Token | . |
Token ID | 13 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ishing | +0.587 |
eking | +0.570 |
breath | +0.567 |
crown | +0.558 |
luck | +0.550 |
Neg logit contributions | |
| -0.739 |
She | -0.700 |
He | -0.685 |
They | -0.675 |
It | -0.664 |
Token | She |
Token ID | 1375 |
Feature activation | -6.11e-01 |
Pos logit contributions | |
job | +0.243 |
luck | +0.242 |
attention | +0.238 |
exam | +0.237 |
problem | +0.236 |
Neg logit contributions | |
Would | -0.065 |
Look | -0.061 |
| -0.060 |
Did | -0.059 |
Wow | -0.059 |
Token | took |
Token ID | 1718 |
Feature activation | -1.48e-02 |
Pos logit contributions | |
| +1.368 |
He | +1.261 |
She | +1.254 |
They | +1.229 |
It | +1.214 |
Neg logit contributions | |
ishing | -1.364 |
eking | -1.324 |
breath | -1.317 |
crown | -1.291 |
veil | -1.289 |
Token | a |
Token ID | 257 |
Feature activation | +5.94e-02 |
Pos logit contributions | |
| +0.038 |
She | +0.036 |
He | +0.035 |
They | +0.035 |
It | +0.035 |
Neg logit contributions | |
ishing | -0.028 |
breath | -0.027 |
eking | -0.027 |
crown | -0.027 |
veil | -0.026 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.94e-01 |
Pos logit contributions | |
ishing | +0.109 |
eking | +0.104 |
ies | +0.095 |
size | +0.092 |
luck | +0.092 |
Neg logit contributions | |
| -0.176 |
| -0.157 |
He | -0.157 |
It | -0.152 |
She | -0.150 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
eking | +0.209 |
ishing | +0.183 |
ies | +0.169 |
airplane | +0.145 |
veil | +0.142 |
Neg logit contributions | |
| -1.091 |
He | -1.080 |
She | -1.048 |
It | -1.024 |
They | -1.011 |
Token | and |
Token ID | 290 |
Feature activation | +5.92e-02 |
Pos logit contributions | |
ishing | +1.207 |
breath | +1.170 |
eking | +1.168 |
crown | +1.163 |
veil | +1.146 |
Neg logit contributions | |
| -1.399 |
She | -1.311 |
He | -1.289 |
They | -1.267 |
It | -1.254 |
Token | slowly |
Token ID | 6364 |
Feature activation | +8.30e-02 |
Pos logit contributions | |
ishing | +0.130 |
breath | +0.126 |
eking | +0.126 |
crown | +0.125 |
veil | +0.124 |
Neg logit contributions | |
| -0.134 |
She | -0.124 |
He | -0.123 |
They | -0.121 |
It | -0.120 |
Token | walked |
Token ID | 6807 |
Feature activation | -1.37e-01 |
Pos logit contributions | |
ishing | +0.189 |
ies | +0.184 |
veil | +0.183 |
eking | +0.182 |
mess | +0.182 |
Neg logit contributions | |
| -0.184 |
He | -0.170 |
She | -0.168 |
They | -0.161 |
It | -0.161 |
Token | up |
Token ID | 510 |
Feature activation | +3.10e-01 |
Pos logit contributions | |
| +0.434 |
She | +0.415 |
He | +0.411 |
They | +0.406 |
It | +0.400 |
Neg logit contributions | |
ishing | -0.181 |
eking | -0.172 |
breath | -0.170 |
crown | -0.169 |
veil | -0.168 |
Token | to |
Token ID | 284 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
ishing | +0.634 |
breath | +0.621 |
eking | +0.613 |
crown | +0.613 |
luck | +0.605 |
Neg logit contributions | |
| -0.742 |
She | -0.697 |
He | -0.688 |
They | -0.681 |
It | -0.672 |
Token | the |
Token ID | 262 |
Feature activation | +3.52e-02 |
Pos logit contributions | |
| +0.824 |
She | +0.788 |
He | +0.769 |
They | +0.764 |
It | +0.758 |
Neg logit contributions | |
breath | -0.626 |
ishing | -0.622 |
luck | -0.615 |
crown | -0.605 |
ies | -0.596 |
Token | bunny |
Token ID | 44915 |
Feature activation | +1.87e-01 |
Pos logit contributions | |
ishing | +0.071 |
eking | +0.067 |
ies | +0.065 |
attention | +0.063 |
et | +0.061 |
Neg logit contributions | |
| -0.094 |
She | -0.084 |
He | -0.084 |
It | -0.082 |
| -0.080 |
Token | took |
Token ID | 1718 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
ishing | +0.404 |
breath | +0.397 |
eking | +0.393 |
crown | +0.390 |
veil | +0.386 |
Neg logit contributions | |
| -0.424 |
She | -0.396 |
He | -0.391 |
They | -0.387 |
It | -0.381 |
Token | a |
Token ID | 257 |
Feature activation | +5.81e-02 |
Pos logit contributions | |
| +0.042 |
He | +0.038 |
She | +0.038 |
It | +0.038 |
They | +0.038 |
Neg logit contributions | |
ishing | -0.035 |
eking | -0.034 |
ies | -0.034 |
veil | -0.034 |
breath | -0.033 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.105 |
eking | +0.101 |
ies | +0.089 |
et | +0.088 |
size | +0.086 |
Neg logit contributions | |
| -0.175 |
| -0.157 |
He | -0.154 |
It | -0.148 |
She | -0.148 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
eking | +0.217 |
ishing | +0.171 |
thieves | +0.156 |
veil | +0.153 |
airplane | +0.152 |
Neg logit contributions | |
| -1.103 |
He | -1.076 |
She | -1.035 |
It | -1.020 |
They | -1.018 |
Token | , |
Token ID | 11 |
Feature activation | +1.35e-01 |
Pos logit contributions | |
ishing | +1.142 |
breath | +1.120 |
crown | +1.104 |
eking | +1.103 |
veil | +1.088 |
Neg logit contributions | |
| -1.431 |
She | -1.345 |
He | -1.328 |
They | -1.314 |
It | -1.299 |
Token | and |
Token ID | 290 |
Feature activation | +5.87e-02 |
Pos logit contributions | |
ishing | +0.265 |
breath | +0.256 |
crown | +0.254 |
eking | +0.254 |
ies | +0.252 |
Neg logit contributions | |
| -0.337 |
She | -0.317 |
He | -0.314 |
They | -0.310 |
It | -0.308 |
Token | gathered |
Token ID | 9272 |
Feature activation | +1.87e-01 |
Pos logit contributions | |
ishing | +0.129 |
breath | +0.126 |
eking | +0.125 |
crown | +0.124 |
veil | +0.123 |
Neg logit contributions | |
| -0.132 |
She | -0.123 |
He | -0.122 |
They | -0.120 |
It | -0.119 |
Token | all |
Token ID | 477 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
eking | +0.339 |
ishing | +0.337 |
veil | +0.328 |
crown | +0.328 |
ies | +0.326 |
Neg logit contributions | |
| -0.499 |
She | -0.467 |
He | -0.466 |
They | -0.460 |
It | -0.457 |
Token | his |
Token ID | 465 |
Feature activation | -3.57e-03 |
Pos logit contributions | |
ishing | +0.429 |
breath | +0.427 |
crown | +0.406 |
luck | +0.406 |
ies | +0.406 |
Neg logit contributions | |
| -0.569 |
She | -0.541 |
He | -0.534 |
They | -0.530 |
It | -0.527 |
Token | , |
Token ID | 11 |
Feature activation | +1.34e-01 |
Pos logit contributions | |
ishing | +0.533 |
breath | +0.524 |
crown | +0.507 |
eking | +0.507 |
veil | +0.507 |
Neg logit contributions | |
| -0.654 |
She | -0.621 |
They | -0.597 |
He | -0.596 |
It | -0.575 |
Token | Mary |
Token ID | 5335 |
Feature activation | -1.63e-01 |
Pos logit contributions | |
breath | +0.329 |
crown | +0.324 |
luck | +0.323 |
ies | +0.323 |
veil | +0.322 |
Neg logit contributions | |
| -0.250 |
He | -0.231 |
She | -0.229 |
It | -0.222 |
They | -0.221 |
Token | took |
Token ID | 1718 |
Feature activation | -1.76e-02 |
Pos logit contributions | |
| +0.356 |
She | +0.333 |
He | +0.328 |
They | +0.324 |
It | +0.320 |
Neg logit contributions | |
ishing | -0.363 |
breath | -0.357 |
crown | -0.352 |
eking | -0.352 |
luck | -0.348 |
Token | a |
Token ID | 257 |
Feature activation | +5.72e-02 |
Pos logit contributions | |
| +0.046 |
She | +0.043 |
They | +0.042 |
It | +0.042 |
He | +0.041 |
Neg logit contributions | |
ishing | -0.035 |
eking | -0.034 |
mess | -0.032 |
ies | -0.032 |
veil | -0.031 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.118 |
eking | +0.105 |
et | +0.099 |
size | +0.098 |
ier | +0.097 |
Neg logit contributions | |
| -0.161 |
| -0.146 |
It | -0.134 |
He | -0.133 |
She | -0.133 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
eking | +0.272 |
ier | +0.233 |
ishing | +0.230 |
attention | +0.193 |
luck | +0.191 |
Neg logit contributions | |
| -1.021 |
He | -0.990 |
She | -0.982 |
They | -0.970 |
It | -0.942 |
Token | before |
Token ID | 878 |
Feature activation | +1.78e-01 |
Pos logit contributions | |
ishing | +1.286 |
eking | +1.235 |
crown | +1.228 |
breath | +1.224 |
luck | +1.213 |
Neg logit contributions | |
| -1.377 |
She | -1.274 |
He | -1.243 |
They | -1.225 |
It | -1.199 |
Token | walking |
Token ID | 6155 |
Feature activation | +1.66e-01 |
Pos logit contributions | |
ishing | +0.371 |
breath | +0.353 |
eking | +0.353 |
crown | +0.352 |
luck | +0.347 |
Neg logit contributions | |
| -0.428 |
She | -0.402 |
He | -0.393 |
They | -0.392 |
It | -0.386 |
Token | inside |
Token ID | 2641 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
ishing | +0.286 |
eking | +0.278 |
mess | +0.266 |
breath | +0.266 |
crown | +0.263 |
Neg logit contributions | |
| -0.470 |
She | -0.448 |
He | -0.440 |
They | -0.437 |
It | -0.422 |
Token | . |
Token ID | 13 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
breath | +0.539 |
ishing | +0.535 |
crown | +0.525 |
veil | +0.521 |
ies | +0.520 |
Neg logit contributions | |
| -0.758 |
He | -0.718 |
She | -0.718 |
They | -0.706 |
It | -0.702 |
Token | All |
Token ID | 1439 |
Feature activation | -4.75e-01 |
Pos logit contributions | |
attention | +0.273 |
calendar | +0.272 |
luck | +0.271 |
job | +0.270 |
size | +0.269 |
Neg logit contributions | |
| -0.085 |
Well | -0.068 |
He | -0.063 |
He | -0.061 |
Don | -0.061 |
Token |
|
Token ID | 198 |
Feature activation | -6.56e-01 |
Pos logit contributions | |
Did | +0.660 |
She | +0.597 |
They | +0.560 |
It | +0.493 |
Her | +0.457 |
Neg logit contributions | |
breath | -1.810 |
adventure | -1.804 |
balance | -1.789 |
- | -1.787 |
mess | -1.744 |
Token | John |
Token ID | 7554 |
Feature activation | -2.51e-01 |
Pos logit contributions | |
She | +1.002 |
| +0.990 |
They | +0.984 |
It | +0.944 |
He | +0.883 |
Neg logit contributions | |
mess | -1.685 |
eking | -1.649 |
airplane | -1.637 |
island | -1.634 |
cream | -1.630 |
Token | took |
Token ID | 1718 |
Feature activation | -1.79e-02 |
Pos logit contributions | |
| +0.543 |
She | +0.511 |
He | +0.500 |
