039 |
way | +0.038 |
drifted | +0.038 |
dark | +0.037 |
Neg logit contributions | |
paramed | -0.048 |
boys | -0.047 |
greedy | -0.046 |
mixer | -0.045 |
mean | -0.045 |
Token | the |
Token ID | 262 |
Feature activation | +8.49e-03 |
Pos logit contributions | |
nights | +0.601 |
Along | +0.590 |
way | +0.569 |
dark | +0.568 |
ingle | +0.559 |
Neg logit contributions | |
boys | -0.294 |
mixer | -0.281 |
mean | -0.274 |
paramed | -0.270 |
roc | -0.261 |
Token | way |
Token ID | 835 |
Feature activation | -7.53e-02 |
Pos logit contributions | |
mean | +0.014 |
boys | +0.013 |
quarrel | +0.012 |
roc | +0.012 |
rude | +0.012 |
Neg logit contributions | |
ingle | -0.013 |
dark | -0.013 |
ment | -0.013 |
ugg | -0.012 |
drifted | -0.012 |
Token | they |
Token ID | 484 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
nights | +0.132 |
Along | +0.131 |
dark | +0.126 |
drifted | +0.124 |
way | +0.122 |
Neg logit contributions | |
mixer | -0.092 |
boys | -0.092 |
mean | -0.091 |
phone | -0.088 |
stern | -0.087 |
Token | talked |
Token ID | 6619 |
Feature activation | -1.67e-02 |
Pos logit contributions | |
Along | +0.042 |
nights | +0.042 |
way | +0.039 |
On | +0.038 |
drifted | +0.038 |
Neg logit contributions | |
paramed | -0.027 |
mean | -0.027 |
mixer | -0.026 |
boys | -0.026 |
liar | -0.026 |
Token | about |
Token ID | 546 |
Feature activation | +1.20e-03 |
Pos logit contributions | |
Along | +0.025 |
nights | +0.024 |
way | +0.023 |
drifted | +0.022 |
On | +0.022 |
Neg logit contributions | |
paramed | -0.028 |
mixer | -0.027 |
liar | -0.027 |
mean | -0.027 |
boys | -0.026 |
Token | to |
Token ID | 284 |
Feature activation | +7.64e-02 |
Pos logit contributions | |
Along | +0.050 |
nights | +0.048 |
way | +0.046 |
whenever | +0.045 |
On | +0.043 |
Neg logit contributions | |
paramed | -0.069 |
mixer | -0.068 |
liar | -0.067 |
wasteful | -0.065 |
roc | -0.065 |
Token | go |
Token ID | 467 |
Feature activation | -6.82e-02 |
Pos logit contributions | |
paramed | +0.139 |
liar | +0.136 |
mixer | +0.136 |
wasteful | +0.134 |
mean | +0.133 |
Neg logit contributions | |
Along | -0.102 |
nights | -0.099 |
way | -0.091 |
On | -0.088 |
drifted | -0.088 |
Token | visit |
Token ID | 3187 |
Feature activation | -2.02e-01 |
Pos logit contributions | |
nights | +0.127 |
way | +0.118 |
Along | +0.117 |
drifted | +0.115 |
lights | +0.111 |
Neg logit contributions | |
mean | -0.090 |
phone | -0.087 |
boys | -0.085 |
paramed | -0.082 |
mad | -0.081 |
Token | Slip |
Token ID | 49988 |
Feature activation | -5.06e-02 |
Pos logit contributions | |
Along | +0.329 |
nights | +0.320 |
way | +0.296 |
On | +0.292 |
whenever | +0.291 |
Neg logit contributions | |
paramed | -0.317 |
liar | -0.306 |
mixer | -0.305 |
wasteful | -0.301 |
mean | -0.296 |
Token | . |
Token ID | 13 |
Feature activation | -3.31e-02 |
Pos logit contributions | |
Along | +0.066 |
nights | +0.065 |
way | +0.060 |
drifted | +0.058 |
On | +0.058 |
Neg logit contributions | |
boys | -0.091 |
mean | -0.089 |
mixer | -0.088 |
greedy | -0.088 |
phone | -0.088 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.94e-01 |
Pos logit contributions | |
nights | +0.052 |
Along | +0.045 |
way | +0.045 |
drifted | +0.045 |
days | +0.044 |
Neg logit contributions | |
paramed | -0.055 |
mixer | -0.051 |
mean | -0.051 |
boys | -0.051 |
greedy | -0.051 |
Token | the |
Token ID | 262 |
Feature activation | +4.49e-03 |
Pos logit contributions | |
nights | +0.604 |
Along | +0.589 |
way | +0.565 |
dark | +0.552 |
more | +0.544 |
Neg logit contributions | |
boys | -0.307 |
paramed | -0.298 |
mixer | -0.296 |
mean | -0.294 |
Lee | -0.293 |
Token | way |
Token ID | 835 |
Feature activation | -7.78e-02 |
Pos logit contributions | |
mean | +0.008 |
phone | +0.007 |
mad | +0.007 |
greedy | +0.007 |
boys | +0.007 |
Neg logit contributions | |
nights | -0.006 |
Along | -0.006 |
dark | -0.006 |
lights | -0.006 |
ingle | -0.006 |
Token | she |
Token ID | 673 |
Feature activation | -1.62e-01 |
Pos logit contributions | |
nights | +0.147 |
Along | +0.146 |
dark | +0.140 |
drifted | +0.137 |
way | +0.134 |
Neg logit contributions | |
mean | -0.086 |
mixer | -0.086 |
Lee | -0.086 |
boys | -0.085 |
phone | -0.085 |
Token | met |
Token ID | 1138 |
Feature activation | -4.24e-02 |
Pos logit contributions | |
nights | +0.301 |
Along | +0.299 |
way | +0.286 |
more | +0.278 |
ingle | +0.274 |
Neg logit contributions | |
paramed | -0.197 |
boys | -0.194 |
mean | -0.193 |
greedy | -0.186 |
mixer | -0.184 |
Token | a |
Token ID | 257 |
Feature activation | -2.84e-02 |
Pos logit contributions | |
Along | +0.069 |
nights | +0.068 |
way | +0.063 |
On | +0.062 |
whenever | +0.061 |
Neg logit contributions | |
paramed | -0.065 |
mixer | -0.063 |
liar | -0.062 |
wasteful | -0.061 |
boys | -0.060 |
Token | make |
Token ID | 787 |
Feature activation | +7.73e-02 |
Pos logit contributions | |
paramed | +0.140 |
mixer | +0.137 |
liar | +0.137 |
mean | +0.136 |
boys | +0.135 |
Neg logit contributions | |
Along | -0.103 |
nights | -0.101 |
way | -0.093 |
On | -0.088 |
drifted | -0.088 |
Token | his |
Token ID | 465 |
Feature activation | +6.52e-03 |
Pos logit contributions | |
paramed | +0.103 |
mixer | +0.101 |
mean | +0.100 |
boys | +0.099 |
mad | +0.099 |
Neg logit contributions | |
Along | -0.142 |
nights | -0.134 |
way | -0.130 |
On | -0.126 |
dark | -0.124 |
Token | way |
Token ID | 835 |
Feature activation | -7.61e-02 |
Pos logit contributions | |
mixer | +0.010 |
mad | +0.009 |
phone | +0.009 |
mean | +0.009 |
paramed | +0.009 |
Neg logit contributions | |
Along | -0.011 |
nights | -0.010 |
way | -0.010 |
dark | -0.010 |
On | -0.010 |
Token | home |
Token ID | 1363 |
Feature activation | -5.08e-02 |
Pos logit contributions | |
Along | +0.156 |
nights | +0.156 |
dark | +0.148 |
growth | +0.147 |
drifted | +0.143 |
Neg logit contributions | |
mean | -0.082 |
mad | -0.082 |
stern | -0.075 |
phone | -0.075 |
angry | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | -3.04e-02 |
Pos logit contributions | |
Along | +0.068 |
nights | +0.063 |
way | +0.059 |
On | +0.058 |
dark | +0.058 |
Neg logit contributions | |
paramed | -0.093 |
boys | -0.091 |
mean | -0.090 |
mixer | -0.090 |
phone | -0.090 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.94e-01 |
Pos logit contributions | |
nights | +0.046 |
lights | +0.044 |
outer | +0.043 |
drifted | +0.042 |
dark | +0.041 |
Neg logit contributions | |
paramed | -0.048 |
roc | -0.044 |
Lee | -0.043 |
mixer | -0.042 |
greedy | -0.041 |
Token | the |
Token ID | 262 |
Feature activation | +7.40e-03 |
Pos logit contributions | |
Along | +0.587 |
nights | +0.571 |
way | +0.554 |
dark | +0.548 |
more | +0.538 |
Neg logit contributions | |
paramed | -0.312 |
mixer | -0.309 |
boys | -0.300 |
mean | -0.293 |
phone | -0.288 |
Token | way |
Token ID | 835 |
Feature activation | -7.62e-02 |
Pos logit contributions | |
mean | +0.012 |
phone | +0.010 |
quarrel | +0.010 |
mad | +0.010 |
boys | +0.010 |
Neg logit contributions | |
dark | -0.012 |
ingle | -0.012 |
Along | -0.012 |
ment | -0.011 |
edly | -0.011 |
Token | , |
Token ID | 11 |
Feature activation | +7.16e-02 |
Pos logit contributions | |
Along | +0.137 |
nights | +0.135 |
dark | +0.128 |
drifted | +0.127 |
On | +0.125 |
Neg logit contributions | |
mixer | -0.094 |
mean | -0.093 |
paramed | -0.093 |
boys | -0.090 |
phone | -0.090 |
Token | he |
Token ID | 339 |
Feature activation | -9.57e-02 |
Pos logit contributions | |
mean | +0.106 |
paramed | +0.106 |
mixer | +0.103 |
mad | +0.103 |
greedy | +0.101 |
Neg logit contributions | |
Along | -0.112 |
nights | -0.111 |
dark | -0.108 |
drifted | -0.105 |
way | -0.102 |
Token | met |
Token ID | 1138 |
Feature activation | -4.21e-02 |
Pos logit contributions | |
nights | +0.158 |
Along | +0.157 |
drifted | +0.145 |
way | +0.144 |
whenever | +0.140 |
Neg logit contributions | |
paramed | -0.146 |
liar | -0.143 |
mixer | -0.142 |
wasteful | -0.140 |
roc | -0.136 |
Token | angel |
Token ID | 18304 |
Feature activation | -9.86e-03 |
Pos logit contributions | |
mean | +0.007 |
phone | +0.007 |
mixer | +0.007 |
mad | +0.007 |
liar | +0.007 |
Neg logit contributions | |
Along | -0.014 |
nights | -0.013 |
On | -0.013 |
way | -0.012 |
drifted | -0.012 |
Token | and |
Token ID | 290 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
Along | +0.018 |
nights | +0.017 |
way | +0.017 |
On | +0.016 |
drifted | +0.016 |
Neg logit contributions | |
boys | -0.013 |
phone | -0.013 |
mean | -0.012 |
stern | -0.012 |
mixer | -0.012 |
Token | followed |
Token ID | 3940 |
Feature activation | -9.12e-02 |
Pos logit contributions | |
Along | +0.045 |
nights | +0.042 |
way | +0.041 |
On | +0.039 |
ingle | +0.039 |
Neg logit contributions | |
paramed | -0.041 |
boys | -0.040 |
mean | -0.040 |
greedy | -0.040 |
phone | -0.039 |
Token | her |
Token ID | 607 |
Feature activation | -7.21e-02 |
Pos logit contributions | |
Along | +0.181 |
nights | +0.178 |
way | +0.169 |
drifted | +0.166 |
On | +0.161 |
Neg logit contributions | |
mean | -0.103 |
paramed | -0.102 |
mad | -0.101 |
phone | -0.100 |
boys | -0.099 |
Token | . |
Token ID | 13 |
Feature activation | -2.99e-02 |
Pos logit contributions | |
dark | +0.141 |
Along | +0.140 |
nights | +0.134 |
way | +0.133 |
ingle | +0.132 |
Neg logit contributions | |
phone | -0.094 |
boys | -0.088 |
stern | -0.084 |
mean | -0.080 |
whining | -0.079 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.94e-01 |
Pos logit contributions | |
nights | +0.055 |
drifted | +0.050 |
ingle | +0.049 |
iest | +0.049 |
days | +0.049 |
Neg logit contributions | |
paramed | -0.040 |
stern | -0.037 |
greedy | -0.037 |
Lee | -0.036 |
boys | -0.036 |
Token | the |
Token ID | 262 |
Feature activation | +6.33e-03 |
