Token | tick |
Token ID | 4378 |
Feature activation | +1.16e+00 |
Pos logit contributions | |
day | +0.061 |
afternoon | +0.053 |
night | +0.049 |
morning | +0.047 |
evening | +0.044 |
Neg logit contributions | |
gru | -0.069 |
Reward | -0.060 |
ution | -0.060 |
god | -0.059 |
rifle | -0.059 |
Token | - |
Token ID | 12 |
Feature activation | -7.78e-01 |
Pos logit contributions | |
day | +0.038 |
ite | +0.034 |
afternoon | +0.034 |
cy | +0.033 |
night | +0.031 |
Neg logit contributions | |
gru | -0.052 |
badge | -0.045 |
ution | -0.043 |
threatening | -0.043 |
warning | -0.042 |
Token | t |
Token ID | 83 |
Feature activation | +1.61e-01 |
Pos logit contributions | |
gru | +0.030 |
badge | +0.025 |
Reward | +0.024 |
ution | +0.023 |
threatening | +0.023 |
Neg logit contributions | |
day | -0.032 |
afternoon | -0.029 |
cy | -0.028 |
night | -0.027 |
ite | -0.027 |
Token | ocked |
Token ID | 3543 |
Feature activation | +3.00e+00 |
Pos logit contributions | |
cy | +0.005 |
day | +0.005 |
olly | +0.005 |
ite | +0.005 |
afternoon | +0.005 |
Neg logit contributions | |
gru | -0.007 |
badge | -0.006 |
warning | -0.006 |
threatening | -0.006 |
probably | -0.006 |
Token | all |
Token ID | 477 |
Feature activation | +6.45e+00 |
Pos logit contributions | |
day | +0.107 |
afternoon | +0.098 |
cy | +0.091 |
night | +0.090 |
ite | +0.089 |
Neg logit contributions | |
gru | -0.130 |
badge | -0.114 |
Reward | -0.108 |
threatening | -0.107 |
rifle | -0.106 |
Token | day |
Token ID | 1110 |
Feature activation | +1.22e+01 |
Pos logit contributions | |
olly | +0.207 |
ite | +0.192 |
ops | +0.191 |
cy | +0.183 |
day | +0.172 |
Neg logit contributions | |
gru | -0.273 |
threatening | -0.229 |
probably | -0.224 |
warning | -0.218 |
menacing | -0.214 |
Token | long |
Token ID | 890 |
Feature activation | +4.67e+00 |
Pos logit contributions | |
day | +0.399 |
afternoon | +0.364 |
cy | +0.363 |
ite | +0.339 |
night | +0.329 |
Neg logit contributions | |
gru | -0.558 |
badge | -0.488 |
threatening | -0.463 |
Reward | -0.461 |
warning | -0.459 |
Token | . |
Token ID | 13 |
Feature activation | +1.32e+00 |
Pos logit contributions | |
cy | +0.151 |
day | +0.148 |
ite | +0.140 |
afternoon | +0.138 |
olly | +0.136 |
Neg logit contributions | |
gru | -0.203 |
badge | -0.182 |
warning | -0.170 |
threatening | -0.169 |
rifle | -0.167 |
Token | One |
Token ID | 1881 |
Feature activation | -1.61e+00 |
Pos logit contributions | |
day | +0.050 |
afternoon | +0.046 |
cy | +0.043 |
night | +0.042 |
ite | +0.041 |
Neg logit contributions | |
gru | -0.054 |
badge | -0.046 |
Reward | -0.046 |
rifle | -0.043 |
threatening | -0.043 |
Token | day |
Token ID | 1110 |
Feature activation | +1.14e+01 |
Pos logit contributions | |
gru | +0.102 |
warning | +0.089 |
threatening | +0.088 |
probably | +0.088 |
menacing | +0.086 |
Neg logit contributions | |
ips | -0.016 |
ops | -0.016 |
he | -0.013 |
cy | -0.012 |
ite | -0.012 |
Token | , |
Token ID | 11 |
Feature activation | +1.65e-01 |
Pos logit contributions | |
day | +0.289 |
afternoon | +0.245 |
night | +0.224 |
cy | +0.215 |
morning | +0.208 |
Neg logit contributions | |
gru | -0.619 |
badge | -0.557 |
Reward | -0.541 |
ution | -0.540 |
rifle | -0.533 |
Token | nap |
Token ID | 25422 |
Feature activation | +3.16e+00 |
Pos logit contributions | |
day | +0.089 |
ite | +0.083 |
cy | +0.082 |
afternoon | +0.079 |
he | +0.078 |
Neg logit contributions | |
gru | -0.223 |
badge | -0.203 |
beard | -0.192 |
warning | -0.191 |
rifle | -0.189 |
Token | ! |
Token ID | 0 |
Feature activation | -2.97e-01 |
Pos logit contributions | |
day | +0.092 |
afternoon | +0.083 |
cy | +0.080 |
ite | +0.079 |
night | +0.076 |
Neg logit contributions | |
gru | -0.154 |
badge | -0.137 |
threatening | -0.128 |
warning | -0.127 |
ution | -0.126 |
Token |
|
Token ID | 198 |
Feature activation | -8.28e-02 |
Pos logit contributions | |
gru | +0.010 |
Reward | +0.009 |
badge | +0.008 |
address | +0.008 |
probably | +0.007 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.012 |
cy | -0.012 |
ite | -0.011 |
night | -0.011 |
Token |
|
Token ID | 198 |
Feature activation | -1.07e-01 |
Pos logit contributions | |
gru | +0.003 |
badge | +0.003 |
Reward | +0.003 |
warning | +0.003 |
threatening | +0.003 |
Neg logit contributions | |
day | -0.003 |
cy | -0.003 |
afternoon | -0.003 |
ite | -0.003 |
night | -0.003 |
Token | The |
Token ID | 464 |
Feature activation | +1.30e+00 |
Pos logit contributions | |
gru | +0.005 |
badge | +0.004 |
Reward | +0.004 |
ution | +0.004 |
rifle | +0.004 |
Neg logit contributions | |
day | -0.004 |
afternoon | -0.003 |
cy | -0.003 |
night | -0.003 |
ite | -0.003 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
cy | +0.034 |
ite | +0.032 |
day | +0.032 |
afternoon | +0.030 |
he | +0.030 |
Neg logit contributions | |
gru | -0.065 |
warning | -0.059 |
badge | -0.058 |
threatening | -0.055 |
message | -0.055 |
Token | . |
Token ID | 13 |
Feature activation | +4.29e-01 |
Pos logit contributions | |
day | +0.365 |
afternoon | +0.330 |
cy | +0.306 |
night | +0.299 |
ite | +0.294 |
Neg logit contributions | |
gru | -0.555 |
badge | -0.487 |
Reward | -0.469 |
threatening | -0.463 |
ution | -0.461 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.23e+00 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.013 |
cy | +0.012 |
night | +0.011 |
ite | +0.011 |
Neg logit contributions | |
gru | -0.020 |
Reward | -0.017 |
badge | -0.017 |
rifle | -0.016 |
threatening | -0.016 |
Token | Once |
Token ID | 7454 |
Feature activation | -6.97e-01 |
Pos logit contributions | |
day | +0.203 |
afternoon | +0.200 |
ite | +0.195 |
olly | +0.191 |
End | +0.184 |
Neg logit contributions | |
gru | -0.165 |
Reward | -0.137 |
Gu | -0.130 |
probably | -0.120 |
Stan | -0.120 |
Token | upon |
Token ID | 2402 |
Feature activation | -4.97e-01 |
Pos logit contributions | |
gru | +0.021 |
Reward | +0.017 |
stern | +0.017 |
hello | +0.017 |
probably | +0.017 |
Neg logit contributions | |
afternoon | -0.027 |
days | -0.025 |
day | -0.024 |
ops | -0.024 |
cy | -0.024 |
Token | a |
Token ID | 257 |
Feature activation | -2.16e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.011 |
Reward | +0.011 |
probably | +0.010 |
spoken | +0.010 |
Neg logit contributions | |
day | -0.024 |
afternoon | -0.023 |
cy | -0.023 |
ite | -0.023 |
olly | -0.022 |
Token | good |
Token ID | 922 |
Feature activation | +2.38e+00 |
Pos logit contributions | |
day | +0.013 |
afternoon | +0.012 |
night | +0.011 |
cy | +0.011 |
morning | +0.010 |
Neg logit contributions | |
gru | -0.013 |
Reward | -0.011 |
badge | -0.011 |
rifle | -0.011 |
ution | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | +4.49e-01 |
Pos logit contributions | |
day | +0.075 |
afternoon | +0.068 |
cy | +0.062 |
night | +0.062 |
ite | +0.060 |
Neg logit contributions | |
gru | -0.113 |
badge | -0.102 |
Reward | -0.096 |
threatening | -0.095 |
ution | -0.095 |
Token |
|
Token ID | 198 |
Feature activation | -2.62e-01 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.016 |
cy | +0.015 |
ite | +0.015 |
night | +0.015 |
Neg logit contributions | |
gru | -0.017 |
Reward | -0.015 |
badge | -0.015 |
address | -0.014 |
threatening | -0.014 |
Token |
|
Token ID | 198 |
Feature activation | -2.76e-01 |
Pos logit contributions | |
gru | +0.010 |
badge | +0.009 |
Reward | +0.009 |
warning | +0.009 |
threatening | +0.009 |
Neg logit contributions | |
day | -0.009 |
cy | -0.009 |
afternoon | -0.009 |
ite | -0.009 |
night | -0.008 |
Token | The |
Token ID | 464 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.012 |
rifle | +0.011 |
Reward | +0.011 |
ution | +0.011 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
he | -0.007 |
night | -0.007 |
cy | -0.007 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ite | +0.043 |
cy | +0.043 |
day | +0.039 |
afternoon | +0.038 |
he | +0.038 |
Neg logit contributions | |
gru | -0.044 |
warning | -0.038 |
badge | -0.037 |
threatening | -0.035 |
stern | -0.034 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.05e+00 |
Pos logit contributions | |
day | +0.361 |
afternoon | +0.329 |
cy | +0.306 |
night | +0.298 |
ite | +0.292 |
Neg logit contributions | |
gru | -0.554 |
badge | -0.487 |
Reward | -0.469 |
threatening | -0.462 |
ution | -0.458 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.49e-01 |
Pos logit contributions | |
ite | +0.210 |
day | +0.208 |
afternoon | +0.207 |
olly | +0.203 |
cy | +0.200 |
Neg logit contributions | |
gru | -0.137 |
Reward | -0.112 |
Gu | -0.108 |
an | -0.099 |
Stan | -0.098 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.20e-01 |
Pos logit contributions | |
gru | +0.020 |
stern | +0.017 |
hello | +0.017 |
Reward | +0.016 |
probably | +0.016 |
Neg logit contributions | |
afternoon | -0.029 |
days | -0.028 |
ops | -0.027 |
akes | -0.027 |
cy | -0.027 |
Token | a |
Token ID | 257 |
Feature activation | -2.15e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.011 |
Reward | +0.010 |
probably | +0.010 |
spoken | +0.010 |
Neg logit contributions | |
day | -0.026 |
ite | -0.025 |
afternoon | -0.025 |
cy | -0.024 |
olly | -0.023 |
Token | time |
Token ID | 640 |
Feature activation | +7.22e+00 |
Pos logit contributions | |
gru | +0.100 |
badge | +0.088 |
probably | +0.087 |
warning | +0.086 |
threatening | +0.086 |
Neg logit contributions | |
ite | -0.055 |
day | -0.053 |
cy | -0.051 |
afternoon | -0.048 |
night | -0.045 |
Token | . |
Token ID | 13 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
gru | +0.328 |
badge | +0.289 |
probably | +0.280 |
warning | +0.280 |
threatening | +0.279 |
Neg logit contributions | |
day | -0.217 |
afternoon | -0.204 |
cy | -0.202 |
ite | -0.199 |
olly | -0.185 |
Token | |
Token ID | 220 |
Feature activation | +5.07e-01 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.011 |
ite | +0.010 |
cy | +0.010 |
night | +0.010 |
Neg logit contributions | |
gru | -0.011 |
Reward | -0.010 |
address | -0.009 |
god | -0.009 |
badge | -0.009 |
Token |
|
Token ID | 198 |
Feature activation | -2.84e-01 |
Pos logit contributions | |
day | +0.020 |
afternoon | +0.018 |
night | +0.017 |
cy | +0.016 |
days | +0.016 |
Neg logit contributions | |
gru | -0.020 |
Reward | -0.017 |
badge | -0.017 |
address | -0.017 |
probably | -0.016 |
Token |
|
Token ID | 198 |
Feature activation | -3.04e-01 |
Pos logit contributions | |
gru | +0.010 |
Reward | +0.008 |
warning | +0.008 |
Frank | +0.008 |
god | +0.008 |
Neg logit contributions | |
day | -0.011 |
cy | -0.011 |
afternoon | -0.011 |
ite | -0.010 |
night | -0.010 |
Token | The |
Token ID | 464 |
Feature activation | +1.10e+00 |
Pos logit contributions | |
gru | +0.012 |
badge | +0.010 |
Reward | +0.010 |
ution | +0.010 |
god | +0.010 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.010 |
cy | -0.010 |
night | -0.010 |
ite | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
day | +0.039 |
ite | +0.039 |
cy | +0.038 |
he | +0.037 |
afternoon | +0.037 |
Neg logit contributions | |
gru | -0.045 |
badge | -0.038 |
warning | -0.037 |
address | -0.036 |
god | -0.036 |
Token | . |
Token ID | 13 |
Feature activation | +2.83e-01 |
Pos logit contributions | |
day | +0.350 |
afternoon | +0.312 |
cy | +0.287 |
night | +0.284 |
ite | +0.276 |
Neg logit contributions | |
gru | -0.571 |
badge | -0.506 |
Reward | -0.487 |
ution | -0.479 |
threatening | -0.478 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.12e+00 |
Pos logit contributions | |
day | +0.009 |
afternoon | +0.009 |
cy | +0.008 |
night | +0.008 |
ite | +0.008 |
Neg logit contributions | |
gru | -0.012 |
Reward | -0.011 |
badge | -0.010 |
god | -0.010 |
address | -0.010 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.18e-01 |
Pos logit contributions | |
day | +0.193 |
afternoon | +0.191 |
ite | +0.188 |
olly | +0.183 |
cy | +0.174 |
Neg logit contributions | |
gru | -0.161 |
Reward | -0.135 |
Gu | -0.123 |
probably | -0.119 |
Start | -0.117 |
Token | upon |
Token ID | 2402 |
Feature activation | -4.93e-01 |
Pos logit contributions | |
gru | +0.021 |
stern | +0.017 |
probably | +0.017 |
hello | +0.016 |
Reward | +0.016 |
Neg logit contributions | |
afternoon | -0.028 |
days | -0.026 |
ops | -0.026 |
day | -0.025 |
akes | -0.025 |
Token | a |
Token ID | 257 |
Feature activation | -2.22e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.011 |
Reward | +0.010 |
probably | +0.010 |
spoken | +0.010 |
Neg logit contributions | |
day | -0.024 |
ite | -0.023 |
afternoon | -0.023 |
cy | -0.023 |
olly | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +3.58e-01 |
Pos logit contributions | |
day | +0.026 |
afternoon | +0.023 |
cy | +0.021 |
night | +0.021 |
ite | +0.020 |
Neg logit contributions | |
gru | -0.045 |
badge | -0.040 |
Reward | -0.038 |
rifle | -0.038 |
threatening | -0.038 |
Token | |
Token ID | 220 |
Feature activation | +5.69e-01 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.013 |
cy | +0.012 |
night | +0.012 |
ite | +0.012 |
Neg logit contributions | |
gru | -0.015 |
badge | -0.012 |
Reward | -0.012 |
rifle | -0.012 |
threatening | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | -3.19e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.019 |
night | +0.017 |
cy | +0.017 |
days | +0.017 |
Neg logit contributions | |
gru | -0.024 |
badge | -0.021 |
Reward | -0.021 |
rifle | -0.020 |
threatening | -0.020 |
Token |
|
Token ID | 198 |
Feature activation | -3.34e-01 |
Pos logit contributions | |
gru | +0.013 |
Reward | +0.011 |
badge | +0.011 |
warning | +0.011 |
god | +0.011 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.011 |
cy | -0.011 |
night | -0.010 |
ite | -0.010 |
Token | The |
Token ID | 464 |
Feature activation | +1.16e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.012 |
Reward | +0.012 |
ution | +0.012 |
rifle | +0.012 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.011 |
cy | -0.011 |
night | -0.011 |
ite | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ite | +0.044 |
day | +0.043 |
he | +0.043 |
cy | +0.043 |
afternoon | +0.042 |
Neg logit contributions | |
gru | -0.044 |
warning | -0.036 |
badge | -0.034 |
threatening | -0.034 |
god | -0.034 |
Token | . |
Token ID | 13 |
Feature activation | +3.22e-01 |
Pos logit contributions | |
day | +0.354 |
afternoon | +0.316 |
cy | +0.291 |
night | +0.288 |
ite | +0.280 |
Neg logit contributions | |
gru | -0.568 |
badge | -0.502 |
Reward | -0.484 |
ution | -0.477 |
threatening | -0.476 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.13e+00 |
Pos logit contributions | |
day | +0.010 |
afternoon | +0.009 |
cy | +0.008 |
night | +0.008 |
ite | +0.008 |
Neg logit contributions | |
gru | -0.016 |
badge | -0.014 |
Reward | -0.013 |
rifle | -0.013 |
threatening | -0.013 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.88e-01 |
Pos logit contributions | |
day | +0.191 |
afternoon | +0.189 |
ite | +0.180 |
olly | +0.173 |
cy | +0.169 |
Neg logit contributions | |
gru | -0.174 |
Reward | -0.143 |
Gu | -0.132 |
Fox | -0.131 |
probably | -0.129 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.34e-01 |
Pos logit contributions | |
gru | +0.023 |
probably | +0.019 |
stern | +0.019 |
Reward | +0.018 |
hello | +0.018 |
Neg logit contributions | |
afternoon | -0.030 |
days | -0.028 |
day | -0.028 |
ops | -0.027 |
cy | -0.027 |
Token | a |
Token ID | 257 |
Feature activation | -2.21e+00 |
Pos logit contributions | |
gru | +0.016 |
badge | +0.013 |
Reward | +0.012 |
threatening | +0.012 |
probably | +0.012 |
Neg logit contributions | |
day | -0.026 |
afternoon | -0.024 |
cy | -0.024 |
ite | -0.024 |
olly | -0.022 |
Token | nice |
Token ID | 3621 |
Feature activation | +2.04e-01 |
Pos logit contributions | |
day | +0.052 |
afternoon | +0.048 |
cy | +0.046 |
ite | +0.045 |
night | +0.044 |
Neg logit contributions | |
gru | -0.042 |
badge | -0.034 |
Reward | -0.033 |
threatening | -0.033 |
warning | -0.032 |
Token | . |
Token ID | 13 |
Feature activation | +3.81e-01 |
Pos logit contributions | |
day | +0.007 |
afternoon | +0.006 |
cy | +0.006 |
ite | +0.006 |
olly | +0.005 |
Neg logit contributions | |
gru | -0.009 |
badge | -0.008 |
threatening | -0.008 |
Reward | -0.008 |
warning | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -2.76e-01 |
Pos logit contributions | |
day | +0.016 |
afternoon | +0.015 |
cy | +0.015 |
ite | +0.015 |
he | +0.015 |
Neg logit contributions | |
gru | -0.013 |
Reward | -0.011 |
address | -0.010 |
badge | -0.010 |
rifle | -0.009 |
Token |
|
Token ID | 198 |
Feature activation | -2.92e-01 |
Pos logit contributions | |
gru | +0.010 |
Reward | +0.009 |
badge | +0.008 |
warning | +0.008 |
threatening | +0.008 |
Neg logit contributions | |
cy | -0.010 |
day | -0.010 |
afternoon | -0.010 |
ite | -0.010 |
night | -0.009 |
Token | The |
Token ID | 464 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
gru | +0.012 |
badge | +0.011 |
Reward | +0.010 |
ution | +0.010 |
rifle | +0.010 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.010 |
cy | -0.009 |
night | -0.009 |
ite | -0.009 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
cy | +0.039 |
day | +0.039 |
ite | +0.038 |
afternoon | +0.037 |
night | +0.034 |
Neg logit contributions | |
gru | -0.053 |
badge | -0.049 |
warning | -0.046 |
address | -0.044 |
Reward | -0.044 |
Token | . |
Token ID | 13 |
Feature activation | +3.30e-01 |
Pos logit contributions | |
day | +0.357 |
afternoon | +0.323 |
cy | +0.298 |
night | +0.292 |
ite | +0.286 |
Neg logit contributions | |
gru | -0.560 |
badge | -0.494 |
Reward | -0.475 |
threatening | -0.468 |
ution | -0.467 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.13e+00 |
Pos logit contributions | |
day | +0.011 |
afternoon | +0.010 |
cy | +0.010 |
ite | +0.009 |
night | +0.009 |
Neg logit contributions | |
gru | -0.014 |
Reward | -0.013 |
badge | -0.012 |
rifle | -0.012 |
address | -0.012 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.45e-01 |
Pos logit contributions | |
afternoon | +0.213 |
ite | +0.212 |
day | +0.211 |
olly | +0.211 |
cy | +0.199 |
Neg logit contributions | |
gru | -0.138 |
Reward | -0.113 |
Gu | -0.106 |
an | -0.098 |
Stan | -0.097 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.76e-01 |
Pos logit contributions | |
gru | +0.021 |
Reward | +0.017 |
hello | +0.017 |
stern | +0.017 |
probably | +0.017 |
Neg logit contributions | |
afternoon | -0.029 |
days | -0.027 |
ops | -0.026 |
day | -0.026 |
cy | -0.026 |
Token | a |
Token ID | 257 |
Feature activation | -2.22e+00 |
Pos logit contributions | |
gru | +0.016 |
badge | +0.013 |
Reward | +0.012 |
probably | +0.012 |
spoken | +0.012 |
Neg logit contributions | |
day | -0.028 |
cy | -0.027 |
ite | -0.027 |
afternoon | -0.027 |
olly | -0.026 |
Token | board |
Token ID | 3096 |
Feature activation | +3.76e+00 |
Pos logit contributions | |
gru | +0.043 |
badge | +0.038 |
warning | +0.037 |
ution | +0.035 |
rifle | +0.034 |
Neg logit contributions | |
cy | -0.044 |
ite | -0.041 |
he | -0.040 |
day | -0.039 |
olly | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | +3.31e-01 |
Pos logit contributions | |
day | +0.116 |
cy | +0.107 |
afternoon | +0.107 |
ite | +0.099 |
night | +0.098 |
Neg logit contributions | |
gru | -0.170 |
badge | -0.152 |
god | -0.143 |
threatening | -0.141 |
Reward | -0.141 |
Token |
|
Token ID | 198 |
Feature activation | -4.25e-01 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.012 |
cy | +0.010 |
night | +0.010 |
ite | +0.010 |
Neg logit contributions | |
gru | -0.013 |
Reward | -0.011 |
god | -0.011 |
badge | -0.011 |
address | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | -4.40e-01 |
Pos logit contributions | |
gru | +0.015 |
Reward | +0.013 |
god | +0.012 |
warning | +0.012 |
badge | +0.012 |
Neg logit contributions | |
day | -0.017 |
cy | -0.016 |
afternoon | -0.016 |
ite | -0.015 |
night | -0.015 |
Token | The |
Token ID | 464 |
Feature activation | +1.07e+00 |
Pos logit contributions | |
gru | +0.018 |
badge | +0.015 |
Reward | +0.015 |
god | +0.014 |
ution | +0.014 |
Neg logit contributions | |
day | -0.017 |
afternoon | -0.015 |
cy | -0.015 |
night | -0.015 |
ite | -0.014 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ite | +0.041 |
cy | +0.040 |
akes | +0.039 |
ites | +0.036 |
ops | +0.036 |
Neg logit contributions | |
gru | -0.041 |
warning | -0.038 |
message | -0.035 |
stern | -0.034 |
badge | -0.034 |
Token | . |
Token ID | 13 |
Feature activation | +2.32e-01 |
Pos logit contributions | |
day | +0.357 |
afternoon | +0.321 |
cy | +0.299 |
night | +0.292 |
ite | +0.288 |
Neg logit contributions | |
gru | -0.561 |
badge | -0.493 |
Reward | -0.474 |
threatening | -0.468 |
ution | -0.466 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.13e+00 |
Pos logit contributions | |
day | +0.008 |
afternoon | +0.007 |
cy | +0.007 |
night | +0.006 |
ite | +0.006 |
Neg logit contributions | |
gru | -0.010 |
Reward | -0.009 |
badge | -0.009 |
god | -0.009 |
address | -0.008 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.66e-01 |
Pos logit contributions | |
day | +0.196 |
afternoon | +0.195 |
ite | +0.188 |
olly | +0.186 |
days | +0.178 |
Neg logit contributions | |
gru | -0.156 |
Reward | -0.131 |
Gu | -0.124 |
probably | -0.115 |
Start | -0.114 |
Token | there |
Token ID | 612 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
gru | +0.022 |
stern | +0.018 |
hello | +0.018 |
probably | +0.017 |
Reward | +0.017 |
Neg logit contributions | |
afternoon | -0.030 |
days | -0.028 |
ops | -0.027 |
day | -0.027 |
akes | -0.027 |
Token | was |
Token ID | 373 |
Feature activation | +1.66e+00 |
Pos logit contributions | |
ite | +0.007 |
afternoon | +0.007 |
olly | +0.007 |
day | +0.007 |
days | +0.007 |
Neg logit contributions | |
gru | -0.009 |
probably | -0.008 |
spoken | -0.008 |
threatening | -0.008 |
ution | -0.007 |
Token | it |
Token ID | 340 |
Feature activation | +4.60e-01 |
Pos logit contributions | |
day | +0.141 |
afternoon | +0.122 |
night | +0.119 |
days | +0.115 |
morning | +0.110 |
Neg logit contributions | |
gru | -0.138 |
Reward | -0.119 |
badge | -0.119 |
rifle | -0.114 |
ution | -0.112 |
Token | . |
Token ID | 13 |
Feature activation | +4.44e-01 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.012 |
cy | +0.012 |
ite | +0.011 |
night | +0.011 |
Neg logit contributions | |
gru | -0.022 |
badge | -0.020 |
Reward | -0.019 |
threatening | -0.019 |
rifle | -0.019 |
Token |
|
Token ID | 198 |
Feature activation | -2.71e-01 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.017 |
cy | +0.015 |
night | +0.015 |
ite | +0.015 |
Neg logit contributions | |
gru | -0.017 |
Reward | -0.014 |
badge | -0.013 |
rifle | -0.013 |
Amir | -0.013 |
Token |
|
Token ID | 198 |
Feature activation | -2.87e-01 |
Pos logit contributions | |
gru | +0.010 |
Reward | +0.008 |
warning | +0.008 |
badge | +0.008 |
threatening | +0.008 |
Neg logit contributions | |
cy | -0.011 |
day | -0.010 |
afternoon | -0.010 |
ite | -0.010 |
night | -0.009 |
Token | The |
Token ID | 464 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
gru | +0.012 |
badge | +0.010 |
Reward | +0.010 |
ution | +0.010 |
threatening | +0.009 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.010 |
cy | -0.010 |
night | -0.009 |
ite | -0.009 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
cy | +0.038 |
olly | +0.037 |
ite | +0.036 |
he | +0.036 |
ops | +0.036 |
Neg logit contributions | |
gru | -0.045 |
warning | -0.038 |
badge | -0.037 |
message | -0.036 |
gen | -0.036 |
Token | . |
Token ID | 13 |
Feature activation | +3.19e-01 |
Pos logit contributions | |
day | +0.349 |
afternoon | +0.310 |
cy | +0.285 |
night | +0.283 |
ite | +0.274 |
Neg logit contributions | |
gru | -0.571 |
badge | -0.506 |
Reward | -0.488 |
ution | -0.480 |