They | +0.497 |
It | +0.493 |
Neg logit contributions | |
ishing | -0.561 |
breath | -0.552 |
crown | -0.543 |
eking | -0.540 |
ies | -0.539 |
Token | a |
Token ID | 257 |
Feature activation | +5.80e-02 |
Pos logit contributions | |
| +0.041 |
He | +0.037 |
It | +0.037 |
She | +0.035 |
They | +0.035 |
Neg logit contributions | |
eking | -0.042 |
ishing | -0.042 |
veil | -0.041 |
ies | -0.041 |
mess | -0.040 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.94e-01 |
Pos logit contributions | |
ishing | +0.095 |
eking | +0.089 |
luck | +0.087 |
ier | +0.078 |
et | +0.078 |
Neg logit contributions | |
| -0.180 |
| -0.161 |
He | -0.160 |
She | -0.160 |
It | -0.159 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
eking | +0.191 |
ishing | +0.188 |
luck | +0.178 |
ier | +0.152 |
thieves | +0.150 |
Neg logit contributions | |
| -1.103 |
He | -1.075 |
She | -1.067 |
They | -1.037 |
It | -1.036 |
Token | and |
Token ID | 290 |
Feature activation | +5.55e-02 |
Pos logit contributions | |
ishing | +1.270 |
eking | +1.229 |
crown | +1.220 |
breath | +1.218 |
luck | +1.205 |
Neg logit contributions | |
| -1.369 |
She | -1.274 |
He | -1.254 |
They | -1.226 |
It | -1.211 |
Token | walked |
Token ID | 6807 |
Feature activation | -1.38e-01 |
Pos logit contributions | |
ishing | +0.119 |
breath | +0.116 |
eking | +0.116 |
crown | +0.115 |
luck | +0.114 |
Neg logit contributions | |
| -0.128 |
She | -0.119 |
He | -0.118 |
They | -0.116 |
It | -0.115 |
Token | into |
Token ID | 656 |
Feature activation | +2.27e-01 |
Pos logit contributions | |
| +0.375 |
She | +0.352 |
He | +0.351 |
They | +0.344 |
It | +0.336 |
Neg logit contributions | |
ishing | -0.251 |
eking | -0.245 |
veil | -0.240 |
crown | -0.238 |
mess | -0.235 |
Token | the |
Token ID | 262 |
Feature activation | +3.36e-02 |
Pos logit contributions | |
breath | +0.459 |
ishing | +0.458 |
crown | +0.446 |
eking | +0.442 |
veil | +0.440 |
Neg logit contributions | |
| -0.541 |
She | -0.511 |
He | -0.502 |
They | -0.499 |
It | -0.492 |
Token | tunnel |
Token ID | 13275 |
Feature activation | +3.35e-01 |
Pos logit contributions | |
ishing | +0.031 |
breath | +0.030 |
eking | +0.029 |
crown | +0.029 |
veil | +0.028 |
Neg logit contributions | |
| -0.118 |
She | -0.113 |
He | -0.112 |
They | -0.111 |
It | -0.110 |
Token | so |
Token ID | 523 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
breath | +0.252 |
crown | +0.249 |
eking | +0.247 |
ishing | +0.246 |
veil | +0.245 |
Neg logit contributions | |
| -0.328 |
He | -0.304 |
She | -0.304 |
They | -0.301 |
It | -0.297 |
Token | she |
Token ID | 673 |
Feature activation | -1.92e-01 |
Pos logit contributions | |
breath | +0.270 |
veil | +0.268 |
crown | +0.264 |
ies | +0.263 |
luck | +0.258 |
Neg logit contributions | |
| -0.211 |
He | -0.204 |
They | -0.199 |
It | -0.198 |
She | -0.195 |
Token | took |
Token ID | 1718 |
Feature activation | -1.77e-02 |
Pos logit contributions | |
| +0.424 |
She | +0.394 |
He | +0.387 |
They | +0.384 |
It | +0.381 |
Neg logit contributions | |
ishing | -0.430 |
luck | -0.416 |
eking | -0.415 |
breath | -0.414 |
attention | -0.407 |
Token | a |
Token ID | 257 |
Feature activation | +5.76e-02 |
Pos logit contributions | |
| +0.040 |
She | +0.037 |
He | +0.036 |
They | +0.036 |
It | +0.036 |
Neg logit contributions | |
ishing | -0.039 |
eking | -0.038 |
breath | -0.038 |
crown | -0.038 |
ies | -0.038 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.094 |
eking | +0.090 |
luck | +0.084 |
size | +0.079 |
ies | +0.079 |
Neg logit contributions | |
| -0.180 |
| -0.161 |
It | -0.159 |
She | -0.157 |
He | -0.153 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
eking | +0.189 |
ishing | +0.162 |
attention | +0.138 |
luck | +0.137 |
ies | +0.133 |
Neg logit contributions | |
| -1.114 |
It | -1.063 |
She | -1.060 |
He | -1.044 |
They | -1.021 |
Token | and |
Token ID | 290 |
Feature activation | +5.71e-02 |
Pos logit contributions | |
ishing | +1.272 |
eking | +1.229 |
crown | +1.220 |
breath | +1.214 |
luck | +1.205 |
Neg logit contributions | |
| -1.385 |
She | -1.282 |
He | -1.247 |
They | -1.221 |
It | -1.206 |
Token | went |
Token ID | 1816 |
Feature activation | -4.11e-02 |
Pos logit contributions | |
ishing | +0.123 |
eking | +0.119 |
luck | +0.117 |
breath | +0.116 |
ies | +0.115 |
Neg logit contributions | |
| -0.134 |
She | -0.124 |
It | -0.121 |
He | -0.120 |
They | -0.120 |
Token | to |
Token ID | 284 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
| +0.108 |
She | +0.101 |
They | +0.097 |
He | +0.097 |
It | +0.095 |
Neg logit contributions | |
ishing | -0.079 |
eking | -0.077 |
amount | -0.076 |
ies | -0.075 |
decision | -0.075 |
Token | her |
Token ID | 607 |
Feature activation | +2.60e-02 |
Pos logit contributions | |
ishing | +0.029 |
eking | +0.028 |
breath | +0.028 |
crown | +0.028 |
ies | +0.028 |
Neg logit contributions | |
| -0.037 |
She | -0.034 |
He | -0.034 |
They | -0.034 |
It | -0.033 |
Token | mom |
Token ID | 1995 |
Feature activation | +7.24e-02 |
Pos logit contributions | |
eking | +0.048 |
ishing | +0.046 |
breath | +0.043 |
luck | +0.043 |
ies | +0.043 |
Neg logit contributions | |
| -0.074 |
She | -0.068 |
They | -0.065 |
He | -0.065 |
It | -0.064 |
Token | sat |
Token ID | 3332 |
Feature activation | -1.20e-02 |
Pos logit contributions | |
veil | +0.122 |
breath | +0.122 |
ies | +0.121 |
crown | +0.120 |
ishing | +0.119 |
Neg logit contributions | |
| -0.126 |
He | -0.121 |
They | -0.119 |
She | -0.118 |
It | -0.116 |
Token | down |
Token ID | 866 |
Feature activation | +1.92e-01 |
Pos logit contributions | |
| +0.035 |
She | +0.034 |
He | +0.033 |
They | +0.033 |
It | +0.033 |
Neg logit contributions | |
ishing | -0.018 |
breath | -0.018 |
crown | -0.017 |
eking | -0.017 |
veil | -0.017 |
Token | to |
Token ID | 284 |
Feature activation | +1.39e-02 |
Pos logit contributions | |
ishing | +0.361 |
breath | +0.353 |
crown | +0.348 |
eking | +0.348 |
veil | +0.342 |
Neg logit contributions | |
| -0.493 |
She | -0.466 |
He | -0.459 |
They | -0.455 |
It | -0.449 |
Token | catch |
Token ID | 4929 |
Feature activation | +1.92e-01 |
Pos logit contributions | |
breath | +0.023 |
crown | +0.022 |
ishing | +0.022 |
veil | +0.022 |
eking | +0.021 |
Neg logit contributions | |
| -0.038 |
She | -0.036 |
He | -0.036 |
They | -0.036 |
It | -0.035 |
Token | her |
Token ID | 607 |
Feature activation | +2.44e-02 |
Pos logit contributions | |
ishing | +0.429 |
eking | +0.417 |
breath | +0.408 |
crown | +0.405 |
airplane | +0.400 |
Neg logit contributions | |
| -0.445 |
She | -0.414 |
It | -0.402 |
They | -0.398 |
He | -0.397 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
ishing | +0.046 |
eking | +0.045 |
plank | +0.043 |
island | +0.042 |
luck | +0.040 |
Neg logit contributions | |
| -0.067 |
She | -0.062 |
He | -0.061 |
They | -0.060 |
It | -0.060 |
Token | . |
Token ID | 13 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ishing | +1.182 |
breath | +1.149 |
eking | +1.141 |
crown | +1.138 |
veil | +1.129 |
Neg logit contributions | |
| -1.452 |
She | -1.359 |
He | -1.336 |
They | -1.314 |
It | -1.296 |
Token | Suddenly |
Token ID | 24975 |
Feature activation | -4.74e-01 |
Pos logit contributions | |
head | +0.249 |
egg | +0.246 |
veil | +0.246 |
home | +0.244 |
crown | +0.243 |
Neg logit contributions | |
He | -0.078 |
Max | -0.078 |
Well | -0.075 |
| -0.068 |
Don | -0.067 |
Token | , |
Token ID | 11 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
| +1.177 |
She | +1.115 |
He | +1.068 |
It | +1.062 |
They | +1.060 |
Neg logit contributions | |
ishing | -0.972 |
eking | -0.922 |
breath | -0.903 |
mess | -0.897 |
crown | -0.893 |
Token | a |
Token ID | 257 |
Feature activation | +5.72e-02 |
Pos logit contributions | |
breath | +0.320 |
veil | +0.312 |
crown | +0.312 |
luck | +0.311 |
ishing | +0.308 |
Neg logit contributions | |
| -0.272 |
He | -0.256 |
They | -0.253 |
She | -0.252 |
It | -0.246 |
Token | deaf |
Token ID | 21815 |
Feature activation | +2.48e-01 |
Pos logit contributions | |
ishing | +0.100 |
eking | +0.093 |
ies | +0.085 |
breath | +0.082 |
attention | +0.081 |
Neg logit contributions | |
| -0.169 |
She | -0.159 |
He | -0.159 |
It | -0.157 |
They | -0.153 |
Token | both |
Token ID | 1111 |
Feature activation | +5.10e-03 |
Pos logit contributions | |
She | +0.420 |
| +0.415 |
He | +0.409 |
They | +0.408 |
It | +0.400 |
Neg logit contributions | |
breath | -0.444 |
crown | -0.433 |
luck | -0.425 |
island | -0.419 |
ies | -0.418 |
Token | finished |
Token ID | 5201 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
ishing | +0.011 |
breath | +0.011 |
crown | +0.011 |
eking | +0.011 |
luck | +0.010 |
Neg logit contributions | |
| -0.012 |
She | -0.011 |
He | -0.011 |
They | -0.011 |
It | -0.011 |
Token | , |
Token ID | 11 |
Feature activation | +1.37e-01 |
Pos logit contributions | |
ishing | +0.496 |
eking | +0.478 |
veil | +0.465 |
crown | +0.461 |
breath | +0.455 |
Neg logit contributions | |
| -0.588 |
She | -0.547 |
They | -0.540 |
He | -0.532 |
It | -0.514 |
Token | they |
Token ID | 484 |
Feature activation | -2.02e-01 |
Pos logit contributions | |
breath | +0.318 |
luck | +0.309 |
ies | +0.308 |
crown | +0.306 |
ishing | +0.305 |
Neg logit contributions | |
| -0.272 |
She | -0.258 |
He | -0.256 |
It | -0.247 |
They | -0.244 |
Token | were |
Token ID | 547 |
Feature activation | -2.13e-01 |
Pos logit contributions | |
| +0.437 |
She | +0.405 |
He | +0.399 |
They | +0.395 |
It | +0.389 |
Neg logit contributions | |