Pos logit contributions | |
Along | +0.642 |
nights | +0.622 |
way | +0.611 |
dark | +0.602 |
more | +0.597 |
Neg logit contributions | |
boys | -0.249 |
mixer | -0.244 |
mean | -0.238 |
phone | -0.234 |
Lee | -0.233 |
Token | way |
Token ID | 835 |
Feature activation | -7.57e-02 |
Pos logit contributions | |
mean | +0.010 |
boys | +0.009 |
phone | +0.009 |
quarrel | +0.008 |
mad | +0.008 |
Neg logit contributions | |
Along | -0.010 |
ingle | -0.010 |
ugg | -0.009 |
dark | -0.009 |
ment | -0.009 |
Token | , |
Token ID | 11 |
Feature activation | +7.13e-02 |
Pos logit contributions | |
Along | +0.150 |
nights | +0.147 |
drifted | +0.143 |
dark | +0.140 |
more | +0.139 |
Neg logit contributions | |
mixer | -0.079 |
mean | -0.076 |
phone | -0.073 |
paramed | -0.071 |
boys | -0.071 |
Token | the |
Token ID | 262 |
Feature activation | +6.63e-03 |
Pos logit contributions | |
mixer | +0.117 |
mean | +0.116 |
boys | +0.115 |
paramed | +0.115 |
greedy | +0.113 |
Neg logit contributions | |
Along | -0.104 |
way | -0.098 |
nights | -0.098 |
ingle | -0.096 |
dark | -0.091 |
Token | angel |
Token ID | 18304 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
mean | +0.007 |
phone | +0.006 |
mad | +0.006 |
greedy | +0.005 |
mixer | +0.005 |
Neg logit contributions | |
Along | -0.016 |
ingle | -0.015 |
On | -0.015 |
ge | -0.014 |
way | -0.014 |
Token | outside |
Token ID | 2354 |
Feature activation | -1.46e-01 |
Pos logit contributions | |
Along | +0.192 |
nights | +0.190 |
way | +0.176 |
drifted | +0.171 |
On | +0.169 |
Neg logit contributions | |
paramed | -0.197 |
mixer | -0.193 |
liar | -0.190 |
mean | -0.189 |
boys | -0.188 |
Token | and |
Token ID | 290 |
Feature activation | -2.34e-02 |
Pos logit contributions | |
Along | +0.239 |
nights | +0.238 |
way | +0.227 |
dark | +0.220 |
drifted | +0.216 |
Neg logit contributions | |
mixer | -0.211 |
paramed | -0.209 |
mad | -0.201 |
mean | -0.201 |
phone | -0.201 |
Token | started |
Token ID | 2067 |
Feature activation | -2.83e-02 |
Pos logit contributions | |
Along | +0.040 |
nights | +0.039 |
way | +0.035 |
drifted | +0.035 |
On | +0.035 |
Neg logit contributions | |
paramed | -0.036 |
liar | -0.035 |
wasteful | -0.034 |
roc | -0.034 |
mixer | -0.033 |
Token | walking |
Token ID | 6155 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
Along | +0.039 |
nights | +0.038 |
whenever | +0.035 |
drifted | +0.035 |
night | +0.034 |
Neg logit contributions | |
liar | -0.051 |
roc | -0.050 |
paramed | -0.050 |
wasteful | -0.050 |
mean | -0.048 |
Token | . |
Token ID | 13 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
Along | +0.246 |
nights | +0.240 |
drifted | +0.227 |
dark | +0.223 |
way | +0.220 |
Neg logit contributions | |
mean | -0.318 |
boys | -0.312 |
mixer | -0.310 |
mad | -0.310 |
paramed | -0.309 |
Token | Along |
Token ID | 17159 |
Feature activation | -2.94e-01 |
Pos logit contributions | |
nights | +0.044 |
ge | +0.041 |
ingle | +0.040 |
drifted | +0.040 |
lifetime | +0.040 |
Neg logit contributions | |
mixer | -0.039 |
paramed | -0.038 |
Lee | -0.036 |
phone | -0.035 |
stern | -0.034 |
Token | the |
Token ID | 262 |
Feature activation | +5.97e-03 |
Pos logit contributions | |
Along | +0.623 |
nights | +0.607 |
way | +0.582 |
dark | +0.579 |
more | +0.572 |
Neg logit contributions | |
mixer | -0.281 |
boys | -0.274 |
mean | -0.266 |
paramed | -0.264 |
phone | -0.258 |
Token | way |
Token ID | 835 |
Feature activation | -7.55e-02 |
Pos logit contributions | |
mean | +0.011 |
phone | +0.010 |
mixer | +0.010 |
paramed | +0.010 |
mad | +0.010 |
Neg logit contributions | |
dark | -0.008 |
Along | -0.008 |
nights | -0.008 |
drifted | -0.007 |
ingle | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | +7.29e-02 |
Pos logit contributions | |
Along | +0.145 |
nights | +0.142 |
drifted | +0.136 |
dark | +0.135 |
more | +0.133 |
Neg logit contributions | |
mixer | -0.087 |
mean | -0.083 |
phone | -0.080 |
paramed | -0.079 |
boys | -0.079 |
Token | she |
Token ID | 673 |
Feature activation | -1.57e-01 |
Pos logit contributions | |
mixer | +0.125 |
boys | +0.122 |
mean | +0.120 |
paramed | +0.119 |
phone | +0.118 |
Neg logit contributions | |
Along | -0.102 |
way | -0.094 |
ingle | -0.093 |
nights | -0.093 |
dark | -0.092 |
Token | ate |
Token ID | 15063 |
Feature activation | -2.55e-02 |
Pos logit contributions | |
nights | +0.256 |
Along | +0.254 |
drifted | +0.243 |
night | +0.231 |
whenever | +0.230 |
Neg logit contributions | |
paramed | -0.245 |
liar | -0.238 |
wasteful | -0.234 |
bluff | -0.232 |
mixer | -0.232 |
Token |
|
Token ID | 198 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
Along | +0.014 |
nights | +0.014 |
way | +0.013 |
drifted | +0.012 |
On | +0.012 |
Neg logit contributions | |
paramed | -0.023 |
mixer | -0.023 |
liar | -0.022 |
mean | -0.022 |
boys | -0.022 |
Token | L |
Token ID | 43 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
Along | +0.016 |
nights | +0.015 |
way | +0.014 |
On | +0.014 |
drifted | +0.014 |
Neg logit contributions | |
paramed | -0.018 |
mixer | -0.017 |
liar | -0.017 |
wasteful | -0.017 |
boys | -0.017 |
Token | ila |
Token ID | 10102 |
Feature activation | -3.47e-02 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.043 |
way | +0.039 |
whenever | +0.039 |
On | +0.039 |
Neg logit contributions | |
paramed | -0.020 |
liar | -0.019 |
mean | -0.019 |
mixer | -0.019 |
roc | -0.019 |
Token | and |
Token ID | 290 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
Along | +0.043 |
On | +0.039 |
drifted | +0.039 |
nights | +0.037 |
way | +0.037 |
Neg logit contributions | |
paramed | -0.068 |
liar | -0.068 |
wasteful | -0.065 |
roc | -0.064 |
mixer | -0.064 |
Token | Ben |
Token ID | 3932 |
Feature activation | +1.05e-01 |
Pos logit contributions | |
Along | +0.019 |
nights | +0.018 |
drifted | +0.018 |
On | +0.018 |
night | +0.018 |
Neg logit contributions | |
paramed | -0.014 |
liar | -0.014 |
wasteful | -0.014 |
mixer | -0.014 |
chefs | -0.013 |
Token | are |
Token ID | 389 |
Feature activation | +1.43e-01 |
Pos logit contributions | |
paramed | +0.153 |
liar | +0.150 |
roc | +0.147 |
wasteful | +0.146 |
mixer | +0.142 |
Neg logit contributions | |
Along | -0.182 |
drifted | -0.171 |
On | -0.168 |
whenever | -0.162 |
night | -0.161 |
Token | happy |
Token ID | 3772 |
Feature activation | +4.21e-02 |
Pos logit contributions | |
roc | +0.265 |
paramed | +0.247 |
liar | +0.241 |
mixer | +0.239 |
ution | +0.232 |
Neg logit contributions | |
nights | -0.187 |
Along | -0.186 |
drifted | -0.183 |
whenever | -0.181 |
night | -0.180 |
Token | . |
Token ID | 13 |
Feature activation | -2.06e-02 |
Pos logit contributions | |
mixer | +0.072 |
paramed | +0.072 |
mean | +0.072 |
mad | +0.071 |
chefs | +0.071 |
Neg logit contributions | |
Along | -0.061 |
nights | -0.058 |
way | -0.055 |
dark | -0.052 |
On | -0.052 |
Token | They |
Token ID | 1119 |
Feature activation | +4.82e-02 |
Pos logit contributions | |
Along | +0.030 |
nights | +0.029 |
way | +0.027 |
On | +0.026 |
drifted | +0.026 |
Neg logit contributions | |
paramed | -0.035 |
liar | -0.035 |
mixer | -0.035 |
mean | -0.034 |
wasteful | -0.034 |
Token | say |
Token ID | 910 |
Feature activation | +7.87e-02 |
Pos logit contributions | |
paramed | +0.066 |
mixer | +0.064 |
liar | +0.064 |
wasteful | +0.062 |
boys | +0.062 |
Neg logit contributions | |
Along | -0.087 |
nights | -0.083 |
On | -0.078 |
drifted | -0.078 |
way | -0.078 |
Token | , |
Token ID | 11 |
Feature activation | +8.95e-02 |
Pos logit contributions | |
roc | +0.174 |
wasteful | +0.171 |
bluff | +0.166 |
boys | +0.163 |
cigarette | +0.162 |
Neg logit contributions | |
whenever | -0.075 |
days | -0.071 |
of | -0.070 |
Along | -0.067 |
day | -0.066 |
Token | hurt |
Token ID | 5938 |
Feature activation | +9.89e-02 |
Pos logit contributions | |
Along | +0.115 |
nights | +0.113 |
way | +0.106 |
On | +0.103 |
drifted | +0.102 |
Neg logit contributions | |
paramed | -0.090 |
liar | -0.087 |
mixer | -0.087 |
mean | -0.086 |
boys | -0.085 |
Token | its |
Token ID | 663 |
Feature activation | -1.11e-02 |
Pos logit contributions | |
paramed | +0.176 |
liar | +0.166 |
mixer | +0.164 |
bluff | +0.163 |
wasteful | +0.162 |
Neg logit contributions | |
nights | -0.137 |
whenever | -0.129 |
Along | -0.128 |
night | -0.128 |
days | -0.126 |
Token | wing |
Token ID | 8539 |
Feature activation | -8.34e-03 |
Pos logit contributions | |
Along | +0.020 |
nights | +0.020 |
way | +0.019 |
On | +0.018 |
drifted | +0.018 |
Neg logit contributions | |
paramed | -0.015 |
mixer | -0.015 |
liar | -0.015 |
mean | -0.015 |
greedy | -0.015 |
Token | . |
Token ID | 13 |
Feature activation | -2.38e-02 |
Pos logit contributions | |
Along | +0.010 |
nights | +0.010 |
way | +0.008 |
On | +0.008 |
drifted | +0.008 |
Neg logit contributions | |
paramed | -0.016 |
mixer | -0.016 |
mean | -0.016 |
liar | -0.016 |
boys | -0.016 |
Token | Lily |
Token ID | 20037 |
Feature activation | +3.14e-02 |
Pos logit contributions | |
Along | +0.035 |
On | +0.032 |
nights | +0.030 |
way | +0.030 |
As | +0.029 |
Neg logit contributions | |
liar | -0.043 |
boys | -0.042 |
paramed | -0.041 |
bluff | -0.041 |
roc | -0.041 |
Token | 's |
Token ID | 338 |
Feature activation | +1.51e-01 |
Pos logit contributions | |
paramed | +0.053 |
liar | +0.053 |
wasteful | +0.052 |
bluff | +0.051 |
mixer | +0.051 |
Neg logit contributions | |
Along | -0.046 |
nights | -0.044 |
drifted | -0.043 |
way | -0.042 |
On | -0.041 |
Token | mom |
Token ID | 1995 |
Feature activation | +8.10e-02 |
Pos logit contributions | |
paramed | +0.207 |
wasteful | +0.203 |
liar | +0.202 |
boys | +0.202 |
roc | +0.201 |
Neg logit contributions | |
nights | -0.261 |
Along | -0.257 |
days | -0.238 |
whenever | -0.237 |
way | -0.237 |