rifle | -0.479 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.08e+00 |
Pos logit contributions | |
day | +0.011 |
afternoon | +0.010 |
cy | +0.009 |
night | +0.009 |
ite | +0.009 |
Neg logit contributions | |
gru | -0.014 |
Reward | -0.012 |
badge | -0.012 |
rifle | -0.012 |
god | -0.011 |
Token | Once |
Token ID | 7454 |
Feature activation | -6.97e-01 |
Pos logit contributions | |
day | +0.192 |
afternoon | +0.190 |
ite | +0.184 |
olly | +0.183 |
cy | +0.175 |
Neg logit contributions | |
gru | -0.158 |
Reward | -0.129 |
Gu | -0.121 |
probably | -0.116 |
Start | -0.113 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.30e-01 |
Pos logit contributions | |
gru | +0.021 |
stern | +0.017 |
hello | +0.016 |
probably | +0.016 |
Reward | +0.016 |
Neg logit contributions | |
afternoon | -0.027 |
days | -0.025 |
ops | -0.025 |
day | -0.024 |
cy | -0.024 |
Token | a |
Token ID | 257 |
Feature activation | -2.19e+00 |
Pos logit contributions | |
gru | +0.015 |
badge | +0.012 |
Reward | +0.012 |
probably | +0.011 |
spoken | +0.011 |
Neg logit contributions | |
day | -0.026 |
afternoon | -0.024 |
ite | -0.024 |
cy | -0.024 |
olly | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | +2.60e-01 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.005 |
cy | +0.005 |
night | +0.005 |
ite | +0.005 |
Neg logit contributions | |
gru | -0.009 |
badge | -0.008 |
Reward | -0.008 |
ution | -0.008 |
rifle | -0.008 |
Token | |
Token ID | 220 |
Feature activation | +4.86e-01 |
Pos logit contributions | |
day | +0.010 |
afternoon | +0.009 |
cy | +0.009 |
ite | +0.008 |
he | +0.008 |
Neg logit contributions | |
gru | -0.010 |
Reward | -0.009 |
badge | -0.008 |
address | -0.008 |
god | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -3.58e-01 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.018 |
night | +0.017 |
cy | +0.016 |
days | +0.016 |
Neg logit contributions | |
gru | -0.019 |
Reward | -0.016 |
badge | -0.016 |
address | -0.015 |
probably | -0.015 |
Token |
|
Token ID | 198 |
Feature activation | -3.78e-01 |
Pos logit contributions | |
gru | +0.014 |
Reward | +0.012 |
badge | +0.011 |
warning | +0.011 |
threatening | +0.011 |
Neg logit contributions | |
cy | -0.013 |
day | -0.013 |
ite | -0.012 |
afternoon | -0.012 |
night | -0.012 |
Token | The |
Token ID | 464 |
Feature activation | +1.20e+00 |
Pos logit contributions | |
gru | +0.017 |
badge | +0.015 |
Reward | +0.015 |
rifle | +0.015 |
ution | +0.015 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.012 |
night | -0.011 |
cy | -0.010 |
he | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
cy | +0.036 |
olly | +0.035 |
ogged | +0.033 |
he | +0.033 |
ite | +0.033 |
Neg logit contributions | |
gru | -0.052 |
warning | -0.048 |
god | -0.046 |
badge | -0.042 |
probably | -0.042 |
Token | . |
Token ID | 13 |
Feature activation | +2.02e-01 |
Pos logit contributions | |
day | +0.368 |
afternoon | +0.336 |
cy | +0.313 |
night | +0.303 |
ite | +0.298 |
Neg logit contributions | |
gru | -0.547 |
badge | -0.478 |
Reward | -0.461 |
threatening | -0.454 |
ution | -0.450 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.08e+00 |
Pos logit contributions | |
day | +0.007 |
afternoon | +0.006 |
cy | +0.006 |
night | +0.005 |
ite | +0.005 |
Neg logit contributions | |
gru | -0.009 |
Reward | -0.008 |
badge | -0.008 |
god | -0.007 |
rifle | -0.007 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.67e-01 |
Pos logit contributions | |
day | +0.202 |
afternoon | +0.201 |
ite | +0.200 |
olly | +0.195 |
cy | +0.187 |
Neg logit contributions | |
gru | -0.153 |
Reward | -0.126 |
Gu | -0.114 |
probably | -0.110 |
Barney | -0.108 |
Token | there |
Token ID | 612 |
Feature activation | +1.46e-01 |
Pos logit contributions | |
gru | +0.023 |
stern | +0.019 |
Reward | +0.019 |
probably | +0.019 |
hello | +0.019 |
Neg logit contributions | |
afternoon | -0.029 |
days | -0.027 |
day | -0.026 |
cy | -0.026 |
ops | -0.026 |
Token | was |
Token ID | 373 |
Feature activation | +1.48e+00 |
Pos logit contributions | |
afternoon | +0.005 |
olly | +0.005 |
ite | +0.005 |
day | +0.005 |
days | +0.004 |
Neg logit contributions | |
gru | -0.005 |
probably | -0.005 |
spoken | -0.005 |
threatening | -0.005 |
ution | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +3.65e-01 |
Pos logit contributions | |
gru | +0.022 |
badge | +0.020 |
god | +0.019 |
address | +0.019 |
warning | +0.019 |
Neg logit contributions | |
day | -0.016 |
afternoon | -0.015 |
cy | -0.014 |
ite | -0.014 |
olly | -0.014 |
Token | |
Token ID | 220 |
Feature activation | +5.38e-01 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.013 |
ite | +0.013 |
cy | +0.012 |
he | +0.012 |
Neg logit contributions | |
gru | -0.013 |
Reward | -0.011 |
address | -0.011 |
god | -0.010 |
probably | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | -3.27e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.020 |
night | +0.018 |
cy | +0.018 |
days | +0.018 |
Neg logit contributions | |
gru | -0.021 |
Reward | -0.018 |
badge | -0.018 |
address | -0.017 |
probably | -0.017 |
Token |
|
Token ID | 198 |
Feature activation | -3.42e-01 |
Pos logit contributions | |
gru | +0.012 |
Reward | +0.011 |
badge | +0.010 |
threatening | +0.010 |
warning | +0.010 |
Neg logit contributions | |
day | -0.012 |
cy | -0.012 |
ite | -0.012 |
afternoon | -0.011 |
night | -0.011 |
Token | The |
Token ID | 464 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
gru | +0.016 |
badge | +0.014 |
Reward | +0.013 |
rifle | +0.013 |
ution | +0.013 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.010 |
night | -0.010 |
cy | -0.010 |
he | -0.009 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
day | +0.031 |
afternoon | +0.027 |
cy | +0.027 |
ite | +0.025 |
night | +0.024 |
Neg logit contributions | |
gru | -0.065 |
badge | -0.058 |
Reward | -0.056 |
warning | -0.055 |
god | -0.055 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.14e+00 |
Pos logit contributions | |
day | +0.360 |
afternoon | +0.326 |
cy | +0.305 |
night | +0.295 |
ite | +0.292 |
Neg logit contributions | |
gru | -0.553 |
badge | -0.485 |
Reward | -0.468 |
threatening | -0.460 |
ution | -0.458 |
Token | Once |
Token ID | 7454 |
Feature activation | -8.18e-01 |
Pos logit contributions | |
day | +0.205 |
ite | +0.203 |
afternoon | +0.200 |
olly | +0.197 |
cy | +0.192 |
Neg logit contributions | |
gru | -0.153 |
Reward | -0.125 |
Gu | -0.121 |
Stan | -0.111 |
probably | -0.111 |
Token | there |
Token ID | 612 |
Feature activation | +2.59e-01 |
Pos logit contributions | |
gru | +0.022 |
stern | +0.018 |
probably | +0.018 |
hello | +0.018 |
Reward | +0.017 |
Neg logit contributions | |
afternoon | -0.032 |
days | -0.031 |
ops | -0.031 |
akes | -0.029 |
oline | -0.029 |
Token | was |
Token ID | 373 |
Feature activation | +1.55e+00 |
Pos logit contributions | |
ite | +0.008 |
olly | +0.008 |
afternoon | +0.008 |
day | +0.008 |
days | +0.007 |
Neg logit contributions | |
gru | -0.010 |
probably | -0.009 |
spoken | -0.008 |
ution | -0.008 |
threatening | -0.008 |
Token | a |
Token ID | 257 |
Feature activation | -2.24e+00 |
Pos logit contributions | |
ite | +0.077 |
day | +0.077 |
cy | +0.075 |
ops | +0.075 |
afternoon | +0.073 |
Neg logit contributions | |
gru | -0.032 |
badge | -0.026 |
probably | -0.024 |
an | -0.024 |
threatening | -0.023 |
Token | wild |
Token ID | 4295 |
Feature activation | -1.91e+00 |
Pos logit contributions | |
day | +0.114 |
afternoon | +0.103 |
cy | +0.096 |
night | +0.095 |
ite | +0.093 |
Neg logit contributions | |
gru | -0.138 |
badge | -0.121 |
Reward | -0.116 |
ution | -0.114 |
rifle | -0.114 |
Token | cube |
Token ID | 23441 |
Feature activation | -1.97e+00 |
Pos logit contributions | |
gru | +0.086 |
badge | +0.075 |
Reward | +0.071 |
rifle | +0.071 |
threatening | +0.071 |
Neg logit contributions | |
day | -0.065 |
afternoon | -0.060 |
cy | -0.055 |
night | -0.055 |
ite | -0.054 |
Token | again |
Token ID | 757 |
Feature activation | +5.47e+00 |
Pos logit contributions | |
gru | +0.095 |
badge | +0.083 |
threatening | +0.080 |
warning | +0.079 |
Reward | +0.079 |
Neg logit contributions | |
day | -0.058 |
afternoon | -0.054 |
cy | -0.050 |
ite | -0.048 |
night | -0.048 |
Token | . |
Token ID | 13 |
Feature activation | +3.67e-01 |
Pos logit contributions | |
day | +0.171 |
afternoon | +0.156 |
cy | +0.145 |
ite | +0.141 |
night | +0.141 |
Neg logit contributions | |
gru | -0.258 |
badge | -0.225 |
threatening | -0.217 |
Reward | -0.217 |
ution | -0.214 |
Token | The |
Token ID | 383 |
Feature activation | -5.83e-01 |
Pos logit contributions | |
day | +0.016 |
afternoon | +0.015 |
cy | +0.014 |
night | +0.014 |
ite | +0.014 |
Neg logit contributions | |
gru | -0.013 |
Reward | -0.011 |
badge | -0.011 |
rifle | -0.010 |
threatening | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
gru | +0.024 |
badge | +0.020 |
warning | +0.019 |
threatening | +0.019 |
rifle | +0.018 |
Neg logit contributions | |
day | -0.021 |
cy | -0.021 |
afternoon | -0.020 |
ite | -0.020 |
he | -0.020 |
Token | . |
Token ID | 13 |
Feature activation | +2.09e-01 |
Pos logit contributions | |
day | +0.309 |
afternoon | +0.257 |
night | +0.236 |
he | +0.225 |
cy | +0.221 |
Neg logit contributions | |
gru | -0.610 |
badge | -0.564 |
Reward | -0.544 |
ution | -0.538 |
rifle | -0.538 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.10e+00 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.006 |
cy | +0.005 |
night | +0.005 |
ite | +0.005 |
Neg logit contributions | |
gru | -0.010 |
badge | -0.009 |
Reward | -0.009 |
rifle | -0.009 |
ution | -0.009 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.57e-01 |
Pos logit contributions | |
day | +0.188 |
afternoon | +0.188 |
olly | +0.178 |
ite | +0.178 |
days | +0.169 |
Neg logit contributions | |
gru | -0.175 |
Reward | -0.145 |
probably | -0.130 |
Gu | -0.130 |
rifle | -0.128 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.46e-01 |
Pos logit contributions | |
gru | +0.024 |
stern | +0.019 |
Reward | +0.019 |
probably | +0.019 |
hello | +0.019 |
Neg logit contributions | |
afternoon | -0.029 |
days | -0.027 |
day | -0.026 |
cy | -0.025 |
ops | -0.025 |
Token | a |
Token ID | 257 |
Feature activation | -2.26e+00 |
Pos logit contributions | |
gru | +0.017 |
badge | +0.014 |
Reward | +0.013 |
threatening | +0.013 |
spoken | +0.013 |
Neg logit contributions | |
day | -0.026 |
afternoon | -0.024 |
cy | -0.023 |
ite | -0.023 |
night | -0.022 |
Token | Bob |
Token ID | 18861 |
Feature activation | -1.05e+00 |
Pos logit contributions | |
day | +0.106 |
afternoon | +0.100 |
ite | +0.095 |
he | +0.094 |
cy | +0.092 |
Neg logit contributions | |
gru | -0.128 |
badge | -0.120 |
god | -0.116 |
probably | -0.114 |
address | -0.114 |
Token | was |
Token ID | 373 |
Feature activation | +2.45e+00 |
Pos logit contributions | |
gru | +0.047 |
badge | +0.040 |
threatening | +0.038 |
rifle | +0.038 |
Reward | +0.038 |
Neg logit contributions | |
day | -0.036 |
afternoon | -0.033 |
ite | -0.030 |
he | -0.030 |
night | -0.030 |
Token | out |
Token ID | 503 |
Feature activation | -1.43e+00 |
Pos logit contributions | |
day | +0.097 |
afternoon | +0.088 |
night | +0.082 |
morning | +0.079 |
cy | +0.077 |
Neg logit contributions | |
gru | -0.098 |
badge | -0.088 |
Reward | -0.086 |
ution | -0.085 |
rifle | -0.082 |
Token | walking |
Token ID | 6155 |
Feature activation | +6.00e+00 |
Pos logit contributions | |
gru | +0.066 |
badge | +0.057 |
threatening | +0.055 |
Reward | +0.055 |
warning | +0.055 |
Neg logit contributions | |
day | -0.047 |
afternoon | -0.042 |
cy | -0.040 |
ite | -0.040 |
night | -0.039 |
Token | one |
Token ID | 530 |
Feature activation | -2.74e+00 |
Pos logit contributions | |
day | +0.187 |
afternoon | +0.171 |
cy | +0.160 |
ite | +0.156 |
olly | +0.156 |
Neg logit contributions | |
gru | -0.283 |
badge | -0.247 |
Reward | -0.234 |
threatening | -0.232 |
rifle | -0.232 |
Token | day |
Token ID | 1110 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
gru | +0.162 |
badge | +0.141 |
warning | +0.138 |
threatening | +0.137 |
god | +0.135 |
Neg logit contributions | |
ops | -0.042 |
cy | -0.036 |
day | -0.035 |
olly | -0.032 |
ite | -0.032 |
Token | , |
Token ID | 11 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
day | +0.322 |
afternoon | +0.285 |
cy | +0.265 |
night | +0.256 |
ite | +0.253 |
Neg logit contributions | |
gru | -0.593 |
badge | -0.528 |
Reward | -0.506 |
threatening | -0.502 |
rifle | -0.500 |
Token | searching |
Token ID | 10342 |
Feature activation | -2.53e+00 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.005 |
night | +0.005 |
cy | +0.005 |
ite | +0.004 |
Neg logit contributions | |
gru | -0.006 |
badge | -0.006 |
Reward | -0.005 |
ution | -0.005 |
rifle | -0.005 |
Token | for |
Token ID | 329 |
Feature activation | -5.69e-01 |
Pos logit contributions | |
gru | +0.117 |
badge | +0.104 |
Reward | +0.100 |
ution | +0.099 |
rifle | +0.098 |
Neg logit contributions | |
day | -0.084 |
afternoon | -0.075 |
night | -0.069 |
cy | -0.069 |
ite | -0.066 |
Token | something |
Token ID | 1223 |
Feature activation | -8.78e+00 |
Pos logit contributions | |
gru | +0.023 |
phant | +0.021 |
badge | +0.021 |
CRE | +0.021 |
sy | +0.021 |
Neg logit contributions | |
day | -0.022 |
days | -0.019 |
night | -0.019 |
afternoon | -0.019 |
morning | -0.017 |
Token | special |
Token ID | 2041 |
Feature activation | -4.20e+00 |
Pos logit contributions | |
gru | +0.377 |
badge | +0.339 |
Reward | +0.329 |
ution | +0.326 |
rifle | +0.318 |
Neg logit contributions | |
day | -0.342 |
afternoon | -0.292 |
night | -0.276 |
morning | -0.259 |
cy | -0.255 |
Token | liked |
Token ID | 8288 |
Feature activation | +5.15e-01 |
Pos logit contributions | |
day | +0.163 |
cy | +0.162 |
he | +0.158 |
ite | +0.155 |
afternoon | +0.151 |
Neg logit contributions | |
gru | -0.136 |
badge | -0.124 |
beard | -0.113 |
address | -0.106 |
rifle | -0.106 |
Token | to |
Token ID | 284 |
Feature activation | -4.32e+00 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.019 |
night | +0.017 |
cy | +0.017 |
morning | +0.017 |
Neg logit contributions | |
gru | -0.021 |
badge | -0.018 |
Reward | -0.017 |
ution | -0.017 |
rifle | -0.017 |
Token | go |
Token ID | 467 |
Feature activation | +8.97e+00 |
Pos logit contributions | |
gru | +0.201 |
badge | +0.178 |
Reward | +0.171 |
ution | +0.168 |
rifle | +0.168 |
Neg logit contributions | |
day | -0.142 |
afternoon | -0.127 |
cy | -0.117 |
night | -0.117 |
ite | -0.113 |
Token | outside |
Token ID | 2354 |
Feature activation | +5.35e+00 |
Pos logit contributions | |
day | +0.291 |
afternoon | +0.254 |
night | +0.237 |
cy | +0.231 |
morning | +0.224 |
Neg logit contributions | |
gru | -0.434 |
badge | -0.379 |
Reward | -0.367 |
ution | -0.365 |
rifle | -0.362 |
Token | every |
Token ID | 790 |
Feature activation | +6.09e+00 |
Pos logit contributions | |
day | +0.166 |
afternoon | +0.157 |
ite | +0.150 |
cy | +0.149 |
olly | +0.148 |
Neg logit contributions | |
gru | -0.232 |
badge | -0.202 |
Reward | -0.193 |
threatening | -0.193 |
warning | -0.191 |
Token | day |
Token ID | 1110 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ops | +0.095 |
olly | +0.081 |
cy | +0.077 |
ite | +0.072 |
ugg | +0.071 |
Neg logit contributions | |
gru | -0.366 |
warning | -0.323 |
threatening | -0.317 |
Reward | -0.317 |
Amen | -0.310 |
Token | . |
Token ID | 13 |
Feature activation | +7.68e-01 |
Pos logit contributions | |
day | +0.367 |
afternoon | +0.337 |
cy | +0.335 |
ite | +0.317 |
night | +0.302 |
Neg logit contributions | |
gru | -0.530 |
badge | -0.464 |
Reward | -0.446 |
threatening | -0.438 |
rifle | -0.434 |
Token | One |
Token ID | 1881 |
Feature activation | -2.08e+00 |
Pos logit contributions | |
day | +0.032 |
afternoon | +0.030 |
cy | +0.028 |
night | +0.028 |
ite | +0.027 |
Neg logit contributions | |
gru | -0.028 |
Reward | -0.023 |
badge | -0.023 |
rifle | -0.022 |
threatening | -0.021 |
Token | day |
Token ID | 1110 |
Feature activation | +1.10e+01 |
Pos logit contributions | |
gru | +0.112 |
warning | +0.097 |
oh | +0.097 |
probably | +0.096 |
message | +0.094 |
Neg logit contributions | |
ips | -0.038 |
ops | -0.035 |
ites | -0.028 |
he | -0.028 |
oms | -0.028 |
Token | he |
Token ID | 339 |
Feature activation | +9.93e-01 |
Pos logit contributions | |
day | +0.263 |
afternoon | +0.222 |
night | +0.199 |
cy | +0.195 |
morning | +0.183 |
Neg logit contributions | |
gru | -0.618 |
badge | -0.557 |
Reward | -0.540 |
ution | -0.537 |
rifle | -0.533 |
Token | heard |
Token ID | 2982 |
Feature activation | -2.29e+00 |
Pos logit contributions | |
day | +0.035 |
afternoon | +0.032 |
ite | +0.031 |
he | +0.030 |
cy | +0.030 |
Neg logit contributions | |
gru | -0.042 |
badge | -0.036 |
Reward | -0.035 |
spoken | -0.035 |
threatening | -0.035 |
Token | decoration |
Token ID | 42050 |
Feature activation | -3.16e+00 |
Pos logit contributions | |
gru | +0.071 |
badge | +0.062 |
Reward | +0.059 |
ution | +0.058 |
rifle | +0.058 |
Neg logit contributions | |
day | -0.061 |
afternoon | -0.056 |
night | -0.052 |
cy | -0.052 |
ite | -0.050 |
Token | . |
Token ID | 13 |
Feature activation | +3.63e-01 |
Pos logit contributions | |
gru | +0.139 |
badge | +0.124 |
god | +0.118 |
Reward | +0.115 |
ution | +0.115 |
Neg logit contributions | |
day | -0.103 |
afternoon | -0.095 |
cy | -0.090 |
ite | -0.087 |
night | -0.086 |
Token |
|
Token ID | 198 |
Feature activation | -4.48e-01 |
Pos logit contributions | |
day | +0.013 |
afternoon | +0.012 |
cy | +0.011 |
night | +0.011 |
ite | +0.011 |
Neg logit contributions | |
gru | -0.016 |
badge | -0.014 |
Reward | -0.013 |
ution | -0.013 |
rifle | -0.013 |
Token |
|
Token ID | 198 |
Feature activation | -4.64e-01 |
Pos logit contributions | |
gru | +0.020 |
badge | +0.016 |
Reward | +0.016 |
god | +0.016 |
warning | +0.016 |
Neg logit contributions | |
day | -0.015 |
afternoon | -0.014 |
cy | -0.014 |
ite | -0.013 |
night | -0.013 |
Token | The |
Token ID | 464 |
Feature activation | +1.11e+00 |
Pos logit contributions | |
gru | +0.018 |
badge | +0.016 |
rifle | +0.015 |
Reward | +0.015 |
ution | +0.015 |
Neg logit contributions | |
day | -0.019 |
afternoon | -0.017 |
night | -0.016 |
cy | -0.016 |
ite | -0.016 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
day | +0.042 |
ite | +0.040 |
afternoon | +0.039 |
cy | +0.039 |
he | +0.038 |
Neg logit contributions | |
gru | -0.044 |
badge | -0.037 |
warning | -0.036 |
god | -0.035 |
threatening | -0.035 |
Token | . |
Token ID | 13 |
Feature activation | +2.99e-01 |
Pos logit contributions | |
day | +0.350 |
afternoon | +0.309 |
night | +0.282 |
cy | +0.282 |
ite | +0.271 |
Neg logit contributions | |
gru | -0.570 |
badge | -0.510 |
Reward | -0.490 |
ution | -0.483 |
rifle | -0.483 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.01e+00 |
Pos logit contributions | |
day | +0.008 |
afternoon | +0.007 |
night | +0.007 |
cy | +0.007 |
ite | +0.006 |
Neg logit contributions | |
gru | -0.015 |
badge | -0.014 |
ution | -0.013 |
Reward | -0.013 |
rifle | -0.013 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.95e-01 |
Pos logit contributions | |
day | +0.184 |
afternoon | +0.181 |
ite | +0.170 |
olly | +0.168 |
cy | +0.163 |
Neg logit contributions | |
gru | -0.176 |
Reward | -0.148 |
Gu | -0.136 |
probably | -0.131 |
rifle | -0.131 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.69e-01 |
Pos logit contributions | |
gru | +0.024 |
probably | +0.020 |
stern | +0.020 |
Reward | +0.019 |
god | +0.019 |
Neg logit contributions | |
afternoon | -0.030 |
days | -0.028 |
day | -0.028 |
cy | -0.027 |
ops | -0.027 |
Token | a |
Token ID | 257 |
Feature activation | -2.26e+00 |
Pos logit contributions | |
gru | +0.018 |
badge | +0.014 |
Reward | +0.014 |
threatening | +0.013 |
spoken | +0.013 |
Neg logit contributions | |
day | -0.027 |
afternoon | -0.025 |
cy | -0.024 |
ite | -0.024 |
night | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | +2.82e-01 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.016 |
cy | +0.015 |
ite | +0.015 |
olly | +0.015 |
Neg logit contributions | |
gru | -0.025 |
badge | -0.022 |
threatening | -0.021 |
warning | -0.021 |
Reward | -0.020 |
Token | |
Token ID | 220 |
Feature activation | +5.49e-01 |
Pos logit contributions | |
day | +0.010 |
afternoon | +0.009 |
cy | +0.008 |
night | +0.008 |
ite | +0.008 |
Neg logit contributions | |
gru | -0.012 |
Reward | -0.010 |
badge | -0.010 |
god | -0.010 |
threatening | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | -2.99e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.020 |
night | +0.018 |
days | +0.017 |
cy | +0.017 |
Neg logit contributions | |
gru | -0.022 |
Reward | -0.019 |
badge | -0.019 |
probably | -0.018 |
address | -0.018 |
Token |
|
Token ID | 198 |
Feature activation | -3.29e-01 |
Pos logit contributions | |
gru | +0.011 |
Reward | +0.010 |
warning | +0.009 |
god | +0.009 |
badge | +0.009 |
Neg logit contributions | |
day | -0.011 |
cy | -0.011 |
afternoon | -0.010 |
ite | -0.010 |
night | -0.010 |
Token | The |
Token ID | 464 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.012 |
Reward | +0.011 |
ution | +0.011 |
god | +0.011 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.011 |
cy | -0.011 |
night | -0.011 |
ite | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ite | +0.028 |
he | +0.028 |
ops | +0.027 |
akes | +0.026 |
cy | +0.026 |
Neg logit contributions | |
gru | -0.057 |
warning | -0.051 |
badge | -0.049 |
god | -0.049 |
message | -0.048 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.14e+00 |
Pos logit contributions | |
day | +0.336 |
afternoon | +0.295 |
night | +0.269 |
cy | +0.267 |
ite | +0.257 |
Neg logit contributions | |
gru | -0.584 |
badge | -0.523 |
Reward | -0.504 |
ution | -0.497 |
rifle | -0.496 |
Token | One |
Token ID | 3198 |
Feature activation | -1.52e-01 |
Pos logit contributions | |
day | +0.193 |
afternoon | +0.187 |
ite | +0.186 |
olly | +0.180 |
cy | +0.177 |
Neg logit contributions | |
gru | -0.165 |
Reward | -0.136 |
Gu | -0.130 |
probably | -0.122 |
Start | -0.120 |
Token | day |
Token ID | 1110 |
Feature activation | +1.08e+01 |
Pos logit contributions | |
gru | +0.008 |
an | +0.007 |
warning | +0.007 |
probably | +0.007 |
message | +0.007 |
Neg logit contributions | |
ops | -0.003 |
ips | -0.003 |
unn | -0.002 |
ites | -0.002 |
akes | -0.002 |
Token | , |
Token ID | 11 |
Feature activation | +5.43e-03 |
Pos logit contributions | |
day | +0.277 |
afternoon | +0.246 |
cy | +0.230 |
ite | +0.224 |
night | +0.217 |
Neg logit contributions | |
gru | -0.566 |
badge | -0.501 |
Reward | -0.482 |
threatening | -0.478 |
rifle | -0.475 |
Token | a |
Token ID | 257 |
Feature activation | -2.23e+00 |
Pos logit contributions | |
day | +0.000 |
afternoon | +0.000 |
ite | +0.000 |
cy | +0.000 |
night | +0.000 |
Neg logit contributions | |
gru | -0.000 |
badge | -0.000 |
threatening | -0.000 |
Reward | -0.000 |
rifle | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | +3.83e-01 |
Pos logit contributions | |
gru | +0.107 |
badge | +0.092 |
Reward | +0.088 |
threatening | +0.088 |
ution | +0.086 |
Neg logit contributions | |
day | -0.099 |
afternoon | -0.091 |
cy | -0.086 |
ite | -0.084 |
night | -0.083 |
Token | |
Token ID | 220 |
Feature activation | +5.51e-01 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.013 |
cy | +0.012 |
night | +0.012 |
ite | +0.012 |
Neg logit contributions | |
gru | -0.015 |
Reward | -0.013 |
badge | -0.013 |
god | -0.012 |
address | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | -3.29e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.019 |
night | +0.018 |
days | +0.017 |
cy | +0.017 |
Neg logit contributions | |