ishing | -0.461 |
breath | -0.451 |
eking | -0.446 |
crown | -0.444 |
luck | -0.440 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
| +0.544 |
She | +0.494 |
He | +0.484 |
They | +0.479 |
| +0.476 |
Neg logit contributions | |
ishing | -0.437 |
veil | -0.404 |
potion | -0.403 |
plank | -0.395 |
breath | -0.394 |
Token | less |
Token ID | 1203 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
ishing | +1.188 |
crown | +1.158 |
veil | +1.146 |
breath | +1.137 |
eking | +1.136 |
Neg logit contributions | |
| -1.469 |
She | -1.354 |
He | -1.337 |
They | -1.323 |
It | -1.284 |
Token | . |
Token ID | 13 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
ishing | +0.442 |
breath | +0.432 |
eking | +0.426 |
crown | +0.425 |
veil | +0.419 |
Neg logit contributions | |
| -0.606 |
She | -0.571 |
He | -0.564 |
They | -0.559 |
It | -0.552 |
Token | G |
Token ID | 402 |
Feature activation | -1.47e-01 |
Pos logit contributions | |
sand | +0.240 |
luck | +0.240 |
attention | +0.238 |
amount | +0.236 |
island | +0.234 |
Neg logit contributions | |
Did | -0.077 |
She | -0.067 |
He | -0.064 |
| -0.061 |
She | -0.061 |
Token | ail |
Token ID | 603 |
Feature activation | +1.85e-01 |
Pos logit contributions | |
| +0.455 |
She | +0.433 |
He | +0.432 |
They | +0.428 |
It | +0.423 |
Neg logit contributions | |
ishing | -0.196 |
breath | -0.187 |
eking | -0.186 |
crown | -0.183 |
ies | -0.181 |
Token | smiled |
Token ID | 13541 |
Feature activation | +4.30e-02 |
Pos logit contributions | |
ishing | +0.428 |
breath | +0.425 |
crown | +0.419 |
eking | +0.416 |
veil | +0.413 |
Neg logit contributions | |
| -0.379 |
She | -0.357 |
He | -0.350 |
They | -0.347 |
It | -0.342 |
Token | had |
Token ID | 550 |
Feature activation | -9.55e-02 |
Pos logit contributions | |
| +0.264 |
She | +0.244 |
He | +0.237 |
They | +0.232 |
It | +0.231 |
Neg logit contributions | |
ishing | -0.367 |
breath | -0.363 |
eking | -0.359 |
luck | -0.356 |
crown | -0.355 |
Token | to |
Token ID | 284 |
Feature activation | +1.66e-02 |
Pos logit contributions | |
| +0.234 |
She | +0.218 |
He | +0.215 |
They | +0.212 |
It | +0.212 |
Neg logit contributions | |
ishing | -0.198 |
breath | -0.188 |
eking | -0.187 |
ies | -0.186 |
crown | -0.184 |
Token | take |
Token ID | 1011 |
Feature activation | +4.19e-02 |
Pos logit contributions | |
ishing | +0.032 |
eking | +0.029 |
ies | +0.029 |
problem | +0.028 |
crown | +0.027 |
Neg logit contributions | |
| -0.045 |
She | -0.042 |
He | -0.042 |
They | -0.041 |
It | -0.041 |
Token | a |
Token ID | 257 |
Feature activation | +5.85e-02 |
Pos logit contributions | |
ishing | +0.088 |
eking | +0.085 |
ies | +0.083 |
island | +0.081 |
veil | +0.080 |
Neg logit contributions | |
| -0.107 |
She | -0.100 |
It | -0.098 |
They | -0.097 |
He | -0.095 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
ishing | +0.123 |
eking | +0.114 |
ps | +0.106 |
et | +0.104 |
isher | +0.103 |
Neg logit contributions | |
| -0.165 |
| -0.146 |
He | -0.137 |
It | -0.136 |
She | -0.133 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
eking | +0.338 |
ier | +0.285 |
ishing | +0.276 |
vell | +0.263 |
attention | +0.250 |
Neg logit contributions | |
| -0.985 |
He | -0.971 |
They | -0.939 |
She | -0.937 |
It | -0.914 |
Token | and |
Token ID | 290 |
Feature activation | +5.77e-02 |
Pos logit contributions | |
ishing | +1.207 |
eking | +1.163 |
crown | +1.156 |
breath | +1.155 |
veil | +1.133 |
Neg logit contributions | |
| -1.458 |
She | -1.363 |
He | -1.332 |
They | -1.295 |
It | -1.277 |
Token | be |
Token ID | 307 |
Feature activation | +6.52e-02 |
Pos logit contributions | |
ishing | +0.111 |
eking | +0.104 |
ies | +0.101 |
breath | +0.101 |
crown | +0.098 |
Neg logit contributions | |
| -0.152 |
She | -0.144 |
He | -0.142 |
They | -0.140 |
It | -0.139 |
Token | brave |
Token ID | 14802 |
Feature activation | +2.98e-01 |
Pos logit contributions | |
decision | +0.118 |
problem | +0.117 |
mess | +0.112 |
veil | +0.112 |
exam | +0.111 |
Neg logit contributions | |
| -0.176 |
She | -0.156 |
He | -0.154 |
| -0.153 |
It | -0.150 |
Token | ! |
Token ID | 0 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
eking | +0.537 |
ishing | +0.515 |
crown | +0.480 |
aches | +0.473 |
veil | +0.472 |
Neg logit contributions | |
| -0.820 |
He | -0.798 |
She | -0.794 |
They | -0.773 |
It | -0.744 |
Token |
|
Token ID | 198 |
Feature activation | -6.54e-01 |
Pos logit contributions | |
home | +0.539 |
job | +0.531 |
exam | +0.486 |
way | +0.485 |
adventure | +0.473 |
Neg logit contributions | |
Would | -0.165 |
Is | -0.162 |
His | -0.154 |
He | -0.150 |
Did | -0.145 |
Token | they |
Token ID | 484 |
Feature activation | -2.01e-01 |
Pos logit contributions | |
ishing | +0.328 |
breath | +0.326 |
crown | +0.320 |
eking | +0.318 |
luck | +0.317 |
Neg logit contributions | |
| -0.272 |
She | -0.254 |
He | -0.250 |
They | -0.246 |
It | -0.243 |
Token | arrived |
Token ID | 5284 |
Feature activation | +1.14e-02 |
Pos logit contributions | |
| +0.468 |
She | +0.441 |
He | +0.434 |
They | +0.430 |
It | +0.425 |
Neg logit contributions | |
ishing | -0.421 |
breath | -0.416 |
crown | -0.410 |
eking | -0.407 |
luck | -0.404 |
Token | home |
Token ID | 1363 |
Feature activation | +4.17e-01 |
Pos logit contributions | |
ishing | +0.018 |
eking | +0.018 |
breath | +0.018 |
veil | +0.018 |
crown | +0.018 |
Neg logit contributions | |
| -0.033 |
She | -0.031 |
He | -0.031 |
They | -0.031 |
It | -0.030 |
Token | out |
Token ID | 503 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
ishing | +0.853 |
veil | +0.829 |
crown | +0.824 |
eking | +0.822 |
breath | +0.816 |
Neg logit contributions | |
| -1.056 |
She | -0.986 |
He | -0.960 |
They | -0.954 |
It | -0.915 |
Token | of |
Token ID | 286 |
Feature activation | +2.65e-02 |
Pos logit contributions | |
ishing | +0.564 |
crown | +0.550 |
veil | +0.548 |
eking | +0.547 |
breath | +0.544 |
Neg logit contributions | |
| -0.632 |
She | -0.580 |
He | -0.570 |
They | -0.564 |
It | -0.549 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
eking | +0.066 |
ishing | +0.063 |
ies | +0.061 |
crown | +0.061 |
an | +0.059 |
Neg logit contributions | |
| -0.059 |
She | -0.052 |
They | -0.051 |
He | -0.049 |
It | -0.049 |
Token | and |
Token ID | 290 |
Feature activation | +6.01e-02 |
Pos logit contributions | |
ishing | +1.217 |
breath | +1.198 |
eking | +1.196 |
crown | +1.194 |
veil | +1.175 |
Neg logit contributions | |
| -1.394 |
She | -1.299 |
He | -1.279 |
They | -1.266 |
It | -1.236 |
Token | laughing |
Token ID | 14376 |
Feature activation | +2.43e-01 |
Pos logit contributions | |
ishing | +0.143 |
eking | +0.139 |
breath | +0.137 |
luck | +0.135 |
crown | +0.134 |
Neg logit contributions | |
| -0.128 |
She | -0.118 |
He | -0.116 |
They | -0.116 |
It | -0.113 |
Token | . |
Token ID | 13 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
eking | +0.502 |
ishing | +0.498 |
breath | +0.480 |
veil | +0.478 |
crown | +0.470 |
Neg logit contributions | |
| -0.608 |
She | -0.570 |
He | -0.570 |
They | -0.568 |
It | -0.550 |
Token | From |
Token ID | 3574 |
Feature activation | -5.59e-01 |
Pos logit contributions | |
mess | +0.120 |
earthquake | +0.115 |
job | +0.114 |
mine | +0.111 |
food | +0.110 |
Neg logit contributions | |
He | -0.187 |
Don | -0.180 |
| -0.180 |
His | -0.178 |
Well | -0.177 |
Token | that |
Token ID | 326 |
Feature activation | +7.13e-02 |
Pos logit contributions | |
| +1.196 |
She | +1.087 |
He | +1.040 |
They | +1.032 |
It | +1.014 |
Neg logit contributions | |
amount | -1.279 |
breath | -1.272 |
ishing | -1.269 |
crown | -1.269 |
eking | -1.264 |
Token | brave |
Token ID | 14802 |
Feature activation | +2.96e-01 |
Pos logit contributions | |
| +0.186 |
She | +0.176 |
He | +0.174 |
They | +0.172 |
It | +0.171 |
Neg logit contributions | |
ishing | -0.121 |
breath | -0.118 |
eking | -0.116 |
crown | -0.116 |
veil | -0.114 |
Token | . |
Token ID | 13 |
Feature activation | +9.46e-02 |
Pos logit contributions | |
eking | +0.486 |
ishing | +0.471 |
crown | +0.464 |
veil | +0.455 |
breath | +0.455 |
Neg logit contributions | |
| -0.854 |
He | -0.836 |
She | -0.817 |
They | -0.800 |
It | -0.795 |
Token | Taking |
Token ID | 20879 |
Feature activation | -3.92e-01 |
Pos logit contributions | |
luck | +0.274 |
job | +0.270 |
problem | +0.267 |
attention | +0.261 |
enemy | +0.261 |
Neg logit contributions | |
| -0.057 |
Well | -0.045 |
She | -0.041 |
Would | -0.039 |
| -0.037 |
Token | a |
Token ID | 257 |
Feature activation | +5.59e-02 |
Pos logit contributions | |
| +0.926 |
He | +0.863 |
It | +0.863 |
She | +0.859 |
They | +0.837 |
Neg logit contributions | |
eking | -0.848 |
ishing | -0.826 |
ies | -0.822 |
breath | -0.808 |
crown | -0.804 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.91e-01 |
Pos logit contributions | |
ishing | +0.100 |
eking | +0.093 |
ies | +0.086 |
island | +0.084 |
et | +0.082 |
Neg logit contributions | |
| -0.166 |
| -0.153 |
He | -0.152 |
She | -0.149 |
It | -0.148 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
eking | +0.234 |
ishing | +0.209 |
ier | +0.185 |
vell | +0.179 |
airplane | +0.178 |
Neg logit contributions | |
He | -1.045 |
| -1.033 |
It | -1.015 |
She | -1.012 |
They | -0.990 |
Token | , |
Token ID | 11 |
Feature activation | +1.32e-01 |
Pos logit contributions | |
ishing | +1.063 |
breath | +1.043 |
crown | +1.025 |
eking | +1.024 |
luck | +1.009 |
Neg logit contributions | |
| -1.497 |
She | -1.413 |
He | -1.396 |
They | -1.383 |
It | -1.368 |
Token | he |