Token | my |
Token ID | 1820 |
Feature activation | -5.28e-02 |
Pos logit contributions | |
liar | +0.148 |
paramed | +0.143 |
bluff | +0.141 |
wasteful | +0.140 |
mixer | +0.139 |
Neg logit contributions | |
Along | -0.109 |
nights | -0.107 |
way | -0.099 |
whenever | -0.097 |
drifted | -0.096 |
Token | picked |
Token ID | 6497 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
drifted | +0.074 |
Along | +0.074 |
nights | +0.069 |
whenever | +0.069 |
way | +0.067 |
Neg logit contributions | |
liar | -0.093 |
bluff | -0.091 |
wasteful | -0.089 |
roc | -0.089 |
chefs | -0.088 |
Token | up |
Token ID | 510 |
Feature activation | -4.64e-02 |
Pos logit contributions | |
Along | +0.021 |
whenever | +0.021 |
night | +0.020 |
nights | +0.019 |
On | +0.019 |
Neg logit contributions | |
liar | -0.037 |
paramed | -0.035 |
wasteful | -0.035 |
roc | -0.034 |
bluff | -0.033 |
Token | the |
Token ID | 262 |
Feature activation | +8.99e-03 |
Pos logit contributions | |
whenever | +0.048 |
Along | +0.043 |
days | +0.043 |
night | +0.042 |
nights | +0.039 |
Neg logit contributions | |
liar | -0.103 |
wasteful | -0.098 |
bluff | -0.097 |
disrespectful | -0.095 |
paramed | -0.095 |
Token | 's |
Token ID | 338 |
Feature activation | +1.51e-01 |
Pos logit contributions | |
Along | +0.085 |
nights | +0.084 |
way | +0.078 |
drifted | +0.078 |
On | +0.076 |
Neg logit contributions | |
paramed | -0.072 |
mixer | -0.070 |
liar | -0.070 |
wasteful | -0.068 |
roc | -0.067 |
Token | friend |
Token ID | 1545 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
paramed | +0.244 |
wasteful | +0.240 |
liar | +0.237 |
mixer | +0.235 |
roc | +0.233 |
Neg logit contributions | |
Along | -0.227 |
nights | -0.224 |
days | -0.201 |
way | -0.200 |
drifted | -0.199 |
Token | , |
Token ID | 11 |
Feature activation | +7.23e-02 |
Pos logit contributions | |
Along | +0.015 |
nights | +0.015 |
whenever | +0.015 |
night | +0.014 |
day | +0.014 |
Neg logit contributions | |
wasteful | -0.024 |
paramed | -0.024 |
bluff | -0.024 |
mixer | -0.023 |
liar | -0.023 |
Token | a |
Token ID | 257 |
Feature activation | -2.64e-02 |
Pos logit contributions | |
wasteful | +0.121 |
mixer | +0.121 |
paramed | +0.121 |
whining | +0.114 |
liar | +0.114 |
Neg logit contributions | |
Along | -0.101 |
whenever | -0.096 |
night | -0.094 |
On | -0.093 |
day | -0.092 |
Token | little |
Token ID | 1310 |
Feature activation | -7.49e-02 |
Pos logit contributions | |
day | +0.028 |
night | +0.025 |
Along | +0.024 |
journey | +0.022 |
evening | +0.022 |
Neg logit contributions | |
wasteful | -0.053 |
roc | -0.052 |
boys | -0.052 |
ution | -0.052 |
bluff | -0.051 |
Token | pig |
Token ID | 12967 |
Feature activation | +1.55e-01 |
Pos logit contributions | |
day | +0.099 |
night | +0.098 |
of | +0.098 |
Along | +0.093 |
nights | +0.091 |
Neg logit contributions | |
bluff | -0.132 |
wasteful | -0.130 |
liar | -0.122 |
disrespectful | -0.121 |
ige | -0.121 |
Token | named |
Token ID | 3706 |
Feature activation | -5.03e-02 |
Pos logit contributions | |
wasteful | +0.230 |
bluff | +0.227 |
paramed | +0.226 |
mixer | +0.222 |
liar | +0.220 |
Neg logit contributions | |
Along | -0.258 |
nights | -0.248 |
way | -0.236 |
whenever | -0.232 |
drifted | -0.231 |
Token | Pink |
Token ID | 14657 |
Feature activation | +8.21e-03 |
Pos logit contributions | |
Along | +0.078 |
nights | +0.071 |
On | +0.070 |
whenever | +0.069 |
way | +0.067 |
Neg logit contributions | |
paramed | -0.080 |
wasteful | -0.080 |
mixer | -0.079 |
liar | -0.079 |
mad | -0.077 |
Token | y |
Token ID | 88 |
Feature activation | +1.44e-03 |
Pos logit contributions | |
boys | +0.013 |
mean | +0.013 |
paramed | +0.012 |
mixer | +0.012 |
liar | +0.012 |
Neg logit contributions | |
Along | -0.013 |
nights | -0.013 |
way | -0.012 |
On | -0.011 |
drifted | -0.011 |
Token | , |
Token ID | 11 |
Feature activation | +7.14e-02 |
Pos logit contributions | |
paramed | +0.003 |
wasteful | +0.003 |
liar | +0.003 |
mixer | +0.003 |
bluff | +0.003 |
Neg logit contributions | |
Along | -0.002 |
nights | -0.002 |
way | -0.002 |
On | -0.002 |
whenever | -0.002 |
Token | came |
Token ID | 1625 |
Feature activation | +4.09e-02 |
Pos logit contributions | |
paramed | +0.108 |
mixer | +0.106 |
liar | +0.105 |
wasteful | +0.104 |
roc | +0.103 |
Neg logit contributions | |
Along | -0.117 |
nights | -0.111 |
On | -0.105 |
way | -0.104 |
drifted | -0.104 |
Token | . |
Token ID | 13 |
Feature activation | -2.24e-02 |
Pos logit contributions | |
mean | +0.037 |
mixer | +0.037 |
greedy | +0.036 |
stern | +0.036 |
liar | +0.036 |
Neg logit contributions | |
Along | -0.033 |
nights | -0.030 |
dark | -0.029 |
On | -0.028 |
drifted | -0.028 |
Token | Together |
Token ID | 17083 |
Feature activation | -5.30e-02 |
Pos logit contributions | |
Along | +0.030 |
nights | +0.028 |
On | +0.026 |
way | +0.026 |
whenever | +0.025 |
Neg logit contributions | |
paramed | -0.042 |
liar | -0.042 |
mixer | -0.041 |
mean | -0.040 |
boys | -0.040 |
Token | they |
Token ID | 484 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
Along | +0.084 |
nights | +0.083 |
way | +0.077 |
On | +0.074 |
drifted | +0.074 |
Neg logit contributions | |
paramed | -0.082 |
mixer | -0.080 |
liar | -0.079 |
mean | -0.079 |
boys | -0.078 |
Token | loaded |
Token ID | 9639 |
Feature activation | +8.69e-02 |
Pos logit contributions | |
Along | +0.029 |
nights | +0.028 |
way | +0.026 |
drifted | +0.026 |
On | +0.025 |
Neg logit contributions | |
paramed | -0.028 |
liar | -0.028 |
mixer | -0.027 |
wasteful | -0.027 |
bluff | -0.027 |
Token | the |
Token ID | 262 |
Feature activation | +1.06e-02 |
Pos logit contributions | |
liar | +0.170 |
paramed | +0.167 |
wasteful | +0.154 |
mean | +0.152 |
greedy | +0.151 |
Neg logit contributions | |
Along | -0.117 |
nights | -0.115 |
whenever | -0.108 |
way | -0.104 |
days | -0.104 |
Token | truck |
Token ID | 7779 |
Feature activation | +1.61e-01 |
Pos logit contributions | |
liar | +0.011 |
boys | +0.011 |
paramed | +0.011 |
wasteful | +0.010 |
bluff | +0.010 |
Neg logit contributions | |
Along | -0.023 |
nights | -0.022 |
way | -0.021 |
days | -0.021 |
day | -0.020 |
Token | with |
Token ID | 351 |
Feature activation | -2.59e-02 |
Pos logit contributions | |
boys | +0.242 |
mixer | +0.241 |
phone | +0.239 |
paramed | +0.235 |
stern | +0.227 |
Neg logit contributions | |
nights | -0.255 |
Along | -0.243 |
drifted | -0.239 |
dark | -0.231 |
outer | -0.228 |
Token | the |
Token ID | 262 |
Feature activation | +1.03e-02 |
Pos logit contributions | |
nights | +0.040 |
Along | +0.038 |
days | +0.036 |
night | +0.036 |
whenever | +0.035 |
Neg logit contributions | |
paramed | -0.043 |
liar | -0.043 |
wasteful | -0.041 |
roc | -0.041 |
bluff | -0.040 |
Token | boxes |
Token ID | 10559 |
Feature activation | +7.32e-02 |
Pos logit contributions | |
boys | +0.012 |
paramed | +0.012 |
liar | +0.012 |
wasteful | +0.012 |
bluff | +0.012 |
Neg logit contributions | |
Along | -0.020 |
nights | -0.020 |
way | -0.019 |
days | -0.018 |
night | -0.018 |
Token | and |
Token ID | 290 |
Feature activation | -2.17e-02 |
Pos logit contributions | |
mixer | +0.114 |
paramed | +0.110 |
stern | +0.109 |
phone | +0.108 |
boys | +0.108 |
Neg logit contributions | |
nights | -0.113 |
Along | -0.112 |
dark | -0.102 |
drifted | -0.101 |
way | -0.100 |
Token | Pete |
Token ID | 17542 |
Feature activation | +3.21e-02 |
Pos logit contributions | |
Along | +0.034 |
nights | +0.034 |
way | +0.031 |
drifted | +0.030 |
On | +0.030 |
Neg logit contributions | |
paramed | -0.034 |
liar | -0.034 |
mixer | -0.033 |
mean | -0.033 |
wasteful | -0.032 |
Token | ," |
Token ID | 553 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
paramed | +0.056 |
liar | +0.054 |
mixer | +0.054 |
wasteful | +0.053 |
mean | +0.053 |
Neg logit contributions | |
Along | -0.057 |
nights | -0.057 |
way | -0.053 |
drifted | -0.051 |
On | -0.050 |
Token | Max |
Token ID | 5436 |
Feature activation | +8.97e-02 |
Pos logit contributions | |
liar | +0.203 |
chefs | +0.202 |
paramed | +0.200 |
wasteful | +0.198 |
roc | +0.196 |
Neg logit contributions | |
Along | -0.196 |
drifted | -0.193 |
whenever | -0.192 |
night | -0.192 |
nights | -0.189 |
Token | said |
Token ID | 531 |
Feature activation | +6.29e-02 |
Pos logit contributions | |
wasteful | +0.164 |
mixer | +0.164 |
liar | +0.162 |
chefs | +0.160 |
paramed | +0.157 |
Neg logit contributions | |
Along | -0.114 |
As | -0.113 |
On | -0.108 |
whenever | -0.108 |
drifted | -0.107 |
Token | , |
Token ID | 11 |
Feature activation | +8.62e-02 |
Pos logit contributions | |
wasteful | +0.125 |
roc | +0.125 |
paramed | +0.124 |
mixer | +0.124 |
liar | +0.123 |
Neg logit contributions | |
Along | -0.069 |
nights | -0.066 |
whenever | -0.065 |
way | -0.064 |
drifted | -0.060 |
Token | feeling |
Token ID | 4203 |
Feature activation | -4.28e-02 |
Pos logit contributions | |
roc | +0.130 |
liar | +0.122 |
wasteful | +0.121 |
paramed | +0.118 |
mixer | +0.117 |
Neg logit contributions | |
Along | -0.150 |
drifted | -0.138 |
nights | -0.136 |
whenever | -0.134 |
On | -0.133 |
Token | sorry |
Token ID | 7926 |
Feature activation | +1.96e-01 |
Pos logit contributions | |
nights | +0.083 |
whenever | +0.083 |
drifted | +0.082 |
night | +0.079 |
Along | +0.078 |
Neg logit contributions | |
mixer | -0.049 |
roc | -0.048 |
paramed | -0.047 |
boys | -0.047 |
Lee | -0.042 |
Token | . |
Token ID | 13 |
Feature activation | -1.84e-02 |
Pos logit contributions | |
mean | +0.317 |
mad | +0.296 |
stern | +0.292 |
sorry | +0.291 |
phone | +0.289 |
Neg logit contributions | |
Along | -0.334 |
nights | -0.319 |
way | -0.293 |
outer | -0.289 |
growth | -0.288 |
Token |
|
Token ID | 198 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