gru | -0.023 |
Reward | -0.020 |
badge | -0.020 |
probably | -0.019 |
address | -0.019 |
Token |
|
Token ID | 198 |
Feature activation | -3.45e-01 |
Pos logit contributions | |
gru | +0.013 |
Reward | +0.011 |
badge | +0.011 |
god | +0.010 |
warning | +0.010 |
Neg logit contributions | |
day | -0.012 |
cy | -0.012 |
afternoon | -0.012 |
ite | -0.011 |
night | -0.011 |
Token | The |
Token ID | 464 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
gru | +0.015 |
badge | +0.013 |
Reward | +0.013 |
ution | +0.012 |
rifle | +0.012 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.011 |
cy | -0.011 |
night | -0.011 |
ite | -0.010 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
ite | +0.033 |
cy | +0.031 |
he | +0.031 |
ogged | +0.029 |
akes | +0.028 |
Neg logit contributions | |
gru | -0.062 |
warning | -0.055 |
message | -0.051 |
god | -0.051 |
threatening | -0.051 |
Token | . |
Token ID | 13 |
Feature activation | +3.76e-01 |
Pos logit contributions | |
day | +0.360 |
afternoon | +0.325 |
cy | +0.302 |
night | +0.295 |
ite | +0.289 |
Neg logit contributions | |
gru | -0.556 |
badge | -0.487 |
Reward | -0.470 |
threatening | -0.462 |
ution | -0.461 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.13e+00 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.010 |
cy | +0.010 |
night | +0.009 |
ite | +0.009 |
Neg logit contributions | |
gru | -0.018 |
badge | -0.016 |
Reward | -0.016 |
rifle | -0.015 |
god | -0.015 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.47e-01 |
Pos logit contributions | |
afternoon | +0.208 |
day | +0.207 |
ite | +0.205 |
olly | +0.204 |
cy | +0.192 |
Neg logit contributions | |
gru | -0.147 |
Reward | -0.120 |
Gu | -0.113 |
probably | -0.105 |
Stan | -0.103 |
Token | there |
Token ID | 612 |
Feature activation | +2.06e-01 |
Pos logit contributions | |
gru | +0.022 |
stern | +0.018 |
hello | +0.018 |
probably | +0.018 |
Reward | +0.018 |
Neg logit contributions | |
afternoon | -0.028 |
days | -0.027 |
day | -0.026 |
ops | -0.025 |
cy | -0.025 |
Token | was |
Token ID | 373 |
Feature activation | +1.56e+00 |
Pos logit contributions | |
afternoon | +0.007 |
olly | +0.006 |
ite | +0.006 |
day | +0.006 |
days | +0.006 |
Neg logit contributions | |
gru | -0.008 |
probably | -0.007 |
spoken | -0.007 |
threatening | -0.007 |
ution | -0.006 |
Token | . |
Token ID | 13 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
gru | +0.000 |
badge | +0.000 |
threatening | +0.000 |
Reward | +0.000 |
spoken | +0.000 |
Neg logit contributions | |
day | -0.000 |
afternoon | -0.000 |
ite | -0.000 |
olly | -0.000 |
cy | -0.000 |
Token | |
Token ID | 220 |
Feature activation | +4.58e-01 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.011 |
cy | +0.011 |
ite | +0.011 |
night | +0.011 |
Neg logit contributions | |
gru | -0.011 |
Reward | -0.009 |
badge | -0.009 |
address | -0.009 |
threatening | -0.009 |
Token |
|
Token ID | 198 |
Feature activation | -3.96e-01 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.017 |
night | +0.016 |
cy | +0.016 |
days | +0.016 |
Neg logit contributions | |
gru | -0.017 |
Reward | -0.014 |
probably | -0.014 |
badge | -0.014 |
address | -0.014 |
Token |
|
Token ID | 198 |
Feature activation | -4.10e-01 |
Pos logit contributions | |
gru | +0.015 |
Reward | +0.013 |
badge | +0.012 |
threatening | +0.012 |
stern | +0.012 |
Neg logit contributions | |
day | -0.015 |
cy | -0.015 |
afternoon | -0.014 |
ite | -0.014 |
night | -0.013 |
Token | The |
Token ID | 464 |
Feature activation | +1.16e+00 |
Pos logit contributions | |
gru | +0.015 |
badge | +0.012 |
Reward | +0.012 |
threatening | +0.011 |
ution | +0.011 |
Neg logit contributions | |
day | -0.017 |
cy | -0.016 |
afternoon | -0.016 |
night | -0.015 |
ite | -0.015 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
cy | +0.029 |
ite | +0.028 |
olly | +0.026 |
he | +0.026 |
day | +0.025 |
Neg logit contributions | |
gru | -0.059 |
warning | -0.053 |
badge | -0.050 |
message | -0.050 |
threatening | -0.049 |
Token | . |
Token ID | 13 |
Feature activation | +2.66e-01 |
Pos logit contributions | |
day | +0.357 |
afternoon | +0.321 |
cy | +0.299 |
night | +0.293 |
ite | +0.288 |
Neg logit contributions | |
gru | -0.559 |
badge | -0.491 |
Reward | -0.472 |
threatening | -0.466 |
ution | -0.465 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.11e+00 |
Pos logit contributions | |
day | +0.009 |
afternoon | +0.008 |
cy | +0.008 |
ite | +0.007 |
night | +0.007 |
Neg logit contributions | |
gru | -0.012 |
Reward | -0.010 |
badge | -0.010 |
address | -0.010 |
god | -0.010 |
Token | Once |
Token ID | 7454 |
Feature activation | -8.26e-01 |
Pos logit contributions | |
day | +0.190 |
afternoon | +0.189 |
ite | +0.181 |
olly | +0.179 |
cy | +0.172 |
Neg logit contributions | |
gru | -0.169 |
Reward | -0.144 |
Gu | -0.131 |
probably | -0.126 |
Start | -0.123 |
Token | upon |
Token ID | 2402 |
Feature activation | -6.01e-01 |
Pos logit contributions | |
gru | +0.023 |
stern | +0.020 |
probably | +0.019 |
hello | +0.018 |
Reward | +0.018 |
Neg logit contributions | |
afternoon | -0.032 |
days | -0.030 |
ops | -0.030 |
cy | -0.029 |
akes | -0.029 |
Token | a |
Token ID | 257 |
Feature activation | -2.26e+00 |
Pos logit contributions | |
gru | +0.016 |
badge | +0.013 |
Reward | +0.012 |
probably | +0.012 |
spoken | +0.012 |
Neg logit contributions | |
day | -0.030 |
ite | -0.029 |
cy | -0.028 |
afternoon | -0.028 |
olly | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | +5.52e-01 |
Pos logit contributions | |
day | +0.058 |
afternoon | +0.050 |
night | +0.048 |
morning | +0.046 |
days | +0.044 |
Neg logit contributions | |
gru | -0.078 |
badge | -0.070 |
Reward | -0.067 |
rifle | -0.065 |
ution | -0.065 |
Token | |
Token ID | 220 |
Feature activation | +4.04e-01 |
Pos logit contributions | |
day | +0.018 |
afternoon | +0.017 |
cy | +0.016 |
ite | +0.016 |
night | +0.015 |
Neg logit contributions | |
gru | -0.024 |
Reward | -0.021 |
badge | -0.021 |
threatening | -0.020 |
ution | -0.020 |
Token |
|
Token ID | 198 |
Feature activation | -1.60e-01 |
Pos logit contributions | |
day | +0.015 |
afternoon | +0.014 |
night | +0.012 |
cy | +0.012 |
days | +0.012 |
Neg logit contributions | |
gru | -0.017 |
Reward | -0.015 |
badge | -0.015 |
rifle | -0.014 |
threatening | -0.014 |
Token |
|
Token ID | 198 |
Feature activation | -1.78e-01 |
Pos logit contributions | |
gru | +0.006 |
badge | +0.005 |
Reward | +0.005 |
warning | +0.005 |
threatening | +0.005 |
Neg logit contributions | |
day | -0.006 |
cy | -0.006 |
afternoon | -0.005 |
ite | -0.005 |
night | -0.005 |
Token | The |
Token ID | 464 |
Feature activation | +1.18e+00 |
Pos logit contributions | |
gru | +0.007 |
badge | +0.006 |
Reward | +0.006 |
ution | +0.006 |
rifle | +0.006 |
Neg logit contributions | |
day | -0.007 |
afternoon | -0.006 |
cy | -0.006 |
night | -0.006 |
ite | -0.006 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
day | +0.037 |
cy | +0.035 |
ite | +0.034 |
afternoon | +0.034 |
he | +0.034 |
Neg logit contributions | |
gru | -0.054 |
badge | -0.045 |
warning | -0.045 |
Reward | -0.044 |
threatening | -0.044 |
Token | . |
Token ID | 13 |
Feature activation | +3.92e-01 |
Pos logit contributions | |
day | +0.365 |
afternoon | +0.335 |
cy | +0.314 |
night | +0.303 |
ite | +0.298 |
Neg logit contributions | |
gru | -0.543 |
badge | -0.471 |
Reward | -0.456 |
threatening | -0.450 |
ution | -0.447 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.18e+00 |
Pos logit contributions | |
day | +0.013 |
afternoon | +0.012 |
cy | +0.011 |
night | +0.011 |
ite | +0.010 |
Neg logit contributions | |
gru | -0.017 |
Reward | -0.015 |
badge | -0.015 |
threatening | -0.014 |
rifle | -0.014 |
Token | Once |
Token ID | 7454 |
Feature activation | -7.53e-01 |
Pos logit contributions | |
day | +0.210 |
ite | +0.210 |
afternoon | +0.208 |
olly | +0.203 |
cy | +0.199 |
Neg logit contributions | |
gru | -0.151 |
Reward | -0.123 |
probably | -0.113 |
Gu | -0.113 |
spoken | -0.110 |
Token | upon |
Token ID | 2402 |
Feature activation | -6.91e-01 |
Pos logit contributions | |
gru | +0.022 |
stern | +0.019 |
hello | +0.019 |
Reward | +0.019 |
probably | +0.018 |
Neg logit contributions | |
afternoon | -0.029 |
days | -0.027 |
day | -0.026 |
ops | -0.025 |
cy | -0.025 |
Token | a |
Token ID | 257 |
Feature activation | -2.10e+00 |
Pos logit contributions | |
gru | +0.019 |
badge | +0.016 |
Reward | +0.015 |
probably | +0.015 |
spoken | +0.015 |
Neg logit contributions | |
day | -0.034 |
afternoon | -0.032 |
cy | -0.031 |
ite | -0.031 |
olly | -0.030 |
Token | he |
Token ID | 339 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
day | +0.033 |
afternoon | +0.031 |
cy | +0.029 |
ite | +0.029 |
night | +0.028 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.024 |
Reward | -0.023 |
threatening | -0.022 |
ution | -0.022 |
Token | was |
Token ID | 373 |
Feature activation | +1.45e+00 |
Pos logit contributions | |
day | +0.052 |
afternoon | +0.050 |
ite | +0.046 |
he | +0.045 |
cy | +0.045 |
Neg logit contributions | |
gru | -0.055 |
probably | -0.053 |
badge | -0.051 |
spoken | -0.051 |
god | -0.049 |
Token | safe |
Token ID | 3338 |
Feature activation | +2.18e+00 |
Pos logit contributions | |
day | +0.056 |
afternoon | +0.051 |
night | +0.048 |
cy | +0.047 |
ite | +0.046 |
Neg logit contributions | |
gru | -0.060 |
badge | -0.052 |
Reward | -0.050 |
ution | -0.049 |
rifle | -0.049 |
Token | . |
Token ID | 13 |
Feature activation | +5.70e-01 |
Pos logit contributions | |
day | +0.065 |
afternoon | +0.058 |
cy | +0.053 |
night | +0.052 |
ite | +0.051 |
Neg logit contributions | |
gru | -0.107 |
badge | -0.096 |
Reward | -0.091 |
threatening | -0.091 |
rifle | -0.090 |
Token | The |
Token ID | 383 |
Feature activation | -3.61e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.020 |
cy | +0.019 |
ite | +0.018 |
night | +0.018 |
Neg logit contributions | |
gru | -0.022 |
Reward | -0.020 |
badge | -0.018 |
address | -0.018 |
rifle | -0.018 |
Token | End |
Token ID | 5268 |
Feature activation | +1.16e+01 |
Pos logit contributions | |
gru | +0.017 |
badge | +0.016 |
warning | +0.015 |
threatening | +0.014 |
rifle | +0.014 |
Neg logit contributions | |
day | -0.011 |
cy | -0.010 |
afternoon | -0.010 |
ite | -0.010 |
olly | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | +4.02e-01 |
Pos logit contributions | |
day | +0.320 |
afternoon | +0.274 |
night | +0.251 |
cy | +0.246 |
he | +0.238 |
Neg logit contributions | |
gru | -0.600 |
badge | -0.543 |
Reward | -0.524 |
ution | -0.518 |
rifle | -0.517 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.19e+00 |
Pos logit contributions | |
day | +0.013 |
afternoon | +0.012 |
cy | +0.012 |
ite | +0.011 |
night | +0.011 |
Neg logit contributions | |
gru | -0.018 |
Reward | -0.016 |
badge | -0.015 |
threatening | -0.015 |
god | -0.015 |
Token | M |
Token ID | 44 |
Feature activation | -1.24e+00 |
Pos logit contributions | |
ite | +0.216 |
afternoon | +0.215 |
day | +0.212 |
olly | +0.205 |
cy | +0.202 |
Neg logit contributions | |
gru | -0.153 |
Reward | -0.123 |
probably | -0.109 |
threatening | -0.108 |
badge | -0.107 |
Token | olly |
Token ID | 5098 |
Feature activation | +6.28e+00 |
Pos logit contributions | |
gru | +0.043 |
god | +0.038 |
Reward | +0.037 |
message | +0.036 |
badge | +0.036 |
Neg logit contributions | |
cy | -0.045 |
day | -0.044 |
afternoon | -0.043 |
ite | -0.041 |
days | -0.039 |
Token | was |
Token ID | 373 |
Feature activation | +1.64e+00 |
Pos logit contributions | |
day | +0.197 |
afternoon | +0.187 |
ite | +0.175 |
he | +0.173 |
End | +0.172 |
Neg logit contributions | |
gru | -0.253 |
badge | -0.234 |
address | -0.231 |
probably | -0.224 |
god | -0.223 |
Token | . |
Token ID | 13 |
Feature activation | +3.36e-01 |
Pos logit contributions | |
gru | +0.232 |
badge | +0.206 |
Reward | +0.197 |
ution | +0.195 |
threatening | +0.194 |
Neg logit contributions | |
day | -0.150 |
afternoon | -0.134 |
cy | -0.124 |
night | -0.122 |
ite | -0.120 |
Token | |
Token ID | 220 |
Feature activation | +5.12e-01 |
Pos logit contributions | |
day | +0.015 |
afternoon | +0.014 |
cy | +0.014 |
ite | +0.014 |
ops | +0.013 |
Neg logit contributions | |
gru | -0.009 |
Reward | -0.008 |
address | -0.007 |
probably | -0.007 |
god | -0.007 |
Token |
|
Token ID | 198 |
Feature activation | -3.44e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.020 |
night | +0.018 |
days | +0.018 |
cy | +0.017 |
Neg logit contributions | |
gru | -0.019 |
Reward | -0.016 |
probably | -0.016 |
address | -0.016 |
badge | -0.015 |
Token |
|
Token ID | 198 |
Feature activation | -3.54e-01 |
Pos logit contributions | |
gru | +0.012 |
Reward | +0.010 |
warning | +0.010 |
god | +0.009 |
badge | +0.009 |
Neg logit contributions | |
cy | -0.013 |
day | -0.013 |
afternoon | -0.013 |
night | -0.012 |
ite | -0.012 |
Token | The |
Token ID | 464 |
Feature activation | +1.15e+00 |
Pos logit contributions | |
gru | +0.014 |
Reward | +0.011 |
badge | +0.011 |
god | +0.011 |
warning | +0.011 |
Neg logit contributions | |
day | -0.014 |
afternoon | -0.013 |
cy | -0.013 |
night | -0.012 |
ite | -0.012 |
Token | End |
Token ID | 5268 |
Feature activation | +1.15e+01 |
Pos logit contributions | |
ite | +0.040 |
cy | +0.040 |
ogged | +0.037 |
akes | +0.037 |
ops | +0.037 |
Neg logit contributions | |
gru | -0.046 |
warning | -0.040 |
god | -0.039 |
badge | -0.039 |
address | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | +3.02e-01 |
Pos logit contributions | |
day | +0.364 |
afternoon | +0.330 |
cy | +0.307 |
night | +0.299 |
ite | +0.295 |
Neg logit contributions | |
gru | -0.547 |
badge | -0.479 |
Reward | -0.462 |
threatening | -0.454 |
ution | -0.453 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.08e+00 |
Pos logit contributions | |
day | +0.010 |
afternoon | +0.010 |
cy | +0.009 |
ite | +0.009 |
night | +0.009 |
Neg logit contributions | |
gru | -0.012 |
Reward | -0.011 |
god | -0.010 |
badge | -0.010 |
probably | -0.010 |
Token | John |
Token ID | 7554 |
Feature activation | -4.94e-01 |
Pos logit contributions | |
day | +0.203 |
afternoon | +0.201 |
ite | +0.198 |
olly | +0.198 |
End | +0.186 |
Neg logit contributions | |
gru | -0.148 |
Reward | -0.121 |
Gu | -0.115 |
probably | -0.104 |
Start | -0.103 |
Token | was |
Token ID | 373 |
Feature activation | +1.51e+00 |
Pos logit contributions | |
probably | +0.017 |
address | +0.016 |
beard | +0.016 |
badge | +0.016 |
gru | +0.016 |
Neg logit contributions | |
day | -0.016 |
afternoon | -0.016 |
olly | -0.016 |
he | -0.015 |
ite | -0.015 |
Token | weak |
Token ID | 4939 |
Feature activation | -8.02e-01 |
Pos logit contributions | |
day | +0.054 |
afternoon | +0.049 |
cy | +0.046 |
night | +0.045 |
ite | +0.045 |
Neg logit contributions | |
gru | -0.065 |
badge | -0.057 |
Reward | -0.054 |
threatening | -0.053 |
rifle | -0.053 |
Token | We |
Token ID | 775 |
Feature activation | -8.51e-02 |
Pos logit contributions | |
day | +0.018 |
afternoon | +0.016 |
cy | +0.015 |
night | +0.015 |
ite | +0.015 |
Neg logit contributions | |
gru | -0.023 |
badge | -0.020 |
ution | -0.019 |
threatening | -0.019 |
spoken | -0.019 |
Token | have |
Token ID | 423 |
Feature activation | +5.61e+00 |
Pos logit contributions | |
gru | +0.004 |
badge | +0.004 |
Reward | +0.004 |
ution | +0.004 |
rifle | +0.004 |
Neg logit contributions | |
day | -0.003 |
afternoon | -0.002 |
cy | -0.002 |
night | -0.002 |
ite | -0.002 |
Token | different |
Token ID | 1180 |
Feature activation | -6.40e+00 |
Pos logit contributions | |
day | +0.230 |
afternoon | +0.209 |
night | +0.198 |
cy | +0.193 |
ite | +0.189 |
Neg logit contributions | |
gru | -0.221 |
badge | -0.186 |
Reward | -0.180 |
ution | -0.176 |
rifle | -0.176 |
Token | likes |
Token ID | 7832 |
Feature activation | -1.41e+00 |
Pos logit contributions | |
gru | +0.250 |
badge | +0.218 |
Reward | +0.206 |
rifle | +0.202 |
ution | +0.202 |
Neg logit contributions | |
day | -0.254 |
afternoon | -0.233 |
cy | -0.221 |
night | -0.217 |
ite | -0.215 |
Token | and |
Token ID | 290 |
Feature activation | +1.88e+00 |
Pos logit contributions | |
gru | +0.054 |
badge | +0.048 |
Reward | +0.045 |
ution | +0.044 |
rifle | +0.044 |
Neg logit contributions | |
day | -0.056 |
afternoon | -0.052 |
cy | -0.049 |
night | -0.048 |
ite | -0.048 |
Token | disl |
Token ID | 19621 |
Feature activation | -2.15e+01 |
Pos logit contributions | |
day | +0.075 |
afternoon | +0.068 |
cy | +0.065 |
night | +0.064 |
ite | +0.063 |
Neg logit contributions | |
gru | -0.074 |
badge | -0.064 |
Reward | -0.061 |
ution | -0.060 |
rifle | -0.060 |
Token | ikes |
Token ID | 7938 |
Feature activation | -3.73e+00 |
Pos logit contributions | |
gru | +0.683 |
badge | +0.585 |
ution | +0.556 |
Reward | +0.546 |
address | +0.535 |
Neg logit contributions | |
day | -0.978 |
afternoon | -0.909 |
night | -0.864 |
cy | -0.859 |
ite | -0.840 |
Token | . |
Token ID | 13 |
Feature activation | +5.07e-01 |
Pos logit contributions | |
gru | +0.133 |
badge | +0.125 |
Reward | +0.118 |
warning | +0.112 |
address | +0.112 |
Neg logit contributions | |
day | -0.133 |
cy | -0.132 |
afternoon | -0.131 |
olly | -0.123 |
ugg | -0.120 |
Token | We |
Token ID | 775 |
Feature activation | -9.93e-02 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.015 |
cy | +0.014 |
ite | +0.014 |
night | +0.014 |
Neg logit contributions | |
gru | -0.023 |
badge | -0.020 |
spoken | -0.020 |
ution | -0.019 |
threatening | -0.019 |
Token | can |
Token ID | 460 |
Feature activation | +8.75e-01 |
Pos logit contributions | |
gru | +0.005 |
badge | +0.004 |
Reward | +0.004 |
rifle | +0.004 |
ution | +0.004 |
Neg logit contributions | |
day | -0.003 |
afternoon | -0.002 |
cy | -0.002 |
night | -0.002 |
ite | -0.002 |
Token | be |
Token ID | 307 |
Feature activation | +4.00e-02 |
Pos logit contributions | |
day | +0.030 |
afternoon | +0.025 |
night | +0.025 |
he | +0.025 |
cy | +0.024 |
Neg logit contributions | |
gru | -0.043 |
badge | -0.035 |
Reward | -0.034 |
rifle | -0.034 |
threatening | -0.034 |
Token | air |
Token ID | 1633 |
Feature activation | +2.56e-01 |
Pos logit contributions | |
gru | +0.143 |
badge | +0.128 |
ution | +0.125 |
rifle | +0.123 |
Reward | +0.121 |
Neg logit contributions | |
day | -0.126 |
night | -0.110 |
afternoon | -0.108 |
he | -0.103 |
morning | -0.102 |
Token | . |
Token ID | 13 |
Feature activation | +9.85e-01 |
Pos logit contributions | |
day | +0.008 |
afternoon | +0.007 |
cy | +0.007 |
ite | +0.007 |
olly | +0.007 |
Neg logit contributions | |
gru | -0.012 |
badge | -0.011 |
threatening | -0.010 |
rifle | -0.010 |
Reward | -0.010 |
Token | He |
Token ID | 679 |
Feature activation | -3.49e+00 |
Pos logit contributions | |
day | +0.039 |
afternoon | +0.035 |
cy | +0.033 |
night | +0.033 |
ite | +0.032 |
Neg logit contributions | |
gru | -0.039 |
badge | -0.034 |
Reward | -0.033 |
rifle | -0.031 |
ution | -0.031 |
Token | watched |
Token ID | 7342 |
Feature activation | -7.80e-01 |
Pos logit contributions | |
gru | +0.157 |
badge | +0.138 |
Reward | +0.133 |
ution | +0.131 |
rifle | +0.130 |
Neg logit contributions | |
day | -0.120 |
afternoon | -0.108 |
night | -0.100 |
cy | -0.100 |
ite | -0.096 |
Token | with |
Token ID | 351 |
Feature activation | -2.95e+00 |
Pos logit contributions | |
gru | +0.031 |
badge | +0.027 |
Reward | +0.026 |
ution | +0.025 |
rifle | +0.025 |
Neg logit contributions | |
day | -0.031 |
afternoon | -0.028 |
night | -0.026 |
cy | -0.026 |
ite | -0.025 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.73e+01 |
Pos logit contributions | |
gru | +0.117 |
badge | +0.101 |
Reward | +0.097 |
ution | +0.095 |
rifle | +0.094 |
Neg logit contributions | |
day | -0.118 |
afternoon | -0.107 |
night | -0.102 |
cy | -0.100 |
morning | -0.097 |
Token | ement |
Token ID | 972 |
Feature activation | -2.32e+00 |
Pos logit contributions | |
gru | +0.759 |
badge | +0.668 |
Reward | +0.654 |
rifle | +0.638 |
threatening | +0.630 |
Neg logit contributions | |
day | -0.613 |
afternoon | -0.552 |
night | -0.513 |
cy | -0.511 |
ite | -0.501 |
Token | as |
Token ID | 355 |
Feature activation | +2.19e+00 |
Pos logit contributions | |
gru | +0.107 |
badge | +0.096 |
threatening | +0.090 |
rifle | +0.090 |
Reward | +0.089 |
Neg logit contributions | |
day | -0.074 |
afternoon | -0.068 |
cy | -0.064 |
ite | -0.061 |
he | -0.061 |
Token | the |
Token ID | 262 |
Feature activation | +2.66e+00 |
Pos logit contributions | |
day | +0.089 |
afternoon | +0.082 |
cy | +0.079 |
ite | +0.077 |
night | +0.076 |
Neg logit contributions | |
gru | -0.084 |
badge | -0.072 |
Reward | -0.068 |
threatening | -0.067 |
rifle | -0.066 |
Token | z |
Token ID | 1976 |
Feature activation | +2.77e+00 |
Pos logit contributions | |
day | +0.107 |
afternoon | +0.099 |
cy | +0.095 |
night | +0.092 |
ite | +0.091 |
Neg logit contributions | |
gru | -0.103 |
badge | -0.088 |
Reward | -0.083 |
rifle | -0.082 |
threatening | -0.081 |
Token | ig |
Token ID | 328 |
Feature activation | +5.48e+00 |
Pos logit contributions | |
day | +0.074 |
afternoon | +0.065 |
night | +0.058 |
cy | +0.053 |
he | +0.052 |
Neg logit contributions | |
gru | -0.146 |
badge | -0.140 |
ution | -0.129 |
god | -0.127 |
phant | -0.126 |
Token | . |
Token ID | 13 |
Feature activation | +3.19e-01 |
Pos logit contributions | |
day | +0.030 |
afternoon | +0.028 |
cy | +0.027 |
he | +0.026 |
ite | +0.026 |
Neg logit contributions | |
gru | -0.024 |
badge | -0.021 |
threatening | -0.020 |
spoken | -0.020 |
warning | -0.019 |
Token | Tim |
Token ID | 5045 |
Feature activation | +2.41e+00 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.011 |
cy | +0.011 |
night | +0.011 |
ite | +0.010 |
Neg logit contributions | |
gru | -0.013 |
badge | -0.011 |
Reward | -0.011 |
rifle | -0.010 |
ution | -0.010 |
Token | my |
Token ID | 1820 |
Feature activation | +4.78e+00 |
Pos logit contributions | |
day | +0.079 |
ite | +0.077 |
afternoon | +0.073 |
days | +0.073 |
olly | +0.070 |
Neg logit contributions | |
gru | -0.089 |
probably | -0.085 |
badge | -0.082 |
threatening | -0.080 |
god | -0.078 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.25e-01 |
Pos logit contributions | |
day | +0.168 |
afternoon | +0.155 |
ite | +0.148 |
cy | +0.145 |
he | +0.144 |
Neg logit contributions | |
gru | -0.198 |
badge | -0.179 |
probably | -0.169 |
threatening | -0.167 |
address | -0.166 |
Token | with |
Token ID | 351 |
Feature activation | -2.52e+00 |
Pos logit contributions | |
gru | +0.012 |
badge | +0.011 |
Reward | +0.010 |
ution | +0.010 |
rifle | +0.010 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.012 |
night | -0.012 |
cy | -0.012 |
ite | -0.011 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.70e+01 |
Pos logit contributions | |
gru | +0.105 |
badge | +0.091 |
Reward | +0.088 |
ution | +0.086 |
rifle | +0.086 |
Neg logit contributions | |
day | -0.096 |
afternoon | -0.087 |
night | -0.082 |
cy | -0.081 |
ite | -0.078 |
Token | ement |
Token ID | 972 |
Feature activation | -2.24e+00 |
Pos logit contributions | |
gru | +0.750 |
badge | +0.656 |
Reward | +0.645 |
rifle | +0.630 |
threatening | +0.620 |
Neg logit contributions | |
day | -0.605 |
afternoon | -0.545 |
night | -0.505 |
cy | -0.501 |
he | -0.494 |
Token | as |
Token ID | 355 |
Feature activation | +2.11e+00 |
Pos logit contributions | |
gru | +0.101 |
badge | +0.090 |
threatening | +0.085 |
Reward | +0.084 |
rifle | +0.084 |
Neg logit contributions | |
day | -0.073 |
afternoon | -0.067 |
cy | -0.064 |
ite | -0.061 |
night | -0.060 |
Token | he |
Token ID | 339 |
Feature activation | +4.89e-01 |
Pos logit contributions | |
day | +0.088 |
afternoon | +0.083 |
cy | +0.080 |