Token ID | 339 |
Feature activation | -2.26e-01 |
Pos logit contributions | |
ishing | +0.342 |
breath | +0.342 |
crown | +0.335 |
luck | +0.335 |
eking | +0.335 |
Neg logit contributions | |
| -0.234 |
She | -0.217 |
He | -0.210 |
They | -0.210 |
It | -0.205 |
Token | walked |
Token ID | 6807 |
Feature activation | -1.40e-01 |
Pos logit contributions | |
| +0.502 |
He | +0.460 |
She | +0.460 |
They | +0.450 |
It | +0.444 |
Neg logit contributions | |
ishing | -0.506 |
eking | -0.489 |
breath | -0.489 |
crown | -0.482 |
luck | -0.481 |
Token | up |
Token ID | 510 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
| +0.415 |
She | +0.395 |
He | +0.393 |
They | +0.386 |
It | +0.381 |
Neg logit contributions | |
ishing | -0.211 |
eking | -0.202 |
breath | -0.202 |
crown | -0.200 |
veil | -0.199 |
Token | to |
Token ID | 284 |
Feature activation | +1.37e-02 |
Pos logit contributions | |
ishing | +0.622 |
breath | +0.617 |
crown | +0.604 |
eking | +0.600 |
luck | +0.595 |
Neg logit contributions | |
| -0.733 |
She | -0.690 |
He | -0.680 |
They | -0.674 |
It | -0.667 |
Token | . |
Token ID | 13 |
Feature activation | +9.73e-02 |
Pos logit contributions | |
ishing | +0.333 |
breath | +0.326 |
crown | +0.324 |
eking | +0.322 |
veil | +0.315 |
Neg logit contributions | |
| -0.452 |
She | -0.421 |
They | -0.411 |
He | -0.410 |
It | -0.402 |
Token | Amy |
Token ID | 14235 |
Feature activation | -1.67e-01 |
Pos logit contributions | |
job | +0.210 |
attention | +0.208 |
calendar | +0.203 |
way | +0.199 |
ways | +0.198 |
Neg logit contributions | |
Are | -0.067 |
He | -0.065 |
Get | -0.065 |
Would | -0.062 |
Is | -0.062 |
Token | took |
Token ID | 1718 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
| +0.358 |
She | +0.333 |
He | +0.328 |
They | +0.324 |
It | +0.320 |
Neg logit contributions | |
ishing | -0.382 |
breath | -0.375 |
eking | -0.370 |
crown | -0.370 |
luck | -0.366 |
Token | a |
Token ID | 257 |
Feature activation | +5.78e-02 |
Pos logit contributions | |
| +0.036 |
She | +0.034 |
It | +0.033 |
He | +0.032 |
They | +0.032 |
Neg logit contributions | |
ishing | -0.034 |
eking | -0.033 |
ies | -0.032 |
island | -0.032 |
mess | -0.031 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.90e-01 |
Pos logit contributions | |
ishing | +0.109 |
eking | +0.099 |
size | +0.091 |
et | +0.091 |
thieves | +0.089 |
Neg logit contributions | |
| -0.172 |
| -0.155 |
It | -0.145 |
She | -0.143 |
They | -0.138 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.225 |
ishing | +0.194 |
thieves | +0.182 |
ier | +0.172 |
ies | +0.162 |
Neg logit contributions | |
| -1.047 |
She | -0.999 |
He | -0.984 |
It | -0.981 |
They | -0.966 |
Token | and |
Token ID | 290 |
Feature activation | +5.73e-02 |
Pos logit contributions | |
ishing | +1.222 |
eking | +1.180 |
crown | +1.165 |
breath | +1.164 |
veil | +1.144 |
Neg logit contributions | |
| -1.415 |
She | -1.326 |
He | -1.287 |
They | -1.260 |
It | -1.250 |
Token | vowed |
Token ID | 19982 |
Feature activation | -4.36e-03 |
Pos logit contributions | |
ishing | +0.128 |
eking | +0.125 |
breath | +0.123 |
crown | +0.122 |
luck | +0.121 |
Neg logit contributions | |
| -0.129 |
She | -0.119 |
He | -0.115 |
They | -0.115 |
It | -0.114 |
Token | to |
Token ID | 284 |
Feature activation | +1.49e-02 |
Pos logit contributions | |
| +0.010 |
He | +0.010 |
She | +0.010 |
They | +0.010 |
It | +0.010 |
Neg logit contributions | |
luck | -0.008 |
ishing | -0.008 |
breath | -0.008 |
crown | -0.008 |
eking | -0.008 |
Token | never |
Token ID | 1239 |
Feature activation | -3.26e-02 |
Pos logit contributions | |
ishing | +0.032 |
eking | +0.029 |
ies | +0.029 |
breath | +0.029 |
mess | +0.028 |
Neg logit contributions | |
| -0.037 |
She | -0.034 |
He | -0.033 |
They | -0.033 |
It | -0.033 |
Token | give |
Token ID | 1577 |
Feature activation | +1.52e-01 |
Pos logit contributions | |
| +0.088 |
She | +0.082 |
He | +0.080 |
They | +0.078 |
It | +0.077 |
Neg logit contributions | |
ishing | -0.062 |
breath | -0.058 |
eking | -0.057 |
veil | -0.056 |
airplane | -0.056 |
Token | . |
Token ID | 13 |
Feature activation | +9.83e-02 |
Pos logit contributions | |
ishing | +0.315 |
crown | +0.304 |
breath | +0.301 |
eking | +0.301 |
veil | +0.300 |
Neg logit contributions | |
| -0.345 |
She | -0.312 |
He | -0.309 |
They | -0.299 |
It | -0.293 |
Token | She |
Token ID | 1375 |
Feature activation | -6.11e-01 |
Pos logit contributions | |
job | +0.276 |
exam | +0.271 |
attention | +0.264 |
luck | +0.263 |
earthquake | +0.262 |
Neg logit contributions | |
| -0.058 |
Would | -0.047 |
Well | -0.046 |
He | -0.041 |
They | -0.038 |
Token | took |
Token ID | 1718 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
| +1.324 |
He | +1.210 |
She | +1.198 |
It | +1.176 |
They | +1.165 |
Neg logit contributions | |
ishing | -1.407 |
eking | -1.364 |
breath | -1.357 |
luck | -1.346 |
crown | -1.343 |
Token | a |
Token ID | 257 |
Feature activation | +5.79e-02 |
Pos logit contributions | |
| +0.041 |
She | +0.039 |
He | +0.038 |
It | +0.038 |
They | +0.037 |
Neg logit contributions | |
ishing | -0.036 |
eking | -0.035 |
ies | -0.035 |
crown | -0.034 |
island | -0.034 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.106 |
eking | +0.099 |
ies | +0.090 |
size | +0.090 |
luck | +0.088 |
Neg logit contributions | |
| -0.170 |
| -0.153 |
He | -0.150 |
It | -0.148 |
She | -0.146 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.199 |
ishing | +0.171 |
ies | +0.149 |
luck | +0.130 |
veil | +0.130 |
Neg logit contributions | |
| -1.095 |
He | -1.082 |
She | -1.051 |
It | -1.031 |
They | -1.018 |
Token | and |
Token ID | 290 |
Feature activation | +5.80e-02 |
Pos logit contributions | |
ishing | +1.182 |
breath | +1.152 |
eking | +1.142 |
crown | +1.140 |
veil | +1.124 |
Neg logit contributions | |
| -1.401 |
She | -1.314 |
He | -1.294 |
They | -1.276 |
It | -1.263 |
Token | stepped |
Token ID | 10764 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
ishing | +0.125 |
breath | +0.122 |
eking | +0.121 |
crown | +0.121 |
luck | +0.119 |
Neg logit contributions | |
| -0.133 |
She | -0.124 |
He | -0.122 |
They | -0.120 |
It | -0.120 |
Token | forward |
Token ID | 2651 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
| +0.036 |
She | +0.034 |
He | +0.034 |
It | +0.033 |
They | +0.033 |
Neg logit contributions | |
ishing | -0.016 |
ies | -0.016 |
mess | -0.016 |
eking | -0.016 |
crown | -0.015 |
Token | . |
Token ID | 13 |
Feature activation | +9.83e-02 |
Pos logit contributions | |
ishing | +0.405 |
veil | +0.390 |
eking | +0.388 |
crown | +0.385 |
mess | +0.381 |
Neg logit contributions | |
| -0.459 |
She | -0.426 |
He | -0.418 |
They | -0.402 |
It | -0.399 |
Token | And |
Token ID | 843 |
Feature activation | -4.28e-01 |
Pos logit contributions | |
job | +0.226 |
attention | +0.222 |
calendar | +0.221 |
earthquake | +0.215 |
exam | +0.215 |
Neg logit contributions | |
He | -0.072 |
Well | -0.068 |
His | -0.058 |
He | -0.058 |
They | -0.057 |
Token | . |
Token ID | 13 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
eking | +0.770 |
ishing | +0.768 |
breath | +0.765 |
crown | +0.742 |
plank | +0.738 |
Neg logit contributions | |
| -0.927 |
She | -0.879 |
They | -0.847 |
He | -0.844 |
It | -0.822 |
Token | Emily |
Token ID | 17608 |
Feature activation | -1.28e-01 |
Pos logit contributions | |
attention | +0.212 |
best | +0.198 |
problem | +0.195 |
wand | +0.195 |
sheets | +0.194 |
Neg logit contributions | |
Rob | -0.116 |
His | -0.104 |
Bob | -0.101 |
He | -0.101 |
He | -0.098 |
Token | took |
Token ID | 1718 |
Feature activation | -1.36e-02 |
Pos logit contributions | |
| +0.265 |
She | +0.245 |
He | +0.240 |
They | +0.237 |
It | +0.234 |
Neg logit contributions | |
ishing | -0.305 |
breath | -0.299 |
eking | -0.296 |
crown | -0.294 |
luck | -0.293 |
Token | a |
Token ID | 257 |
Feature activation | +6.07e-02 |
Pos logit contributions | |
| +0.032 |
She | +0.029 |
It | +0.029 |
He | +0.028 |
They | +0.028 |
Neg logit contributions | |
ishing | -0.031 |
eking | -0.031 |
mess | -0.030 |
ies | -0.030 |
island | -0.029 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.94e-01 |
Pos logit contributions | |
ishing | +0.110 |
eking | +0.102 |
size | +0.096 |
thieves | +0.095 |
luck | +0.093 |
Neg logit contributions | |
| -0.182 |
| -0.163 |
It | -0.153 |
She | -0.152 |
He | -0.151 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.217 |
thieves | +0.193 |
ishing | +0.187 |
enemy | +0.172 |
ier | +0.170 |
Neg logit contributions | |
| -1.086 |
She | -1.025 |
He | -1.025 |
It | -0.996 |
They | -0.994 |
Token | and |
Token ID | 290 |
Feature activation | +6.00e-02 |
Pos logit contributions | |
ishing | +1.306 |
eking | +1.259 |
crown | +1.234 |
breath | +1.232 |
luck | +1.220 |
Neg logit contributions | |
| -1.359 |
She | -1.263 |
He | -1.220 |
They | -1.188 |
It | -1.179 |
Token | smiled |
Token ID | 13541 |
Feature activation | +4.32e-02 |
Pos logit contributions | |
ishing | +0.130 |
breath | +0.126 |
eking | +0.126 |
crown | +0.124 |
luck | +0.123 |
Neg logit contributions | |
| -0.138 |
She | -0.128 |
He | -0.126 |
They | -0.125 |
It | -0.124 |
Token | . |
Token ID | 13 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
ishing | +0.081 |
eking | +0.080 |
breath | +0.079 |
mess | +0.077 |
crown | +0.077 |
Neg logit contributions | |
| -0.114 |
She | -0.108 |
He | -0.106 |
They | -0.104 |
It | -0.102 |
Token | |
Token ID | 220 |
Feature activation | -5.47e-01 |
Pos logit contributions | |