nights | +0.024 |
Along | +0.023 |
way | +0.021 |
drifted | +0.020 |
ingle | +0.020 |
Neg logit contributions | |
paramed | -0.035 |
liar | -0.034 |
mean | -0.033 |
mixer | -0.033 |
greedy | -0.033 |
Token |
|
Token ID | 198 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
nights | +0.022 |
Along | +0.021 |
drifted | +0.019 |
way | +0.019 |
filled | +0.019 |
Neg logit contributions | |
paramed | -0.020 |
stern | -0.020 |
mean | -0.020 |
mad | -0.020 |
mixer | -0.019 |
Token | The |
Token ID | 464 |
Feature activation | -4.16e-02 |
Pos logit contributions | |
Along | +0.020 |
nights | +0.020 |
way | +0.018 |
drifted | +0.018 |
On | +0.018 |
Neg logit contributions | |
paramed | -0.020 |
liar | -0.019 |
mixer | -0.019 |
mean | -0.019 |
boys | -0.019 |
Token | dog |
Token ID | 3290 |
Feature activation | +1.24e-01 |
Pos logit contributions | |
Along | +0.067 |
nights | +0.067 |
way | +0.062 |
drifted | +0.060 |
days | +0.060 |
Neg logit contributions | |
paramed | -0.063 |
mixer | -0.060 |
liar | -0.060 |
boys | -0.060 |
wasteful | -0.059 |
Token | new |
Token ID | 649 |
Feature activation | -1.17e-02 |
Pos logit contributions | |
paramed | +0.019 |
liar | +0.019 |
roc | +0.018 |
boys | +0.018 |
mixer | +0.018 |
Neg logit contributions | |
way | -0.010 |
nights | -0.010 |
Along | -0.009 |
whenever | -0.009 |
day | -0.009 |
Token | and |
Token ID | 290 |
Feature activation | -9.42e-03 |
Pos logit contributions | |
Along | +0.015 |
nights | +0.014 |
way | +0.013 |
drifted | +0.013 |
On | +0.012 |
Neg logit contributions | |
paramed | -0.022 |
mixer | -0.022 |
liar | -0.022 |
mean | -0.022 |
boys | -0.021 |
Token | had |
Token ID | 550 |
Feature activation | -5.59e-02 |
Pos logit contributions | |
nights | +0.015 |
Along | +0.014 |
whenever | +0.013 |
days | +0.013 |
way | +0.013 |
Neg logit contributions | |
paramed | -0.016 |
mixer | -0.015 |
liar | -0.015 |
roc | -0.014 |
boys | -0.014 |
Token | fun |
Token ID | 1257 |
Feature activation | +9.71e-04 |
Pos logit contributions | |
nights | +0.083 |
Along | +0.077 |
days | +0.074 |
drifted | +0.073 |
way | +0.072 |
Neg logit contributions | |
paramed | -0.099 |
mixer | -0.094 |
roc | -0.091 |
liar | -0.091 |
wasteful | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | -1.44e-02 |
Pos logit contributions | |
mean | +0.002 |
mixer | +0.002 |
phone | +0.002 |
mad | +0.002 |
chefs | +0.001 |
Neg logit contributions | |
Along | -0.002 |
nights | -0.001 |
way | -0.001 |
On | -0.001 |
ingle | -0.001 |
Token | They |
Token ID | 1119 |
Feature activation | +4.71e-02 |
Pos logit contributions | |
nights | +0.024 |
Along | +0.023 |
way | +0.022 |
ingle | +0.021 |
drifted | +0.021 |
Neg logit contributions | |
paramed | -0.022 |
greedy | -0.021 |
mixer | -0.021 |
liar | -0.020 |
stern | -0.020 |
Token | looked |
Token ID | 3114 |
Feature activation | -1.11e-02 |
Pos logit contributions | |
paramed | +0.074 |
mixer | +0.073 |
liar | +0.071 |
roc | +0.071 |
wasteful | +0.070 |
Neg logit contributions | |
Along | -0.074 |
nights | -0.071 |
drifted | -0.070 |
On | -0.067 |
whenever | -0.066 |
Token | at |
Token ID | 379 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
Along | +0.016 |
nights | +0.015 |
way | +0.014 |
whenever | +0.014 |
On | +0.014 |
Neg logit contributions | |
paramed | -0.020 |
liar | -0.019 |
mixer | -0.019 |
boys | -0.018 |
roc | -0.018 |
Token | each |
Token ID | 1123 |
Feature activation | +8.13e-02 |
Pos logit contributions | |
paramed | +0.013 |
liar | +0.013 |
mixer | +0.012 |
boys | +0.012 |
wasteful | +0.012 |
Neg logit contributions | |
nights | -0.019 |
Along | -0.019 |
whenever | -0.018 |
night | -0.017 |
way | -0.017 |
Token | other |
Token ID | 584 |
Feature activation | -8.40e-04 |
Pos logit contributions | |
mean | +0.111 |
mixer | +0.110 |
paramed | +0.108 |
phone | +0.108 |
liar | +0.108 |
Neg logit contributions | |
Along | -0.144 |
nights | -0.141 |
way | -0.133 |
drifted | -0.131 |
On | -0.129 |
Token | and |
Token ID | 290 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
Along | +0.001 |
nights | +0.001 |
way | +0.001 |
drifted | +0.001 |
On | +0.001 |
Neg logit contributions | |
paramed | -0.002 |
liar | -0.002 |
mixer | -0.001 |
wasteful | -0.001 |
mean | -0.001 |
Token | laugh |
Token ID | 6487 |
Feature activation | +8.83e-02 |
Pos logit contributions | |
Along | +0.016 |
nights | +0.016 |
way | +0.015 |
drifted | +0.015 |
On | +0.015 |
Neg logit contributions | |
paramed | -0.015 |
mixer | -0.015 |
liar | -0.015 |
mean | -0.014 |
boys | -0.014 |
Token | still |
Token ID | 991 |
Feature activation | +4.53e-03 |
Pos logit contributions | |
paramed | +0.158 |
mixer | +0.154 |
liar | +0.154 |
mean | +0.152 |
boys | +0.151 |
Neg logit contributions | |
Along | -0.120 |
nights | -0.117 |
way | -0.108 |
On | -0.103 |
drifted | -0.103 |
Token | rang |
Token ID | 28077 |
Feature activation | -6.51e-02 |
Pos logit contributions | |
paramed | +0.008 |
mixer | +0.007 |
liar | +0.007 |
mean | +0.007 |
boys | +0.007 |
Neg logit contributions | |
Along | -0.007 |
nights | -0.007 |
way | -0.006 |
On | -0.006 |
drifted | -0.006 |
Token | in |
Token ID | 287 |
Feature activation | -5.74e-02 |
Pos logit contributions | |
Along | +0.086 |
nights | +0.086 |
way | +0.078 |
whenever | +0.075 |
drifted | +0.075 |
Neg logit contributions | |
paramed | -0.118 |
liar | -0.115 |
mixer | -0.115 |
wasteful | -0.113 |
boys | -0.112 |
Token | his |
Token ID | 465 |
Feature activation | +1.50e-02 |
Pos logit contributions | |
Along | +0.096 |
nights | +0.094 |
way | +0.088 |
On | +0.085 |
drifted | +0.085 |
Neg logit contributions | |
paramed | -0.086 |
mixer | -0.083 |
liar | -0.083 |
boys | -0.081 |
wasteful | -0.081 |
Token | ears |
Token ID | 11368 |
Feature activation | +4.85e-02 |
Pos logit contributions | |
mean | +0.021 |
paramed | +0.021 |
mixer | +0.021 |
phone | +0.021 |
greedy | +0.021 |
Neg logit contributions | |
Along | -0.026 |
nights | -0.025 |
way | -0.025 |
On | -0.024 |
ingle | -0.023 |
Token | as |
Token ID | 355 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
paramed | +0.083 |
mixer | +0.082 |
mean | +0.081 |
liar | +0.081 |
boys | +0.080 |
Neg logit contributions | |
Along | -0.068 |
nights | -0.067 |
way | -0.061 |
On | -0.059 |
drifted | -0.059 |
Token | he |
Token ID | 339 |
Feature activation | -8.84e-02 |
Pos logit contributions | |
Along | +0.284 |
nights | +0.264 |
way | +0.251 |
On | +0.243 |
ingle | +0.239 |
Neg logit contributions | |
mean | -0.298 |
liar | -0.295 |
greedy | -0.290 |
mixer | -0.286 |
stern | -0.283 |
Token | vanished |
Token ID | 23717 |
Feature activation | -4.47e-02 |
Pos logit contributions | |
Along | +0.154 |
nights | +0.151 |
way | +0.144 |
On | +0.137 |
dark | +0.134 |
Neg logit contributions | |
boys | -0.122 |
mean | -0.122 |
greedy | -0.119 |
liar | -0.118 |
paramed | -0.117 |
Token | into |
Token ID | 656 |
Feature activation | -6.85e-02 |
Pos logit contributions | |
Along | +0.064 |
nights | +0.060 |
On | +0.055 |
way | +0.055 |
dark | +0.054 |
Neg logit contributions | |
mean | -0.077 |
boys | -0.075 |
mad | -0.075 |
phone | -0.075 |
stern | -0.075 |
Token | the |
Token ID | 262 |
Feature activation | +1.74e-02 |
Pos logit contributions | |
Along | +0.132 |
nights | +0.123 |
way | +0.122 |
On | +0.117 |
dark | +0.116 |
Neg logit contributions | |
mean | -0.085 |
liar | -0.084 |
paramed | -0.084 |
mixer | -0.083 |
mad | -0.082 |
Token | they |
Token ID | 484 |
Feature activation | -1.97e-02 |
Pos logit contributions | |
paramed | +0.133 |
mixer | +0.125 |
liar | +0.125 |
roc | +0.123 |
boys | +0.120 |
Neg logit contributions | |
Along | -0.103 |
whenever | -0.096 |
nights | -0.096 |
way | -0.095 |
On | -0.091 |
Token | saw |
Token ID | 2497 |
Feature activation | +3.68e-02 |
Pos logit contributions | |
Along | +0.027 |
drifted | +0.026 |
nights | +0.026 |
whenever | +0.025 |
night | +0.025 |
Neg logit contributions | |
paramed | -0.036 |
bluff | -0.033 |
liar | -0.033 |
mixer | -0.033 |
roc | -0.032 |
Token | a |
Token ID | 257 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
paramed | +0.077 |
liar | +0.072 |
roc | +0.071 |
mixer | +0.071 |
whining | +0.071 |
Neg logit contributions | |
nights | -0.037 |
whenever | -0.037 |
way | -0.036 |
days | -0.036 |
night | -0.035 |
Token | sign |
Token ID | 1051 |
Feature activation | +7.74e-02 |
Pos logit contributions | |
day | +0.030 |
way | +0.028 |
night | +0.028 |
Along | +0.027 |
journey | +0.026 |
Neg logit contributions | |
boys | -0.051 |
paramed | -0.049 |
eper | -0.047 |
chefs | -0.047 |
bluff | -0.046 |
Token | that |
Token ID | 326 |
Feature activation | -3.65e-02 |
Pos logit contributions | |
paramed | +0.144 |
roc | +0.143 |
liar | +0.142 |
wasteful | +0.138 |
boys | +0.137 |
Neg logit contributions | |
Along | -0.098 |
nights | -0.094 |
way | -0.092 |
On | -0.086 |
whenever | -0.084 |
Token | said |
Token ID | 531 |
Feature activation | +5.28e-02 |
Pos logit contributions | |
nights | +0.058 |
Along | +0.057 |
way | +0.056 |
On | +0.054 |
night | +0.054 |
Neg logit contributions | |
paramed | -0.057 |
liar | -0.055 |
roc | -0.053 |
mixer | -0.053 |
wasteful | -0.051 |
Token | , |
Token ID | 11 |
Feature activation | +7.18e-02 |
Pos logit contributions | |
paramed | +0.082 |
roc | +0.081 |
wasteful | +0.079 |
boys | +0.078 |
whining | +0.078 |
Neg logit contributions | |
Along | -0.079 |
whenever | -0.076 |
way | -0.076 |
On | -0.074 |
before | -0.073 |
Token | " |
Token ID | 366 |
Feature activation | +2.00e-01 |
Pos logit contributions | |
roc | +0.129 |
liar | +0.125 |
boys | +0.117 |
wasteful | +0.117 |
paramed | +0.115 |
Neg logit contributions | |
Along | -0.103 |
On | -0.098 |
As | -0.095 |
before | -0.092 |
whenever | -0.092 |
Token | Do |
Token ID | 5211 |