ite | +0.077 |
olly | +0.076 |
Neg logit contributions | |
gru | -0.077 |
badge | -0.066 |
Reward | -0.061 |
threatening | -0.060 |
rifle | -0.059 |
Token | saw |
Token ID | 2497 |
Feature activation | -4.08e+00 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.017 |
he | +0.016 |
cy | +0.016 |
olly | +0.016 |
Neg logit contributions | |
gru | -0.019 |
badge | -0.017 |
god | -0.017 |
spoken | -0.016 |
ution | -0.016 |
Token | stars |
Token ID | 5788 |
Feature activation | +9.20e-01 |
Pos logit contributions | |
gru | +0.159 |
badge | +0.136 |
Reward | +0.134 |
ution | +0.131 |
rifle | +0.130 |
Neg logit contributions | |
day | -0.165 |
afternoon | -0.149 |
night | -0.142 |
cy | -0.135 |
morning | -0.135 |
Token | go |
Token ID | 467 |
Feature activation | +8.99e+00 |
Pos logit contributions | |
gru | +0.229 |
badge | +0.199 |
Reward | +0.192 |
ution | +0.191 |
rifle | +0.189 |
Neg logit contributions | |
day | -0.172 |
afternoon | -0.149 |
night | -0.142 |
morning | -0.135 |
cy | -0.134 |
Token | ! |
Token ID | 0 |
Feature activation | +1.47e-01 |
Pos logit contributions | |
day | +0.323 |
afternoon | +0.296 |
cy | +0.275 |
night | +0.271 |
ite | +0.266 |
Neg logit contributions | |
gru | -0.387 |
badge | -0.342 |
Reward | -0.325 |
rifle | -0.319 |
threatening | -0.317 |
Token | Everyone |
Token ID | 11075 |
Feature activation | -7.44e-01 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.005 |
night | +0.005 |
cy | +0.005 |
ite | +0.005 |
Neg logit contributions | |
gru | -0.006 |
badge | -0.005 |
Reward | -0.005 |
rifle | -0.005 |
threatening | -0.005 |
Token | watched |
Token ID | 7342 |
Feature activation | -1.85e-01 |
Pos logit contributions | |
gru | +0.034 |
badge | +0.029 |
threatening | +0.028 |
spoken | +0.028 |
probably | +0.028 |
Neg logit contributions | |
day | -0.023 |
afternoon | -0.022 |
cy | -0.020 |
olly | -0.020 |
ite | -0.020 |
Token | in |
Token ID | 287 |
Feature activation | +4.79e+00 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
Reward | +0.007 |
threatening | +0.007 |
rifle | +0.007 |
Neg logit contributions | |
day | -0.007 |
afternoon | -0.006 |
cy | -0.006 |
ite | -0.006 |
night | -0.005 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.70e+01 |
Pos logit contributions | |
day | +0.170 |
afternoon | +0.150 |
night | +0.143 |
morning | +0.134 |
cy | +0.131 |
Neg logit contributions | |
gru | -0.216 |
badge | -0.191 |
Reward | -0.186 |
rifle | -0.182 |
ution | -0.182 |
Token | ement |
Token ID | 972 |
Feature activation | -2.08e+00 |
Pos logit contributions | |
gru | +0.686 |
badge | +0.594 |
Reward | +0.562 |
ution | +0.558 |
threatening | +0.550 |
Neg logit contributions | |
day | -0.659 |
afternoon | -0.600 |
cy | -0.563 |
night | -0.561 |
ite | -0.546 |
Token | as |
Token ID | 355 |
Feature activation | +2.65e+00 |
Pos logit contributions | |
gru | +0.092 |
badge | +0.082 |
threatening | +0.077 |
warning | +0.077 |
rifle | +0.076 |
Neg logit contributions | |
day | -0.068 |
afternoon | -0.063 |
cy | -0.060 |
ite | -0.058 |
night | -0.056 |
Token | it |
Token ID | 340 |
Feature activation | +6.57e-01 |
Pos logit contributions | |
day | +0.098 |
afternoon | +0.091 |
cy | +0.087 |
ite | +0.085 |
night | +0.083 |
Neg logit contributions | |
gru | -0.109 |
badge | -0.093 |
threatening | -0.088 |
Reward | -0.088 |
warning | -0.087 |
Token | took |
Token ID | 1718 |
Feature activation | +2.22e+00 |
Pos logit contributions | |
day | +0.022 |
afternoon | +0.020 |
cy | +0.018 |
night | +0.018 |
ite | +0.018 |
Neg logit contributions | |
gru | -0.030 |
badge | -0.026 |
god | -0.025 |
threatening | -0.025 |
Reward | -0.025 |
Token | off |
Token ID | 572 |
Feature activation | +2.68e+00 |
Pos logit contributions | |
day | +0.087 |
afternoon | +0.079 |
night | +0.076 |
morning | +0.072 |
cy | +0.072 |
Neg logit contributions | |
gru | -0.091 |
badge | -0.079 |
Reward | -0.076 |
ution | -0.075 |
rifle | -0.074 |
Token | friends |
Token ID | 2460 |
Feature activation | +9.88e-01 |
Pos logit contributions | |
gru | +0.155 |
badge | +0.136 |
rifle | +0.125 |
Reward | +0.125 |
warning | +0.124 |
Neg logit contributions | |
day | -0.148 |
afternoon | -0.137 |
cy | -0.132 |
night | -0.127 |
he | -0.126 |
Token | stopped |
Token ID | 5025 |
Feature activation | +2.63e+00 |
Pos logit contributions | |
day | +0.033 |
afternoon | +0.030 |
cy | +0.027 |
night | +0.027 |
ite | +0.026 |
Neg logit contributions | |
gru | -0.045 |
badge | -0.040 |
Reward | -0.038 |
ution | -0.038 |
rifle | -0.038 |
Token | and |
Token ID | 290 |
Feature activation | +2.25e+00 |
Pos logit contributions | |
day | +0.093 |
afternoon | +0.081 |
night | +0.077 |
cy | +0.073 |
morning | +0.072 |
Neg logit contributions | |
gru | -0.120 |
badge | -0.105 |
Reward | -0.105 |
ution | -0.101 |
rifle | -0.100 |
Token | watched |
Token ID | 7342 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
day | +0.076 |
afternoon | +0.066 |
night | +0.064 |
cy | +0.063 |
morning | +0.059 |
Neg logit contributions | |
gru | -0.104 |
badge | -0.092 |
Reward | -0.091 |
rifle | -0.089 |
ution | -0.089 |
Token | in |
Token ID | 287 |
Feature activation | +4.72e+00 |
Pos logit contributions | |
day | +0.005 |
afternoon | +0.005 |
night | +0.005 |
cy | +0.004 |
morning | +0.004 |
Neg logit contributions | |
gru | -0.007 |
badge | -0.006 |
Reward | -0.006 |
ution | -0.006 |
rifle | -0.006 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.69e+01 |
Pos logit contributions | |
day | +0.199 |
afternoon | +0.179 |
night | +0.172 |
morning | +0.163 |
days | +0.158 |
Neg logit contributions | |
gru | -0.181 |
Reward | -0.158 |
badge | -0.155 |
phant | -0.153 |
ution | -0.152 |
Token | ement |
Token ID | 972 |
Feature activation | -1.92e+00 |
Pos logit contributions | |
gru | +0.719 |
badge | +0.627 |
Reward | +0.607 |
rifle | +0.594 |
threatening | +0.589 |
Neg logit contributions | |
day | -0.624 |
afternoon | -0.566 |
night | -0.527 |
cy | -0.524 |
ite | -0.510 |
Token | . |
Token ID | 13 |
Feature activation | +5.73e-01 |
Pos logit contributions | |
gru | +0.090 |
badge | +0.080 |
threatening | +0.075 |
Reward | +0.075 |
rifle | +0.075 |
Neg logit contributions | |
day | -0.061 |
afternoon | -0.055 |
cy | -0.052 |
ite | -0.050 |
night | -0.050 |
Token | The |
Token ID | 383 |
Feature activation | -7.43e-01 |
Pos logit contributions | |
day | +0.020 |
afternoon | +0.018 |
cy | +0.017 |
night | +0.017 |
morning | +0.016 |
Neg logit contributions | |
gru | -0.025 |
badge | -0.022 |
ution | -0.021 |
Reward | -0.021 |
rifle | -0.021 |
Token | dog |
Token ID | 3290 |
Feature activation | -9.94e-01 |
Pos logit contributions | |
gru | +0.030 |
badge | +0.026 |
rifle | +0.024 |
warning | +0.024 |
threatening | +0.024 |
Neg logit contributions | |
day | -0.028 |
afternoon | -0.026 |
cy | -0.025 |
he | -0.025 |
ite | -0.024 |
Token | too |
Token ID | 1165 |
Feature activation | +2.66e-02 |
Pos logit contributions | |
gru | +0.048 |
badge | +0.043 |
Reward | +0.042 |
ution | +0.041 |
rifle | +0.041 |
Neg logit contributions | |
day | -0.030 |
afternoon | -0.026 |
night | -0.024 |
cy | -0.024 |
morning | -0.023 |
Token | the |
Token ID | 262 |
Feature activation | +2.91e+00 |
Pos logit contributions | |
gru | +0.106 |
badge | +0.093 |
Reward | +0.092 |
ution | +0.089 |
rifle | +0.089 |
Neg logit contributions | |
day | -0.072 |
afternoon | -0.063 |
night | -0.059 |
cy | -0.058 |
morning | -0.056 |
Token | tree |
Token ID | 5509 |
Feature activation | +3.18e+00 |
Pos logit contributions | |
day | +0.097 |
cy | +0.091 |
afternoon | +0.090 |
olly | +0.086 |
he | +0.085 |
Neg logit contributions | |
gru | -0.126 |
badge | -0.109 |
warning | -0.105 |
threatening | -0.105 |
probably | -0.102 |
Token | to |
Token ID | 284 |
Feature activation | -5.16e+00 |
Pos logit contributions | |
day | +0.101 |
afternoon | +0.091 |
cy | +0.086 |
night | +0.084 |
ite | +0.082 |
Neg logit contributions | |
gru | -0.149 |
badge | -0.131 |
rifle | -0.124 |
threatening | -0.124 |
Reward | -0.124 |
Token | its |
Token ID | 663 |
Feature activation | -5.08e+00 |
Pos logit contributions | |
gru | +0.223 |
badge | +0.192 |
Reward | +0.186 |
rifle | +0.183 |
ution | +0.182 |
Neg logit contributions | |
day | -0.196 |
afternoon | -0.173 |
night | -0.166 |
cy | -0.157 |
he | -0.156 |
Token | once |
Token ID | 1752 |
Feature activation | +4.21e+00 |
Pos logit contributions | |
gru | +0.214 |
badge | +0.187 |
Reward | +0.180 |
ution | +0.176 |
rifle | +0.175 |
Neg logit contributions | |
day | -0.193 |
afternoon | -0.174 |
night | -0.163 |
cy | -0.158 |
morning | -0.155 |
Token | former |
Token ID | 1966 |
Feature activation | -1.69e+01 |
Pos logit contributions | |
day | +0.138 |
olly | +0.137 |
ite | +0.132 |
cy | +0.130 |
afternoon | +0.128 |
Neg logit contributions | |
gru | -0.174 |
god | -0.146 |
warning | -0.144 |
badge | -0.142 |
threatening | -0.141 |
Token | beauty |
Token ID | 8737 |
Feature activation | -5.37e+00 |
Pos logit contributions | |
gru | +0.708 |
badge | +0.606 |
ution | +0.574 |
god | +0.572 |
Reward | +0.572 |
Neg logit contributions | |
day | -0.613 |
afternoon | -0.554 |
cy | -0.547 |
ite | -0.533 |
he | -0.530 |
Token | . |
Token ID | 13 |
Feature activation | +2.67e-01 |
Pos logit contributions | |
gru | +0.232 |
badge | +0.198 |
threatening | +0.191 |
god | +0.189 |
warning | +0.189 |
Neg logit contributions | |
day | -0.172 |
afternoon | -0.163 |
cy | -0.159 |
ite | -0.153 |
night | -0.145 |
Token | He |
Token ID | 679 |
Feature activation | -3.34e+00 |
Pos logit contributions | |
day | +0.011 |
afternoon | +0.010 |
cy | +0.010 |
night | +0.010 |
ite | +0.009 |
Neg logit contributions | |
gru | -0.010 |
badge | -0.008 |
Reward | -0.008 |
rifle | -0.008 |
ution | -0.008 |
Token | was |
Token ID | 373 |
Feature activation | +8.66e-01 |
Pos logit contributions | |
gru | +0.149 |
badge | +0.131 |
god | +0.130 |
probably | +0.127 |
threatening | +0.125 |
Neg logit contributions | |
day | -0.107 |
afternoon | -0.099 |
cy | -0.093 |
ite | -0.092 |
he | -0.091 |
Token | happy |
Token ID | 3772 |
Feature activation | +2.52e+00 |
Pos logit contributions | |
day | +0.034 |
afternoon | +0.031 |
night | +0.029 |
cy | +0.028 |
morning | +0.028 |
Neg logit contributions | |
gru | -0.035 |
badge | -0.031 |
Reward | -0.029 |
ution | -0.029 |
rifle | -0.029 |
Token | ered |
Token ID | 1068 |
Feature activation | -4.83e-01 |
Pos logit contributions | |
gru | +0.289 |
badge | +0.247 |
Reward | +0.243 |
rifle | +0.233 |
threatening | +0.230 |
Neg logit contributions | |
day | -0.311 |
afternoon | -0.285 |
cy | -0.276 |
ite | -0.271 |
night | -0.265 |
Token | in |
Token ID | 287 |
Feature activation | +4.61e+00 |
Pos logit contributions | |
gru | +0.025 |
badge | +0.023 |
ution | +0.022 |
Reward | +0.022 |
phant | +0.022 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.012 |
night | -0.011 |
cy | -0.010 |
morning | -0.010 |
Token | , |
Token ID | 11 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
day | +0.146 |
afternoon | +0.122 |
night | +0.120 |
morning | +0.110 |
cy | +0.104 |
Neg logit contributions | |
gru | -0.215 |
badge | -0.206 |
phant | -0.202 |
ution | -0.199 |
Reward | -0.198 |
Token | blinking |
Token ID | 43196 |
Feature activation | -2.39e+00 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
Reward | +0.007 |
ution | +0.007 |
god | +0.007 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.005 |
night | -0.005 |
morning | -0.005 |
cy | -0.004 |
Token | in |
Token ID | 287 |
Feature activation | +4.60e+00 |
Pos logit contributions | |
gru | +0.114 |
Reward | +0.110 |
ution | +0.106 |
fracture | +0.106 |
badge | +0.105 |
Neg logit contributions | |
day | -0.068 |
afternoon | -0.059 |
night | -0.056 |
days | -0.052 |
morning | -0.052 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.68e+01 |
Pos logit contributions | |
day | +0.152 |
afternoon | +0.133 |
night | +0.126 |
morning | +0.117 |
cy | +0.115 |
Neg logit contributions | |
gru | -0.213 |
badge | -0.198 |
Reward | -0.191 |
ution | -0.189 |
rifle | -0.188 |
Token | ement |
Token ID | 972 |
Feature activation | -1.76e+00 |
Pos logit contributions | |
gru | +0.751 |
badge | +0.672 |
Reward | +0.663 |
rifle | +0.648 |
threatening | +0.636 |
Neg logit contributions | |
day | -0.574 |
afternoon | -0.515 |
night | -0.480 |
he | -0.474 |
cy | -0.469 |
Token | . |
Token ID | 13 |
Feature activation | +2.27e-01 |
Pos logit contributions | |
gru | +0.088 |
badge | +0.078 |
Reward | +0.076 |
ution | +0.075 |
rifle | +0.074 |
Neg logit contributions | |
day | -0.052 |
afternoon | -0.046 |
night | -0.042 |
cy | -0.041 |
ite | -0.040 |
Token |
|
Token ID | 198 |
Feature activation | -4.38e-01 |
Pos logit contributions | |
day | +0.008 |
afternoon | +0.007 |
night | +0.007 |
cy | +0.007 |
ite | +0.006 |
Neg logit contributions | |
gru | -0.010 |
badge | -0.009 |
ution | -0.008 |
Reward | -0.008 |
threatening | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -4.77e-01 |
Pos logit contributions | |
gru | +0.021 |
badge | +0.018 |
Reward | +0.018 |
threatening | +0.017 |
rifle | +0.017 |
Neg logit contributions | |
day | -0.014 |
afternoon | -0.012 |
cy | -0.012 |
night | -0.011 |
ite | -0.011 |
Token | But |
Token ID | 1537 |
Feature activation | -4.05e+00 |
Pos logit contributions | |
gru | +0.019 |
badge | +0.017 |
Reward | +0.016 |
ution | +0.016 |
rifle | +0.016 |
Neg logit contributions | |
day | -0.018 |
afternoon | -0.017 |
cy | -0.015 |
night | -0.015 |
ite | -0.015 |
Token | egg |
Token ID | 5935 |
Feature activation | -4.78e+00 |
Pos logit contributions | |
gru | +0.198 |
badge | +0.170 |
Reward | +0.161 |
rifle | +0.158 |
ution | +0.158 |
Neg logit contributions | |
day | -0.210 |
afternoon | -0.193 |
cy | -0.182 |
night | -0.181 |
ite | -0.177 |
Token | ! |
Token ID | 0 |
Feature activation | -8.12e-02 |
Pos logit contributions | |
gru | +0.226 |
badge | +0.200 |
Reward | +0.192 |
rifle | +0.189 |
threatening | +0.188 |
Neg logit contributions | |
day | -0.151 |
afternoon | -0.136 |
cy | -0.125 |
night | -0.124 |
ite | -0.120 |
Token | He |
Token ID | 679 |
Feature activation | -3.08e+00 |
Pos logit contributions | |
gru | +0.003 |
badge | +0.003 |
Reward | +0.003 |
ution | +0.003 |
rifle | +0.003 |
Neg logit contributions | |
day | -0.003 |
afternoon | -0.003 |
cy | -0.003 |
night | -0.003 |
ite | -0.003 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
gru | +0.149 |
badge | +0.133 |
Reward | +0.128 |
ution | +0.126 |
rifle | +0.126 |
Neg logit contributions | |
day | -0.095 |
afternoon | -0.084 |
night | -0.077 |
cy | -0.077 |
ite | -0.074 |
Token | in |
Token ID | 287 |
Feature activation | +4.68e+00 |
Pos logit contributions | |
gru | +0.001 |
badge | +0.001 |
ution | +0.001 |
Reward | +0.001 |
rifle | +0.001 |
Neg logit contributions | |
day | -0.001 |
afternoon | -0.001 |
night | -0.001 |
morning | -0.001 |
cy | -0.001 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.68e+01 |
Pos logit contributions | |
day | +0.188 |
afternoon | +0.167 |
night | +0.164 |
morning | +0.153 |
days | +0.150 |
Neg logit contributions | |
gru | -0.185 |
badge | -0.168 |
Reward | -0.161 |
ution | -0.160 |
phant | -0.157 |
Token | ement |
Token ID | 972 |
Feature activation | -1.75e+00 |
Pos logit contributions | |
gru | +0.756 |
badge | +0.674 |
Reward | +0.662 |
rifle | +0.647 |
threatening | +0.635 |
Neg logit contributions | |
day | -0.574 |
afternoon | -0.516 |
night | -0.481 |
he | -0.475 |
cy | -0.475 |
Token | as |
Token ID | 355 |
Feature activation | +2.44e+00 |
Pos logit contributions | |
gru | +0.085 |
badge | +0.076 |
Reward | +0.073 |
rifle | +0.072 |
threatening | +0.072 |
Neg logit contributions | |
day | -0.053 |
afternoon | -0.047 |
cy | -0.044 |
night | -0.043 |
ite | -0.042 |
Token | the |
Token ID | 262 |
Feature activation | +2.70e+00 |
Pos logit contributions | |
day | +0.095 |
afternoon | +0.088 |
cy | +0.084 |
ite | +0.082 |
night | +0.081 |
Neg logit contributions | |
gru | -0.098 |
badge | -0.083 |
Reward | -0.079 |
threatening | -0.078 |
rifle | -0.077 |
Token | egg |
Token ID | 5935 |
Feature activation | -5.22e+00 |
Pos logit contributions | |
day | +0.117 |
afternoon | +0.107 |
night | +0.101 |
cy | +0.100 |
morning | +0.097 |
Neg logit contributions | |
gru | -0.099 |
badge | -0.086 |
Reward | -0.081 |
ution | -0.080 |
rifle | -0.079 |
Token | began |
Token ID | 2540 |
Feature activation | +3.13e+00 |
Pos logit contributions | |
gru | +0.235 |
badge | +0.207 |
Reward | +0.196 |
rifle | +0.194 |
threatening | +0.194 |
Neg logit contributions | |
day | -0.176 |
afternoon | -0.160 |
cy | -0.149 |
night | -0.146 |
ite | -0.144 |
Token | her |
Token ID | 607 |
Feature activation | -1.91e+00 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.020 |
ite | +0.019 |
cy | +0.019 |
he | +0.019 |
Neg logit contributions | |
gru | -0.022 |
badge | -0.019 |
threatening | -0.018 |
Reward | -0.018 |
spoken | -0.018 |
Token | mom |
Token ID | 1995 |
Feature activation | +2.71e+00 |
Pos logit contributions | |
gru | +0.076 |
badge | +0.068 |
address | +0.062 |
Reward | +0.062 |
warning | +0.062 |
Neg logit contributions | |
day | -0.070 |
afternoon | -0.065 |
cy | -0.064 |
ite | -0.063 |
he | -0.062 |
Token | my |
Token ID | 1820 |
Feature activation | +4.93e+00 |
Pos logit contributions | |
day | +0.092 |
afternoon | +0.087 |
ite | +0.084 |
olly | +0.081 |
cy | +0.080 |
Neg logit contributions | |
gru | -0.109 |
badge | -0.099 |
god | -0.097 |
address | -0.096 |
probably | -0.095 |
Token | watched |
Token ID | 7342 |
Feature activation | +7.00e-01 |
Pos logit contributions | |
day | +0.169 |
afternoon | +0.159 |
ite | +0.149 |
olly | +0.145 |
cy | +0.144 |
Neg logit contributions | |
gru | -0.202 |
badge | -0.185 |
address | -0.177 |
god | -0.175 |
probably | -0.173 |
Token | in |
Token ID | 287 |
Feature activation | +5.11e+00 |
Pos logit contributions | |
day | +0.026 |
afternoon | +0.024 |
cy | +0.023 |
ite | +0.022 |
night | +0.022 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.025 |
Reward | -0.024 |
ution | -0.024 |
threatening | -0.024 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.68e+01 |
Pos logit contributions | |
day | +0.170 |
afternoon | +0.151 |
night | +0.141 |
cy | +0.134 |
morning | +0.132 |
Neg logit contributions | |
gru | -0.240 |
badge | -0.213 |
Reward | -0.207 |
rifle | -0.204 |
ution | -0.202 |
Token | ement |
Token ID | 972 |
Feature activation | -1.94e+00 |
Pos logit contributions | |
gru | +0.682 |
badge | +0.591 |
Reward | +0.560 |
ution | +0.555 |
rifle | +0.549 |
Neg logit contributions | |
day | -0.651 |
afternoon | -0.592 |
cy | -0.555 |
night | -0.553 |
ite | -0.539 |
Token | as |
Token ID | 355 |
Feature activation | +2.23e+00 |
Pos logit contributions | |
gru | +0.084 |
badge | +0.074 |
warning | +0.069 |
threatening | +0.069 |
address | +0.069 |
Neg logit contributions | |
day | -0.064 |
afternoon | -0.059 |
cy | -0.057 |
ite | -0.056 |
night | -0.054 |
Token | the |
Token ID | 262 |
Feature activation | +2.90e+00 |
Pos logit contributions | |
day | +0.082 |
afternoon | +0.075 |
cy | +0.071 |
ite | +0.070 |
night | +0.069 |
Neg logit contributions | |
gru | -0.093 |
badge | -0.081 |
Reward | -0.077 |
threatening | -0.076 |
ution | -0.075 |
Token | machine |
Token ID | 4572 |
Feature activation | -4.86e+00 |
Pos logit contributions | |
day | +0.113 |
afternoon | +0.105 |
cy | +0.102 |
ite | +0.100 |
he | +0.098 |
Neg logit contributions | |
gru | -0.115 |
badge | -0.097 |
Reward | -0.090 |
warning | -0.090 |
threatening | -0.090 |
Token | flew |
Token ID | 13112 |
Feature activation | +4.60e+00 |
Pos logit contributions | |
gru | +0.211 |
badge | +0.189 |
god | +0.177 |
Reward | +0.175 |
threatening | +0.175 |
Neg logit contributions | |
day | -0.161 |
afternoon | -0.149 |
cy | -0.143 |
ite | -0.138 |
olly | -0.138 |
Token | . |
Token ID | 13 |
Feature activation | +6.81e-01 |
Pos logit contributions | |
gru | +0.211 |
badge | +0.188 |
Reward | +0.178 |
threatening | +0.177 |
rifle | +0.176 |
Neg logit contributions | |
day | -0.142 |
afternoon | -0.129 |
cy | -0.123 |
ite | -0.117 |
night | -0.117 |
Token |
|
Token ID | 198 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
day | +0.026 |
afternoon | +0.024 |
cy | +0.022 |
night | +0.022 |
ite | +0.022 |
Neg logit contributions | |
gru | -0.028 |
badge | -0.024 |
Reward | -0.023 |
ution | -0.023 |
rifle | -0.023 |
Token |
|
Token ID | 198 |
Feature activation | -2.28e-01 |
Pos logit contributions | |
gru | +0.005 |
badge | +0.004 |
Reward | +0.004 |
threatening | +0.004 |
ution | +0.004 |
Neg logit contributions | |
day | -0.003 |
afternoon | -0.003 |
cy | -0.003 |
night | -0.003 |
ite | -0.003 |
Token | To |
Token ID | 2514 |
Feature activation | -5.59e+00 |
Pos logit contributions | |
gru | +0.010 |
badge | +0.008 |
Reward | +0.008 |
ution | +0.008 |
rifle | +0.008 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
cy | -0.007 |
night | -0.007 |
ite | -0.007 |
Token | her |
Token ID | 607 |
Feature activation | -1.71e+00 |
Pos logit contributions | |
gru | +0.227 |
badge | +0.199 |
Reward | +0.192 |
god | +0.186 |
threatening | +0.184 |
Neg logit contributions | |
day | -0.208 |
afternoon | -0.192 |
ite | -0.181 |
cy | -0.180 |
night | -0.173 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.68e+01 |
Pos logit contributions | |
gru | +0.054 |
badge | +0.045 |
warning | +0.041 |
god | +0.041 |
threatening | +0.041 |
Neg logit contributions | |
day | -0.079 |
afternoon | -0.075 |
he | -0.072 |
cy | -0.072 |
ite | -0.071 |
Token | ement |
Token ID | 972 |
Feature activation | -1.76e+00 |
Pos logit contributions | |
gru | +0.690 |
badge | +0.600 |
Reward | +0.570 |
ution | +0.562 |
rifle | +0.559 |
Neg logit contributions | |
day | -0.641 |
afternoon | -0.583 |
cy | -0.546 |
night | -0.545 |
ite | -0.529 |
Token | , |
Token ID | 11 |
Feature activation | -7.28e-01 |
Pos logit contributions | |
gru | +0.088 |
badge | +0.080 |
rifle | +0.075 |
Reward | +0.075 |
god | +0.075 |
Neg logit contributions | |
day | -0.048 |
afternoon | -0.044 |
cy | -0.040 |
night | -0.038 |
ite | -0.038 |
Token | the |
Token ID | 262 |
Feature activation | +2.92e+00 |
Pos logit contributions | |
gru | +0.034 |
badge | +0.031 |
ution | +0.030 |
rifle | +0.030 |
Reward | +0.030 |
Neg logit contributions | |
day | -0.024 |
afternoon | -0.020 |
night | -0.019 |
morning | -0.018 |
cy | -0.018 |
Token | clothes |
Token ID | 8242 |
Feature activation | -3.66e+00 |
Pos logit contributions | |
day | +0.119 |
afternoon | +0.109 |
cy | +0.104 |
night | +0.102 |
ite | +0.101 |
Neg logit contributions | |
gru | -0.113 |
badge | -0.097 |
Reward | -0.092 |
rifle | -0.090 |
ution | -0.090 |
Token | came |
Token ID | 1625 |
Feature activation | -2.17e+00 |
Pos logit contributions | |
gru | +0.175 |
badge | +0.163 |
address | +0.152 |
spoken | +0.151 |
probably | +0.151 |
Neg logit contributions | |
day | -0.103 |
afternoon | -0.096 |
cy | -0.086 |
night | -0.082 |
olly | -0.080 |
Token |
|
Token ID | 198 |
Feature activation | -2.41e-01 |
Pos logit contributions | |
gru | +0.009 |
badge | +0.008 |
Reward | +0.007 |
ution | +0.007 |
threatening | +0.007 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.006 |
cy | -0.005 |
night | -0.005 |
ite | -0.005 |
Token | L |
Token ID | 43 |
Feature activation | +8.60e-01 |
Pos logit contributions | |
gru | +0.010 |
badge | +0.009 |
Reward | +0.009 |
ution | +0.008 |
rifle | +0.008 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
cy | -0.008 |
night | -0.007 |
ite | -0.007 |
Token | ucky |
Token ID | 5309 |
Feature activation | +3.33e+00 |
Pos logit contributions | |
day | +0.034 |
afternoon | +0.031 |
night | +0.030 |