home | +0.222 |
attention | +0.218 |
luck | +0.210 |
calendar | +0.204 |
wand | +0.202 |
Neg logit contributions | |
His | -0.073 |
Rob | -0.071 |
He | -0.065 |
Tom | -0.063 |
Bob | -0.062 |
Token |
|
Token ID | 198 |
Feature activation | -6.53e-01 |
Pos logit contributions | |
Did | +0.353 |
He | +0.326 |
Well | +0.291 |
Would | +0.280 |
Hi | +0.260 |
Neg logit contributions | |
attention | -1.535 |
home | -1.523 |
ies | -1.515 |
luck | -1.512 |
eking | -1.504 |
Token | € |
Token ID | 26391 |
Feature activation | -6.22e-02 |
Pos logit contributions | |
| +0.398 |
She | +0.380 |
They | +0.373 |
He | +0.368 |
It | +0.358 |
Neg logit contributions | |
ishing | -0.328 |
eking | -0.325 |
luck | -0.308 |
breath | -0.304 |
et | -0.302 |
Token | After |
Token ID | 2293 |
Feature activation | -4.68e-01 |
Pos logit contributions | |
She | +0.088 |
He | +0.085 |
They | +0.083 |
From | +0.080 |
| +0.078 |
Neg logit contributions | |
mess | -0.151 |
problem | -0.150 |
job | -0.149 |
ies | -0.148 |
airplane | -0.148 |
Token | taking |
Token ID | 2263 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
| +1.153 |
She | +1.069 |
He | +1.051 |
They | +1.042 |
It | +1.031 |
Neg logit contributions | |
ishing | -0.967 |
breath | -0.913 |
ies | -0.905 |
eking | -0.903 |
veil | -0.900 |
Token | a |
Token ID | 257 |
Feature activation | +5.59e-02 |
Pos logit contributions | |
ishing | +0.298 |
eking | +0.290 |
ies | +0.285 |
island | +0.282 |
veil | +0.281 |
Neg logit contributions | |
| -0.380 |
She | -0.353 |
He | -0.351 |
It | -0.349 |
They | -0.347 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.90e-01 |
Pos logit contributions | |
ishing | +0.104 |
eking | +0.094 |
ies | +0.088 |
et | +0.083 |
ts | +0.081 |
Neg logit contributions | |
| -0.167 |
| -0.152 |
He | -0.146 |
She | -0.144 |
It | -0.142 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.311 |
vell | +0.299 |
ishing | +0.280 |
ier | +0.268 |
thieves | +0.253 |
Neg logit contributions | |
| -0.965 |
He | -0.939 |
She | -0.906 |
They | -0.905 |
It | -0.888 |
Token | Billy |
Token ID | 15890 |
Feature activation | -1.70e-01 |
Pos logit contributions | |
ishing | +1.023 |
breath | +1.004 |
crown | +0.986 |
eking | +0.984 |
luck | +0.970 |
Neg logit contributions | |
| -1.529 |
She | -1.445 |
He | -1.428 |
They | -1.416 |
It | -1.401 |
Token | replied |
Token ID | 8712 |
Feature activation | -1.23e-01 |
Pos logit contributions | |
| +0.361 |
She | +0.334 |
He | +0.330 |
They | +0.324 |
It | +0.320 |
Neg logit contributions | |
ishing | -0.396 |
breath | -0.389 |
eking | -0.383 |
crown | -0.383 |
veil | -0.378 |
Token | , |
Token ID | 11 |
Feature activation | +1.34e-01 |
Pos logit contributions | |
| +0.319 |
She | +0.300 |
He | +0.297 |
It | +0.292 |
They | +0.291 |
Neg logit contributions | |
ishing | -0.234 |
breath | -0.227 |
eking | -0.224 |
ies | -0.218 |
crown | -0.218 |
Token | � |
Token ID | 6184 |
Feature activation | -8.98e-02 |
Pos logit contributions | |
crown | +0.362 |
veil | +0.355 |
luck | +0.353 |
airplane | +0.353 |
breath | +0.349 |
Neg logit contributions | |
| -0.160 |
They | -0.158 |
She | -0.150 |
He | -0.145 |
It | -0.140 |
Token | � |
Token ID | 95 |
Feature activation | -3.54e-02 |
Pos logit contributions | |
| +0.224 |
She | +0.210 |
They | +0.207 |
He | +0.205 |
It | +0.199 |
Neg logit contributions | |
ishing | -0.168 |
eking | -0.165 |
breath | -0.161 |
crown | -0.160 |
luck | -0.159 |
Token | other |
Token ID | 584 |
Feature activation | +2.02e-01 |
Pos logit contributions | |
| +0.840 |
He | +0.743 |
They | +0.739 |
She | +0.738 |
It | +0.693 |
Neg logit contributions | |
ishing | -0.751 |
eking | -0.706 |
breath | -0.673 |
grip | -0.665 |
potion | -0.654 |
Token | child |
Token ID | 1200 |
Feature activation | +9.59e-02 |
Pos logit contributions | |
breath | +0.339 |
ishing | +0.332 |
eking | +0.332 |
veil | +0.325 |
crown | +0.321 |
Neg logit contributions | |
| -0.576 |
She | -0.541 |
He | -0.541 |
They | -0.533 |
It | -0.514 |
Token | took |
Token ID | 1718 |
Feature activation | -1.94e-02 |
Pos logit contributions | |
ishing | +0.230 |
breath | +0.225 |
eking | +0.222 |
luck | +0.221 |
veil | +0.220 |
Neg logit contributions | |
| -0.199 |
She | -0.185 |
He | -0.181 |
They | -0.178 |
It | -0.176 |
Token | a |
Token ID | 257 |
Feature activation | +5.69e-02 |
Pos logit contributions | |
| +0.045 |
They | +0.041 |
It | +0.041 |
She | +0.041 |
He | +0.041 |
Neg logit contributions | |
eking | -0.045 |
ishing | -0.044 |
veil | -0.043 |
ies | -0.043 |
crown | -0.042 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
ishing | +0.096 |
eking | +0.093 |
luck | +0.083 |
ies | +0.081 |
island | +0.081 |
Neg logit contributions | |
| -0.174 |
| -0.157 |
She | -0.156 |
He | -0.154 |
It | -0.153 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.170 |
ishing | +0.140 |
airplane | +0.134 |
island | +0.131 |
ies | +0.130 |
Neg logit contributions | |
| -1.114 |
He | -1.081 |
She | -1.079 |
It | -1.055 |
They | -1.038 |
Token | and |
Token ID | 290 |
Feature activation | +5.68e-02 |
Pos logit contributions | |
ishing | +1.197 |
breath | +1.164 |
eking | +1.158 |
crown | +1.155 |
veil | +1.139 |
Neg logit contributions | |
| -1.395 |
She | -1.304 |
He | -1.286 |
They | -1.270 |
It | -1.249 |
Token | said |
Token ID | 531 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
ishing | +0.133 |
breath | +0.130 |
eking | +0.129 |
crown | +0.128 |
luck | +0.127 |
Neg logit contributions | |
| -0.121 |
She | -0.112 |
He | -0.110 |
They | -0.109 |
It | -0.108 |
Token | , |
Token ID | 11 |
Feature activation | +1.34e-01 |
Pos logit contributions | |
| +0.265 |
She | +0.251 |
He | +0.248 |
They | +0.246 |
It | +0.243 |
Neg logit contributions | |
ishing | -0.176 |
breath | -0.173 |
crown | -0.171 |
eking | -0.170 |
luck | -0.168 |
Token | " |
Token ID | 366 |
Feature activation | -3.82e-01 |
Pos logit contributions | |
eking | +0.351 |
breath | +0.346 |
crown | +0.344 |
airplane | +0.338 |
mess | +0.336 |
Neg logit contributions | |
| -0.168 |
She | -0.163 |
He | -0.156 |
They | -0.156 |
It | -0.149 |
Token | Sharing |
Token ID | 36644 |
Feature activation | -3.53e-01 |
Pos logit contributions | |
| +0.657 |
They | +0.640 |
He | +0.636 |
She | +0.620 |
From | +0.608 |
Neg logit contributions | |
ishing | -0.926 |
eking | -0.914 |
ater | -0.898 |
forth | -0.876 |
breath | -0.863 |
Token | eyes |
Token ID | 2951 |
Feature activation | +2.16e-01 |
Pos logit contributions | |
| +0.004 |
It | +0.004 |
He | +0.004 |
They | +0.004 |
She | +0.004 |
Neg logit contributions | |
ishing | -0.003 |
eking | -0.003 |
ies | -0.003 |
ier | -0.003 |
island | -0.003 |
Token | , |
Token ID | 11 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
ishing | +0.479 |
eking | +0.464 |
breath | +0.462 |
crown | +0.458 |
veil | +0.454 |
Neg logit contributions | |
| -0.498 |
She | -0.466 |
He | -0.459 |
They | -0.448 |
It | -0.443 |
Token | took |
Token ID | 1718 |
Feature activation | -1.43e-02 |
Pos logit contributions | |
ishing | +0.269 |
breath | +0.262 |
eking | +0.258 |
luck | +0.256 |
crown | +0.256 |
Neg logit contributions | |
| -0.338 |
She | -0.317 |
He | -0.314 |
They | -0.309 |
It | -0.306 |
Token | a |
Token ID | 257 |
Feature activation | +6.01e-02 |
Pos logit contributions | |
| +0.036 |
He | +0.034 |
It | +0.034 |
She | +0.033 |
They | +0.033 |
Neg logit contributions | |
ishing | -0.028 |
eking | -0.028 |
ies | -0.028 |
veil | -0.027 |
mess | -0.027 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.94e-01 |
Pos logit contributions | |
ishing | +0.112 |
eking | +0.107 |
ies | +0.097 |
size | +0.091 |
luck | +0.091 |
Neg logit contributions | |
| -0.178 |
| -0.160 |
It | -0.158 |
He | -0.157 |
She | -0.152 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.248 |
ishing | +0.218 |
ies | +0.196 |
airplane | +0.193 |
ier | +0.191 |
Neg logit contributions | |
| -1.049 |
He | -1.035 |
It | -1.012 |
She | -0.997 |
They | -0.992 |
Token | and |
Token ID | 290 |
Feature activation | +5.88e-02 |
Pos logit contributions | |
ishing | +1.237 |
eking | +1.191 |
breath | +1.190 |
crown | +1.188 |
veil | +1.172 |
Neg logit contributions | |
| -1.388 |
She | -1.290 |
He | -1.271 |
They | -1.247 |
It | -1.230 |
Token | opened |
Token ID | 4721 |
Feature activation | +5.42e-02 |
Pos logit contributions | |
ishing | +0.130 |
eking | +0.125 |
breath | +0.124 |
ies | +0.122 |
crown | +0.122 |
Neg logit contributions | |
| -0.135 |
She | -0.124 |
He | -0.124 |
They | -0.122 |
It | -0.121 |
Token | the |
Token ID | 262 |
Feature activation | +3.75e-02 |
Pos logit contributions | |
ishing | +0.116 |
eking | +0.113 |
breath | +0.113 |
crown | +0.112 |
veil | +0.111 |
Neg logit contributions | |
| -0.125 |
She | -0.117 |
He | -0.116 |
They | -0.114 |
It | -0.113 |
Token | gate |
Token ID | 8946 |
Feature activation | +3.50e-01 |
Pos logit contributions | |
ishing | +0.070 |
eking | +0.065 |
ies | +0.064 |
ie | +0.057 |
cking | +0.055 |
Neg logit contributions | |
| -0.106 |
It | -0.103 |
She | -0.098 |
They | -0.098 |
He | -0.097 |
Token | . |
Token ID | 13 |
Feature activation | +9.85e-02 |
Pos logit contributions | |
ishing | +0.651 |
breath | +0.636 |
crown | +0.624 |
eking | +0.624 |
veil | +0.617 |
Neg logit contributions | |
| -0.903 |
She | -0.852 |
He | -0.841 |
They | -0.832 |
It | -0.823 |
Token | feathers |
Token ID | 27400 |
Feature activation | +4.14e-01 |
Pos logit contributions | |
eking | +0.705 |
ieve | +0.672 |
exam | +0.655 |
ault | +0.649 |