Feature activation | +3.56e-02 |
Pos logit contributions | |
paramed | +0.294 |
mixer | +0.285 |
liar | +0.284 |
mean | +0.280 |
boys | +0.279 |
Neg logit contributions | |
Along | -0.337 |
nights | -0.332 |
way | -0.310 |
drifted | -0.300 |
On | -0.300 |
Token | not |
Token ID | 407 |
Feature activation | +6.35e-02 |
Pos logit contributions | |
paramed | +0.036 |
mixer | +0.034 |
liar | +0.034 |
boys | +0.033 |
wasteful | +0.033 |
Neg logit contributions | |
Along | -0.077 |
nights | -0.076 |
way | -0.072 |
On | -0.070 |
drifted | -0.070 |
Token | eat |
Token ID | 4483 |
Feature activation | -3.83e-02 |
Pos logit contributions | |
paramed | +0.101 |
mixer | +0.098 |
liar | +0.098 |
boys | +0.095 |
wasteful | +0.095 |
Neg logit contributions | |
Along | -0.100 |
nights | -0.097 |
way | -0.091 |
On | -0.088 |
drifted | -0.087 |
Token | wonderful |
Token ID | 7932 |
Feature activation | -4.28e-02 |
Pos logit contributions | |
Along | +0.048 |
nights | +0.045 |
way | +0.043 |
On | +0.042 |
drifted | +0.042 |
Neg logit contributions | |
mixer | -0.045 |
paramed | -0.045 |
mean | -0.044 |
boys | -0.043 |
greedy | -0.043 |
Token | surprise |
Token ID | 5975 |
Feature activation | +3.88e-02 |
Pos logit contributions | |
Along | +0.076 |
nights | +0.073 |
way | +0.069 |
On | +0.067 |
drifted | +0.067 |
Neg logit contributions | |
paramed | -0.060 |
mixer | -0.059 |
liar | -0.057 |
mean | -0.057 |
boys | -0.056 |
Token | � |
Token ID | 6184 |
Feature activation | -7.85e-03 |
Pos logit contributions | |
paramed | +0.070 |
mixer | +0.070 |
mean | +0.068 |
liar | +0.068 |
boys | +0.068 |
Neg logit contributions | |
Along | -0.052 |
nights | -0.050 |
way | -0.046 |
drifted | -0.044 |
On | -0.044 |
Token | � |
Token ID | 95 |
Feature activation | +6.06e-02 |
Pos logit contributions | |
nights | +0.012 |
Along | +0.011 |
drifted | +0.011 |
whenever | +0.011 |
way | +0.011 |
Neg logit contributions | |
paramed | -0.013 |
mixer | -0.012 |
boys | -0.012 |
liar | -0.012 |
phone | -0.011 |
Token | € |
Token ID | 26391 |
Feature activation | +7.72e-02 |
Pos logit contributions | |
paramed | +0.096 |
mixer | +0.093 |
liar | +0.092 |
boys | +0.091 |
mean | +0.091 |
Neg logit contributions | |
nights | -0.094 |
Along | -0.093 |
way | -0.085 |
drifted | -0.084 |
whenever | -0.083 |
Token | � |
Token ID | 447 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
paramed | +0.115 |
boys | +0.112 |
mixer | +0.112 |
liar | +0.110 |
mean | +0.109 |
Neg logit contributions | |
nights | -0.126 |
Along | -0.125 |
way | -0.116 |
drifted | -0.112 |
On | -0.111 |
Token | � |
Token ID | 250 |
Feature activation | +1.62e-01 |
Pos logit contributions | |
paramed | +0.112 |
liar | +0.109 |
mixer | +0.108 |
wasteful | +0.107 |
mean | +0.107 |
Neg logit contributions | |
Along | -0.086 |
nights | -0.082 |
way | -0.076 |
On | -0.074 |
drifted | -0.073 |
Token | a |
Token ID | 257 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
paramed | +0.215 |
mixer | +0.195 |
greedy | +0.188 |
mean | +0.187 |
liar | +0.186 |
Neg logit contributions | |
nights | -0.292 |
Along | -0.277 |
way | -0.270 |
drifted | -0.267 |
lights | -0.258 |
Token | drawing |
Token ID | 8263 |
Feature activation | -3.12e-02 |
Pos logit contributions | |
Along | +0.029 |
nights | +0.029 |
night | +0.027 |
day | +0.027 |
way | +0.026 |
Neg logit contributions | |
wasteful | -0.032 |
roc | -0.031 |
boys | -0.031 |
paramed | -0.031 |
liar | -0.030 |
Token | from |
Token ID | 422 |
Feature activation | -2.91e-02 |
Pos logit contributions | |
Along | +0.043 |
nights | +0.042 |
way | +0.039 |
On | +0.037 |
drifted | +0.037 |
Neg logit contributions | |
paramed | -0.055 |
mixer | -0.054 |
liar | -0.054 |
mean | -0.053 |
boys | -0.052 |
Token | a |
Token ID | 257 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
Along | +0.054 |
nights | +0.053 |
way | +0.050 |
drifted | +0.049 |
On | +0.048 |
Neg logit contributions | |
paramed | -0.037 |
mixer | -0.037 |
liar | -0.036 |
mean | -0.036 |
boys | -0.035 |
Token | Lucy |
Token ID | 22162 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
mixer | +0.106 |
paramed | +0.104 |
mean | +0.101 |
boys | +0.100 |
chefs | +0.100 |
Neg logit contributions | |
Along | -0.111 |
nights | -0.110 |
way | -0.102 |
drifted | -0.102 |
ingle | -0.100 |
Token | saw |
Token ID | 2497 |
Feature activation | +3.46e-02 |
Pos logit contributions | |
Along | +0.031 |
nights | +0.030 |
way | +0.028 |
On | +0.028 |
drifted | +0.028 |
Neg logit contributions | |
paramed | -0.023 |
liar | -0.022 |
mixer | -0.022 |
wasteful | -0.022 |
mean | -0.021 |
Token | a |
Token ID | 257 |
Feature activation | -2.80e-02 |
Pos logit contributions | |
liar | +0.070 |
roc | +0.070 |
boys | +0.069 |
whining | +0.068 |
bluff | +0.068 |
Neg logit contributions | |
nights | -0.035 |
whenever | -0.034 |
days | -0.034 |
Along | -0.033 |
night | -0.032 |
Token | big |
Token ID | 1263 |
Feature activation | -3.65e-03 |
Pos logit contributions | |
day | +0.029 |
night | +0.026 |
journey | +0.025 |
nights | +0.025 |
path | +0.025 |
Neg logit contributions | |
boys | -0.057 |
roc | -0.056 |
wasteful | -0.055 |
ishly | -0.053 |
paramed | -0.053 |
Token | red |
Token ID | 2266 |
Feature activation | +5.80e-02 |
Pos logit contributions | |
day | +0.005 |
night | +0.005 |
journey | +0.005 |
view | +0.004 |
days | +0.004 |
Neg logit contributions | |
roc | -0.006 |
Fighting | -0.006 |
bluff | -0.006 |
wasteful | -0.006 |
boys | -0.006 |
Token | ball |
Token ID | 2613 |
Feature activation | +5.86e-02 |
Pos logit contributions | |
wasteful | +0.079 |
bluff | +0.077 |
boys | +0.076 |
paramed | +0.075 |
liar | +0.075 |
Neg logit contributions | |
Along | -0.104 |
nights | -0.097 |
way | -0.097 |
night | -0.093 |
On | -0.093 |
Token | . |
Token ID | 13 |
Feature activation | -3.47e-02 |
Pos logit contributions | |
mixer | +0.104 |
mean | +0.102 |
paramed | +0.099 |
phone | +0.099 |
greedy | +0.097 |
Neg logit contributions | |
nights | -0.086 |
Along | -0.081 |
drifted | -0.073 |
more | -0.070 |
way | -0.070 |
Token | She |
Token ID | 1375 |
Feature activation | -6.65e-02 |
Pos logit contributions | |
Along | +0.049 |
nights | +0.049 |
way | +0.045 |
drifted | +0.043 |
more | +0.042 |
Neg logit contributions | |
paramed | -0.060 |
mixer | -0.059 |
liar | -0.058 |
mean | -0.057 |
boys | -0.057 |
Token | wanted |
Token ID | 2227 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
Along | +0.105 |
nights | +0.100 |
way | +0.094 |
On | +0.093 |
drifted | +0.092 |
Neg logit contributions | |
paramed | -0.108 |
liar | -0.106 |
wasteful | -0.104 |
mixer | -0.102 |
bluff | -0.102 |
Token | to |
Token ID | 284 |
Feature activation | +7.63e-02 |
Pos logit contributions | |
Along | +0.020 |
nights | +0.019 |
whenever | +0.019 |
days | +0.018 |
night | +0.018 |
Neg logit contributions | |
paramed | -0.045 |
liar | -0.045 |
roc | -0.044 |
wasteful | -0.043 |
bluff | -0.043 |
Token | play |
Token ID | 711 |
Feature activation | -1.11e-01 |
Pos logit contributions | |
paramed | +0.124 |
mixer | +0.122 |
mean | +0.120 |
liar | +0.119 |
boys | +0.117 |
Neg logit contributions | |
nights | -0.118 |
Along | -0.115 |
way | -0.107 |
drifted | -0.104 |
more | -0.100 |
Token | . |
Token ID | 13 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
Along | +0.049 |
nights | +0.048 |
way | +0.043 |
whenever | +0.040 |
On | +0.040 |
Neg logit contributions | |
paramed | -0.103 |
liar | -0.100 |
mixer | -0.100 |
wasteful | -0.099 |
mean | -0.098 |
Token | We |
Token ID | 775 |
Feature activation | +2.03e-02 |
Pos logit contributions | |
nights | +0.044 |
Along | +0.041 |
drifted | +0.040 |
way | +0.039 |
ingle | +0.038 |
Neg logit contributions | |
paramed | -0.041 |
boys | -0.040 |
mixer | -0.040 |
greedy | -0.039 |
roc | -0.039 |
Token | can |
Token ID | 460 |
Feature activation | +1.26e-01 |
Pos logit contributions | |
paramed | +0.029 |
liar | +0.028 |
mixer | +0.028 |
wasteful | +0.027 |
bluff | +0.027 |
Neg logit contributions | |
Along | -0.036 |
nights | -0.034 |
way | -0.032 |
On | -0.032 |
whenever | -0.032 |
Token | restore |
Token ID | 11169 |
Feature activation | -3.17e-02 |
Pos logit contributions | |
paramed | +0.172 |
mixer | +0.169 |
boys | +0.167 |
mean | +0.165 |
liar | +0.162 |
Neg logit contributions | |
nights | -0.227 |
Along | -0.223 |
way | -0.208 |
drifted | -0.203 |
days | -0.199 |
Token | it |
Token ID | 340 |
Feature activation | +3.25e-02 |
Pos logit contributions | |
nights | +0.044 |
whenever | +0.042 |
Along | +0.040 |
night | +0.040 |
way | +0.040 |
Neg logit contributions | |
paramed | -0.060 |
liar | -0.057 |
mixer | -0.053 |
whining | -0.052 |
wasteful | -0.052 |
Token | and |
Token ID | 290 |
Feature activation | -2.45e-02 |
Pos logit contributions | |
boys | +0.047 |
mean | +0.044 |
phone | +0.044 |
chefs | +0.042 |
mixer | +0.041 |
Neg logit contributions | |
Along | -0.057 |
nights | -0.055 |
drifted | -0.054 |
way | -0.051 |
On | -0.050 |
Token | make |
Token ID | 787 |
Feature activation | +8.26e-02 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.044 |
way | +0.041 |
drifted | +0.040 |
On | +0.039 |
Neg logit contributions | |
mean | -0.033 |
boys | -0.032 |
paramed | -0.032 |
mixer | -0.032 |
phone | -0.031 |
Token | it |
Token ID | 340 |
Feature activation | +3.21e-02 |
Pos logit contributions | |
paramed | +0.133 |
liar | +0.129 |
mixer | +0.126 |
wasteful | +0.123 |
greedy | +0.121 |
Neg logit contributions | |
nights | -0.130 |
Along | -0.126 |
way | -0.119 |
whenever | -0.117 |
days | -0.114 |
Token | look |
Token ID | 804 |
Feature activation | +3.57e-02 |
Pos logit contributions | |
mean | +0.040 |
boys | +0.036 |
mad | +0.035 |
phone | +0.035 |
sorry | +0.034 |
Neg logit contributions | |