ite | +0.030 |
morning | +0.029 |
Neg logit contributions | |
gru | -0.031 |
probably | -0.027 |
spoken | -0.026 |
badge | -0.025 |
warning | -0.025 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.53e-01 |
Pos logit contributions | |
day | +0.114 |
afternoon | +0.104 |
ite | +0.100 |
cy | +0.099 |
he | +0.098 |
Neg logit contributions | |
gru | -0.146 |
badge | -0.126 |
god | -0.122 |
threatening | -0.121 |
warning | -0.120 |
Token | in |
Token ID | 287 |
Feature activation | +4.30e+00 |
Pos logit contributions | |
gru | +0.017 |
badge | +0.015 |
Reward | +0.014 |
ution | +0.014 |
rifle | +0.014 |
Neg logit contributions | |
day | -0.012 |
afternoon | -0.010 |
night | -0.010 |
cy | -0.009 |
morning | -0.009 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.68e+01 |
Pos logit contributions | |
day | +0.178 |
afternoon | +0.161 |
night | +0.154 |
cy | +0.148 |
morning | +0.146 |
Neg logit contributions | |
gru | -0.166 |
badge | -0.143 |
Reward | -0.138 |
ution | -0.136 |
rifle | -0.134 |
Token | ement |
Token ID | 972 |
Feature activation | -1.91e+00 |
Pos logit contributions | |
gru | +0.751 |
badge | +0.659 |
Reward | +0.656 |
rifle | +0.638 |
threatening | +0.624 |
Neg logit contributions | |
day | -0.580 |
afternoon | -0.526 |
night | -0.485 |
he | -0.478 |
cy | -0.478 |
Token | as |
Token ID | 355 |
Feature activation | +2.25e+00 |
Pos logit contributions | |
gru | +0.088 |
badge | +0.078 |
threatening | +0.074 |
Reward | +0.074 |
rifle | +0.073 |
Neg logit contributions | |
day | -0.062 |
afternoon | -0.056 |
cy | -0.053 |
night | -0.051 |
ite | -0.051 |
Token | the |
Token ID | 262 |
Feature activation | +2.69e+00 |
Pos logit contributions | |
day | +0.088 |
afternoon | +0.083 |
cy | +0.079 |
ite | +0.077 |
night | +0.075 |
Neg logit contributions | |
gru | -0.089 |
badge | -0.076 |
threatening | -0.071 |
Reward | -0.070 |
warning | -0.069 |
Token | frog |
Token ID | 21264 |
Feature activation | +7.63e-03 |
Pos logit contributions | |
day | +0.105 |
afternoon | +0.099 |
cy | +0.099 |
ite | +0.098 |
he | +0.097 |
Neg logit contributions | |
gru | -0.103 |
badge | -0.085 |
warning | -0.081 |
threatening | -0.080 |
rifle | -0.077 |
Token | zoom |
Token ID | 19792 |
Feature activation | +3.94e+00 |
Pos logit contributions | |
day | +0.000 |
afternoon | +0.000 |
cy | +0.000 |
he | +0.000 |
night | +0.000 |
Neg logit contributions | |
gru | -0.000 |
badge | -0.000 |
threatening | -0.000 |
Reward | -0.000 |
god | -0.000 |
Token | eddy |
Token ID | 21874 |
Feature activation | +2.34e+00 |
Pos logit contributions | |
gru | +0.308 |
badge | +0.275 |
Reward | +0.264 |
rifle | +0.261 |
ution | +0.259 |
Neg logit contributions | |
day | -0.224 |
afternoon | -0.202 |
night | -0.184 |
cy | -0.178 |
morning | -0.176 |
Token | bear |
Token ID | 6842 |
Feature activation | -1.84e-01 |
Pos logit contributions | |
day | +0.103 |
afternoon | +0.096 |
cy | +0.091 |
ite | +0.090 |
night | +0.089 |
Neg logit contributions | |
gru | -0.080 |
badge | -0.069 |
Reward | -0.064 |
rifle | -0.064 |
god | -0.062 |
Token | and |
Token ID | 290 |
Feature activation | +2.19e+00 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
Reward | +0.007 |
rifle | +0.007 |
ution | +0.007 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.006 |
cy | -0.005 |
night | -0.005 |
ite | -0.005 |
Token | watched |
Token ID | 7342 |
Feature activation | +3.25e-01 |
Pos logit contributions | |
day | +0.082 |
afternoon | +0.072 |
night | +0.070 |
morning | +0.066 |
cy | +0.066 |
Neg logit contributions | |
gru | -0.094 |
ution | -0.079 |
badge | -0.078 |
Reward | -0.078 |
rifle | -0.076 |
Token | in |
Token ID | 287 |
Feature activation | +4.89e+00 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.010 |
night | +0.010 |
morning | +0.010 |
he | +0.009 |
Neg logit contributions | |
gru | -0.014 |
badge | -0.013 |
Reward | -0.012 |
ution | -0.012 |
rifle | -0.012 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.67e+01 |
Pos logit contributions | |
day | +0.184 |
afternoon | +0.161 |
night | +0.158 |
morning | +0.145 |
days | +0.142 |
Neg logit contributions | |
gru | -0.211 |
badge | -0.187 |
Reward | -0.183 |
ution | -0.181 |
rifle | -0.181 |
Token | ement |
Token ID | 972 |
Feature activation | -1.70e+00 |
Pos logit contributions | |
gru | +0.708 |
badge | +0.619 |
Reward | +0.596 |
rifle | +0.585 |
threatening | +0.579 |
Neg logit contributions | |
day | -0.618 |
afternoon | -0.560 |
night | -0.523 |
cy | -0.521 |
ite | -0.506 |
Token | as |
Token ID | 355 |
Feature activation | +2.57e+00 |
Pos logit contributions | |
gru | +0.078 |
badge | +0.069 |
Reward | +0.066 |
threatening | +0.065 |
warning | +0.065 |
Neg logit contributions | |
day | -0.055 |
afternoon | -0.050 |
cy | -0.047 |
ite | -0.045 |
night | -0.045 |
Token | it |
Token ID | 340 |
Feature activation | +8.05e-01 |
Pos logit contributions | |
day | +0.098 |
afternoon | +0.091 |
cy | +0.086 |
ite | +0.084 |
night | +0.083 |
Neg logit contributions | |
gru | -0.104 |
badge | -0.090 |
Reward | -0.085 |
threatening | -0.084 |
warning | -0.082 |
Token | shrunk |
Token ID | 38900 |
Feature activation | -2.36e+00 |
Pos logit contributions | |
day | +0.027 |
afternoon | +0.024 |
night | +0.022 |
cy | +0.022 |
ite | +0.021 |
Neg logit contributions | |
gru | -0.037 |
badge | -0.033 |
Reward | -0.032 |
rifle | -0.031 |
ution | -0.031 |
Token | an |
Token ID | 281 |
Feature activation | -8.80e+00 |
Pos logit contributions | |
gru | +0.111 |
badge | +0.099 |
Reward | +0.095 |
ution | +0.093 |
rifle | +0.093 |
Neg logit contributions | |
day | -0.076 |
afternoon | -0.068 |
night | -0.063 |
cy | -0.063 |
ite | -0.060 |
Token | as |
Token ID | 355 |
Feature activation | +2.53e+00 |
Pos logit contributions | |
day | +0.162 |
afternoon | +0.149 |
cy | +0.146 |
ite | +0.138 |
night | +0.138 |
Neg logit contributions | |
gru | -0.249 |
badge | -0.224 |
rifle | -0.206 |
threatening | -0.203 |
armor | -0.201 |
Token | the |
Token ID | 262 |
Feature activation | +3.05e+00 |
Pos logit contributions | |
day | +0.097 |
afternoon | +0.097 |
cy | +0.095 |
ite | +0.092 |
ops | +0.090 |
Neg logit contributions | |
gru | -0.092 |
threatening | -0.074 |
badge | -0.073 |
menacing | -0.072 |
warning | -0.071 |
Token | woman |
Token ID | 2415 |
Feature activation | -4.69e+00 |
Pos logit contributions | |
day | +0.110 |
cy | +0.107 |
he | +0.106 |
afternoon | +0.104 |
ite | +0.103 |
Neg logit contributions | |
gru | -0.123 |
badge | -0.105 |
warning | -0.099 |
rifle | -0.098 |
threatening | -0.097 |
Token | watched |
Token ID | 7342 |
Feature activation | -1.11e-01 |
Pos logit contributions | |
gru | +0.189 |
badge | +0.169 |
god | +0.168 |
threatening | +0.157 |
warning | +0.156 |
Neg logit contributions | |
day | -0.160 |
afternoon | -0.152 |
cy | -0.142 |
he | -0.141 |
olly | -0.138 |
Token | in |
Token ID | 287 |
Feature activation | +4.48e+00 |
Pos logit contributions | |
gru | +0.005 |
badge | +0.004 |
Reward | +0.004 |
ution | +0.004 |
rifle | +0.004 |
Neg logit contributions | |
day | -0.004 |
afternoon | -0.004 |
night | -0.003 |
cy | -0.003 |
morning | -0.003 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.65e+01 |
Pos logit contributions | |
day | +0.178 |
afternoon | +0.161 |
night | +0.152 |
cy | +0.147 |
morning | +0.145 |
Neg logit contributions | |
gru | -0.180 |
badge | -0.157 |
Reward | -0.151 |
ution | -0.148 |
rifle | -0.148 |
Token | ement |
Token ID | 972 |
Feature activation | -1.63e+00 |
Pos logit contributions | |
gru | +0.706 |
badge | +0.618 |
Reward | +0.599 |
rifle | +0.586 |
threatening | +0.580 |
Neg logit contributions | |
day | -0.605 |
afternoon | -0.549 |
night | -0.511 |
cy | -0.508 |
ite | -0.495 |
Token | . |
Token ID | 13 |
Feature activation | +1.90e-01 |
Pos logit contributions | |
gru | +0.073 |
badge | +0.065 |
threatening | +0.061 |
rifle | +0.060 |
warning | +0.060 |
Neg logit contributions | |
day | -0.054 |
afternoon | -0.049 |
cy | -0.047 |
ite | -0.045 |
night | -0.045 |
Token | She |
Token ID | 1375 |
Feature activation | -2.26e+00 |
Pos logit contributions | |
day | +0.007 |
afternoon | +0.007 |
cy | +0.006 |
night | +0.006 |
ite | +0.006 |
Neg logit contributions | |
gru | -0.008 |
badge | -0.007 |
Reward | -0.006 |
rifle | -0.006 |
ution | -0.006 |
Token | saw |
Token ID | 2497 |
Feature activation | -4.08e+00 |
Pos logit contributions | |
gru | +0.098 |
badge | +0.084 |
god | +0.083 |
threatening | +0.081 |
spoken | +0.081 |
Neg logit contributions | |
day | -0.077 |
afternoon | -0.070 |
cy | -0.068 |
he | -0.066 |
ite | -0.065 |
Token | the |
Token ID | 262 |
Feature activation | +2.86e+00 |
Pos logit contributions | |
gru | +0.174 |
badge | +0.152 |
Reward | +0.148 |
ution | +0.145 |
rifle | +0.144 |
Neg logit contributions | |
day | -0.151 |
afternoon | -0.135 |
night | -0.127 |
cy | -0.123 |
morning | -0.121 |
Token | animals |
Token ID | 4695 |
Feature activation | -8.38e-01 |
Pos logit contributions | |
day | +0.170 |
he | +0.163 |
afternoon | +0.162 |
ite | +0.158 |
cy | +0.155 |
Neg logit contributions | |
gru | -0.134 |
threatening | -0.106 |
badge | -0.104 |
Reward | -0.103 |
god | -0.102 |
Token | were |
Token ID | 547 |
Feature activation | +1.60e+00 |
Pos logit contributions | |
gru | +0.039 |
badge | +0.035 |
threatening | +0.034 |
rifle | +0.034 |
spoken | +0.033 |
Neg logit contributions | |
day | -0.024 |
afternoon | -0.023 |
cy | -0.020 |
night | -0.020 |
olly | -0.020 |
Token | watching |
Token ID | 4964 |
Feature activation | -7.20e-01 |
Pos logit contributions | |
day | +0.062 |
afternoon | +0.054 |
night | +0.051 |
morning | +0.048 |
days | +0.046 |
Neg logit contributions | |
gru | -0.065 |
Reward | -0.059 |
badge | -0.059 |
ution | -0.059 |
rifle | -0.056 |
Token | him |
Token ID | 683 |
Feature activation | -2.10e+00 |
Pos logit contributions | |
gru | +0.032 |
badge | +0.029 |
Reward | +0.028 |
ution | +0.027 |
rifle | +0.027 |
Neg logit contributions | |
day | -0.025 |
afternoon | -0.022 |
night | -0.021 |
cy | -0.020 |
morning | -0.020 |
Token | with |
Token ID | 351 |
Feature activation | -2.49e+00 |
Pos logit contributions | |
gru | +0.094 |
badge | +0.083 |
Reward | +0.080 |
threatening | +0.079 |
rifle | +0.079 |
Neg logit contributions | |
day | -0.070 |
afternoon | -0.065 |
cy | -0.060 |
night | -0.059 |
ite | -0.058 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.65e+01 |
Pos logit contributions | |
gru | +0.104 |
ution | +0.091 |
badge | +0.091 |
Reward | +0.090 |
CRE | +0.087 |
Neg logit contributions | |
day | -0.095 |
afternoon | -0.082 |
night | -0.081 |
morning | -0.075 |
days | -0.074 |
Token | ement |
Token ID | 972 |
Feature activation | -1.62e+00 |
Pos logit contributions | |
gru | +0.688 |
badge | +0.600 |
Reward | +0.575 |
ution | +0.563 |
rifle | +0.563 |
Neg logit contributions | |
day | -0.622 |
afternoon | -0.564 |
cy | -0.527 |
night | -0.526 |
ite | -0.512 |
Token | . |
Token ID | 13 |
Feature activation | +1.40e-01 |
Pos logit contributions | |
gru | +0.071 |
badge | +0.064 |
rifle | +0.060 |
address | +0.060 |
threatening | +0.060 |
Neg logit contributions | |
day | -0.052 |
afternoon | -0.049 |
cy | -0.047 |
he | -0.045 |
ite | -0.044 |
Token | |
Token ID | 220 |
Feature activation | -3.15e-01 |
Pos logit contributions | |
day | +0.005 |
afternoon | +0.005 |
cy | +0.004 |
night | +0.004 |
ite | +0.004 |
Neg logit contributions | |
gru | -0.006 |
badge | -0.005 |
Reward | -0.005 |
ution | -0.005 |
rifle | -0.005 |
Token |
|
Token ID | 198 |
Feature activation | -4.39e-01 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.012 |
Reward | +0.012 |
rifle | +0.012 |
threatening | +0.012 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.010 |
night | -0.009 |
cy | -0.009 |
days | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -4.74e-01 |
Pos logit contributions | |
gru | +0.020 |
Reward | +0.017 |
badge | +0.017 |
threatening | +0.016 |
ution | +0.016 |
Neg logit contributions | |
day | -0.015 |
afternoon | -0.013 |
cy | -0.012 |
night | -0.012 |
ite | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | +3.26e-01 |
Pos logit contributions | |
day | +0.100 |
afternoon | +0.085 |
night | +0.078 |
cy | +0.073 |
morning | +0.072 |
Neg logit contributions | |
gru | -0.195 |
badge | -0.176 |
Reward | -0.174 |
rifle | -0.171 |
ution | -0.171 |
Token | She |
Token ID | 1375 |
Feature activation | -2.06e+00 |
Pos logit contributions | |
day | +0.013 |
afternoon | +0.012 |
cy | +0.011 |
ite | +0.011 |
night | +0.011 |
Neg logit contributions | |
gru | -0.013 |
badge | -0.010 |
Reward | -0.010 |
address | -0.010 |
threatening | -0.010 |
Token | grabbed |
Token ID | 13176 |
Feature activation | -1.04e+00 |
Pos logit contributions | |
gru | +0.095 |
badge | +0.084 |
Reward | +0.080 |
ution | +0.079 |
rifle | +0.079 |
Neg logit contributions | |
day | -0.069 |
afternoon | -0.062 |
cy | -0.057 |
night | -0.057 |
ite | -0.055 |
Token | the |
Token ID | 262 |
Feature activation | +3.00e+00 |
Pos logit contributions | |
gru | +0.047 |
badge | +0.042 |
Reward | +0.040 |
ution | +0.040 |
rifle | +0.039 |
Neg logit contributions | |
day | -0.036 |
afternoon | -0.032 |
night | -0.030 |
cy | -0.030 |
morning | -0.029 |
Token | fire |
Token ID | 2046 |
Feature activation | -6.88e+00 |
Pos logit contributions | |
day | +0.147 |
afternoon | +0.131 |
night | +0.131 |
morning | +0.125 |
evening | +0.119 |
Neg logit contributions | |
gru | -0.097 |
ution | -0.083 |
Reward | -0.081 |
badge | -0.079 |
rets | -0.078 |
Token | extingu |
Token ID | 24995 |
Feature activation | -1.65e+01 |
Pos logit contributions | |
gru | +0.316 |
badge | +0.281 |
Reward | +0.266 |
rifle | +0.265 |
threatening | +0.263 |
Neg logit contributions | |
day | -0.226 |
afternoon | -0.204 |
cy | -0.185 |
night | -0.185 |
ite | -0.181 |
Token | isher |
Token ID | 4828 |
Feature activation | -5.77e+00 |
Pos logit contributions | |
gru | +0.778 |
badge | +0.675 |
Reward | +0.672 |
spoken | +0.658 |
threatening | +0.657 |
Neg logit contributions | |
day | -0.526 |
cy | -0.502 |
afternoon | -0.472 |
ite | -0.459 |
night | -0.442 |
Token | and |
Token ID | 290 |
Feature activation | +1.61e+00 |
Pos logit contributions | |
gru | +0.297 |
badge | +0.267 |
Reward | +0.254 |
rifle | +0.254 |
threatening | +0.252 |
Neg logit contributions | |
day | -0.158 |
afternoon | -0.140 |
cy | -0.128 |
night | -0.125 |
ite | -0.121 |
Token | put |
Token ID | 1234 |
Feature activation | +4.38e+00 |
Pos logit contributions | |
day | +0.051 |
afternoon | +0.045 |
night | +0.043 |
cy | +0.042 |
ite | +0.040 |
Neg logit contributions | |
gru | -0.077 |
badge | -0.067 |
Reward | -0.066 |
ution | -0.065 |
rifle | -0.064 |
Token | out |
Token ID | 503 |
Feature activation | -2.07e+00 |
Pos logit contributions | |
day | +0.187 |
afternoon | +0.173 |
cy | +0.163 |
night | +0.162 |
ite | +0.159 |
Neg logit contributions | |
gru | -0.160 |
badge | -0.137 |
Reward | -0.129 |
ution | -0.126 |
rifle | -0.126 |
Token | the |
Token ID | 262 |
Feature activation | +3.05e+00 |
Pos logit contributions | |
gru | +0.089 |
badge | +0.078 |
Reward | +0.075 |
ution | +0.074 |
rifle | +0.073 |
Neg logit contributions | |
day | -0.076 |
afternoon | -0.068 |
night | -0.064 |
cy | -0.063 |
ite | -0.061 |
Token | little |
Token ID | 1310 |
Feature activation | +3.75e+00 |
Pos logit contributions | |
cy | +0.034 |
day | +0.033 |
he | +0.032 |
ite | +0.032 |
afternoon | +0.032 |
Neg logit contributions | |
gru | -0.042 |
badge | -0.036 |
warning | -0.034 |
god | -0.033 |
message | -0.033 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
day | +0.115 |
he | +0.115 |
afternoon | +0.108 |
ops | +0.104 |
olly | +0.101 |
Neg logit contributions | |
gru | -0.165 |
badge | -0.143 |
god | -0.138 |
rifle | -0.133 |
warning | -0.130 |
Token | looked |
Token ID | 3114 |
Feature activation | -2.79e-01 |
Pos logit contributions | |
day | +0.041 |
afternoon | +0.038 |
he | +0.037 |
ite | +0.036 |
cy | +0.036 |
Neg logit contributions | |
gru | -0.051 |
god | -0.045 |
probably | -0.043 |
badge | -0.043 |
address | -0.043 |
Token | up |
Token ID | 510 |
Feature activation | +3.62e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.012 |
Reward | +0.012 |
ution | +0.012 |
rifle | +0.012 |
Neg logit contributions | |
day | -0.008 |
afternoon | -0.007 |
night | -0.007 |
cy | -0.007 |
morning | -0.006 |
Token | in |
Token ID | 287 |
Feature activation | +4.75e+00 |
Pos logit contributions | |
day | +0.112 |
afternoon | +0.100 |
night | +0.092 |
cy | +0.091 |
ite | +0.087 |
Neg logit contributions | |
gru | -0.176 |
badge | -0.157 |
Reward | -0.152 |
ution | -0.149 |
rifle | -0.149 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.65e+01 |
Pos logit contributions | |
day | +0.169 |
afternoon | +0.149 |
night | +0.144 |
morning | +0.135 |
cy | +0.132 |
Neg logit contributions | |
gru | -0.215 |
badge | -0.193 |
Reward | -0.188 |
ution | -0.186 |
phant | -0.185 |
Token | ement |
Token ID | 972 |
Feature activation | -1.65e+00 |
Pos logit contributions | |
gru | +0.733 |
badge | +0.647 |
Reward | +0.639 |
rifle | +0.624 |
threatening | +0.614 |
Neg logit contributions | |
day | -0.572 |
afternoon | -0.517 |
night | -0.479 |
he | -0.475 |
cy | -0.473 |
Token | and |
Token ID | 290 |
Feature activation | +1.56e+00 |
Pos logit contributions | |
gru | +0.078 |
badge | +0.069 |
threatening | +0.065 |
rifle | +0.065 |
Reward | +0.065 |
Neg logit contributions | |
day | -0.051 |
afternoon | -0.046 |
cy | -0.044 |
ite | -0.042 |
night | -0.042 |
Token | said |
Token ID | 531 |
Feature activation | -3.64e+00 |
Pos logit contributions | |
day | +0.056 |
afternoon | +0.051 |
night | +0.047 |
cy | +0.047 |
morning | +0.045 |
Neg logit contributions | |
gru | -0.068 |
badge | -0.060 |
Reward | -0.058 |
rifle | -0.057 |
ution | -0.057 |
Token | , |
Token ID | 11 |
Feature activation | -4.50e-01 |
Pos logit contributions | |
gru | +0.207 |
badge | +0.193 |
Reward | +0.192 |
ution | +0.192 |
rifle | +0.190 |
Neg logit contributions | |
day | -0.076 |
afternoon | -0.057 |
night | -0.053 |
morning | -0.051 |
olly | -0.046 |
Token | " |
Token ID | 366 |
Feature activation | -3.06e+00 |
Pos logit contributions | |
gru | +0.017 |
badge | +0.015 |
ution | +0.015 |
Reward | +0.014 |
rifle | +0.014 |
Neg logit contributions | |
day | -0.018 |
afternoon | -0.016 |
night | -0.015 |
morning | -0.015 |
cy | -0.014 |
Token | they |
Token ID | 484 |
Feature activation | +1.77e+00 |
Pos logit contributions | |
day | +0.000 |
afternoon | +0.000 |
cy | +0.000 |
night | +0.000 |
ite | +0.000 |
Neg logit contributions | |
gru | -0.000 |
badge | -0.000 |
Reward | -0.000 |
threatening | -0.000 |
rifle | -0.000 |
Token | both |
Token ID | 1111 |
Feature activation | +5.90e+00 |
Pos logit contributions | |
day | +0.062 |
afternoon | +0.057 |
cy | +0.052 |
night | +0.052 |
ite | +0.051 |
Neg logit contributions | |
gru | -0.076 |
badge | -0.068 |
Reward | -0.064 |
threatening | -0.063 |
rifle | -0.063 |
Token | looked |
Token ID | 3114 |
Feature activation | -2.27e-01 |
Pos logit contributions | |
day | +0.215 |
afternoon | +0.201 |
ite | +0.188 |
cy | +0.187 |
he | +0.185 |
Neg logit contributions | |
gru | -0.240 |
badge | -0.203 |
spoken | -0.196 |
threatening | -0.194 |
threatened | -0.189 |
Token | around |
Token ID | 1088 |
Feature activation | +4.40e+00 |
Pos logit contributions | |
gru | +0.010 |
badge | +0.009 |
Reward | +0.009 |
ution | +0.009 |
rifle | +0.009 |
Neg logit contributions | |
day | -0.008 |
afternoon | -0.007 |
night | -0.007 |
cy | -0.006 |
morning | -0.006 |
Token | in |
Token ID | 287 |
Feature activation | +4.74e+00 |
Pos logit contributions | |
day | +0.119 |
afternoon | +0.102 |
night | +0.093 |
cy | +0.088 |
morning | +0.086 |
Neg logit contributions | |
gru | -0.233 |
badge | -0.212 |
Reward | -0.208 |
ution | -0.204 |
rifle | -0.203 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.65e+01 |
Pos logit contributions | |
day | +0.167 |
afternoon | +0.148 |
night | +0.139 |
cy | +0.133 |
morning | +0.130 |
Neg logit contributions | |
gru | -0.214 |
badge | -0.193 |
Reward | -0.186 |
rifle | -0.183 |
ution | -0.183 |
Token | ement |
Token ID | 972 |
Feature activation | -1.64e+00 |
Pos logit contributions | |
gru | +0.717 |
badge | +0.630 |
Reward | +0.616 |
rifle | +0.601 |
threatening | +0.594 |
Neg logit contributions | |
day | -0.590 |
afternoon | -0.533 |
night | -0.494 |
cy | -0.491 |
he | -0.483 |
Token | . |
Token ID | 13 |
Feature activation | +3.96e-01 |
Pos logit contributions | |
gru | +0.076 |
badge | +0.068 |
threatening | +0.064 |
rifle | +0.063 |
warning | +0.063 |
Neg logit contributions | |
day | -0.052 |
afternoon | -0.047 |
cy | -0.045 |
ite | -0.043 |
night | -0.043 |
Token | The |
Token ID | 383 |
Feature activation | -7.45e-01 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.013 |
night | +0.012 |
cy | +0.012 |
ite | +0.011 |
Neg logit contributions | |
gru | -0.017 |
badge | -0.015 |
Reward | -0.015 |
ution | -0.014 |
rifle | -0.014 |
Token | lake |
Token ID | 13546 |
Feature activation | -1.79e+00 |
Pos logit contributions | |
gru | +0.030 |
badge | +0.025 |
rifle | +0.023 |
threatening | +0.023 |
warning | +0.023 |
Neg logit contributions | |
day | -0.028 |
afternoon | -0.026 |
cy | -0.026 |
he | -0.025 |
ite | -0.025 |
Token | was |
Token ID | 373 |
Feature activation | +8.22e-01 |
Pos logit contributions | |
gru | +0.093 |
badge | +0.083 |
Reward | +0.079 |
ution | +0.079 |
rifle | +0.078 |
Neg logit contributions | |
day | -0.049 |
afternoon | -0.043 |
cy | -0.039 |
night | -0.039 |
ite | -0.038 |
Token | The |
Token ID | 383 |
Feature activation | -6.51e-01 |
Pos logit contributions | |
day | +0.020 |
afternoon | +0.019 |
cy | +0.017 |
night | +0.017 |
ite | +0.017 |
Neg logit contributions | |
gru | -0.022 |
badge | -0.019 |
Reward | -0.018 |
ution | -0.017 |
rifle | -0.017 |
Token | little |
Token ID | 1310 |
Feature activation | +4.07e+00 |
Pos logit contributions | |
gru | +0.025 |
badge | +0.022 |
warning | +0.020 |
address | +0.020 |
Reward | +0.020 |
Neg logit contributions | |
day | -0.024 |
cy | -0.023 |
ite | -0.023 |
afternoon | -0.023 |
he | -0.022 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.42e+00 |
Pos logit contributions | |
day | +0.123 |
afternoon | +0.112 |
he | +0.110 |
ite | +0.104 |
night | +0.102 |
Neg logit contributions | |
gru | -0.190 |
badge | -0.168 |
god | -0.158 |
Reward | -0.155 |
rifle | -0.154 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.03e-01 |
Pos logit contributions | |
day | +0.048 |
afternoon | +0.044 |
ite | +0.041 |
cy | +0.041 |
night | +0.041 |
Neg logit contributions | |
gru | -0.062 |
badge | -0.054 |
god | -0.052 |
threatening | -0.051 |
probably | -0.051 |
Token | with |
Token ID | 351 |
Feature activation | -2.76e+00 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
Reward | +0.007 |
threatening | +0.007 |
ution | +0.007 |
Neg logit contributions | |
day | -0.008 |
afternoon | -0.007 |
cy | -0.007 |
ite | -0.007 |
night | -0.006 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.64e+01 |
Pos logit contributions | |
gru | +0.114 |
badge | +0.099 |
Reward | +0.095 |
ution | +0.093 |
rifle | +0.093 |
Neg logit contributions | |
day | -0.107 |
afternoon | -0.097 |
night | -0.092 |
cy | -0.090 |
morning | -0.087 |
Token | ement |
Token ID | 972 |
Feature activation | -1.70e+00 |
Pos logit contributions | |
gru | +0.707 |
badge | +0.618 |
Reward | +0.601 |
rifle | +0.589 |
threatening | +0.581 |
Neg logit contributions | |
day | -0.598 |
afternoon | -0.541 |
night | -0.502 |
cy | -0.500 |
he | -0.489 |
Token | when |
Token ID | 618 |
Feature activation | +1.84e+00 |