vell | +0.634 |
Neg logit contributions | |
| -1.132 |
They | -1.115 |
He | -1.073 |
She | -1.030 |
It | -1.018 |
Token | and |
Token ID | 290 |
Feature activation | +5.50e-02 |
Pos logit contributions | |
ishing | +0.936 |
breath | +0.908 |
crown | +0.891 |
plank | +0.891 |
eking | +0.888 |
Neg logit contributions | |
| -0.951 |
She | -0.902 |
He | -0.859 |
They | -0.840 |
It | -0.830 |
Token | took |
Token ID | 1718 |
Feature activation | -2.09e-02 |
Pos logit contributions | |
ishing | +0.122 |
luck | +0.117 |
eking | +0.115 |
problem | +0.114 |
decision | +0.113 |
Neg logit contributions | |
| -0.125 |
She | -0.116 |
They | -0.113 |
He | -0.113 |
It | -0.111 |
Token | a |
Token ID | 257 |
Feature activation | +5.57e-02 |
Pos logit contributions | |
| +0.050 |
She | +0.046 |
It | +0.046 |
He | +0.044 |
They | +0.044 |
Neg logit contributions | |
ishing | -0.046 |
eking | -0.045 |
ies | -0.044 |
mess | -0.044 |
island | -0.043 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.108 |
eking | +0.098 |
luck | +0.091 |
et | +0.090 |
size | +0.090 |
Neg logit contributions | |
| -0.164 |
| -0.144 |
It | -0.137 |
She | -0.135 |
He | -0.135 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.243 |
ishing | +0.213 |
thieves | +0.204 |
ies | +0.187 |
attention | +0.183 |
Neg logit contributions | |
| -1.073 |
He | -1.008 |
She | -1.000 |
It | -0.985 |
They | -0.983 |
Token | of |
Token ID | 286 |
Feature activation | +2.56e-02 |
Pos logit contributions | |
ishing | +1.217 |
eking | +1.163 |
crown | +1.161 |
breath | +1.158 |
luck | +1.145 |
Neg logit contributions | |
| -1.440 |
She | -1.330 |
He | -1.301 |
They | -1.277 |
It | -1.256 |
Token | fresh |
Token ID | 4713 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
ishing | +0.045 |
eking | +0.043 |
ies | +0.041 |
ater | +0.039 |
plank | +0.038 |
Neg logit contributions | |
| -0.074 |
It | -0.068 |
She | -0.068 |
He | -0.067 |
They | -0.066 |
Token | air |
Token ID | 1633 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
eking | +0.580 |
exam | +0.530 |
mistake | +0.526 |
ishing | +0.525 |
airplane | +0.520 |
Neg logit contributions | |
| -0.827 |
They | -0.762 |
She | -0.761 |
He | -0.756 |
It | -0.747 |
Token | . |
Token ID | 13 |
Feature activation | +9.40e-02 |
Pos logit contributions | |
ishing | +0.570 |
problem | +0.553 |
decision | +0.551 |
exam | +0.550 |
luck | +0.549 |
Neg logit contributions | |
| -0.492 |
She | -0.445 |
They | -0.431 |
He | -0.415 |
It | -0.391 |
Token | |
Token ID | 220 |
Feature activation | -5.50e-01 |
Pos logit contributions | |
home | +0.207 |
way | +0.180 |
job | +0.179 |
honey | +0.178 |
cob | +0.174 |
Neg logit contributions | |
Did | -0.081 |
Well | -0.080 |
He | -0.079 |
Is | -0.076 |
His | -0.074 |
Token | , |
Token ID | 11 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
ishing | +0.494 |
eking | +0.475 |
veil | +0.475 |
breath | +0.473 |
crown | +0.467 |
Neg logit contributions | |
| -0.575 |
She | -0.536 |
He | -0.534 |
They | -0.526 |
It | -0.513 |
Token | it |
Token ID | 340 |
Feature activation | -5.43e-02 |
Pos logit contributions | |
ishing | +0.282 |
breath | +0.278 |
crown | +0.274 |
eking | +0.273 |
luck | +0.270 |
Neg logit contributions | |
| -0.297 |
She | -0.278 |
He | -0.274 |
They | -0.271 |
It | -0.268 |
Token | blew |
Token ID | 17948 |
Feature activation | +1.79e-01 |
Pos logit contributions | |
| +0.112 |
She | +0.102 |
He | +0.101 |
They | +0.099 |
It | +0.098 |
Neg logit contributions | |
ishing | -0.131 |
breath | -0.128 |
eking | -0.127 |
crown | -0.126 |
luck | -0.125 |
Token | a |
Token ID | 257 |
Feature activation | +5.57e-02 |
Pos logit contributions | |
eking | +0.332 |
ishing | +0.331 |
veil | +0.318 |
crown | +0.313 |
ies | +0.312 |
Neg logit contributions | |
| -0.483 |
He | -0.452 |
She | -0.451 |
They | -0.447 |
It | -0.444 |
Token | big |
Token ID | 1263 |
Feature activation | +2.80e-01 |
Pos logit contributions | |
ishing | +0.103 |
eking | +0.091 |
ies | +0.087 |
thieves | +0.085 |
luck | +0.081 |
Neg logit contributions | |
| -0.167 |
| -0.150 |
It | -0.146 |
He | -0.144 |
She | -0.140 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.562 |
ishing | +0.557 |
ies | +0.524 |
thieves | +0.481 |
ils | +0.479 |
Neg logit contributions | |
| -0.740 |
It | -0.699 |
He | -0.684 |
| -0.673 |
She | -0.661 |
Token | of |
Token ID | 286 |
Feature activation | +2.59e-02 |
Pos logit contributions | |
ishing | +1.206 |
eking | +1.154 |
breath | +1.149 |
crown | +1.148 |
veil | +1.134 |
Neg logit contributions | |
| -1.435 |
She | -1.329 |
He | -1.309 |
They | -1.284 |
It | -1.281 |
Token | air |
Token ID | 1633 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
ishing | +0.053 |
eking | +0.050 |
ies | +0.050 |
ater | +0.046 |
rying | +0.045 |
Neg logit contributions | |
| -0.068 |
It | -0.063 |
He | -0.062 |
She | -0.060 |
They | -0.060 |
Token | out |
Token ID | 503 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
ishing | +0.488 |
eking | +0.460 |
luck | +0.454 |
plank | +0.454 |
crown | +0.451 |
Neg logit contributions | |
| -0.582 |
She | -0.536 |
He | -0.521 |
They | -0.517 |
It | -0.508 |
Token | , |
Token ID | 11 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
ishing | +0.605 |
eking | +0.583 |
veil | +0.579 |
crown | +0.574 |
ies | +0.566 |
Neg logit contributions | |
| -0.588 |
He | -0.546 |
She | -0.545 |
They | -0.527 |
It | -0.522 |
Token | and |
Token ID | 290 |
Feature activation | +5.41e-02 |
Pos logit contributions | |
ishing | +0.251 |
breath | +0.244 |
crown | +0.240 |
eking | +0.240 |
luck | +0.238 |
Neg logit contributions | |
| -0.335 |
She | -0.314 |
He | -0.310 |
They | -0.308 |
It | -0.305 |
Token | The |
Token ID | 464 |
Feature activation | -3.44e-01 |
Pos logit contributions | |
She | +1.064 |
They | +1.048 |
It | +1.034 |
He | +0.969 |
| +0.949 |
Neg logit contributions | |
plank | -1.620 |
airplane | -1.614 |
crown | -1.609 |
island | -1.562 |
breath | -1.549 |
Token | soldier |
Token ID | 10686 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
| +0.946 |
He | +0.891 |
They | +0.847 |
It | +0.840 |
She | +0.840 |
Neg logit contributions | |
eking | -0.641 |
ishing | -0.641 |
breath | -0.598 |
ies | -0.597 |
grip | -0.585 |
Token | took |
Token ID | 1718 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
ishing | +0.472 |
breath | +0.466 |
crown | +0.459 |
eking | +0.458 |
luck | +0.453 |
Neg logit contributions | |
| -0.471 |
She | -0.441 |
He | -0.434 |
They | -0.430 |
It | -0.424 |
Token | a |
Token ID | 257 |
Feature activation | +6.02e-02 |
Pos logit contributions | |
| +0.038 |
He | +0.035 |
She | +0.035 |
It | +0.035 |
They | +0.035 |
Neg logit contributions | |
ishing | -0.031 |
eking | -0.030 |
ies | -0.029 |
veil | -0.029 |
mess | -0.029 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.112 |
eking | +0.103 |
ies | +0.094 |
luck | +0.093 |
size | +0.091 |
Neg logit contributions | |
| -0.177 |
| -0.162 |
He | -0.155 |
She | -0.154 |
It | -0.154 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.190 |
ishing | +0.166 |
ies | +0.149 |
luck | +0.125 |
crown | +0.124 |
Neg logit contributions | |
| -1.108 |
He | -1.092 |
She | -1.066 |
They | -1.041 |
It | -1.041 |
Token | and |
Token ID | 290 |
Feature activation | +5.96e-02 |
Pos logit contributions | |
ishing | +1.146 |
breath | +1.121 |
crown | +1.106 |
eking | +1.106 |
luck | +1.090 |
Neg logit contributions | |
| -1.421 |
She | -1.335 |
He | -1.318 |
They | -1.303 |
It | -1.288 |
Token | said |
Token ID | 531 |
Feature activation | -9.56e-02 |
Pos logit contributions | |
ishing | +0.129 |
breath | +0.126 |
eking | +0.125 |
crown | +0.125 |
luck | +0.123 |
Neg logit contributions | |
| -0.136 |
She | -0.126 |
He | -0.125 |
They | -0.123 |
It | -0.122 |
Token | , |
Token ID | 11 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
| +0.254 |
She | +0.241 |
He | +0.238 |
They | +0.236 |
It | +0.233 |
Neg logit contributions | |
ishing | -0.168 |
breath | -0.165 |
crown | -0.162 |
eking | -0.162 |
luck | -0.160 |
Token | " |
Token ID | 366 |
Feature activation | -3.80e-01 |
Pos logit contributions | |
eking | +0.357 |
breath | +0.357 |
crown | +0.353 |
airplane | +0.345 |
veil | +0.342 |
Neg logit contributions | |
| -0.167 |
She | -0.167 |
They | -0.163 |
He | -0.155 |
It | -0.152 |
Token | Yes |
Token ID | 5297 |
Feature activation | -2.64e-01 |
Pos logit contributions | |
They | +0.652 |
| +0.637 |
She | +0.621 |
He | +0.614 |
It | +0.603 |
Neg logit contributions | |
luck | -0.912 |
ishing | -0.902 |
breath | -0.881 |
veil | -0.879 |
plank | -0.874 |
Token | . |
Token ID | 13 |
Feature activation | +9.07e-02 |
Pos logit contributions | |
ishing | +0.699 |
breath | +0.689 |
eking | +0.676 |
mess | +0.663 |
veil | +0.652 |
Neg logit contributions | |
| -0.831 |
He | -0.774 |
She | -0.767 |
They | -0.754 |
It | -0.748 |
Token | He |
Token ID | 679 |
Feature activation | -6.07e-01 |
Pos logit contributions | |
home | +0.197 |
way | +0.192 |
place | +0.186 |
adventure | +0.185 |
spot | +0.184 |
Neg logit contributions | |
Don | -0.091 |
Well | -0.090 |
Did | -0.083 |
Since | -0.081 |
From | -0.077 |
Token | took |
Token ID | 1718 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
| +1.342 |
He | +1.206 |
They | +1.183 |
It | +1.150 |
She | +1.119 |
Neg logit contributions | |
ishing | -1.367 |
attention | -1.358 |
eking | -1.355 |
luck | -1.314 |
veil | -1.301 |
Token | a |
Token ID | 257 |
Feature activation | +5.50e-02 |
Pos logit contributions | |
| +0.043 |