Along | -0.068 |
nights | -0.065 |
way | -0.062 |
growth | -0.061 |
On | -0.060 |
Token | like |
Token ID | 588 |
Feature activation | -5.87e-03 |
Pos logit contributions | |
mean | +0.047 |
greedy | +0.044 |
mad | +0.043 |
wasteful | +0.043 |
boys | +0.042 |
Neg logit contributions | |
Along | -0.069 |
nights | -0.066 |
way | -0.063 |
drifted | -0.062 |
On | -0.061 |
Token | new |
Token ID | 649 |
Feature activation | -1.76e-02 |
Pos logit contributions | |
Along | +0.011 |
nights | +0.011 |
way | +0.010 |
ingle | +0.010 |
On | +0.010 |
Neg logit contributions | |
mean | -0.007 |
mixer | -0.007 |
liar | -0.007 |
greedy | -0.007 |
boys | -0.007 |
Token | family |
Token ID | 1641 |
Feature activation | -4.49e-02 |
Pos logit contributions | |
Along | +0.103 |
nights | +0.099 |
way | +0.093 |
On | +0.091 |
drifted | +0.089 |
Neg logit contributions | |
paramed | -0.108 |
mixer | -0.106 |
liar | -0.105 |
mean | -0.104 |
boys | -0.104 |
Token | and |
Token ID | 290 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
nights | +0.070 |
Along | +0.070 |
way | +0.064 |
drifted | +0.063 |
whenever | +0.062 |
Neg logit contributions | |
paramed | -0.072 |
liar | -0.070 |
mixer | -0.069 |
wasteful | -0.069 |
mean | -0.068 |
Token | garden |
Token ID | 11376 |
Feature activation | -8.99e-02 |
Pos logit contributions | |
nights | +0.032 |
Along | +0.031 |
way | +0.028 |
whenever | +0.028 |
days | +0.028 |
Neg logit contributions | |
paramed | -0.033 |
mixer | -0.032 |
liar | -0.032 |
wasteful | -0.031 |
roc | -0.030 |
Token | heard |
Token ID | 2982 |
Feature activation | +3.34e-02 |
Pos logit contributions | |
nights | +0.127 |
Along | +0.122 |
way | +0.116 |
whenever | +0.114 |
drifted | +0.112 |
Neg logit contributions | |
paramed | -0.160 |
liar | -0.156 |
wasteful | -0.152 |
mixer | -0.151 |
bluff | -0.150 |
Token | her |
Token ID | 607 |
Feature activation | -6.70e-02 |
Pos logit contributions | |
paramed | +0.047 |
mixer | +0.045 |
liar | +0.045 |
mean | +0.045 |
boys | +0.044 |
Neg logit contributions | |
Along | -0.059 |
nights | -0.057 |
way | -0.054 |
drifted | -0.052 |
On | -0.052 |
Token | beautiful |
Token ID | 4950 |
Feature activation | -3.12e-02 |
Pos logit contributions | |
Along | +0.116 |
nights | +0.115 |
way | +0.106 |
drifted | +0.103 |
On | +0.103 |
Neg logit contributions | |
paramed | -0.094 |
boys | -0.093 |
mean | -0.093 |
mad | -0.092 |
phone | -0.092 |
Token | singing |
Token ID | 13777 |
Feature activation | -8.98e-02 |
Pos logit contributions | |
Along | +0.052 |
nights | +0.049 |
way | +0.046 |
On | +0.046 |
drifted | +0.045 |
Neg logit contributions | |
paramed | -0.047 |
mixer | -0.047 |
mean | -0.046 |
phone | -0.045 |
boys | -0.045 |
Token | . |
Token ID | 13 |
Feature activation | -2.79e-02 |
Pos logit contributions | |
Along | +0.119 |
nights | +0.114 |
way | +0.105 |
On | +0.102 |
drifted | +0.100 |
Neg logit contributions | |
paramed | -0.163 |
mixer | -0.160 |
mean | -0.160 |
boys | -0.158 |
liar | -0.158 |
Token | |
Token ID | 220 |
Feature activation | +4.55e-02 |
Pos logit contributions | |
nights | +0.040 |
Along | +0.037 |
drifted | +0.035 |
way | +0.035 |
ingle | +0.034 |
Neg logit contributions | |
paramed | -0.046 |
mixer | -0.046 |
stern | -0.045 |
mean | -0.045 |
liar | -0.045 |
Token |
|
Token ID | 198 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
paramed | +0.077 |
boys | +0.077 |
mixer | +0.076 |
stern | +0.074 |
mean | +0.074 |
Neg logit contributions | |
nights | -0.066 |
Along | -0.058 |
drifted | -0.057 |
days | -0.055 |
way | -0.055 |
Token |
|
Token ID | 198 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
nights | +0.031 |
Along | +0.029 |
drifted | +0.027 |
way | +0.027 |
ingle | +0.026 |
Neg logit contributions | |
mean | -0.039 |
paramed | -0.039 |
stern | -0.039 |
mixer | -0.039 |
liar | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | -3.22e-02 |
Pos logit contributions | |
Along | +0.069 |
nights | +0.067 |
drifted | +0.061 |
On | +0.060 |
dark | +0.059 |
Neg logit contributions | |
boys | -0.095 |
phone | -0.091 |
mixer | -0.091 |
paramed | -0.091 |
mean | -0.089 |
Token | The |
Token ID | 383 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
nights | +0.049 |
Along | +0.046 |
drifted | +0.044 |
way | +0.044 |
more | +0.043 |
Neg logit contributions | |
paramed | -0.054 |
mixer | -0.051 |
greedy | -0.050 |
liar | -0.050 |
mean | -0.050 |
Token | ghost |
Token ID | 10905 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
Along | +0.023 |
nights | +0.021 |
way | +0.021 |
On | +0.020 |
dark | +0.020 |
Neg logit contributions | |
paramed | -0.017 |
mean | -0.017 |
mixer | -0.017 |
liar | -0.017 |
phone | -0.017 |
Token | was |
Token ID | 373 |
Feature activation | +2.16e-02 |
Pos logit contributions | |
Along | +0.036 |
nights | +0.035 |
way | +0.032 |
drifted | +0.031 |
On | +0.031 |
Neg logit contributions | |
paramed | -0.037 |
liar | -0.036 |
mixer | -0.036 |
wasteful | -0.035 |
mean | -0.035 |
Token | curious |
Token ID | 11040 |
Feature activation | +8.04e-02 |
Pos logit contributions | |
roc | +0.036 |
liar | +0.035 |
mixer | +0.035 |
paramed | +0.034 |
boys | +0.033 |
Neg logit contributions | |
Along | -0.033 |
nights | -0.033 |
night | -0.030 |
day | -0.030 |
of | -0.029 |
Token | and |
Token ID | 290 |
Feature activation | -2.76e-02 |
Pos logit contributions | |
paramed | +0.140 |
mixer | +0.137 |
liar | +0.136 |
mean | +0.134 |
boys | +0.134 |
Neg logit contributions | |
Along | -0.114 |
nights | -0.111 |
way | -0.102 |
drifted | -0.098 |
On | -0.098 |
Token | decided |
Token ID | 3066 |
Feature activation | -3.67e-02 |
Pos logit contributions | |
Along | +0.044 |
nights | +0.044 |
way | +0.040 |
drifted | +0.039 |
On | +0.039 |
Neg logit contributions | |
paramed | -0.044 |
mixer | -0.042 |
liar | -0.042 |
boys | -0.041 |
wasteful | -0.041 |
Token | to |
Token ID | 284 |
Feature activation | +7.64e-02 |
Pos logit contributions | |
Along | +0.048 |
nights | +0.046 |
way | +0.045 |
whenever | +0.043 |
more | +0.041 |
Neg logit contributions | |
paramed | -0.069 |
mixer | -0.068 |
liar | -0.067 |
wasteful | -0.065 |
roc | -0.065 |
Token | explore |
Token ID | 7301 |
Feature activation | -2.24e-01 |
Pos logit contributions | |
paramed | +0.138 |
mixer | +0.135 |
mean | +0.135 |
liar | +0.134 |
boys | +0.133 |
Neg logit contributions | |
nights | -0.101 |
Along | -0.101 |
way | -0.091 |
drifted | -0.088 |
On | -0.086 |
Token | . |
Token ID | 13 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
Along | +0.367 |
nights | +0.355 |
way | +0.332 |
drifted | +0.326 |
On | +0.320 |
Neg logit contributions | |
paramed | -0.333 |
mixer | -0.331 |
mean | -0.329 |
boys | -0.324 |
liar | -0.323 |
Token | As |
Token ID | 1081 |
Feature activation | -2.09e-01 |
Pos logit contributions | |
nights | +0.053 |
drifted | +0.049 |
ingle | +0.047 |
lights | +0.047 |
more | +0.046 |
Neg logit contributions | |
paramed | -0.049 |
mixer | -0.046 |
greedy | -0.046 |
Lee | -0.045 |
stern | -0.045 |
Token | for |
Token ID | 329 |
Feature activation | +1.74e-02 |
Pos logit contributions | |
paramed | +0.041 |
liar | +0.040 |
mixer | +0.039 |
boys | +0.039 |
mean | +0.038 |
Neg logit contributions | |
Along | -0.026 |
nights | -0.026 |
way | -0.023 |
whenever | -0.022 |
drifted | -0.022 |
Token | the |
Token ID | 262 |
Feature activation | +1.59e-02 |
Pos logit contributions | |
paramed | +0.032 |
roc | +0.030 |
liar | +0.030 |
boys | +0.030 |
mixer | +0.028 |
Neg logit contributions | |
nights | -0.024 |
days | -0.022 |
whenever | -0.022 |
Along | -0.022 |
night | -0.021 |
Token | third |
Token ID | 2368 |
Feature activation | +5.09e-02 |
Pos logit contributions | |
paramed | +0.020 |
liar | +0.019 |
boys | +0.019 |
mixer | +0.019 |
wasteful | +0.019 |
Neg logit contributions | |
nights | -0.029 |
Along | -0.029 |
way | -0.028 |
drifted | -0.027 |
days | -0.027 |
Token | time |
Token ID | 640 |
Feature activation | -1.35e-01 |
Pos logit contributions | |
mixer | +0.061 |
paramed | +0.058 |
mean | +0.056 |
liar | +0.055 |
greedy | +0.054 |
Neg logit contributions | |
Along | -0.105 |
nights | -0.101 |
drifted | -0.095 |
On | -0.093 |
way | -0.093 |
Token | . |
Token ID | 13 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
Along | +0.170 |
nights | +0.159 |
way | +0.147 |
On | +0.144 |
drifted | +0.139 |
Neg logit contributions | |
mixer | -0.256 |
paramed | -0.254 |
mean | -0.254 |
liar | -0.253 |
greedy | -0.251 |
Token | She |
Token ID | 1375 |
Feature activation | -4.96e-02 |
Pos logit contributions | |
Along | +0.027 |
nights | +0.026 |
On | +0.024 |
way | +0.024 |
drifted | +0.023 |
Neg logit contributions | |
paramed | -0.041 |
liar | -0.040 |
mixer | -0.040 |
boys | -0.040 |
mean | -0.039 |
Token | wanted |
Token ID | 2227 |
Feature activation | -1.57e-02 |
Pos logit contributions | |
drifted | +0.076 |
nights | +0.068 |
Along | +0.068 |
way | +0.065 |
dark | +0.064 |
Neg logit contributions | |
paramed | -0.080 |
bluff | -0.079 |
roc | -0.078 |
liar | -0.076 |
boys | -0.072 |
Token | water |
Token ID | 1660 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
nights | +0.020 |
whenever | +0.019 |
night | +0.018 |
days | +0.018 |
Along | +0.018 |
Neg logit contributions | |
paramed | -0.030 |
liar | -0.029 |
boys | -0.028 |
roc | -0.028 |
bluff | -0.028 |
Token | . |
Token ID | 13 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
whenever | +0.039 |
nights | +0.036 |
night | +0.036 |
day | +0.035 |
of | +0.034 |
Neg logit contributions | |
paramed | -0.052 |
liar | -0.052 |
bluff | -0.051 |
roc | -0.049 |
boys | -0.049 |
Token | She |
Token ID | 1375 |
Feature activation | -4.98e-02 |
Pos logit contributions | |
Along | +0.024 |
nights | +0.022 |
On | +0.021 |
way | +0.021 |