Pos logit contributions | |
gru | +0.069 |
badge | +0.062 |
warning | +0.058 |
threatening | +0.058 |
address | +0.058 |
Neg logit contributions | |
day | -0.059 |
afternoon | -0.055 |
cy | -0.054 |
ite | -0.053 |
he | -0.051 |
Token | her |
Token ID | 607 |
Feature activation | -1.71e+00 |
Pos logit contributions | |
day | +0.063 |
afternoon | +0.058 |
cy | +0.055 |
ite | +0.053 |
night | +0.052 |
Neg logit contributions | |
gru | -0.082 |
badge | -0.071 |
Reward | -0.068 |
threatening | -0.067 |
ution | -0.067 |
Token | mom |
Token ID | 1995 |
Feature activation | +2.81e+00 |
Pos logit contributions | |
gru | +0.062 |
badge | +0.055 |
address | +0.052 |
warning | +0.051 |
god | +0.050 |
Neg logit contributions | |
day | -0.067 |
afternoon | -0.064 |
cy | -0.063 |
he | -0.063 |
ite | -0.063 |
Token | my |
Token ID | 1820 |
Feature activation | +4.63e+00 |
Pos logit contributions | |
day | +0.110 |
afternoon | +0.103 |
ite | +0.094 |
night | +0.093 |
olly | +0.091 |
Neg logit contributions | |
gru | -0.107 |
badge | -0.095 |
god | -0.092 |
address | -0.089 |
Reward | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | +6.26e-01 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.011 |
threatening | +0.011 |
warning | +0.011 |
ution | +0.010 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.009 |
ite | -0.008 |
olly | -0.008 |
cy | -0.008 |
Token | He |
Token ID | 679 |
Feature activation | -3.07e+00 |
Pos logit contributions | |
day | +0.024 |
afternoon | +0.021 |
night | +0.020 |
cy | +0.020 |
ite | +0.019 |
Neg logit contributions | |
gru | -0.025 |
badge | -0.023 |
ution | -0.022 |
Reward | -0.021 |
threatening | -0.021 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.84e-01 |
Pos logit contributions | |
gru | +0.148 |
badge | +0.133 |
Reward | +0.128 |
ution | +0.127 |
rifle | +0.125 |
Neg logit contributions | |
day | -0.095 |
afternoon | -0.083 |
cy | -0.077 |
night | -0.077 |
ite | -0.074 |
Token | it |
Token ID | 340 |
Feature activation | +8.97e-01 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.012 |
Reward | +0.011 |
ution | +0.011 |
rifle | +0.011 |
Neg logit contributions | |
day | -0.010 |
afternoon | -0.008 |
night | -0.008 |
cy | -0.008 |
morning | -0.008 |
Token | in |
Token ID | 287 |
Feature activation | +4.51e+00 |
Pos logit contributions | |
day | +0.033 |
afternoon | +0.030 |
cy | +0.029 |
ite | +0.028 |
night | +0.028 |
Neg logit contributions | |
gru | -0.037 |
badge | -0.033 |
Reward | -0.031 |
threatening | -0.031 |
rifle | -0.031 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.63e+01 |
Pos logit contributions | |
day | +0.160 |
afternoon | +0.138 |
night | +0.137 |
morning | +0.123 |
days | +0.122 |
Neg logit contributions | |
gru | -0.203 |
badge | -0.185 |
Reward | -0.179 |
ution | -0.179 |
CRE | -0.176 |
Token | ement |
Token ID | 972 |
Feature activation | -1.64e+00 |
Pos logit contributions | |
gru | +0.704 |
badge | +0.618 |
Reward | +0.600 |
rifle | +0.587 |
threatening | +0.580 |
Neg logit contributions | |
day | -0.589 |
afternoon | -0.534 |
night | -0.497 |
cy | -0.496 |
he | -0.485 |
Token | as |
Token ID | 355 |
Feature activation | +2.51e+00 |
Pos logit contributions | |
gru | +0.076 |
badge | +0.068 |
threatening | +0.064 |
Reward | +0.064 |
rifle | +0.063 |
Neg logit contributions | |
day | -0.053 |
afternoon | -0.048 |
cy | -0.045 |
ite | -0.043 |
night | -0.043 |
Token | it |
Token ID | 340 |
Feature activation | +8.27e-01 |
Pos logit contributions | |
day | +0.108 |
afternoon | +0.103 |
cy | +0.099 |
ite | +0.096 |
olly | +0.094 |
Neg logit contributions | |
gru | -0.089 |
badge | -0.074 |
threatening | -0.069 |
Reward | -0.069 |
warning | -0.067 |
Token | went |
Token ID | 1816 |
Feature activation | +8.45e+00 |
Pos logit contributions | |
day | +0.028 |
afternoon | +0.025 |
cy | +0.023 |
night | +0.023 |
ite | +0.023 |
Neg logit contributions | |
gru | -0.038 |
badge | -0.033 |
Reward | -0.032 |
ution | -0.031 |
rifle | -0.031 |
Token | higher |
Token ID | 2440 |
Feature activation | +3.40e+00 |
Pos logit contributions | |
day | +0.254 |
afternoon | +0.211 |
night | +0.202 |
morning | +0.193 |
ite | +0.182 |
Neg logit contributions | |
gru | -0.418 |
badge | -0.380 |
ution | -0.371 |
spoken | -0.368 |
address | -0.367 |
Token |
|
Token ID | 198 |
Feature activation | -7.76e-02 |
Pos logit contributions | |
day | +0.063 |
afternoon | +0.059 |
night | +0.054 |
cy | +0.053 |
morning | +0.051 |
Neg logit contributions | |
gru | -0.068 |
badge | -0.058 |
Reward | -0.057 |
rifle | -0.055 |
ution | -0.055 |
Token |
|
Token ID | 198 |
Feature activation | -1.93e-01 |
Pos logit contributions | |
gru | +0.004 |
badge | +0.003 |
Reward | +0.003 |
ution | +0.003 |
threatening | +0.003 |
Neg logit contributions | |
day | -0.002 |
afternoon | -0.002 |
cy | -0.002 |
night | -0.002 |
ite | -0.002 |
Token | Sam |
Token ID | 16305 |
Feature activation | -9.21e-01 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
rifle | +0.007 |
Reward | +0.007 |
ution | +0.007 |
Neg logit contributions | |
day | -0.007 |
afternoon | -0.006 |
cy | -0.006 |
ite | -0.006 |
he | -0.006 |
Token | gasped |
Token ID | 45236 |
Feature activation | -1.77e+00 |
Pos logit contributions | |
gru | +0.039 |
badge | +0.034 |
god | +0.032 |
Reward | +0.032 |
ution | +0.031 |
Neg logit contributions | |
day | -0.034 |
afternoon | -0.031 |
night | -0.029 |
ite | -0.028 |
days | -0.028 |
Token | in |
Token ID | 287 |
Feature activation | +4.16e+00 |
Pos logit contributions | |
gru | +0.090 |
badge | +0.081 |
Reward | +0.079 |
ution | +0.078 |
rifle | +0.077 |
Neg logit contributions | |
day | -0.050 |
afternoon | -0.044 |
night | -0.040 |
cy | -0.040 |
he | -0.037 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.63e+01 |
Pos logit contributions | |
day | +0.167 |
afternoon | +0.147 |
night | +0.143 |
morning | +0.135 |
evening | +0.130 |
Neg logit contributions | |
gru | -0.157 |
phant | -0.142 |
badge | -0.141 |
rier | -0.140 |
ution | -0.137 |
Token | ement |
Token ID | 972 |
Feature activation | -1.43e+00 |
Pos logit contributions | |
gru | +0.709 |
badge | +0.621 |
Reward | +0.610 |
rifle | +0.598 |
threatening | +0.589 |
Neg logit contributions | |
day | -0.579 |
afternoon | -0.524 |
night | -0.485 |
cy | -0.483 |
he | -0.479 |
Token | . |
Token ID | 13 |
Feature activation | +5.02e-01 |
Pos logit contributions | |
gru | +0.070 |
badge | +0.063 |
Reward | +0.060 |
rifle | +0.059 |
threatening | +0.059 |
Neg logit contributions | |
day | -0.043 |
afternoon | -0.038 |
cy | -0.036 |
night | -0.035 |
ite | -0.034 |
Token | He |
Token ID | 679 |
Feature activation | -3.38e+00 |
Pos logit contributions | |
day | +0.022 |
afternoon | +0.021 |
cy | +0.020 |
night | +0.020 |
ite | +0.019 |
Neg logit contributions | |
gru | -0.017 |
badge | -0.015 |
Reward | -0.014 |
ution | -0.014 |
rifle | -0.013 |
Token | had |
Token ID | 550 |
Feature activation | +4.99e+00 |
Pos logit contributions | |
gru | +0.153 |
badge | +0.135 |
Reward | +0.129 |
ution | +0.127 |
rifle | +0.127 |
Neg logit contributions | |
day | -0.115 |
afternoon | -0.104 |
cy | -0.096 |
night | -0.096 |
ite | -0.093 |
Token | never |
Token ID | 1239 |
Feature activation | +8.72e-01 |
Pos logit contributions | |
day | +0.186 |
afternoon | +0.164 |
night | +0.162 |
morning | +0.148 |
days | +0.144 |
Neg logit contributions | |
gru | -0.215 |
badge | -0.186 |
Reward | -0.181 |
rifle | -0.180 |
ution | -0.179 |
Token | Her |
Token ID | 2332 |
Feature activation | -5.37e+00 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.019 |
night | +0.017 |
cy | +0.017 |
morning | +0.017 |
Neg logit contributions | |
gru | -0.030 |
badge | -0.026 |
ution | -0.025 |
rifle | -0.025 |
Reward | -0.025 |
Token | friend |
Token ID | 1545 |
Feature activation | -9.01e-01 |
Pos logit contributions | |
gru | +0.210 |
badge | +0.184 |
Reward | +0.171 |
ution | +0.170 |
rifle | +0.170 |
Neg logit contributions | |
day | -0.214 |
afternoon | -0.197 |
cy | -0.185 |
night | -0.184 |
he | -0.180 |
Token | started |
Token ID | 2067 |
Feature activation | +3.99e+00 |
Pos logit contributions | |
gru | +0.042 |
badge | +0.037 |
Reward | +0.036 |
ution | +0.035 |
rifle | +0.035 |
Neg logit contributions | |
day | -0.030 |
afternoon | -0.027 |
cy | -0.025 |
night | -0.025 |
ite | -0.024 |
Token | crying |
Token ID | 13774 |
Feature activation | -1.67e+00 |
Pos logit contributions | |
day | +0.129 |
afternoon | +0.103 |
night | +0.096 |
morning | +0.092 |
cy | +0.086 |
Neg logit contributions | |
gru | -0.201 |
Reward | -0.178 |
god | -0.171 |
Amen | -0.171 |
rier | -0.170 |
Token | and |
Token ID | 290 |
Feature activation | +1.98e+00 |
Pos logit contributions | |
gru | +0.077 |
badge | +0.070 |
threatening | +0.067 |
warning | +0.065 |
Reward | +0.064 |
Neg logit contributions | |
day | -0.052 |
afternoon | -0.048 |
ite | -0.046 |
cy | -0.045 |
olly | -0.043 |
Token | ran |
Token ID | 4966 |
Feature activation | +6.42e+00 |
Pos logit contributions | |
day | +0.069 |
afternoon | +0.060 |
night | +0.057 |
cy | +0.055 |
morning | +0.054 |
Neg logit contributions | |
gru | -0.090 |
badge | -0.080 |
rifle | -0.077 |
ution | -0.076 |
Reward | -0.076 |
Token | away |
Token ID | 1497 |
Feature activation | -3.06e+00 |
Pos logit contributions | |
day | +0.220 |
afternoon | +0.190 |
night | +0.181 |
cy | +0.174 |
morning | +0.168 |
Neg logit contributions | |
gru | -0.306 |
god | -0.260 |
Reward | -0.259 |
beard | -0.258 |
badge | -0.256 |
Token | . |
Token ID | 13 |
Feature activation | +6.03e-01 |
Pos logit contributions | |
gru | +0.170 |
badge | +0.154 |
Reward | +0.152 |
rifle | +0.150 |
ution | +0.150 |
Neg logit contributions | |
day | -0.077 |
afternoon | -0.061 |
night | -0.058 |
cy | -0.054 |
he | -0.052 |
Token |
|
Token ID | 198 |
Feature activation | -2.59e-02 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.017 |
cy | +0.015 |
night | +0.015 |
ite | +0.015 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.026 |
Reward | -0.025 |
ution | -0.024 |
threatening | -0.024 |
Token |
|
Token ID | 198 |
Feature activation | -5.31e-02 |
Pos logit contributions | |
gru | +0.001 |
Reward | +0.001 |
badge | +0.001 |
threatening | +0.001 |
ution | +0.001 |
Neg logit contributions | |
day | -0.001 |
afternoon | -0.001 |
ite | -0.001 |
cy | -0.001 |
night | -0.001 |
Token | L |
Token ID | 43 |
Feature activation | +1.44e+00 |
Pos logit contributions | |
gru | +0.002 |
badge | +0.002 |
Reward | +0.002 |
threatening | +0.002 |
ution | +0.002 |
Neg logit contributions | |
day | -0.002 |
afternoon | -0.002 |
cy | -0.002 |
ite | -0.002 |
night | -0.002 |
Token | it |
Token ID | 340 |
Feature activation | +2.06e-01 |
Pos logit contributions | |
gru | +0.048 |
badge | +0.042 |
Reward | +0.041 |
ution | +0.041 |
threatening | +0.041 |
Neg logit contributions | |
day | -0.030 |
afternoon | -0.025 |
night | -0.025 |
cy | -0.025 |
ite | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | +9.64e-01 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.006 |
cy | +0.005 |
night | +0.005 |
ite | +0.005 |
Neg logit contributions | |
gru | -0.010 |
badge | -0.009 |
Reward | -0.009 |
ution | -0.009 |
rifle | -0.008 |
Token | They |
Token ID | 1119 |
Feature activation | -8.07e-01 |
Pos logit contributions | |
day | +0.030 |
afternoon | +0.025 |
night | +0.024 |
cy | +0.023 |
ite | +0.023 |
Neg logit contributions | |
gru | -0.048 |
badge | -0.043 |
ution | -0.041 |
Reward | -0.040 |
god | -0.040 |
Token | see |
Token ID | 766 |
Feature activation | -9.76e-01 |
Pos logit contributions | |
gru | +0.031 |
badge | +0.028 |
Reward | +0.026 |
address | +0.025 |
rifle | +0.025 |
Neg logit contributions | |
day | -0.032 |
afternoon | -0.030 |
night | -0.027 |
olly | -0.027 |
cy | -0.027 |
Token | other |
Token ID | 584 |
Feature activation | +4.41e+00 |
Pos logit contributions | |
gru | +0.042 |
badge | +0.036 |
Reward | +0.035 |
ution | +0.034 |
rifle | +0.034 |
Neg logit contributions | |
day | -0.037 |
afternoon | -0.033 |
night | -0.031 |
cy | -0.030 |
ite | -0.029 |
Token | kids |
Token ID | 3988 |
Feature activation | +4.24e+00 |
Pos logit contributions | |
day | +0.151 |
afternoon | +0.137 |
cy | +0.129 |
night | +0.126 |
ite | +0.126 |
Neg logit contributions | |
gru | -0.197 |
badge | -0.175 |
Reward | -0.167 |
threatening | -0.163 |
ution | -0.163 |
Token | running |
Token ID | 2491 |
Feature activation | +5.37e+00 |
Pos logit contributions | |
day | +0.161 |
afternoon | +0.152 |
night | +0.139 |
cy | +0.138 |
olly | +0.137 |
Neg logit contributions | |
gru | -0.167 |
badge | -0.150 |
threatening | -0.141 |
warning | -0.141 |
Reward | -0.139 |
Token | to |
Token ID | 284 |
Feature activation | -5.57e+00 |
Pos logit contributions | |
day | +0.175 |
afternoon | +0.159 |
cy | +0.146 |
night | +0.145 |
ite | +0.144 |
Neg logit contributions | |
gru | -0.248 |
badge | -0.220 |
Reward | -0.209 |
rifle | -0.207 |
threatening | -0.207 |
Token | the |
Token ID | 262 |
Feature activation | +3.04e+00 |
Pos logit contributions | |
gru | +0.261 |
badge | +0.230 |
Reward | +0.220 |
rifle | +0.219 |
ution | +0.216 |
Neg logit contributions | |
day | -0.188 |
afternoon | -0.164 |
night | -0.155 |
cy | -0.153 |
morning | -0.145 |
Token | park |
Token ID | 3952 |
Feature activation | +3.81e+00 |
Pos logit contributions | |
day | +0.100 |
afternoon | +0.088 |
night | +0.082 |
cy | +0.079 |
morning | +0.078 |
Neg logit contributions | |
gru | -0.145 |
badge | -0.127 |
Reward | -0.123 |
rifle | -0.122 |
ution | -0.121 |
Token | too |
Token ID | 1165 |
Feature activation | -1.60e-01 |
Pos logit contributions | |
day | +0.118 |
afternoon | +0.109 |
cy | +0.102 |
ite | +0.099 |
olly | +0.099 |
Neg logit contributions | |
gru | -0.178 |
badge | -0.158 |
Reward | -0.151 |
threatening | -0.149 |
ution | -0.149 |
Token | fly |
Token ID | 6129 |
Feature activation | +3.88e+00 |
Pos logit contributions | |
gru | +0.287 |
threatening | +0.233 |
menacing | +0.228 |
rifle | +0.226 |
stern | +0.226 |
Neg logit contributions | |
day | -0.181 |
night | -0.143 |
afternoon | -0.142 |
morning | -0.133 |
ite | -0.128 |
Token | in |
Token ID | 287 |
Feature activation | +4.10e+00 |
Pos logit contributions | |
day | +0.125 |
afternoon | +0.106 |
night | +0.102 |
cy | +0.098 |
ite | +0.096 |
Neg logit contributions | |
gru | -0.192 |
badge | -0.166 |
Reward | -0.163 |
ution | -0.163 |
warning | -0.160 |
Token | one |
Token ID | 530 |
Feature activation | -3.65e+00 |
Pos logit contributions | |
day | +0.127 |
afternoon | +0.111 |
night | +0.104 |
cy | +0.098 |
morning | +0.096 |
Neg logit contributions | |
gru | -0.203 |
badge | -0.180 |
Reward | -0.173 |
ution | -0.172 |
rifle | -0.171 |
Token | . |
Token ID | 13 |
Feature activation | +5.61e-01 |
Pos logit contributions | |
gru | +0.139 |
god | +0.131 |
badge | +0.128 |
Reward | +0.126 |
message | +0.125 |
Neg logit contributions | |
ops | -0.111 |
day | -0.110 |
afternoon | -0.107 |
ugg | -0.104 |
cy | -0.104 |
Token | One |
Token ID | 1881 |
Feature activation | -2.02e+00 |
Pos logit contributions | |
day | +0.024 |
afternoon | +0.022 |
cy | +0.021 |
night | +0.021 |
ite | +0.020 |
Neg logit contributions | |
gru | -0.021 |
badge | -0.018 |
Reward | -0.017 |
ution | -0.016 |
rifle | -0.016 |
Token | day |
Token ID | 1110 |
Feature activation | +1.07e+01 |
Pos logit contributions | |
gru | +0.107 |
message | +0.102 |
letter | +0.099 |
warning | +0.099 |
badge | +0.098 |
Neg logit contributions | |
ips | -0.033 |
oms | -0.028 |
ops | -0.025 |
olly | -0.024 |
ites | -0.023 |
Token | , |
Token ID | 11 |
Feature activation | +1.33e-02 |
Pos logit contributions | |
day | +0.299 |
afternoon | +0.284 |
cy | +0.264 |
he | +0.260 |
ite | +0.253 |
Neg logit contributions | |
gru | -0.502 |
badge | -0.464 |
Reward | -0.440 |
threatening | -0.428 |
rifle | -0.428 |
Token | Tim |
Token ID | 5045 |
Feature activation | +2.82e+00 |
Pos logit contributions | |
day | +0.001 |
afternoon | +0.000 |
night | +0.000 |
cy | +0.000 |
morning | +0.000 |
Neg logit contributions | |
gru | -0.001 |
badge | -0.000 |
Reward | -0.000 |
god | -0.000 |
ution | -0.000 |
Token | my |
Token ID | 1820 |
Feature activation | +4.57e+00 |
Pos logit contributions | |
olly | +0.098 |
ite | +0.093 |
ugg | +0.091 |
day | +0.091 |
oms | +0.091 |
Neg logit contributions | |
probably | -0.089 |
spoken | -0.087 |
gru | -0.085 |
badge | -0.082 |
god | -0.081 |
Token | 's |
Token ID | 338 |
Feature activation | -4.16e+00 |
Pos logit contributions | |
day | +0.180 |
afternoon | +0.172 |
he | +0.169 |
olly | +0.163 |
cy | +0.163 |
Neg logit contributions | |
gru | -0.153 |
badge | -0.137 |
rifle | -0.130 |
spoken | -0.130 |
god | -0.129 |
Token | mom |
Token ID | 1995 |
Feature activation | +3.14e+00 |
Pos logit contributions | |
gru | +0.138 |
badge | +0.134 |
address | +0.120 |
rifle | +0.117 |
message | +0.117 |
Neg logit contributions | |
day | -0.171 |
afternoon | -0.166 |
cy | -0.163 |
he | -0.154 |
night | -0.152 |
Token | that |
Token ID | 326 |
Feature activation | -1.88e+00 |
Pos logit contributions | |
gru | +0.166 |
badge | +0.144 |
Reward | +0.141 |
rifle | +0.140 |
ution | +0.138 |
Neg logit contributions | |
day | -0.130 |
afternoon | -0.114 |
night | -0.108 |
cy | -0.102 |
ite | -0.102 |
Token | she |
Token ID | 673 |
Feature activation | +2.64e+00 |
Pos logit contributions | |
gru | +0.080 |
badge | +0.070 |
Reward | +0.068 |
threatening | +0.067 |
ution | +0.066 |
Neg logit contributions | |
day | -0.067 |
afternoon | -0.061 |
cy | -0.058 |
ite | -0.057 |
night | -0.056 |
Token | was |
Token ID | 373 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
day | +0.093 |
afternoon | +0.089 |
morning | +0.078 |
ite | +0.078 |
he | +0.077 |
Neg logit contributions | |
gru | -0.107 |
probably | -0.100 |
badge | -0.095 |
threatening | -0.094 |
spoken | -0.093 |
Token | lucky |
Token ID | 9670 |
Feature activation | -6.08e-01 |
Pos logit contributions | |
day | +0.045 |
afternoon | +0.039 |
night | +0.037 |
morning | +0.035 |
days | +0.035 |
Neg logit contributions | |
gru | -0.048 |
Reward | -0.043 |
badge | -0.043 |
ution | -0.043 |
rifle | -0.042 |
Token | to |
Token ID | 284 |
Feature activation | -5.36e+00 |
Pos logit contributions | |
gru | +0.023 |
badge | +0.020 |
threatening | +0.019 |
Reward | +0.018 |
warning | +0.018 |
Neg logit contributions | |
day | -0.024 |
afternoon | -0.023 |
cy | -0.022 |
ite | -0.021 |
he | -0.021 |
Token | have |
Token ID | 423 |
Feature activation | +5.89e+00 |
Pos logit contributions | |
gru | +0.267 |
badge | +0.238 |
Reward | +0.231 |
ution | +0.228 |
threatening | +0.226 |
Neg logit contributions | |
day | -0.156 |
afternoon | -0.140 |
cy | -0.129 |
night | -0.126 |
ite | -0.126 |
Token | such |
Token ID | 884 |
Feature activation | -6.14e+00 |
Pos logit contributions | |
day | +0.236 |
afternoon | +0.213 |
night | +0.200 |
days | +0.194 |
cy | +0.192 |
Neg logit contributions | |
gru | -0.242 |
badge | -0.208 |
Reward | -0.201 |
rifle | -0.199 |
ution | -0.191 |
Token | good |
Token ID | 922 |
Feature activation | +2.40e+00 |
Pos logit contributions | |
gru | +0.285 |
ution | +0.259 |
Reward | +0.255 |
rifle | +0.249 |
rets | +0.246 |
Neg logit contributions | |
day | -0.208 |
afternoon | -0.172 |
night | -0.168 |
days | -0.161 |
morning | -0.150 |
Token | friends |
Token ID | 2460 |
Feature activation | +1.70e+00 |
Pos logit contributions | |
day | +0.082 |
afternoon | +0.075 |
cy | +0.070 |
he | +0.070 |
night | +0.069 |
Neg logit contributions | |
gru | -0.104 |
badge | -0.094 |
rifle | -0.087 |
threatening | -0.087 |
warning | -0.087 |
Token | who |
Token ID | 508 |
Feature activation | -1.87e+00 |
Pos logit contributions | |
day | +0.052 |
afternoon | +0.047 |
cy | +0.043 |
night | +0.043 |
ite | +0.042 |
Neg logit contributions | |
gru | -0.081 |
badge | -0.072 |
Reward | -0.069 |
ution | -0.068 |
threatening | -0.068 |
Token | would |
Token ID | 561 |
Feature activation | +3.54e+00 |
Pos logit contributions | |
gru | +0.081 |
badge | +0.072 |
Reward | +0.069 |
threatening | +0.068 |
rifle | +0.067 |
Neg logit contributions | |
day | -0.065 |
afternoon | -0.060 |
cy | -0.055 |
night | -0.054 |
ite | -0.054 |
Token | big |
Token ID | 1263 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
gru | +0.099 |
badge | +0.085 |
Reward | +0.081 |
ution | +0.080 |
rifle | +0.079 |
Neg logit contributions | |
day | -0.099 |
afternoon | -0.090 |
cy | -0.085 |
night | -0.085 |
ite | -0.082 |
Token | plate |
Token ID | 7480 |
Feature activation | -2.76e+00 |
Pos logit contributions | |
day | +0.046 |
afternoon | +0.043 |
cy | +0.041 |
ite | +0.040 |
night | +0.039 |
Neg logit contributions | |
gru | -0.052 |
badge | -0.045 |
rifle | -0.043 |
warning | -0.042 |
threatening | -0.041 |
Token | and |
Token ID | 290 |
Feature activation | +1.67e+00 |
Pos logit contributions | |
gru | +0.142 |
badge | +0.128 |
Reward | +0.122 |
rifle | +0.122 |
threatening | +0.121 |
Neg logit contributions | |
day | -0.074 |
afternoon | -0.065 |
cy | -0.061 |
night | -0.058 |
ite | -0.057 |
Token | took |
Token ID | 1718 |
Feature activation | +1.95e+00 |
Pos logit contributions | |
day | +0.059 |
afternoon | +0.053 |
cy | +0.049 |
night | +0.049 |
ite | +0.048 |
Neg logit contributions | |
gru | -0.074 |
badge | -0.065 |
Reward | -0.062 |
ution | -0.061 |
rifle | -0.061 |
Token | it |
Token ID | 340 |
Feature activation | +5.73e-01 |
Pos logit contributions | |
day | +0.083 |
afternoon | +0.076 |
night | +0.072 |
cy | +0.072 |
ite | +0.069 |
Neg logit contributions | |
gru | -0.072 |
badge | -0.062 |
Reward | -0.059 |
ution | -0.058 |
rifle | -0.057 |
Token | outside |
Token ID | 2354 |
Feature activation | +4.31e+00 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.019 |
cy | +0.018 |
ite | +0.018 |
night | +0.018 |
Neg logit contributions | |
gru | -0.023 |
badge | -0.021 |
rifle | -0.019 |
Reward | -0.019 |
threatening | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | +4.63e-01 |
Pos logit contributions | |
day | +0.141 |
afternoon | +0.126 |
cy | +0.117 |
night | +0.116 |
ite | +0.113 |
Neg logit contributions | |
gru | -0.201 |
badge | -0.178 |
Reward | -0.171 |
ution | -0.168 |
rifle | -0.168 |
Token | When |
Token ID | 1649 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.018 |
cy | +0.017 |
night | +0.017 |
ite | +0.016 |
Neg logit contributions | |
gru | -0.018 |
badge | -0.015 |
Reward | -0.014 |
rifle | -0.014 |
threatening | -0.014 |
Token | his |
Token ID | 465 |
Feature activation | -4.13e+00 |
Pos logit contributions | |
day | +0.051 |
afternoon | +0.047 |
cy | +0.046 |
ite | +0.045 |
he | +0.044 |
Neg logit contributions | |
gru | -0.046 |
badge | -0.039 |
Reward | -0.038 |
threatening | -0.037 |
rifle | -0.036 |
Token | friend |
Token ID | 1545 |
Feature activation | -9.44e-01 |
Pos logit contributions | |
gru | +0.163 |
badge | +0.144 |
rifle | +0.135 |
address | +0.132 |
warning | +0.132 |
Neg logit contributions | |
day | -0.157 |
afternoon | -0.146 |
he | -0.141 |
cy | -0.140 |
ite | -0.138 |
Token | arrived |
Token ID | 5284 |
Feature activation | -2.67e+00 |
Pos logit contributions | |
gru | +0.040 |
badge | +0.035 |
Reward | +0.034 |
god | +0.033 |
threatening | +0.033 |
Neg logit contributions | |
day | -0.033 |
afternoon | -0.031 |
ite | -0.028 |
cy | -0.028 |
night | -0.028 |
Token | Yes |
Token ID | 5297 |
Feature activation | -1.05e+00 |
Pos logit contributions | |
gru | +0.144 |
badge | +0.127 |
Reward | +0.120 |
rifle | +0.119 |
threatening | +0.118 |
Neg logit contributions | |
day | -0.119 |
afternoon | -0.107 |
cy | -0.099 |
night | -0.099 |
ite | -0.097 |
Token | , |
Token ID | 11 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