He | +0.040 |
It | +0.039 |
She | +0.037 |
They | +0.037 |
Neg logit contributions | |
eking | -0.049 |
ishing | -0.049 |
ies | -0.047 |
mess | -0.046 |
veil | -0.046 |
Token | deep |
Token ID | 2769 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
ishing | +0.111 |
eking | +0.103 |
ier | +0.094 |
luck | +0.093 |
et | +0.091 |
Neg logit contributions | |
| -0.155 |
| -0.138 |
He | -0.133 |
It | -0.130 |
They | -0.125 |
Token | breath |
Token ID | 8033 |
Feature activation | +5.80e-01 |
Pos logit contributions | |
eking | +0.257 |
ier | +0.233 |
exam | +0.222 |
permission | +0.209 |
ishing | +0.201 |
Neg logit contributions | |
| -1.021 |
He | -1.015 |
They | -0.962 |
It | -0.954 |
She | -0.936 |
Token | and |
Token ID | 290 |
Feature activation | +5.32e-02 |
Pos logit contributions | |
ishing | +1.318 |
eking | +1.284 |
crown | +1.264 |
breath | +1.259 |
veil | +1.256 |
Neg logit contributions | |
| -1.331 |
She | -1.225 |
He | -1.211 |
They | -1.177 |
It | -1.160 |
Token | jumped |
Token ID | 11687 |
Feature activation | +1.13e-01 |
Pos logit contributions | |
ishing | +0.118 |
eking | +0.115 |
breath | +0.113 |
luck | +0.112 |
mess | +0.112 |
Neg logit contributions | |
| -0.121 |
He | -0.111 |
She | -0.109 |
They | -0.109 |
It | -0.108 |
Token | in |
Token ID | 287 |
Feature activation | +2.51e-01 |
Pos logit contributions | |
ishing | +0.208 |
eking | +0.196 |
veil | +0.191 |
ies | +0.191 |
airplane | +0.185 |
Neg logit contributions | |
| -0.319 |
He | -0.301 |
She | -0.294 |
They | -0.291 |
It | -0.288 |
Token | . |
Token ID | 13 |
Feature activation | +9.11e-02 |
Pos logit contributions | |
ishing | +0.578 |
eking | +0.568 |
ies | +0.553 |
veil | +0.530 |
crown | +0.527 |
Neg logit contributions | |
| -0.601 |
He | -0.543 |
She | -0.538 |
They | -0.527 |
It | -0.526 |
Token |
|
Token ID | 198 |
Feature activation | -6.60e-01 |
Pos logit contributions | |
sand | +0.196 |
home | +0.195 |
balance | +0.192 |
way | +0.191 |
job | +0.189 |
Neg logit contributions | |
Don | -0.085 |
Well | -0.083 |
Did | -0.075 |
Kids | -0.073 |
She | -0.072 |
Token | . |
Token ID | 13 |
Feature activation | +2.77e-02 |
Pos logit contributions | |
ishing | +0.094 |
eking | +0.088 |
veil | +0.086 |
airplane | +0.085 |
breath | +0.085 |
Neg logit contributions | |
| -0.093 |
He | -0.087 |
She | -0.086 |
They | -0.086 |
It | -0.083 |
Token | Tim |
Token ID | 5045 |
Feature activation | -3.01e-01 |
Pos logit contributions | |
airplane | +0.075 |
crown | +0.073 |
mine | +0.071 |
castle | +0.070 |
veil | +0.070 |
Neg logit contributions | |
Well | -0.011 |
Even | -0.009 |
| -0.009 |
Even | -0.007 |
Don | -0.007 |
Token | cries |
Token ID | 24691 |
Feature activation | +7.27e-02 |
Pos logit contributions | |
| +0.676 |
She | +0.634 |
He | +0.624 |
They | +0.620 |
It | +0.614 |
Neg logit contributions | |
ishing | -0.648 |
breath | -0.635 |
crown | -0.626 |
eking | -0.624 |
ies | -0.619 |
Token | . |
Token ID | 13 |
Feature activation | +5.68e-02 |
Pos logit contributions | |
ishing | +0.143 |
breath | +0.139 |
eking | +0.138 |
crown | +0.137 |
veil | +0.136 |
Neg logit contributions | |
| -0.182 |
She | -0.172 |
He | -0.170 |
They | -0.168 |
It | -0.166 |
Token |
|
Token ID | 198 |
Feature activation | -6.92e-01 |
Pos logit contributions | |
breath | +0.126 |
falling | +0.125 |
phone | +0.121 |
crown | +0.121 |
earthquake | +0.117 |
Neg logit contributions | |
Well | -0.060 |
Would | -0.059 |
Since | -0.046 |
Even | -0.045 |
Here | -0.045 |
Token |
|
Token ID | 198 |
Feature activation | -6.90e-01 |
Pos logit contributions | |
Did | +0.761 |
She | +0.555 |
Well | +0.478 |
It | +0.455 |
There | +0.444 |
Neg logit contributions | |
mistake | -1.822 |
breath | -1.810 |
ep | -1.809 |
- | -1.780 |
dangerous | -1.762 |
Token | M |
Token ID | 44 |
Feature activation | -3.59e-01 |
Pos logit contributions | |
Did | +1.270 |
It | +1.256 |
She | +1.242 |
He | +1.185 |
They | +1.178 |
Neg logit contributions | |
breath | -1.445 |
ep | -1.404 |
eking | -1.395 |
eping | -1.346 |
ishing | -1.344 |
Token | ia |
Token ID | 544 |
Feature activation | +2.04e-02 |
Pos logit contributions | |
| +0.858 |
She | +0.803 |
He | +0.801 |
They | +0.776 |
It | +0.761 |
Neg logit contributions | |
ishing | -0.743 |
breath | -0.731 |
veil | -0.711 |
crown | -0.706 |
eking | -0.702 |
Token | looks |
Token ID | 3073 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
ishing | +0.042 |
breath | +0.041 |
eking | +0.041 |
crown | +0.040 |
luck | +0.040 |
Neg logit contributions | |
| -0.048 |
She | -0.046 |
He | -0.045 |
They | -0.044 |
It | -0.044 |
Token | at |
Token ID | 379 |
Feature activation | +1.95e-01 |
Pos logit contributions | |
ishing | +0.136 |
plank | +0.132 |
eking | +0.132 |
island | +0.128 |
breath | +0.128 |
Neg logit contributions | |
| -0.250 |
She | -0.235 |
He | -0.232 |
They | -0.228 |
It | -0.227 |
Token | her |
Token ID | 607 |
Feature activation | +5.33e-03 |
Pos logit contributions | |
breath | +0.455 |
ishing | +0.452 |
crown | +0.441 |
eking | +0.438 |
veil | +0.436 |
Neg logit contributions | |
| -0.402 |
She | -0.377 |
He | -0.373 |
They | -0.369 |
It | -0.364 |
Token | away |
Token ID | 1497 |
Feature activation | +3.78e-01 |
Pos logit contributions | |
| +0.386 |
She | +0.356 |
He | +0.351 |
They | +0.341 |
It | +0.328 |
Neg logit contributions | |
ishing | -0.386 |
breath | -0.369 |
eking | -0.365 |
plank | -0.364 |
veil | -0.362 |
Token | from |
Token ID | 422 |
Feature activation | +1.12e-01 |
Pos logit contributions | |
ishing | +0.745 |
eking | +0.705 |
breath | +0.698 |
crown | +0.695 |
veil | +0.693 |
Neg logit contributions | |
| -0.966 |
She | -0.901 |
He | -0.892 |
They | -0.881 |
It | -0.853 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.88e-01 |
Pos logit contributions | |
ishing | +0.252 |
ies | +0.239 |
eking | +0.237 |
veil | +0.235 |
crown | +0.234 |
Neg logit contributions | |
| -0.254 |
She | -0.235 |
He | -0.232 |
They | -0.226 |
It | -0.222 |
Token | . |
Token ID | 13 |
Feature activation | +5.86e-02 |
Pos logit contributions | |
| +0.718 |
She | +0.668 |
He | +0.660 |
They | +0.651 |
It | +0.636 |
Neg logit contributions | |
ishing | -0.571 |
breath | -0.557 |
crown | -0.551 |
eking | -0.548 |
veil | -0.543 |
Token |
|
Token ID | 198 |
Feature activation | -6.92e-01 |
Pos logit contributions | |
mine | +0.130 |
crown | +0.128 |
mess | +0.127 |
mistake | +0.123 |
hiding | +0.123 |
Neg logit contributions | |
Well | -0.034 |
Would | -0.030 |
Even | -0.027 |
Don | -0.026 |
That | -0.023 |
Token |
|
Token ID | 198 |
Feature activation | -6.90e-01 |
Pos logit contributions | |
Did | +0.844 |
There | +0.552 |
She | +0.549 |
All | +0.535 |
It | +0.531 |
Neg logit contributions | |
mistake | -1.764 |
mess | -1.750 |
ep | -1.721 |
home | -1.675 |
dangerous | -1.658 |
Token | " |
Token ID | 1 |
Feature activation | -2.63e-01 |
Pos logit contributions | |
Did | +1.370 |
It | +1.363 |
She | +1.280 |
There | +1.269 |
He | +1.264 |
Neg logit contributions | |
ep | -1.224 |
plank | -1.182 |
ater | -1.178 |
cream | -1.176 |
mess | -1.133 |
Token | Tom |
Token ID | 13787 |
Feature activation | -3.15e-01 |
Pos logit contributions | |
She | +0.478 |
From | +0.466 |
They | +0.462 |
It | +0.461 |
| +0.461 |
Neg logit contributions | |
cream | -0.606 |
veil | -0.555 |
plank | -0.549 |
ishing | -0.548 |
crown | -0.546 |
Token | , |
Token ID | 11 |
Feature activation | +1.09e-01 |
Pos logit contributions | |
| +0.822 |
She | +0.767 |
He | +0.759 |
They | +0.746 |
It | +0.742 |
Neg logit contributions | |
ishing | -0.592 |
breath | -0.579 |
crown | -0.569 |
veil | -0.569 |
eking | -0.569 |
Token | go |
Token ID | 467 |
Feature activation | +1.47e-01 |
Pos logit contributions | |
ishing | +0.259 |
breath | +0.252 |
eking | +0.250 |
crown | +0.247 |
luck | +0.246 |
Neg logit contributions | |
| -0.229 |
She | -0.211 |
He | -0.209 |
They | -0.206 |
It | -0.205 |
Token | away |
Token ID | 1497 |
Feature activation | +4.04e-01 |
Pos logit contributions | |
ishing | +0.185 |
veil | +0.175 |
eking | +0.172 |
ies | +0.170 |
crown | +0.166 |
Neg logit contributions | |
| -0.495 |
He | -0.459 |
She | -0.456 |
They | -0.445 |
It | -0.443 |
Token | starting |
Token ID | 3599 |
Feature activation | +2.68e-01 |
Pos logit contributions | |
| +0.008 |
He | +0.007 |
She | +0.007 |
They | +0.007 |
| +0.007 |
Neg logit contributions | |
eking | -0.006 |
ishing | -0.006 |
ier | -0.006 |
ts | -0.005 |
plank | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
ishing | +0.567 |
crown | +0.545 |
eking | +0.545 |
veil | +0.545 |
breath | +0.535 |
Neg logit contributions | |
| -0.644 |
She | -0.595 |
He | -0.586 |
They | -0.580 |
It | -0.560 |
Token | feel |
Token ID | 1254 |
Feature activation | +6.08e-02 |
Pos logit contributions | |
| +0.044 |
She | +0.042 |
He | +0.041 |
They | +0.041 |
It | +0.041 |
Neg logit contributions | |
ishing | -0.029 |
eking | -0.027 |
ies | -0.026 |
crown | -0.026 |
luck | -0.026 |
Token | sad |
Token ID | 6507 |
Feature activation | +2.30e-01 |
Pos logit contributions | |
ishing | +0.103 |
eking | +0.098 |
ies | +0.097 |
breath | +0.096 |
crown | +0.096 |
Neg logit contributions | |
| -0.173 |
She | -0.163 |
He | -0.161 |
It | -0.159 |
They | -0.159 |
Token | . |
Token ID | 13 |
Feature activation | +5.72e-02 |
Pos logit contributions | |
ishing | +0.494 |
eking | +0.490 |
breath | +0.483 |
crown | +0.480 |
veil | +0.475 |
Neg logit contributions | |
| -0.548 |
She | -0.519 |
He | -0.511 |
They | -0.507 |