whenever | +0.020 |
Neg logit contributions | |
paramed | -0.045 |
liar | -0.043 |
mixer | -0.043 |
boys | -0.043 |
wasteful | -0.043 |
Token | cried |
Token ID | 16896 |
Feature activation | -2.48e-03 |
Pos logit contributions | |
drifted | +0.078 |
nights | +0.074 |
Along | +0.074 |
way | +0.069 |
On | +0.069 |
Neg logit contributions | |
paramed | -0.079 |
bluff | -0.075 |
roc | -0.075 |
liar | -0.074 |
mixer | -0.073 |
Token | Once |
Token ID | 7454 |
Feature activation | -1.26e-01 |
Pos logit contributions | |
nights | +0.155 |
Along | +0.142 |
way | +0.138 |
drifted | +0.133 |
filled | +0.131 |
Neg logit contributions | |
mixer | -0.152 |
Lee | -0.146 |
mean | -0.145 |
chefs | -0.144 |
paramed | -0.141 |
Token | upon |
Token ID | 2402 |
Feature activation | -1.42e-01 |
Pos logit contributions | |
nights | +0.151 |
Along | +0.146 |
growth | +0.133 |
drifted | +0.131 |
more | +0.127 |
Neg logit contributions | |
mean | -0.249 |
mad | -0.246 |
mixer | -0.246 |
stern | -0.241 |
sorry | -0.237 |
Token | a |
Token ID | 257 |
Feature activation | -2.43e-02 |
Pos logit contributions | |
nights | +0.336 |
Along | +0.331 |
way | +0.327 |
horizon | +0.316 |
ingle | +0.309 |
Neg logit contributions | |
mean | -0.128 |
mad | -0.111 |
greedy | -0.110 |
liar | -0.107 |
chefs | -0.106 |
Token | time |
Token ID | 640 |
Feature activation | -1.36e-01 |
Pos logit contributions | |
nights | +0.024 |
Along | +0.024 |
way | +0.022 |
dark | +0.021 |
drifted | +0.020 |
Neg logit contributions | |
mean | -0.052 |
greedy | -0.050 |
mad | -0.050 |
mixer | -0.050 |
liar | -0.050 |
Token | , |
Token ID | 11 |
Feature activation | +8.10e-02 |
Pos logit contributions | |
nights | +0.242 |
Along | +0.241 |
drifted | +0.222 |
way | +0.220 |
dark | +0.215 |
Neg logit contributions | |
mixer | -0.183 |
mean | -0.182 |
paramed | -0.181 |
liar | -0.177 |
greedy | -0.174 |
Token | there |
Token ID | 612 |
Feature activation | -7.97e-02 |
Pos logit contributions | |
mean | +0.112 |
mixer | +0.107 |
greedy | +0.105 |
mad | +0.105 |
stern | +0.101 |
Neg logit contributions | |
nights | -0.150 |
Along | -0.144 |
drifted | -0.136 |
way | -0.133 |
ingle | -0.130 |
Token | lived |
Token ID | 5615 |
Feature activation | -8.00e-02 |
Pos logit contributions | |
Along | +0.116 |
nights | +0.113 |
way | +0.104 |
drifted | +0.100 |
On | +0.100 |
Neg logit contributions | |
paramed | -0.135 |
mixer | -0.133 |
liar | -0.131 |
mean | -0.130 |
boys | -0.129 |
Token | a |
Token ID | 257 |
Feature activation | -2.45e-02 |
Pos logit contributions | |
Along | +0.128 |
nights | +0.121 |
On | +0.112 |
way | +0.112 |
whenever | +0.109 |
Neg logit contributions | |
paramed | -0.128 |
liar | -0.125 |
mixer | -0.124 |
wasteful | -0.124 |
boys | -0.120 |
Token | hero |
Token ID | 4293 |
Feature activation | +1.20e-02 |
Pos logit contributions | |
Along | +0.023 |
nights | +0.022 |
day | +0.020 |
night | +0.020 |
days | +0.019 |
Neg logit contributions | |
paramed | -0.052 |
wasteful | -0.052 |
roc | -0.051 |
boys | -0.050 |
bluff | -0.050 |
Token | . |
Token ID | 13 |
Feature activation | -1.81e-02 |
Pos logit contributions | |
liar | +0.012 |
mean | +0.011 |
greedy | +0.011 |
paramed | +0.011 |
phone | +0.011 |
Neg logit contributions | |
Along | -0.026 |
nights | -0.026 |
drifted | -0.025 |
On | -0.024 |
way | -0.024 |
Token | He |
Token ID | 679 |
Feature activation | +3.23e-03 |
Pos logit contributions | |
nights | +0.033 |
drifted | +0.029 |
lights | +0.029 |
way | +0.029 |
lifetime | +0.028 |
Neg logit contributions | |
Lee | -0.025 |
paramed | -0.024 |
mean | -0.024 |
liar | -0.023 |
roc | -0.023 |
Token | and |
Token ID | 290 |
Feature activation | -2.64e-02 |
Pos logit contributions | |
Along | +0.063 |
nights | +0.060 |
way | +0.055 |
On | +0.053 |
drifted | +0.053 |
Neg logit contributions | |
paramed | -0.083 |
mixer | -0.083 |
liar | -0.082 |
mean | -0.082 |
boys | -0.081 |
Token | fun |
Token ID | 1257 |
Feature activation | -1.01e-02 |
Pos logit contributions | |
nights | +0.039 |
Along | +0.039 |
way | +0.035 |
drifted | +0.035 |
On | +0.034 |
Neg logit contributions | |
paramed | -0.044 |
liar | -0.043 |
mixer | -0.042 |
boys | -0.042 |
wasteful | -0.041 |
Token | . |
Token ID | 13 |
Feature activation | -2.01e-02 |
Pos logit contributions | |
Along | +0.015 |
nights | +0.014 |
On | +0.013 |
way | +0.013 |
dark | +0.013 |
Neg logit contributions | |
mixer | -0.017 |
mean | -0.016 |
phone | -0.016 |
boys | -0.016 |
chefs | -0.016 |
Token | Can |
Token ID | 1680 |
Feature activation | +1.23e-01 |
Pos logit contributions | |
nights | +0.031 |
way | +0.029 |
Along | +0.028 |
drifted | +0.028 |
ingle | +0.027 |
Neg logit contributions | |
paramed | -0.032 |
mixer | -0.031 |
liar | -0.030 |
greedy | -0.030 |
mean | -0.030 |
Token | I |
Token ID | 314 |
Feature activation | -5.09e-03 |
Pos logit contributions | |
paramed | +0.237 |
liar | +0.233 |
mixer | +0.229 |
wasteful | +0.228 |
boys | +0.225 |
Neg logit contributions | |
Along | -0.152 |
nights | -0.143 |
way | -0.134 |
On | -0.132 |
whenever | -0.131 |
Token | wear |
Token ID | 5806 |
Feature activation | -1.91e-01 |
Pos logit contributions | |
Along | +0.008 |
nights | +0.008 |
way | +0.007 |
drifted | +0.007 |
On | +0.007 |
Neg logit contributions | |
paramed | -0.008 |
mixer | -0.008 |
liar | -0.008 |
mean | -0.008 |
boys | -0.008 |
Token | them |
Token ID | 606 |
Feature activation | +8.70e-02 |
Pos logit contributions | |
Along | +0.319 |
nights | +0.312 |
whenever | +0.293 |
On | +0.291 |
night | +0.287 |
Neg logit contributions | |
paramed | -0.291 |
liar | -0.287 |
wasteful | -0.274 |
roc | -0.271 |
greedy | -0.267 |
Token | to |
Token ID | 284 |
Feature activation | +7.77e-02 |
Pos logit contributions | |
paramed | +0.119 |
liar | +0.115 |
mixer | +0.114 |
wasteful | +0.112 |
boys | +0.112 |
Neg logit contributions | |
Along | -0.155 |
nights | -0.153 |
way | -0.143 |
On | -0.140 |
drifted | -0.138 |
Token | school |
Token ID | 1524 |
Feature activation | -1.20e-01 |
Pos logit contributions | |
paramed | +0.119 |
liar | +0.116 |
mixer | +0.114 |
wasteful | +0.112 |
greedy | +0.111 |
Neg logit contributions | |
Along | -0.125 |
nights | -0.124 |
whenever | -0.114 |
On | -0.113 |
way | -0.112 |
Token | ?" |
Token ID | 1701 |
Feature activation | +7.64e-02 |
Pos logit contributions | |
Along | +0.207 |
nights | +0.203 |
way | +0.190 |
On | +0.187 |
whenever | +0.186 |
Neg logit contributions | |
paramed | -0.171 |
liar | -0.166 |
mixer | -0.161 |
wasteful | -0.160 |
boys | -0.160 |
Token |
|
Token ID | 198 |
Feature activation | -1.99e-02 |
Pos logit contributions | |
paramed | +0.147 |
mixer | +0.144 |
liar | +0.142 |
mean | +0.141 |
boys | +0.140 |
Neg logit contributions | |
Along | -0.093 |
nights | -0.093 |
way | -0.084 |
drifted | -0.080 |
On | -0.078 |
Token | that |
Token ID | 326 |
Feature activation | -3.29e-02 |
Pos logit contributions | |
mixer | +0.162 |
mean | +0.161 |
paramed | +0.161 |
boys | +0.158 |
phone | +0.158 |
Neg logit contributions | |
Along | -0.136 |
nights | -0.133 |
way | -0.118 |
drifted | -0.118 |
On | -0.117 |
Token | made |
Token ID | 925 |
Feature activation | +7.28e-02 |
Pos logit contributions | |
Along | +0.052 |
nights | +0.051 |
way | +0.048 |
On | +0.046 |
drifted | +0.046 |
Neg logit contributions | |
paramed | -0.051 |
mixer | -0.050 |
liar | -0.050 |
mean | -0.049 |
boys | -0.049 |
Token | new |
Token ID | 649 |
Feature activation | -1.83e-02 |
Pos logit contributions | |
paramed | +0.093 |
mixer | +0.092 |
mean | +0.090 |
phone | +0.088 |
mad | +0.086 |
Neg logit contributions | |
Along | -0.138 |
nights | -0.130 |
way | -0.126 |
On | -0.121 |
more | -0.120 |
Token | pills |
Token ID | 19521 |
Feature activation | +8.02e-02 |
Pos logit contributions | |
Along | +0.029 |
nights | +0.028 |
way | +0.026 |
On | +0.026 |
drifted | +0.025 |
Neg logit contributions | |
paramed | -0.029 |
mixer | -0.028 |
boys | -0.028 |
mean | -0.028 |
liar | -0.028 |
Token | every |
Token ID | 790 |
Feature activation | -2.06e-01 |
Pos logit contributions | |
mean | +0.137 |
mixer | +0.134 |
boys | +0.132 |
phone | +0.132 |
liar | +0.131 |
Neg logit contributions | |
Along | -0.117 |
nights | -0.112 |
way | -0.101 |
On | -0.101 |
drifted | -0.096 |
Token | day |
Token ID | 1110 |
Feature activation | -2.16e-01 |
Pos logit contributions | |
Along | +0.182 |
On | +0.145 |
lights | +0.141 |
Throughout | +0.140 |
Everywhere | +0.138 |
Neg logit contributions | |
mixer | -0.469 |
mean | -0.469 |
mad | -0.452 |
phone | -0.450 |
liar | -0.446 |
Token | . |
Token ID | 13 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
Along | +0.274 |
nights | +0.261 |
way | +0.241 |
On | +0.233 |
drifted | +0.227 |
Neg logit contributions | |
paramed | -0.401 |
mixer | -0.398 |
mean | -0.396 |
liar | -0.394 |
boys | -0.393 |
Token | Sal |
Token ID | 4849 |
Feature activation | +9.99e-02 |
Pos logit contributions | |
nights | +0.027 |
Along | +0.026 |
way | +0.025 |
drifted | +0.024 |
dark | +0.023 |
Neg logit contributions | |
paramed | -0.031 |
mixer | -0.030 |
liar | -0.029 |
roc | -0.029 |
wasteful | -0.028 |
Token | gave |
Token ID | 2921 |
Feature activation | +3.35e-02 |
Pos logit contributions | |
liar | +0.180 |
paramed | +0.179 |
bluff | +0.175 |
wasteful | +0.173 |
mean | +0.169 |
Neg logit contributions | |
nights | -0.136 |
Along | -0.133 |
drifted | -0.123 |
way | -0.120 |
days | -0.117 |
Token | Joe |
Token ID | 5689 |
Feature activation | -8.59e-03 |
Pos logit contributions | |
paramed | +0.051 |
liar | +0.050 |
wasteful | +0.049 |
mixer | +0.048 |
mean | +0.048 |
Neg logit contributions | |
Along | -0.055 |
nights | -0.055 |
way | -0.050 |
whenever | -0.049 |
On | -0.049 |
Token | a |