gru | +0.047 |
badge | +0.041 |
Reward | +0.039 |
ution | +0.039 |
rifle | +0.038 |
Neg logit contributions | |
day | -0.037 |
afternoon | -0.033 |
cy | -0.031 |
night | -0.030 |
ite | -0.030 |
Token | I |
Token ID | 314 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
gru | +0.007 |
badge | +0.006 |
Reward | +0.006 |
ution | +0.006 |
rifle | +0.006 |
Neg logit contributions | |
day | -0.008 |
afternoon | -0.007 |
night | -0.007 |
cy | -0.006 |
morning | -0.006 |
Token | would |
Token ID | 561 |
Feature activation | +3.57e+00 |
Pos logit contributions | |
day | +0.045 |
afternoon | +0.041 |
night | +0.038 |
cy | +0.038 |
ite | +0.036 |
Neg logit contributions | |
gru | -0.057 |
badge | -0.050 |
Reward | -0.047 |
ution | -0.047 |
rifle | -0.047 |
Token | love |
Token ID | 1842 |
Feature activation | +2.74e-01 |
Pos logit contributions | |
day | +0.141 |
afternoon | +0.125 |
night | +0.120 |
cy | +0.116 |
morning | +0.114 |
Neg logit contributions | |
gru | -0.148 |
badge | -0.127 |
Reward | -0.122 |
rifle | -0.120 |
ution | -0.120 |
Token | some |
Token ID | 617 |
Feature activation | -3.58e+00 |
Pos logit contributions | |
day | +0.011 |
afternoon | +0.010 |
cy | +0.010 |
ite | +0.010 |
he | +0.010 |
Neg logit contributions | |
gru | -0.010 |
badge | -0.009 |
Reward | -0.008 |
ution | -0.008 |
threatening | -0.008 |
Token | !" |
Token ID | 2474 |
Feature activation | +2.52e-02 |
Pos logit contributions | |
gru | +0.113 |
god | +0.097 |
warning | +0.096 |
message | +0.094 |
badge | +0.094 |
Neg logit contributions | |
cy | -0.147 |
ite | -0.138 |
day | -0.137 |
he | -0.136 |
afternoon | -0.133 |
Token |
|
Token ID | 198 |
Feature activation | -3.05e-01 |
Pos logit contributions | |
day | +0.001 |
afternoon | +0.001 |
cy | +0.001 |
ite | +0.001 |
End | +0.001 |
Neg logit contributions | |
gru | -0.001 |
badge | -0.001 |
Reward | -0.001 |
address | -0.001 |
rifle | -0.001 |
Token |
|
Token ID | 198 |
Feature activation | -3.15e-01 |
Pos logit contributions | |
gru | +0.012 |
Reward | +0.010 |
badge | +0.009 |
god | +0.009 |
threatening | +0.009 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.011 |
cy | -0.011 |
ite | -0.010 |
night | -0.010 |
Token | The |
Token ID | 464 |
Feature activation | +1.08e+00 |
Pos logit contributions | |
gru | +0.014 |
badge | +0.013 |
Reward | +0.012 |
rifle | +0.012 |
ution | +0.012 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.010 |
night | -0.009 |
cy | -0.009 |
ite | -0.009 |
Token | bird |
Token ID | 6512 |
Feature activation | -1.60e+00 |
Pos logit contributions | |
olly | +0.047 |
cy | +0.047 |
afternoon | +0.045 |
he | +0.045 |
day | +0.045 |
Neg logit contributions | |
gru | -0.033 |
warning | -0.027 |
badge | -0.025 |
threatening | -0.025 |
message | -0.025 |
Token | dress |
Token ID | 6576 |
Feature activation | -1.14e+00 |
Pos logit contributions | |
gru | +0.143 |
badge | +0.125 |
Reward | +0.122 |
ution | +0.121 |
threatening | +0.121 |
Neg logit contributions | |
day | -0.082 |
afternoon | -0.069 |
night | -0.065 |
cy | -0.061 |
morning | -0.061 |
Token | would |
Token ID | 561 |
Feature activation | +3.59e+00 |
Pos logit contributions | |
gru | +0.047 |
badge | +0.042 |
probably | +0.041 |
god | +0.041 |
address | +0.040 |
Neg logit contributions | |
day | -0.038 |
afternoon | -0.036 |
night | -0.032 |
cy | -0.032 |
he | -0.032 |
Token | be |
Token ID | 307 |
Feature activation | +1.46e-01 |
Pos logit contributions | |
day | +0.132 |
afternoon | +0.112 |
night | +0.111 |
morning | +0.102 |
cy | +0.101 |
Neg logit contributions | |
gru | -0.166 |
badge | -0.143 |
rifle | -0.138 |
Reward | -0.134 |
ution | -0.134 |
Token | perfect |
Token ID | 2818 |
Feature activation | -9.99e-01 |
Pos logit contributions | |
day | +0.006 |
afternoon | +0.005 |
night | +0.005 |
cy | +0.005 |
morning | +0.005 |
Neg logit contributions | |
gru | -0.006 |
badge | -0.005 |
Reward | -0.005 |
ution | -0.005 |
rifle | -0.005 |
Token | for |
Token ID | 329 |
Feature activation | -1.68e+00 |
Pos logit contributions | |
gru | +0.044 |
badge | +0.038 |
Reward | +0.037 |
threatening | +0.036 |
rifle | +0.036 |
Neg logit contributions | |
day | -0.034 |
afternoon | -0.031 |
cy | -0.029 |
ite | -0.028 |
night | -0.028 |
Token | you |
Token ID | 345 |
Feature activation | -6.11e-01 |
Pos logit contributions | |
gru | +0.067 |
badge | +0.058 |
Reward | +0.056 |
ution | +0.054 |
threatening | +0.054 |
Neg logit contributions | |
day | -0.066 |
afternoon | -0.060 |
cy | -0.058 |
ite | -0.056 |
night | -0.056 |
Token | . |
Token ID | 13 |
Feature activation | +5.75e-01 |
Pos logit contributions | |
gru | +0.026 |
badge | +0.023 |
Reward | +0.022 |
threatening | +0.022 |
rifle | +0.022 |
Neg logit contributions | |
day | -0.021 |
afternoon | -0.020 |
cy | -0.019 |
ite | -0.018 |
night | -0.018 |
Token | Would |
Token ID | 10928 |
Feature activation | -2.05e+00 |
Pos logit contributions | |
day | +0.023 |
afternoon | +0.021 |
cy | +0.020 |
night | +0.020 |
ite | +0.019 |
Neg logit contributions | |
gru | -0.023 |
badge | -0.020 |
Reward | -0.019 |
ution | -0.018 |
rifle | -0.018 |
Token | you |
Token ID | 345 |
Feature activation | -6.23e-01 |
Pos logit contributions | |
gru | +0.080 |
Reward | +0.068 |
badge | +0.067 |
ution | +0.066 |
god | +0.065 |
Neg logit contributions | |
day | -0.080 |
afternoon | -0.075 |
cy | -0.074 |
ite | -0.071 |
night | -0.068 |
Token | like |
Token ID | 588 |
Feature activation | -7.50e-01 |
Pos logit contributions | |
gru | +0.023 |
badge | +0.020 |
Reward | +0.019 |
ution | +0.019 |
rifle | +0.019 |
Neg logit contributions | |
day | -0.026 |
afternoon | -0.024 |
night | -0.023 |
cy | -0.023 |
ite | -0.022 |
Token | to |
Token ID | 284 |
Feature activation | -5.40e+00 |
Pos logit contributions | |
gru | +0.034 |
badge | +0.029 |
rifle | +0.028 |
Reward | +0.028 |
ution | +0.028 |
Neg logit contributions | |
day | -0.028 |
afternoon | -0.024 |
night | -0.024 |
morning | -0.023 |
days | -0.022 |
Token | his |
Token ID | 465 |
Feature activation | -3.96e+00 |
Pos logit contributions | |
cy | +0.047 |
ite | +0.047 |
olly | +0.047 |
day | +0.045 |
afternoon | +0.045 |
Neg logit contributions | |
gru | -0.025 |
badge | -0.021 |
threatening | -0.020 |
Reward | -0.019 |
spoken | -0.018 |
Token | mom |
Token ID | 1995 |
Feature activation | +3.07e+00 |
Pos logit contributions | |
gru | +0.118 |
badge | +0.108 |
address | +0.100 |
warning | +0.098 |
message | +0.097 |
Neg logit contributions | |
cy | -0.165 |
ite | -0.164 |
afternoon | -0.163 |
he | -0.163 |
day | -0.162 |
Token | came |
Token ID | 1625 |
Feature activation | -2.13e+00 |
Pos logit contributions | |
day | +0.119 |
afternoon | +0.110 |
ite | +0.100 |
cy | +0.099 |
night | +0.099 |
Neg logit contributions | |
gru | -0.122 |
badge | -0.107 |
Reward | -0.104 |
god | -0.102 |
threatening | -0.100 |
Token | back |
Token ID | 736 |
Feature activation | +5.07e+00 |
Pos logit contributions | |
gru | +0.097 |
badge | +0.086 |
Reward | +0.082 |
rifle | +0.081 |
threatening | +0.081 |
Neg logit contributions | |
day | -0.071 |
afternoon | -0.065 |
ite | -0.058 |
cy | -0.058 |
night | -0.058 |
Token | from |
Token ID | 422 |
Feature activation | -2.76e+00 |
Pos logit contributions | |
day | +0.155 |
afternoon | +0.147 |
ite | +0.128 |
night | +0.125 |
cy | +0.123 |
Neg logit contributions | |
gru | -0.233 |
badge | -0.210 |
Reward | -0.201 |
warning | -0.200 |
threatening | -0.199 |
Token | the |
Token ID | 262 |
Feature activation | +3.20e+00 |
Pos logit contributions | |
gru | +0.126 |
badge | +0.113 |
ution | +0.106 |
threatening | +0.106 |
warning | +0.105 |
Neg logit contributions | |
day | -0.089 |
afternoon | -0.082 |
ite | -0.079 |
cy | -0.078 |
olly | -0.076 |
Token | store |
Token ID | 3650 |
Feature activation | +4.66e-01 |
Pos logit contributions | |
day | +0.110 |
afternoon | +0.093 |
night | +0.089 |
morning | +0.084 |
he | +0.083 |
Neg logit contributions | |
gru | -0.156 |
badge | -0.134 |
rifle | -0.131 |
ution | -0.131 |
threatening | -0.130 |
Token | , |
Token ID | 11 |
Feature activation | -1.67e-01 |
Pos logit contributions | |
day | +0.010 |
afternoon | +0.009 |
night | +0.008 |
cy | +0.008 |
ite | +0.007 |
Neg logit contributions | |
gru | -0.027 |
badge | -0.024 |
Reward | -0.024 |
ution | -0.023 |
rifle | -0.023 |
Token | she |
Token ID | 673 |
Feature activation | +2.64e+00 |
Pos logit contributions | |
gru | +0.007 |
badge | +0.006 |
rifle | +0.006 |
Reward | +0.006 |
ution | +0.006 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.006 |
night | -0.005 |
cy | -0.005 |
he | -0.005 |
Token | saw |
Token ID | 2497 |
Feature activation | -3.57e+00 |
Pos logit contributions | |
day | +0.088 |
afternoon | +0.077 |
cy | +0.074 |
night | +0.073 |
he | +0.070 |
Neg logit contributions | |
gru | -0.121 |
badge | -0.106 |
rifle | -0.103 |
Reward | -0.103 |
ution | -0.102 |
Token | the |
Token ID | 262 |
Feature activation | +3.21e+00 |
Pos logit contributions | |
gru | +0.149 |
badge | +0.129 |
Reward | +0.123 |
rifle | +0.121 |
ution | +0.121 |
Neg logit contributions | |
day | -0.138 |
afternoon | -0.125 |
night | -0.117 |
cy | -0.117 |
ite | -0.113 |
Token |
|
Token ID | 198 |
Feature activation | -3.66e-01 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.018 |
night | +0.017 |
morning | +0.017 |
cy | +0.017 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.026 |
ution | -0.025 |
rifle | -0.025 |
threatening | -0.025 |
Token |
|
Token ID | 198 |
Feature activation | -4.01e-01 |
Pos logit contributions | |
gru | +0.015 |
Reward | +0.013 |
badge | +0.013 |
threatening | +0.012 |
warning | +0.012 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.013 |
cy | -0.012 |
ite | -0.012 |
night | -0.012 |
Token | " |
Token ID | 1 |
Feature activation | -2.33e+00 |
Pos logit contributions | |
gru | +0.018 |
badge | +0.015 |
Reward | +0.015 |
ution | +0.014 |
threatening | +0.014 |
Neg logit contributions | |
day | -0.014 |
afternoon | -0.013 |
night | -0.012 |
cy | -0.012 |
ite | -0.012 |
Token | Hi |
Token ID | 17250 |
Feature activation | -4.80e+00 |
Pos logit contributions | |
gru | +0.100 |
badge | +0.088 |
rifle | +0.087 |
threatening | +0.085 |
Reward | +0.084 |
Neg logit contributions | |
day | -0.082 |
afternoon | -0.076 |
cy | -0.071 |
he | -0.070 |
night | -0.069 |
Token | , |
Token ID | 11 |
Feature activation | -2.76e-01 |
Pos logit contributions | |
gru | +0.239 |
badge | +0.213 |
Reward | +0.205 |
god | +0.201 |
rifle | +0.201 |
Neg logit contributions | |
day | -0.138 |
afternoon | -0.128 |
he | -0.111 |
night | -0.110 |
ite | -0.109 |
Token | Max |
Token ID | 5436 |
Feature activation | +2.70e-01 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.011 |
ution | +0.011 |
rifle | +0.011 |
Reward | +0.011 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
night | -0.008 |
cy | -0.007 |
morning | -0.007 |
Token | ," |
Token ID | 553 |
Feature activation | +6.63e-01 |
Pos logit contributions | |
day | +0.009 |
afternoon | +0.009 |
cy | +0.008 |
night | +0.008 |
ite | +0.008 |
Neg logit contributions | |
gru | -0.012 |
badge | -0.011 |
Reward | -0.010 |
ution | -0.010 |
rifle | -0.010 |
Token | Lily |
Token ID | 20037 |
Feature activation | +3.67e+00 |
Pos logit contributions | |
day | +0.018 |
afternoon | +0.014 |
night | +0.014 |
morning | +0.013 |
days | +0.012 |
Neg logit contributions | |
gru | -0.034 |
rifle | -0.032 |
CRE | -0.031 |
ution | -0.031 |
threatening | -0.031 |
Token | said |
Token ID | 531 |
Feature activation | -3.92e+00 |
Pos logit contributions | |
day | +0.109 |
afternoon | +0.092 |
night | +0.090 |
cy | +0.086 |
morning | +0.082 |
Neg logit contributions | |
gru | -0.175 |
badge | -0.158 |
Reward | -0.155 |
ution | -0.155 |
rifle | -0.153 |
Token | softly |
Token ID | 26625 |
Feature activation | +6.02e-01 |
Pos logit contributions | |
gru | +0.210 |
badge | +0.190 |
ution | +0.187 |
Reward | +0.186 |
rifle | +0.183 |
Neg logit contributions | |
day | -0.100 |
afternoon | -0.083 |
night | -0.078 |
cy | -0.076 |
morning | -0.071 |
Token | . |
Token ID | 13 |
Feature activation | +6.04e-01 |
Pos logit contributions | |
day | +0.012 |
afternoon | +0.009 |
night | +0.009 |
cy | +0.008 |
morning | +0.008 |
Neg logit contributions | |
gru | -0.034 |
ution | -0.032 |
Reward | -0.032 |
badge | -0.032 |
rets | -0.031 |
Token | mineral |
Token ID | 18352 |
Feature activation | -2.64e+00 |
Pos logit contributions | |
day | +0.127 |
afternoon | +0.120 |
cy | +0.114 |
ite | +0.112 |
night | +0.112 |
Neg logit contributions | |
gru | -0.105 |
badge | -0.091 |
rifle | -0.083 |
Reward | -0.083 |
threatening | -0.083 |
Token | all |
Token ID | 477 |
Feature activation | +4.88e+00 |
Pos logit contributions | |
gru | +0.105 |
badge | +0.098 |
threatening | +0.093 |
rifle | +0.093 |
warning | +0.089 |
Neg logit contributions | |
day | -0.081 |
afternoon | -0.078 |
cy | -0.078 |
ite | -0.074 |
End | -0.074 |
Token | day |
Token ID | 1110 |
Feature activation | +1.06e+01 |
Pos logit contributions | |
olly | +0.055 |
ite | +0.054 |
ops | +0.053 |
text | +0.042 |
ugg | +0.040 |
Neg logit contributions | |
gru | -0.271 |
probably | -0.252 |
threatening | -0.250 |
badge | -0.244 |
unlike | -0.241 |
Token | , |
Token ID | 11 |
Feature activation | -2.37e-01 |
Pos logit contributions | |
day | +0.361 |
afternoon | +0.329 |
cy | +0.316 |
ite | +0.303 |
night | +0.303 |
Neg logit contributions | |
gru | -0.466 |
badge | -0.412 |
rifle | -0.388 |
Reward | -0.388 |
threatening | -0.388 |
Token | bending |
Token ID | 30808 |
Feature activation | +3.18e+00 |
Pos logit contributions | |
gru | +0.010 |
badge | +0.008 |
Reward | +0.008 |
ution | +0.008 |
rifle | +0.008 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
night | -0.008 |
cy | -0.008 |
morning | -0.008 |
Token | and |
Token ID | 290 |
Feature activation | +1.77e+00 |
Pos logit contributions | |
day | +0.127 |
afternoon | +0.117 |
olly | +0.110 |
cy | +0.109 |
ite | +0.109 |
Neg logit contributions | |
gru | -0.120 |
badge | -0.107 |
rifle | -0.100 |
ution | -0.099 |
threatening | -0.098 |
Token | flex |
Token ID | 7059 |
Feature activation | +8.93e-01 |
Pos logit contributions | |
day | +0.072 |
afternoon | +0.066 |
cy | +0.063 |
ite | +0.062 |
night | +0.062 |
Neg logit contributions | |
gru | -0.067 |
badge | -0.058 |
Reward | -0.055 |
ution | -0.054 |
threatening | -0.054 |
Token | ing |
Token ID | 278 |
Feature activation | +5.70e+00 |
Pos logit contributions | |
day | +0.032 |
afternoon | +0.029 |
night | +0.027 |
morning | +0.026 |
ite | +0.026 |
Neg logit contributions | |
gru | -0.036 |
badge | -0.033 |
rifle | -0.031 |
threatening | -0.030 |
Reward | -0.030 |
Token | it |
Token ID | 340 |
Feature activation | +3.46e-01 |
Pos logit contributions | |
day | +0.206 |
afternoon | +0.187 |
cy | +0.174 |
night | +0.173 |
ite | +0.169 |
Neg logit contributions | |
gru | -0.246 |
badge | -0.216 |
Reward | -0.206 |
ution | -0.203 |
rifle | -0.203 |
Token | . |
Token ID | 13 |
Feature activation | +4.73e-01 |
Pos logit contributions | |
day | +0.011 |
afternoon | +0.010 |
cy | +0.009 |
night | +0.009 |
ite | +0.009 |
Neg logit contributions | |
gru | -0.016 |
badge | -0.015 |
Reward | -0.014 |
threatening | -0.014 |
rifle | -0.014 |
Token | But |
Token ID | 887 |
Feature activation | -5.23e+00 |
Pos logit contributions | |
day | +0.018 |
afternoon | +0.017 |
cy | +0.016 |
ite | +0.016 |
night | +0.016 |
Neg logit contributions | |
gru | -0.019 |
badge | -0.016 |
Reward | -0.016 |
threatening | -0.015 |
ution | -0.015 |
Token | she |
Token ID | 673 |
Feature activation | +2.63e+00 |
Pos logit contributions | |
gru | +0.011 |
rifle | +0.010 |
badge | +0.010 |
ution | +0.009 |
threatening | +0.009 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.004 |
morning | -0.004 |
night | -0.004 |
evening | -0.004 |
Token | asked |
Token ID | 1965 |
Feature activation | -3.81e+00 |
Pos logit contributions | |
day | +0.072 |
afternoon | +0.057 |
night | +0.056 |
went | +0.053 |
morning | +0.050 |
Neg logit contributions | |
gru | -0.125 |
badge | -0.120 |
extingu | -0.117 |
rier | -0.116 |
rifle | -0.116 |
Token | . |
Token ID | 13 |
Feature activation | +6.57e-01 |
Pos logit contributions | |
gru | +0.206 |
badge | +0.186 |
Reward | +0.182 |
ution | +0.181 |
rifle | +0.181 |
Neg logit contributions | |
day | -0.097 |
afternoon | -0.080 |
night | -0.076 |
cy | -0.074 |
morning | -0.069 |
Token |
|
Token ID | 198 |
Feature activation | -5.10e-02 |
Pos logit contributions | |
day | +0.020 |
afternoon | +0.017 |
morning | +0.016 |
night | +0.016 |
cy | +0.015 |
Neg logit contributions | |
gru | -0.032 |
badge | -0.028 |
rifle | -0.028 |
ution | -0.028 |
rets | -0.028 |
Token |
|
Token ID | 198 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
gru | +0.002 |
badge | +0.002 |
Reward | +0.002 |
ution | +0.002 |
threatening | +0.002 |
Neg logit contributions | |
day | -0.002 |
afternoon | -0.001 |
cy | -0.001 |
night | -0.001 |
ite | -0.001 |
Token | Her |
Token ID | 9360 |
Feature activation | -4.96e+00 |
Pos logit contributions | |
gru | +0.005 |
badge | +0.004 |
rifle | +0.004 |
Reward | +0.004 |
threatening | +0.004 |
Neg logit contributions | |
day | -0.004 |
afternoon | -0.004 |
cy | -0.003 |
he | -0.003 |
night | -0.003 |
Token | mom |
Token ID | 1995 |
Feature activation | +3.08e+00 |
Pos logit contributions | |
gru | +0.205 |
badge | +0.182 |
Reward | +0.167 |
address | +0.165 |
rifle | +0.165 |
Neg logit contributions | |
day | -0.188 |
afternoon | -0.174 |
cy | -0.164 |
night | -0.161 |
he | -0.160 |
Token | replied |
Token ID | 8712 |
Feature activation | -3.95e+00 |
Pos logit contributions | |
day | +0.096 |
afternoon | +0.083 |
cy | +0.082 |
night | +0.077 |
olly | +0.075 |
Neg logit contributions | |
gru | -0.142 |
Reward | -0.129 |
badge | -0.127 |
rifle | -0.125 |
ution | -0.122 |
Token | , |
Token ID | 11 |
Feature activation | -2.49e-01 |
Pos logit contributions | |
gru | +0.231 |
badge | +0.210 |
Reward | +0.206 |
rifle | +0.203 |
ution | +0.203 |
Neg logit contributions | |
day | -0.083 |
afternoon | -0.067 |
night | -0.059 |
cy | -0.059 |
morning | -0.054 |
Token | " |
Token ID | 366 |
Feature activation | -3.35e+00 |
Pos logit contributions | |
gru | +0.010 |
Reward | +0.009 |
rifle | +0.009 |
badge | +0.009 |
rets | +0.008 |
Neg logit contributions | |
day | -0.010 |
afternoon | -0.008 |
night | -0.008 |
morning | -0.008 |
days | -0.008 |
Token | The |
Token ID | 464 |
Feature activation | +9.54e-01 |
Pos logit contributions | |
gru | +0.155 |
badge | +0.144 |
rifle | +0.136 |
Reward | +0.130 |
warning | +0.130 |
Neg logit contributions | |
day | -0.113 |
afternoon | -0.101 |
ite | -0.098 |
cy | -0.098 |
he | -0.091 |
Token | I |
Token ID | 314 |
Feature activation | +6.60e-01 |
Pos logit contributions | |
day | +0.016 |
afternoon | +0.015 |
cy | +0.014 |
night | +0.014 |
ite | +0.014 |
Neg logit contributions | |
gru | -0.017 |
badge | -0.015 |
Reward | -0.014 |
ution | -0.014 |
rifle | -0.014 |
Token | don |
Token ID | 836 |
Feature activation | +9.96e-01 |
Pos logit contributions | |
day | +0.022 |
afternoon | +0.020 |
cy | +0.018 |
night | +0.018 |
ite | +0.018 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.026 |
Reward | -0.025 |
threatening | -0.024 |
ution | -0.024 |
Token | 't |
Token ID | 470 |
Feature activation | -3.23e+00 |
Pos logit contributions | |
day | +0.048 |
afternoon | +0.045 |
cy | +0.044 |
night | +0.042 |
morning | +0.041 |
Neg logit contributions | |
gru | -0.031 |
badge | -0.025 |
Reward | -0.025 |
threatening | -0.023 |
rifle | -0.023 |
Token | know |
Token ID | 760 |
Feature activation | -3.90e+00 |
Pos logit contributions | |
gru | +0.125 |
badge | +0.109 |
Reward | +0.103 |
threatening | +0.101 |
ution | +0.101 |
Neg logit contributions | |
day | -0.128 |
afternoon | -0.119 |
cy | -0.112 |
night | -0.110 |
ite | -0.109 |
Token | the |
Token ID | 262 |
Feature activation | +3.33e+00 |
Pos logit contributions | |
gru | +0.131 |
badge | +0.112 |
Reward | +0.106 |
threatening | +0.104 |
warning | +0.102 |
Neg logit contributions | |
day | -0.173 |
afternoon | -0.165 |
cy | -0.159 |
he | -0.152 |
night | -0.151 |
Token | number |
Token ID | 1271 |
Feature activation | -4.67e+00 |
Pos logit contributions | |
day | +0.141 |
afternoon | +0.126 |
night | +0.120 |
morning | +0.115 |
cy | +0.113 |
Neg logit contributions | |
gru | -0.127 |
badge | -0.109 |
Reward | -0.106 |
ution | -0.105 |
rifle | -0.102 |
Token | !" |
Token ID | 2474 |
Feature activation | -2.66e-01 |
Pos logit contributions | |
gru | +0.204 |
badge | +0.182 |
Reward | +0.175 |
address | +0.171 |
god | +0.170 |
Neg logit contributions | |
day | -0.152 |
afternoon | -0.145 |
cy | -0.133 |
night | -0.129 |
olly | -0.128 |
Token | |
Token ID | 220 |
Feature activation | -5.35e-02 |
Pos logit contributions | |
gru | +0.013 |
badge | +0.012 |
ution | +0.011 |
rifle | +0.011 |
Reward | +0.011 |
Neg logit contributions | |
day | -0.008 |
afternoon | -0.007 |
night | -0.007 |
cy | -0.006 |
morning | -0.006 |
Token |
|
Token ID | 198 |
Feature activation | -5.41e-02 |
Pos logit contributions | |
gru | +0.002 |
Reward | +0.002 |
badge | +0.002 |
address | +0.002 |
ution | +0.002 |
Neg logit contributions | |
day | -0.002 |
afternoon | -0.002 |
cy | -0.002 |
night | -0.002 |
days | -0.002 |
Token |
|
Token ID | 198 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
gru | +0.003 |
badge | +0.002 |
Reward | +0.002 |
ution | +0.002 |
threatening | +0.002 |
Neg logit contributions | |
day | -0.002 |
afternoon | -0.002 |
night | -0.001 |
cy | -0.001 |
ite | -0.001 |
Token | B |
Token ID | 33 |
Feature activation | +2.91e+00 |
Pos logit contributions | |
gru | +0.007 |
badge | +0.006 |
rifle | +0.006 |
Reward | +0.006 |
threatening | +0.006 |
Neg logit contributions | |
day | -0.005 |
afternoon | -0.005 |
he | -0.004 |
cy | -0.004 |
ite | -0.004 |
Token | says |
Token ID | 1139 |
Feature activation | -4.98e+00 |
Pos logit contributions | |
day | +0.092 |
afternoon | +0.085 |
night | +0.079 |
he | +0.077 |
morning | +0.075 |
Neg logit contributions | |
gru | -0.100 |
badge | -0.088 |
Reward | -0.085 |
threatening | -0.084 |
rifle | -0.084 |
Token | , |
Token ID | 11 |
Feature activation | -1.73e-01 |
Pos logit contributions | |
gru | +0.282 |
badge | +0.255 |
Reward | +0.248 |
ution | +0.246 |
rifle | +0.245 |
Neg logit contributions | |
day | -0.116 |
afternoon | -0.094 |
night | -0.085 |
cy | -0.083 |
ite | -0.082 |
Token | " |
Token ID | 366 |
Feature activation | -3.32e+00 |
Pos logit contributions | |
gru | +0.006 |
badge | +0.005 |
ution | +0.005 |
Reward | +0.005 |
rifle | +0.005 |
Neg logit contributions | |
day | -0.007 |
afternoon | -0.007 |
night | -0.006 |
morning | -0.006 |
cy | -0.006 |
Token | What |
Token ID | 2061 |
Feature activation | -3.94e+00 |
Pos logit contributions | |
gru | +0.115 |
Reward | +0.099 |
badge | +0.097 |
ution | +0.095 |
rifle | +0.091 |
Neg logit contributions | |
day | -0.140 |
afternoon | -0.129 |
night | -0.124 |
olly | -0.122 |
days | -0.119 |
Token | are |
Token ID | 389 |
Feature activation | +2.48e-01 |
Pos logit contributions | |
gru | +0.181 |
badge | +0.160 |
Reward | +0.156 |
ution | +0.151 |
address | +0.151 |
Neg logit contributions | |
day | -0.127 |
afternoon | -0.117 |
cy | -0.108 |
night | -0.106 |
he | -0.103 |
Token | those |
Token ID | 883 |
Feature activation | -1.04e+01 |
Pos logit contributions | |
day | +0.008 |
afternoon | +0.008 |
night | +0.007 |
cy | +0.007 |
ite | +0.007 |
Neg logit contributions | |
gru | -0.011 |
badge | -0.010 |
Reward | -0.009 |
rifle | -0.009 |
threatening | -0.009 |
Token | ?" |
Token ID | 1701 |
Feature activation | -2.31e-01 |
Pos logit contributions | |
gru | +0.394 |
badge | +0.351 |
Reward | +0.346 |
ution | +0.330 |
rifle | +0.322 |
Neg logit contributions | |
day | -0.397 |
afternoon | -0.381 |
night | -0.366 |
cy | -0.355 |
morning | -0.342 |
Token |
|
Token ID | 198 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
gru | +0.009 |
badge | +0.008 |
Reward | +0.008 |
ution | +0.007 |
address | +0.007 |
Neg logit contributions | |
day | -0.009 |
afternoon | -0.008 |
cy | -0.008 |
ite | -0.008 |
night | -0.007 |
Token |
|
Token ID | 198 |
Feature activation | -1.81e-01 |
Pos logit contributions | |
gru | +0.007 |
badge | +0.006 |
Reward | +0.006 |
ution | +0.006 |
threatening | +0.006 |
Neg logit contributions | |
day | -0.005 |
afternoon | -0.005 |
cy | -0.004 |
night | -0.004 |
ite | -0.004 |
Token | Ben |
Token ID | 11696 |
Feature activation | +1.89e+00 |
Pos logit contributions | |
gru | +0.008 |
badge | +0.007 |
Reward | +0.007 |
ution | +0.007 |
rifle | +0.007 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.006 |
cy | -0.005 |
night | -0.005 |
ite | -0.005 |
Token | says |
Token ID | 1139 |
Feature activation | -4.91e+00 |
Pos logit contributions | |
day | +0.071 |
afternoon | +0.064 |
cy | +0.060 |
night | +0.060 |
ite | +0.058 |
Neg logit contributions | |
gru | -0.079 |
badge | -0.069 |
Reward | -0.066 |
ution | -0.065 |
rifle | -0.065 |
Token | said |
Token ID | 531 |
Feature activation | -3.67e+00 |
Pos logit contributions | |
day | +0.069 |
afternoon | +0.058 |
night | +0.058 |
cy | +0.056 |
morning | +0.053 |
Neg logit contributions | |
gru | -0.093 |
rifle | -0.079 |
badge | -0.078 |
ution | -0.076 |
Reward | -0.076 |
Token | , |
Token ID | 11 |
Feature activation | -2.93e-01 |
Pos logit contributions | |
gru | +0.218 |
badge | +0.199 |
ution | +0.196 |
rifle | +0.196 |
Reward | +0.196 |
Neg logit contributions | |
day | -0.074 |
afternoon | -0.056 |
night | -0.052 |
morning | -0.047 |
days | -0.044 |
Token | " |
Token ID | 366 |
Feature activation | -3.08e+00 |
Pos logit contributions | |
gru | +0.012 |
ution | +0.011 |
badge | +0.010 |
rifle | +0.010 |
Reward | +0.010 |
Neg logit contributions | |
day | -0.011 |
afternoon | -0.009 |
night | -0.009 |
morning | -0.009 |
he | -0.008 |
Token | Mom |
Token ID | 29252 |
Feature activation | +3.06e+00 |
Pos logit contributions | |
gru | +0.161 |
badge | +0.147 |
rifle | +0.141 |
Reward | +0.138 |
threatening | +0.136 |
Neg logit contributions | |
day | -0.090 |
afternoon | -0.079 |
cy | -0.074 |
night | -0.071 |
he | -0.069 |
Token | , |
Token ID | 11 |
Feature activation | -2.70e-01 |
Pos logit contributions | |
day | +0.091 |
afternoon | +0.088 |
he | +0.079 |
ite | +0.077 |
olly | +0.076 |
Neg logit contributions | |
gru | -0.136 |
badge | -0.124 |
address | -0.119 |
probably | -0.118 |
god | -0.118 |
Token | my |
Token ID | 616 |
Feature activation | -6.82e+00 |
Pos logit contributions | |
gru | +0.012 |
badge | +0.011 |
rifle | +0.010 |
Reward | +0.010 |
ution | +0.010 |
Neg logit contributions | |
day | -0.010 |
afternoon | -0.008 |
night | -0.008 |
cy | -0.008 |
morning | -0.008 |
Token | car |
Token ID | 1097 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
gru | +0.256 |
badge | +0.238 |
address | +0.209 |
rifle | +0.207 |
Reward | +0.205 |
Neg logit contributions | |
day | -0.266 |
afternoon | -0.257 |
cy | -0.243 |
he | -0.238 |
night | -0.235 |
Token | is |
Token ID | 318 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
day | +0.003 |
afternoon | +0.003 |
cy | +0.003 |
olly | +0.003 |
night | +0.003 |
Neg logit contributions | |
gru | -0.004 |
badge | -0.004 |
Reward | -0.004 |
probably | -0.004 |
ution | -0.004 |
Token | broken |
Token ID | 5445 |
Feature activation | -2.64e+00 |
Pos logit contributions | |
gru | +0.037 |
badge | +0.030 |
Reward | +0.029 |
ution | +0.029 |
rifle | +0.028 |
Neg logit contributions | |
day | -0.051 |
afternoon | -0.047 |
night | -0.045 |
morning | -0.043 |
cy | -0.043 |
Token | . |
Token ID | 13 |
Feature activation | +8.03e-01 |
Pos logit contributions | |
gru | +0.107 |
badge | +0.096 |
threatening | +0.091 |
probably | +0.090 |
warning | +0.089 |
Neg logit contributions | |
day | -0.091 |
afternoon | -0.086 |
olly | -0.083 |
ite | -0.082 |
cy | -0.082 |
Token | Can |
Token ID | 1680 |
Feature activation | -5.86e+00 |
Pos logit contributions | |
day | +0.030 |
afternoon | +0.027 |
cy | +0.026 |
ite | +0.026 |
night | +0.026 |
Neg logit contributions | |
gru | -0.032 |
badge | -0.029 |
ution | -0.027 |
threatening | -0.027 |
spoken | -0.027 |
Token | match |
Token ID | 2872 |
Feature activation | -3.24e+00 |
Pos logit contributions | |
gru | +0.196 |
badge | +0.176 |
Reward | +0.170 |
ution | +0.168 |
rifle | +0.167 |
Neg logit contributions | |
day | -0.098 |
afternoon | -0.085 |
cy | -0.076 |
night | -0.076 |
ite | -0.073 |
Token | the |
Token ID | 262 |
Feature activation | +2.46e+00 |
Pos logit contributions | |
gru | +0.134 |
badge | +0.118 |
Reward | +0.113 |
ution | +0.111 |
rifle | +0.110 |
Neg logit contributions | |
day | -0.121 |
afternoon | -0.111 |
cy | -0.103 |
night | -0.102 |
ite | -0.100 |
Token | lions |
Token ID | 31420 |
Feature activation | -2.18e+00 |
Pos logit contributions | |
day | +0.122 |
afternoon | +0.114 |
night | +0.108 |
cy | +0.107 |
ite | +0.105 |
Neg logit contributions | |
gru | -0.074 |
badge | -0.060 |
Reward | -0.056 |
ution | -0.055 |
threatening | -0.055 |
Token | . |
Token ID | 13 |
Feature activation | +7.12e-01 |
Pos logit contributions | |
gru | +0.086 |
badge | +0.081 |
Reward | +0.075 |
address | +0.074 |
god | +0.074 |
Neg logit contributions | |
day | -0.073 |
afternoon | -0.070 |
cy | -0.068 |
olly | -0.066 |
ugg | -0.062 |
Token | It |
Token ID | 632 |
Feature activation | -3.34e+00 |
Pos logit contributions | |
day | +0.025 |
afternoon | +0.022 |
ite | +0.021 |
cy | +0.021 |
night | +0.021 |
Neg logit contributions | |
gru | -0.031 |
badge | -0.027 |
ution | -0.026 |
threatening | -0.026 |
spoken | -0.026 |
Token | 's |
Token ID | 338 |
Feature activation | -4.74e+00 |
Pos logit contributions | |
gru | +0.178 |
badge | +0.159 |
ution | +0.151 |
Reward | +0.151 |
rifle | +0.151 |
Neg logit contributions | |
day | -0.091 |
afternoon | -0.077 |
night | -0.071 |
cy | -0.071 |
ite | -0.068 |
Token | okay |
Token ID | 8788 |
Feature activation | -3.47e+00 |
Pos logit contributions | |
gru | +0.200 |
badge | +0.169 |
ution | +0.167 |
Reward | +0.166 |
rifle | +0.161 |
Neg logit contributions | |
day | -0.184 |
afternoon | -0.164 |
night | -0.156 |
morning | -0.147 |
cy | -0.143 |
Token | . |
Token ID | 13 |
Feature activation | +6.65e-01 |
Pos logit contributions | |
gru | +0.148 |
badge | +0.130 |
Reward | +0.126 |
warning | +0.124 |
threatening | +0.123 |
Neg logit contributions | |
day | -0.117 |
afternoon | -0.113 |
he | -0.105 |
cy | -0.105 |
ite | -0.101 |
Token | We |
Token ID | 775 |
Feature activation | -1.84e-01 |
Pos logit contributions | |
day | +0.023 |
afternoon | +0.020 |
ite | +0.020 |
cy | +0.020 |
night | +0.019 |
Neg logit contributions | |
gru | -0.029 |
badge | -0.025 |
threatening | -0.025 |
spoken | -0.025 |
ution | -0.024 |
Token | can |
Token ID | 460 |
Feature activation | +8.22e-01 |
Pos logit contributions | |
gru | +0.009 |
badge | +0.008 |
Reward | +0.008 |
ution | +0.007 |
rifle | +0.007 |
Neg logit contributions | |
day | -0.006 |
afternoon | -0.005 |
cy | -0.004 |
night | -0.004 |
ite | -0.004 |
Token | clean |
Token ID | 3424 |
Feature activation | +3.29e+00 |
Pos logit contributions | |
day | +0.028 |
afternoon | +0.024 |
cy | +0.023 |
night | +0.023 |
he | +0.023 |
Neg logit contributions | |
gru | -0.040 |
badge | -0.033 |
rifle | -0.032 |
Reward | -0.032 |
ution | -0.032 |
Token | before |
Token ID | 878 |
Feature activation | +5.09e+00 |
Pos logit contributions | |
day | +0.044 |
afternoon | +0.040 |
cy | +0.040 |
ite | +0.038 |
night | +0.037 |
Neg logit contributions | |
gru | -0.061 |
badge | -0.055 |
address | -0.050 |
Reward | -0.050 |
threatening | -0.050 |
Token | dinner |
Token ID | 8073 |
Feature activation | +8.33e-01 |
Pos logit contributions | |
day | +0.170 |
afternoon | +0.156 |
cy | +0.152 |
ite | +0.148 |
olly | +0.141 |
Neg logit contributions | |
gru | -0.223 |
badge | -0.197 |
Reward | -0.190 |
threatening | -0.185 |
warning | -0.184 |
Token | . |
Token ID | 13 |
Feature activation | +4.50e-01 |
Pos logit contributions | |
day | +0.024 |
cy | +0.023 |
ite | +0.022 |
afternoon | +0.022 |
night | +0.020 |
Neg logit contributions | |
gru | -0.038 |
badge | -0.034 |
Reward | -0.032 |
address | -0.032 |
warning | -0.032 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +5.41e+00 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.016 |
cy | +0.015 |
night | +0.015 |
ite | +0.015 |
Neg logit contributions | |
gru | -0.018 |
Reward | -0.015 |
badge | -0.015 |
threatening | -0.014 |
address | -0.014 |
Token | S |
Token ID | 50 |
Feature activation | -4.23e+00 |
Pos logit contributions | |
afternoon | +0.223 |
day | +0.219 |
olly | +0.210 |
ite | +0.209 |
days | +0.205 |
Neg logit contributions | |
gru | -0.153 |
Reward | -0.128 |
Gu | -0.110 |
Fox | -0.109 |
probably | -0.109 |
Token | aman |
Token ID | 10546 |
Feature activation | -1.59e+01 |
Pos logit contributions | |
gru | +0.155 |
god | +0.143 |
Reward | +0.138 |
address | +0.137 |
badge | +0.136 |
Neg logit contributions | |
day | -0.154 |
afternoon | -0.144 |
days | -0.136 |
ite | -0.134 |
cy | -0.128 |
Token | tha |
Token ID | 12898 |
Feature activation | -2.92e-01 |
Pos logit contributions | |
gru | +0.557 |
badge | +0.469 |
ution | +0.448 |
Reward | +0.435 |
god | +0.433 |
Neg logit contributions | |
day | -0.696 |
afternoon | -0.645 |
night | -0.608 |
cy | -0.597 |
ite | -0.595 |
Token | was |
Token ID | 373 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
probably | +0.009 |
address | +0.009 |
god | +0.009 |
beard | +0.008 |
gru | +0.008 |
Neg logit contributions | |
afternoon | -0.011 |
day | -0.011 |
ite | -0.010 |
olly | -0.010 |
days | -0.010 |
Token | a |
Token ID | 257 |
Feature activation | -2.17e+00 |
Pos logit contributions | |
day | +0.036 |
cy | +0.034 |
ite | +0.033 |
afternoon | +0.032 |
ops | +0.031 |
Neg logit contributions | |
gru | -0.041 |
badge | -0.035 |
probably | -0.033 |
threatening | -0.033 |
Reward | -0.033 |
Token | lazy |
Token ID | 16931 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
gru | +0.080 |
badge | +0.069 |
god | +0.063 |
threatening | +0.062 |
beard | +0.062 |
Neg logit contributions | |
day | -0.083 |
ops | -0.080 |
afternoon | -0.079 |
ite | -0.075 |
he | -0.074 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.56e+00 |
Pos logit contributions | |
day | +0.042 |
afternoon | +0.040 |
he | +0.038 |
ite | +0.038 |
cy | +0.038 |
Neg logit contributions | |
gru | -0.044 |
badge | -0.039 |
threatening | -0.035 |
god | -0.035 |
rifle | -0.035 |
Token | pond |
Token ID | 16723 |
Feature activation | +1.78e-01 |
Pos logit contributions | |
day | +0.113 |
afternoon | +0.107 |
ite | +0.107 |
olly | +0.105 |
cy | +0.105 |
Neg logit contributions | |
gru | -0.114 |
threatening | -0.095 |
warning | -0.092 |
Reward | -0.091 |
badge | -0.090 |
Token | . |
Token ID | 13 |
Feature activation | +5.13e-01 |
Pos logit contributions | |
day | +0.005 |
afternoon | +0.005 |
cy | +0.005 |
ite | +0.005 |
night | +0.005 |
Neg logit contributions | |
gru | -0.009 |
badge | -0.007 |
Reward | -0.007 |
threatening | -0.007 |
rifle | -0.007 |
Token | It |
Token ID | 632 |
Feature activation | -2.94e+00 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.016 |
night | +0.016 |
morning | +0.016 |
olly | +0.015 |
Neg logit contributions | |
gru | -0.020 |
badge | -0.019 |
god | -0.019 |
ution | -0.019 |
spoken | -0.018 |
Token | was |
Token ID | 373 |
Feature activation | +1.22e+00 |
Pos logit contributions | |
gru | +0.179 |
badge | +0.174 |
Reward | +0.167 |
ution | +0.165 |
address | +0.165 |
Neg logit contributions | |
day | -0.052 |
afternoon | -0.038 |
cy | -0.034 |
night | -0.033 |
morning | -0.026 |
Token | a |
Token ID | 257 |
Feature activation | -2.54e+00 |
Pos logit contributions | |
day | +0.041 |
afternoon | +0.035 |
night | +0.035 |
morning | +0.031 |
days | +0.030 |
Neg logit contributions | |
gru | -0.054 |
ution | -0.053 |
rets | -0.053 |
phant | -0.052 |
CRE | -0.051 |
Token | tort |
Token ID | 7619 |
Feature activation | -1.33e+01 |
Pos logit contributions | |
gru | +0.105 |
ution | +0.092 |
Reward | +0.092 |
badge | +0.091 |
spoken | +0.087 |
Neg logit contributions | |
day | -0.099 |
afternoon | -0.086 |
night | -0.082 |
cy | -0.077 |
morning | -0.077 |
Token | oise |
Token ID | 25678 |
Feature activation | +5.87e-01 |
Pos logit contributions | |
gru | +0.652 |
badge | +0.581 |
Reward | +0.558 |
ution | +0.550 |
rifle | +0.548 |
Neg logit contributions | |
day | -0.408 |
afternoon | -0.362 |
cy | -0.334 |
night | -0.330 |
ite | -0.321 |
Token | ! |
Token ID | 0 |
Feature activation | -4.83e-01 |
Pos logit contributions | |
day | +0.019 |
afternoon | +0.016 |
night | +0.015 |
cy | +0.015 |
ite | +0.015 |
Neg logit contributions | |
gru | -0.028 |
badge | -0.025 |
Reward | -0.024 |
ution | -0.024 |
rifle | -0.023 |
Token | The |
Token ID | 383 |
Feature activation | -7.13e-01 |
Pos logit contributions | |
gru | +0.023 |
badge | +0.021 |
ution | +0.020 |
god | +0.020 |
spoken | +0.020 |
Neg logit contributions | |
day | -0.015 |
afternoon | -0.013 |
night | -0.012 |
cy | -0.012 |
morning | -0.012 |
Token | all |
Token ID | 477 |
Feature activation | +4.80e+00 |
Pos logit contributions | |
gru | +0.027 |
badge | +0.023 |
Reward | +0.022 |
threatening | +0.021 |
ution | +0.021 |
Neg logit contributions | |
day | -0.030 |
afternoon | -0.028 |
cy | -0.026 |
night | -0.026 |
ite | -0.026 |
Token | igator |
Token ID | 23823 |
Feature activation | -3.27e+00 |
Pos logit contributions | |
day | +0.173 |
afternoon | +0.165 |
cy | +0.161 |
olly | +0.157 |
ite | +0.157 |
Neg logit contributions | |
gru | -0.192 |
Reward | -0.156 |
badge | -0.154 |
threatening | -0.153 |
rifle | -0.149 |
Token | all |
Token ID | 477 |
Feature activation | +4.98e+00 |
Pos logit contributions | |
day | +0.113 |
afternoon | +0.103 |
cy | +0.098 |
night | +0.095 |
ite | +0.095 |
Neg logit contributions | |
gru | -0.139 |
badge | -0.122 |
Reward | -0.116 |
threatening | -0.114 |
ution | -0.114 |
Token | . |
Token ID | 13 |
Feature activation | +3.97e-01 |
Pos logit contributions | |
day | +0.145 |
cy | +0.139 |
ite | +0.139 |
afternoon | +0.136 |
olly | +0.136 |
Neg logit contributions | |
gru | -0.220 |
threatening | -0.184 |
probably | -0.184 |
god | -0.183 |
warning | -0.183 |
Token | She |
Token ID | 1375 |
Feature activation | -2.08e+00 |
Pos logit contributions | |
day | +0.017 |
afternoon | +0.015 |
cy | +0.014 |
ite | +0.014 |
night | +0.014 |
Neg logit contributions | |
gru | -0.014 |
Reward | -0.012 |
badge | -0.011 |
address | -0.011 |
probably | -0.011 |
Token | blinked |
Token ID | 40360 |
Feature activation | +6.52e-01 |
Pos logit contributions | |
gru | +0.081 |
god | +0.074 |
probably | +0.071 |
badge | +0.071 |
spoken | +0.069 |
Neg logit contributions | |
day | -0.071 |
afternoon | -0.066 |
he | -0.065 |
ite | -0.064 |
cy | -0.064 |
Token | in |
Token ID | 287 |
Feature activation | +4.11e+00 |
Pos logit contributions | |
day | +0.021 |
afternoon | +0.018 |
night | +0.017 |
cy | +0.017 |
morning | +0.016 |
Neg logit contributions | |
gru | -0.031 |
Reward | -0.028 |
badge | -0.028 |
ution | -0.027 |
rifle | -0.026 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.53e+01 |
Pos logit contributions | |
day | +0.181 |
afternoon | +0.165 |
night | +0.158 |
cy | +0.151 |
morning | +0.151 |
Neg logit contributions | |
gru | -0.149 |
badge | -0.129 |
Reward | -0.122 |
ution | -0.121 |
rifle | -0.120 |
Token | ement |
Token ID | 972 |
Feature activation | -9.68e-01 |
Pos logit contributions | |
gru | +0.712 |
Reward | +0.632 |
badge | +0.627 |
rifle | +0.617 |
threatening | +0.599 |
Neg logit contributions | |
day | -0.499 |
afternoon | -0.450 |
night | -0.415 |
he | -0.413 |
days | -0.404 |
Token | as |
Token ID | 355 |
Feature activation | +2.43e+00 |
Pos logit contributions | |
gru | +0.045 |
badge | +0.040 |
threatening | +0.037 |
rifle | +0.037 |
Reward | +0.037 |
Neg logit contributions | |
day | -0.031 |
afternoon | -0.028 |
cy | -0.026 |
ite | -0.026 |
night | -0.026 |
Token | the |
Token ID | 262 |
Feature activation | +2.97e+00 |
Pos logit contributions | |
day | +0.094 |
afternoon | +0.092 |
cy | +0.091 |
ite | +0.089 |
ops | +0.084 |
Neg logit contributions | |
gru | -0.091 |
badge | -0.074 |
threatening | -0.071 |
warning | -0.071 |
address | -0.071 |
Token | pet |
Token ID | 4273 |
Feature activation | -4.93e+00 |
Pos logit contributions | |
day | +0.096 |
he | +0.090 |
cy | +0.089 |
ite | +0.089 |
afternoon | +0.088 |
Neg logit contributions | |
gru | -0.130 |
badge | -0.109 |
warning | -0.108 |
address | -0.107 |
threatening | -0.106 |
Token | als |
Token ID | 874 |
Feature activation | -3.27e+00 |
Pos logit contributions | |
gru | +0.210 |
badge | +0.187 |
god | +0.182 |
Reward | +0.176 |
address | +0.174 |
Neg logit contributions | |
day | -0.167 |
afternoon | -0.156 |
cy | -0.140 |
night | -0.137 |
he | -0.137 |
Token | pirate |
Token ID | 25868 |
Feature activation | -4.79e+00 |
Pos logit contributions | |
gru | +0.090 |
ution | +0.080 |
Reward | +0.079 |
spoken | +0.075 |
rifle | +0.074 |
Neg logit contributions | |
day | -0.091 |
afternoon | -0.079 |
night | -0.076 |
morning | -0.072 |
cy | -0.069 |
Token | . |
Token ID | 13 |
Feature activation | +8.18e-01 |
Pos logit contributions | |
gru | +0.196 |
Reward | +0.164 |
badge | +0.161 |
ution | +0.161 |
threatening | +0.159 |
Neg logit contributions | |
day | -0.191 |
afternoon | -0.170 |
night | -0.161 |
cy | -0.156 |
ite | -0.155 |
Token | " |
Token ID | 366 |
Feature activation | -2.88e+00 |
Pos logit contributions | |
day | +0.032 |
afternoon | +0.029 |
night | +0.027 |
cy | +0.027 |
ite | +0.027 |
Neg logit contributions | |
gru | -0.033 |
badge | -0.028 |
Reward | -0.027 |
rifle | -0.026 |
ution | -0.026 |
Token | Ah |
Token ID | 10910 |
Feature activation | -4.06e+00 |
Pos logit contributions | |
gru | +0.118 |
badge | +0.105 |
rifle | +0.098 |
Reward | +0.097 |
threatening | +0.097 |
Neg logit contributions | |
day | -0.115 |
afternoon | -0.105 |
cy | -0.100 |
night | -0.096 |
he | -0.095 |
Token | oy |
Token ID | 726 |
Feature activation | -4.89e+00 |
Pos logit contributions | |
gru | +0.150 |
badge | +0.128 |
Reward | +0.117 |
ution | +0.117 |
god | +0.116 |
Neg logit contributions | |
day | -0.172 |
afternoon | -0.160 |
cy | -0.154 |
night | -0.150 |
he | -0.150 |
Token | mate |
Token ID | 16133 |
Feature activation | -1.42e+01 |
Pos logit contributions | |
gru | +0.229 |
badge | +0.207 |
Reward | +0.196 |
rifle | +0.195 |
ution | +0.192 |
Neg logit contributions | |
day | -0.154 |
afternoon | -0.144 |
night | -0.127 |
cy | -0.124 |
he | -0.121 |
Token | y |
Token ID | 88 |
Feature activation | +1.98e+00 |
Pos logit contributions | |
gru | +0.604 |
badge | +0.573 |
spoken | +0.549 |
rifle | +0.541 |
address | +0.541 |
Neg logit contributions | |
day | -0.392 |
afternoon | -0.390 |
days | -0.339 |
End | -0.329 |
olly | -0.329 |
Token | !" |
Token ID | 2474 |
Feature activation | +4.06e-01 |
Pos logit contributions | |
day | +0.064 |
afternoon | +0.060 |
cy | +0.055 |
he | +0.054 |
night | +0.053 |
Neg logit contributions | |
gru | -0.088 |
badge | -0.080 |
rifle | -0.076 |
Reward | -0.076 |
god | -0.074 |
Token | said |
Token ID | 531 |
Feature activation | -3.63e+00 |
Pos logit contributions | |
day | +0.014 |
afternoon | +0.012 |
night | +0.011 |
morning | +0.011 |
cy | +0.010 |
Neg logit contributions | |
gru | -0.019 |
badge | -0.017 |
ution | -0.017 |
fracture | -0.016 |
threatening | -0.016 |
Token | Lily |
Token ID | 20037 |
Feature activation | +3.81e+00 |
Pos logit contributions | |
gru | +0.180 |
badge | +0.169 |
ution | +0.168 |
Reward | +0.166 |
fracture | +0.165 |
Neg logit contributions | |
day | -0.110 |
night | -0.089 |
afternoon | -0.088 |
morning | -0.085 |
cy | -0.080 |
Token | . |
Token ID | 13 |
Feature activation | +6.23e-01 |
Pos logit contributions | |
day | +0.109 |
afternoon | +0.096 |
cy | +0.087 |
night | +0.087 |
ite | +0.084 |
Neg logit contributions | |
gru | -0.194 |
badge | -0.173 |
Reward | -0.166 |
rifle | -0.164 |
ution | -0.164 |
Token | god |
Token ID | 5770 |
Feature activation | -1.42e+01 |
Pos logit contributions | |
gru | +0.221 |
badge | +0.193 |
Reward | +0.188 |
ution | +0.186 |
rifle | +0.183 |
Neg logit contributions | |
day | -0.173 |
afternoon | -0.154 |
night | -0.144 |
cy | -0.141 |
ite | -0.138 |
Token | mother |
Token ID | 13552 |
Feature activation | -5.42e+00 |
Pos logit contributions | |
gru | +0.498 |
badge | +0.414 |
ution | +0.392 |
Reward | +0.391 |
rifle | +0.391 |
Neg logit contributions | |
day | -0.648 |
afternoon | -0.591 |
night | -0.567 |
cy | -0.554 |
he | -0.549 |
Token | . |
Token ID | 13 |
Feature activation | +5.15e-01 |
Pos logit contributions | |
gru | +0.237 |
badge | +0.210 |
god | +0.200 |
Reward | +0.198 |
rifle | +0.196 |
Neg logit contributions | |
day | -0.186 |
afternoon | -0.171 |
cy | -0.160 |
night | -0.155 |
ite | -0.153 |
Token | The |
Token ID | 383 |
Feature activation | -2.77e-01 |
Pos logit contributions | |
day | +0.018 |
afternoon | +0.016 |
cy | +0.015 |
night | +0.015 |
ite | +0.014 |
Neg logit contributions | |
gru | -0.023 |
badge | -0.020 |
ution | -0.019 |
rifle | -0.019 |
threatening | -0.019 |
Token | fairy |
Token ID | 25607 |
Feature activation | -4.92e+00 |
Pos logit contributions | |
gru | +0.009 |
badge | +0.007 |
Reward | +0.007 |
ution | +0.007 |
rifle | +0.007 |
Neg logit contributions | |
day | -0.013 |
afternoon | -0.012 |
night | -0.012 |
cy | -0.012 |
ite | -0.011 |
Token | god |
Token ID | 5770 |
Feature activation | -1.42e+01 |
Pos logit contributions | |
gru | +0.235 |
badge | +0.209 |
Reward | +0.202 |
ution | +0.199 |
rifle | +0.197 |
Neg logit contributions | |
day | -0.156 |
afternoon | -0.138 |
night | -0.127 |
cy | -0.127 |
ite | -0.122 |
Token | mother |
Token ID | 13552 |
Feature activation | -5.39e+00 |
Pos logit contributions | |
gru | +0.423 |
badge | +0.349 |
Reward | +0.327 |
god | +0.320 |
ution | +0.313 |
Neg logit contributions | |
day | -0.700 |
afternoon | -0.655 |
cy | -0.619 |
night | -0.616 |
ite | -0.605 |
Token | smiled |
Token ID | 13541 |
Feature activation | +1.55e+00 |
Pos logit contributions | |
gru | +0.248 |
badge | +0.219 |
Reward | +0.209 |
god | +0.208 |
rifle | +0.206 |
Neg logit contributions | |
day | -0.178 |
afternoon | -0.161 |
cy | -0.148 |
night | -0.147 |
ite | -0.142 |
Token | and |
Token ID | 290 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
day | +0.032 |
afternoon | +0.025 |
night | +0.024 |
cy | +0.023 |
morning | +0.021 |
Neg logit contributions | |
gru | -0.091 |
ution | -0.084 |
Reward | -0.083 |
badge | -0.083 |
rifle | -0.081 |
Token | said |
Token ID | 531 |
Feature activation | -3.89e+00 |
Pos logit contributions | |
day | +0.052 |
afternoon | +0.046 |
night | +0.044 |
cy | +0.042 |
morning | +0.041 |
Neg logit contributions | |
gru | -0.065 |
badge | -0.058 |
rifle | -0.056 |
ution | -0.056 |
Reward | -0.056 |
Token | , |
Token ID | 11 |
Feature activation | -5.40e-01 |
Pos logit contributions | |
gru | +0.229 |
badge | +0.212 |
rifle | +0.210 |
ution | +0.208 |
Reward | +0.207 |
Neg logit contributions | |
day | -0.076 |
afternoon | -0.056 |
night | -0.053 |
morning | -0.049 |
olly | -0.046 |