Token ID | 257 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
Along | +0.016 |
nights | +0.016 |
way | +0.015 |
On | +0.014 |
drifted | +0.014 |
Neg logit contributions | |
paramed | -0.011 |
mixer | -0.011 |
mean | -0.011 |
greedy | -0.010 |
boys | -0.010 |
Token | He |
Token ID | 679 |
Feature activation | +7.62e-04 |
Pos logit contributions | |
nights | +0.042 |
drifted | +0.038 |
way | +0.037 |
ingle | +0.036 |
filled | +0.036 |
Neg logit contributions | |
mixer | -0.037 |
stern | -0.036 |
greedy | -0.036 |
sorry | -0.035 |
bluff | -0.035 |
Token | knew |
Token ID | 2993 |
Feature activation | -1.94e-03 |
Pos logit contributions | |
mean | +0.001 |
boys | +0.001 |
greedy | +0.001 |
paramed | +0.001 |
mad | +0.001 |
Neg logit contributions | |
Along | -0.001 |
nights | -0.001 |
way | -0.001 |
On | -0.001 |
ingle | -0.001 |
Token | that |
Token ID | 326 |
Feature activation | -3.56e-02 |
Pos logit contributions | |
Along | +0.003 |
nights | +0.003 |
way | +0.003 |
On | +0.003 |
drifted | +0.003 |
Neg logit contributions | |
mean | -0.003 |
mixer | -0.003 |
liar | -0.003 |
greedy | -0.003 |
paramed | -0.003 |
Token | as |
Token ID | 355 |
Feature activation | -1.87e-01 |
Pos logit contributions | |
Along | +0.050 |
nights | +0.047 |
way | +0.044 |
On | +0.043 |
drifted | +0.042 |
Neg logit contributions | |
paramed | -0.062 |
mixer | -0.062 |
mean | -0.061 |
liar | -0.061 |
boys | -0.060 |
Token | long |
Token ID | 890 |
Feature activation | -9.24e-02 |
Pos logit contributions | |
Along | +0.367 |
way | +0.339 |
ingle | +0.319 |
On | +0.319 |
drifted | +0.312 |
Neg logit contributions | |
mean | -0.249 |
greedy | -0.226 |
mad | -0.213 |
sorry | -0.210 |
chefs | -0.209 |
Token | as |
Token ID | 355 |
Feature activation | -1.87e-01 |
Pos logit contributions | |
Along | +0.107 |
ingle | +0.103 |
dark | +0.100 |
way | +0.094 |
On | +0.093 |
Neg logit contributions | |
mean | -0.174 |
fighting | -0.171 |
greedy | -0.170 |
mixer | -0.170 |
chefs | -0.169 |
Token | he |
Token ID | 339 |
Feature activation | -9.31e-02 |
Pos logit contributions | |
Along | +0.309 |
ingle | +0.273 |
way | +0.273 |
On | +0.267 |
drifted | +0.259 |
Neg logit contributions | |
mean | -0.308 |
greedy | -0.288 |
sorry | -0.282 |
chefs | -0.282 |
mad | -0.277 |
Token | had |
Token ID | 550 |
Feature activation | -6.23e-02 |
Pos logit contributions | |
Along | +0.177 |
way | +0.164 |
nights | +0.160 |
On | +0.158 |
ingle | +0.153 |
Neg logit contributions | |
mean | -0.112 |
boys | -0.107 |
greedy | -0.102 |
qu | -0.100 |
phone | -0.100 |
Token | this |
Token ID | 428 |
Feature activation | -4.71e-03 |
Pos logit contributions | |
Along | +0.110 |
nights | +0.106 |
way | +0.100 |
On | +0.098 |
dark | +0.097 |
Neg logit contributions | |
mean | -0.085 |
paramed | -0.084 |
liar | -0.082 |
mixer | -0.082 |
greedy | -0.082 |
Token | door |
Token ID | 3420 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
Along | +0.007 |
nights | +0.007 |
drifted | +0.006 |
ingle | +0.006 |
On | +0.006 |
Neg logit contributions | |
mean | -0.008 |
mixer | -0.008 |
phone | -0.008 |
stern | -0.008 |
paramed | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | +7.71e-02 |
Pos logit contributions | |
Along | +0.018 |
nights | +0.017 |
way | +0.016 |
ingle | +0.015 |
On | +0.015 |
Neg logit contributions | |
mean | -0.016 |
phone | -0.015 |
stern | -0.015 |
mixer | -0.015 |
paramed | -0.015 |
Token | dance |
Token ID | 9280 |
Feature activation | -1.28e-01 |
Pos logit contributions | |
paramed | +0.144 |
mixer | +0.141 |
liar | +0.140 |
mean | +0.139 |
boys | +0.138 |
Neg logit contributions | |
Along | -0.113 |
nights | -0.112 |
way | -0.103 |
drifted | -0.098 |
On | -0.098 |
Token | . |
Token ID | 13 |
Feature activation | -3.25e-02 |
Pos logit contributions | |
Along | +0.166 |
nights | +0.163 |
way | +0.150 |
drifted | +0.142 |
On | +0.141 |
Neg logit contributions | |
paramed | -0.234 |
mixer | -0.231 |
mean | -0.229 |
liar | -0.228 |
boys | -0.226 |
Token | She |
Token ID | 1375 |
Feature activation | -6.16e-02 |
Pos logit contributions | |
nights | +0.055 |
way | +0.051 |
drifted | +0.050 |
lifetime | +0.048 |
Along | +0.047 |
Neg logit contributions | |
paramed | -0.048 |
mixer | -0.048 |
mean | -0.046 |
Lee | -0.046 |
liar | -0.044 |
Token | dances |
Token ID | 38207 |
Feature activation | -8.74e-02 |
Pos logit contributions | |
Along | +0.115 |
way | +0.110 |
nights | +0.109 |
ingle | +0.102 |
more | +0.102 |
Neg logit contributions | |
mean | -0.085 |
greedy | -0.076 |
liar | -0.075 |
mixer | -0.075 |
boys | -0.073 |
Token | every |
Token ID | 790 |
Feature activation | -2.07e-01 |
Pos logit contributions | |
Along | +0.123 |
nights | +0.121 |
way | +0.115 |
drifted | +0.110 |
dark | +0.107 |
Neg logit contributions | |
mean | -0.149 |
mixer | -0.144 |
paramed | -0.143 |
liar | -0.142 |
phone | -0.141 |
Token | day |
Token ID | 1110 |
Feature activation | -2.16e-01 |
Pos logit contributions | |
Along | +0.187 |
On | +0.154 |
nights | +0.151 |
more | +0.142 |
Throughout | +0.141 |
Neg logit contributions | |
mean | -0.486 |
mixer | -0.482 |
phone | -0.462 |
mad | -0.452 |
greedy | -0.441 |
Token | in |
Token ID | 287 |
Feature activation | -6.18e-02 |
Pos logit contributions | |
Along | +0.363 |
nights | +0.346 |
way | +0.342 |
ingle | +0.331 |
drifted | +0.327 |
Neg logit contributions | |
mean | -0.315 |
mixer | -0.297 |
greedy | -0.294 |
phone | -0.287 |
fighting | -0.287 |
Token | her |
Token ID | 607 |
Feature activation | -6.86e-02 |
Pos logit contributions | |
Along | +0.134 |
way | +0.134 |
nights | +0.127 |
drifted | +0.123 |
dark | +0.122 |
Neg logit contributions | |
mean | -0.063 |
mixer | -0.059 |
mad | -0.058 |
phone | -0.056 |
greedy | -0.053 |
Token | room |
Token ID | 2119 |
Feature activation | -6.86e-03 |
Pos logit contributions | |
Along | +0.124 |
nights | +0.121 |
way | +0.120 |
more | +0.113 |
On | +0.111 |
Neg logit contributions | |
mixer | -0.093 |
mean | -0.090 |
phone | -0.090 |
paramed | -0.087 |
greedy | -0.086 |
Token | . |
Token ID | 13 |
Feature activation | -3.30e-02 |
Pos logit contributions | |
Along | +0.011 |
nights | +0.010 |
way | +0.010 |
On | +0.009 |
drifted | +0.009 |
Neg logit contributions | |
mean | -0.011 |
mixer | -0.011 |
phone | -0.011 |
greedy | -0.011 |
liar | -0.011 |
Token | She |
Token ID | 1375 |
Feature activation | -6.19e-02 |
Pos logit contributions | |
nights | +0.056 |
way | +0.053 |
lifetime | +0.050 |
ingle | +0.049 |
drifted | +0.049 |
Neg logit contributions | |
mixer | -0.050 |
paramed | -0.049 |
Lee | -0.047 |
mean | -0.046 |
liar | -0.045 |
Token | stories |
Token ID | 3923 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
Along | +0.046 |
nights | +0.045 |
way | +0.042 |
On | +0.041 |
drifted | +0.041 |
Neg logit contributions | |
paramed | -0.039 |
mixer | -0.037 |
liar | -0.037 |
mean | -0.037 |
boys | -0.037 |
Token | . |
Token ID | 13 |
Feature activation | -2.96e-02 |
Pos logit contributions | |
Along | +0.018 |
nights | +0.018 |
way | +0.016 |
drifted | +0.016 |
ingle | +0.015 |
Neg logit contributions | |
mean | -0.022 |
mixer | -0.021 |
phone | -0.021 |
boys | -0.021 |
paramed | -0.021 |
Token |
|
Token ID | 198 |
Feature activation | -2.69e-02 |
Pos logit contributions | |
nights | +0.043 |
way | +0.041 |
Along | +0.040 |
drifted | +0.038 |
ingle | +0.037 |
Neg logit contributions | |
mixer | -0.050 |
paramed | -0.048 |
mean | -0.047 |
greedy | -0.047 |
stern | -0.047 |
Token |
|
Token ID | 198 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
nights | +0.042 |
drifted | +0.038 |
Along | +0.038 |
way | +0.036 |
filled | +0.035 |
Neg logit contributions | |
mixer | -0.042 |
mean | -0.041 |
stern | -0.040 |
paramed | -0.039 |
phone | -0.038 |
Token | One |
Token ID | 3198 |
Feature activation | -1.53e-01 |
Pos logit contributions | |
nights | +0.033 |
Along | +0.033 |
drifted | +0.030 |
way | +0.030 |
ingle | +0.029 |
Neg logit contributions | |
paramed | -0.048 |
mean | -0.048 |
mixer | -0.048 |
liar | -0.046 |
greedy | -0.046 |
Token | day |
Token ID | 1110 |
Feature activation | -2.17e-01 |
Pos logit contributions | |
Along | +0.167 |
growth | +0.156 |
drifted | +0.149 |
way | +0.146 |
nights | +0.144 |
Neg logit contributions | |
mean | -0.346 |
mixer | -0.312 |
mad | -0.309 |
greedy | -0.303 |
rude | -0.301 |
Token | , |
Token ID | 11 |
Feature activation | +7.61e-02 |
Pos logit contributions | |
Along | +0.396 |
nights | +0.386 |
way | +0.367 |
dark | +0.349 |
drifted | +0.349 |
Neg logit contributions | |
mixer | -0.288 |
mean | -0.282 |
paramed | -0.282 |
liar | -0.277 |
stern | -0.273 |
Token | she |
Token ID | 673 |
Feature activation | -1.52e-01 |
Pos logit contributions | |
mixer | +0.083 |
stern | +0.081 |
mean | +0.080 |
chefs | +0.080 |
mad | +0.077 |
Neg logit contributions | |
Along | -0.154 |
nights | -0.152 |
way | -0.149 |
ingle | -0.142 |
dark | -0.142 |
Token | goes |
Token ID | 2925 |
Feature activation | -6.62e-02 |
Pos logit contributions | |
Along | +0.271 |
nights | +0.268 |
way | +0.257 |
more | +0.241 |
On | +0.240 |
Neg logit contributions | |
mean | -0.208 |
boys | -0.201 |
paramed | -0.201 |
liar | -0.201 |
greedy | -0.200 |
Token | to |
Token ID | 284 |
Feature activation | +8.23e-02 |
Pos logit contributions | |
nights | +0.116 |
Along | +0.113 |
way | +0.108 |
drifted | +0.107 |
more | +0.104 |
Neg logit contributions | |
mean | -0.090 |
phone | -0.090 |
paramed | -0.089 |
mixer | -0.089 |
boys | -0.087 |
Token | the |
Token ID | 262 |
Feature activation | +1.16e-02 |
Pos logit contributions | |
paramed | +0.110 |
mixer | +0.108 |
mean | +0.107 |
liar | +0.107 |
boys | +0.105 |
Neg logit contributions | |
Along | -0.147 |
nights | -0.147 |
way | -0.139 |
drifted | -0.134 |
On | -0.130 |