It | -0.483 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
mess | +0.134 |
home | +0.134 |
mine | +0.133 |
problem | +0.133 |
airplane | +0.132 |
Neg logit contributions | |
Well | -0.024 |
Even | -0.020 |
That | -0.018 |
Did | -0.018 |
Don | -0.017 |
Token |
|
Token ID | 198 |
Feature activation | -6.87e-01 |
Pos logit contributions | |
Did | +0.774 |
They | +0.670 |
She | +0.647 |
It | +0.629 |
He | +0.604 |
Neg logit contributions | |
mistake | -1.742 |
mess | -1.666 |
home | -1.662 |
breath | -1.655 |
adventure | -1.652 |
Token | Suddenly |
Token ID | 38582 |
Feature activation | -2.81e-01 |
Pos logit contributions | |
It | +1.209 |
They | +1.201 |
He | +1.195 |
She | +1.161 |
Did | +1.159 |
Neg logit contributions | |
ater | -1.306 |
eking | -1.301 |
game | -1.292 |
veil | -1.271 |
breath | -1.271 |
Token | , |
Token ID | 11 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
| +0.707 |
She | +0.669 |
He | +0.659 |
They | +0.646 |
It | +0.644 |
Neg logit contributions | |
ishing | -0.566 |
eking | -0.543 |
breath | -0.529 |
crown | -0.517 |
mess | -0.515 |
Token | the |
Token ID | 262 |
Feature activation | +1.06e-02 |
Pos logit contributions | |
ishing | +0.259 |
breath | +0.251 |
eking | +0.251 |
crown | +0.247 |
luck | +0.245 |
Neg logit contributions | |
| -0.221 |
She | -0.206 |
He | -0.202 |
They | -0.199 |
It | -0.196 |
Token | bald |
Token ID | 28273 |
Feature activation | +2.83e-01 |
Pos logit contributions | |
ishing | +0.019 |
eking | +0.018 |
ies | +0.016 |
breath | +0.016 |
cking | +0.015 |
Neg logit contributions | |
| -0.029 |
He | -0.028 |
She | -0.027 |
It | -0.027 |
They | -0.026 |
Token | too |
Token ID | 1165 |
Feature activation | +1.17e-01 |
Pos logit contributions | |
Did | +0.723 |
Would | +0.511 |
She | +0.501 |
Well | +0.496 |
He | +0.465 |
Neg logit contributions | |
mistake | -1.936 |
mess | -1.900 |
breath | -1.892 |
- | -1.870 |
ies | -1.826 |
Token | high |
Token ID | 1029 |
Feature activation | +2.54e-01 |
Pos logit contributions | |
ishing | +0.231 |
eking | +0.213 |
crown | +0.208 |
airplane | +0.205 |
veil | +0.204 |
Neg logit contributions | |
| -0.320 |
He | -0.292 |
She | -0.288 |
They | -0.286 |
It | -0.275 |
Token | for |
Token ID | 329 |
Feature activation | +3.05e-01 |
Pos logit contributions | |
ishing | +0.516 |
veil | +0.514 |
crown | +0.512 |
eking | +0.506 |
breath | +0.491 |
Neg logit contributions | |
| -0.625 |
He | -0.581 |
She | -0.573 |
They | -0.562 |
It | -0.539 |
Token | them |
Token ID | 606 |
Feature activation | -4.63e-02 |
Pos logit contributions | |
ishing | +0.720 |
ies | +0.681 |
crown | +0.676 |
eking | +0.675 |
veil | +0.675 |
Neg logit contributions | |
| -0.671 |
She | -0.609 |
He | -0.606 |
They | -0.597 |
It | -0.586 |
Token | . |
Token ID | 13 |
Feature activation | +5.43e-02 |
Pos logit contributions | |
| +0.112 |
He | +0.100 |
She | +0.100 |
They | +0.100 |
It | +0.094 |
Neg logit contributions | |
ishing | -0.099 |
eking | -0.094 |
veil | -0.093 |
breath | -0.093 |
crown | -0.093 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
mine | +0.129 |
crown | +0.125 |
airplane | +0.125 |
egg | +0.123 |
castle | +0.121 |
Neg logit contributions | |
Well | -0.026 |
Even | -0.022 |
Don | -0.021 |
That | -0.018 |
Did | -0.015 |
Token |
|
Token ID | 198 |
Feature activation | -6.86e-01 |
Pos logit contributions | |
Did | +0.816 |
She | +0.637 |
It | +0.603 |
There | +0.598 |
He | +0.555 |
Neg logit contributions | |
mistake | -1.733 |
mess | -1.700 |
home | -1.661 |
breath | -1.651 |
adventure | -1.612 |
Token | " |
Token ID | 1 |
Feature activation | -2.58e-01 |
Pos logit contributions | |
It | +1.256 |
He | +1.208 |
Did | +1.186 |
She | +1.186 |
They | +1.166 |
Neg logit contributions | |
plank | -1.349 |
ater | -1.332 |
ep | -1.321 |
breath | -1.314 |
eking | -1.312 |
Token | Let |
Token ID | 5756 |
Feature activation | -4.28e-01 |
Pos logit contributions | |
She | +0.443 |
| +0.435 |
He | +0.431 |
It | +0.421 |
They | +0.418 |
Neg logit contributions | |
cream | -0.618 |
veil | -0.614 |
plank | -0.611 |
ishing | -0.610 |
breath | -0.608 |
Token | 's |
Token ID | 338 |
Feature activation | -6.11e-02 |
Pos logit contributions | |
| +0.907 |
She | +0.830 |
He | +0.814 |
They | +0.806 |
It | +0.805 |
Neg logit contributions | |
ishing | -1.029 |
breath | -0.998 |
eking | -0.995 |
crown | -0.986 |
veil | -0.969 |
Token | get |
Token ID | 651 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
| +0.160 |
They | +0.145 |
She | +0.144 |
It | +0.144 |
He | +0.144 |
Neg logit contributions | |
ishing | -0.125 |
eking | -0.118 |
ies | -0.113 |
luck | -0.108 |
veil | -0.108 |
Token | Lily |
Token ID | 20037 |
Feature activation | -2.54e-01 |
Pos logit contributions | |
ishing | +0.102 |
breath | +0.095 |
luck | +0.095 |
veil | +0.094 |
crown | +0.094 |
Neg logit contributions | |
| -0.102 |
She | -0.091 |
He | -0.088 |
They | -0.087 |
It | -0.084 |
Token | and |
Token ID | 290 |
Feature activation | +3.12e-02 |
Pos logit contributions | |
| +0.582 |
She | +0.546 |
He | +0.540 |
They | +0.533 |
It | +0.528 |
Neg logit contributions | |
ishing | -0.535 |
breath | -0.528 |
crown | -0.521 |
eking | -0.518 |
luck | -0.513 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.83e-01 |
Pos logit contributions | |
ishing | +0.083 |
breath | +0.081 |
eking | +0.080 |
crown | +0.079 |
luck | +0.079 |
Neg logit contributions | |
| -0.057 |
She | -0.052 |
He | -0.051 |
They | -0.051 |
It | -0.049 |
Token | ask |
Token ID | 1265 |
Feature activation | -5.54e-02 |
Pos logit contributions | |
| +0.651 |
She | +0.609 |
He | +0.599 |
They | +0.595 |
It | +0.584 |
Neg logit contributions | |
ishing | -0.605 |
breath | -0.593 |
eking | -0.584 |
crown | -0.583 |
veil | -0.576 |
Token | . |
Token ID | 13 |
Feature activation | +5.60e-02 |
Pos logit contributions | |
| +0.131 |
She | +0.121 |
He | +0.119 |
They | +0.118 |
It | +0.115 |
Neg logit contributions | |
ishing | -0.119 |
breath | -0.115 |
eking | -0.115 |
crown | -0.113 |
veil | -0.113 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
crown | +0.117 |
hiding | +0.117 |
old | +0.116 |
ring | +0.115 |
mistake | +0.114 |
Neg logit contributions | |
Well | -0.049 |
Even | -0.038 |
Even | -0.033 |
Don | -0.032 |
There | -0.032 |
Token |
|
Token ID | 198 |
Feature activation | -6.87e-01 |
Pos logit contributions | |
Did | +0.827 |
Well | +0.573 |
He | +0.503 |
Finally | +0.497 |
There | +0.497 |
Neg logit contributions | |
mistake | -1.754 |
home | -1.744 |
- | -1.743 |
ep | -1.727 |
mess | -1.723 |
Token | " |
Token ID | 1 |
Feature activation | -2.61e-01 |
Pos logit contributions | |
It | +1.457 |
He | +1.412 |
Did | +1.385 |
She | +1.347 |
Even | +1.324 |
Neg logit contributions | |
breath | -1.236 |
plank | -1.230 |
ep | -1.227 |
cream | -1.214 |
ater | -1.211 |
Token | Maybe |
Token ID | 13300 |
Feature activation | -3.97e-01 |
Pos logit contributions | |
He | +0.442 |
She | +0.438 |
From | +0.434 |
It | +0.429 |
| +0.427 |
Neg logit contributions | |
cream | -0.645 |
veil | -0.619 |
ishing | -0.607 |
crown | -0.597 |
breath | -0.594 |
Token | , |
Token ID | 11 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
| +0.838 |
She | +0.756 |
He | +0.745 |
They | +0.744 |
It | +0.733 |
Neg logit contributions | |
ishing | -0.966 |
breath | -0.941 |
eking | -0.930 |
crown | -0.914 |
veil | -0.906 |
Token | if |
Token ID | 611 |
Feature activation | -3.29e-02 |
Pos logit contributions | |
ishing | +0.290 |
breath | +0.281 |
eking | +0.279 |
crown | +0.270 |
ies | +0.270 |
Neg logit contributions | |
| -0.228 |
She | -0.210 |
He | -0.205 |
They | -0.204 |
It | -0.203 |
Token | soft |
Token ID | 2705 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
| +0.006 |
She | +0.006 |
He | +0.006 |
They | +0.006 |
It | +0.005 |
Neg logit contributions | |
ishing | -0.005 |
eking | -0.005 |
breath | -0.005 |
luck | -0.005 |
crown | -0.005 |
Token | ," |
Token ID | 553 |
Feature activation | +5.01e-02 |
Pos logit contributions | |
eking | +0.460 |
ishing | +0.451 |
plank | +0.431 |
crown | +0.420 |
airplane | +0.419 |
Neg logit contributions | |
| -0.583 |
She | -0.549 |
They | -0.546 |
He | -0.536 |
It | -0.525 |
Token | replied |
Token ID | 8712 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
crown | +0.108 |
breath | +0.106 |
ies | +0.105 |
veil | +0.105 |
egg | +0.099 |
Neg logit contributions | |
It | -0.087 |
| -0.084 |
Even | -0.084 |
They | -0.083 |
From | -0.082 |
Token | Lily |
Token ID | 20037 |
Feature activation | -2.32e-01 |
Pos logit contributions | |
| +0.340 |
She | +0.327 |
He | +0.323 |
They | +0.322 |
It | +0.314 |
Neg logit contributions | |
breath | -0.326 |
crown | -0.321 |
ishing | -0.321 |
veil | -0.315 |
luck | -0.313 |
Token | . |
Token ID | 13 |
Feature activation | +6.08e-02 |
Pos logit contributions | |
| +0.615 |
She | +0.587 |
He | +0.583 |
It | +0.582 |
They | +0.582 |
Neg logit contributions | |
ishing | -0.390 |
ies | -0.384 |
breath | -0.382 |
crown | -0.381 |
luck | -0.376 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
home | +0.131 |
crown | +0.131 |
drawing | +0.127 |
breath | +0.126 |
hiding | +0.125 |
Neg logit contributions | |
Well | -0.041 |
Did | -0.036 |
Don | -0.029 |
Where | -0.028 |
He | -0.027 |
Token |
|
Token ID | 198 |
Feature activation | -6.87e-01 |
Pos logit contributions | |
Did | +0.749 |
There | +0.620 |
He | +0.601 |
Maybe | +0.541 |
She | +0.523 |
Neg logit contributions | |
- | -1.928 |
ep | -1.888 |
ies | -1.852 |
ie | -1.821 |
is | -1.795 |
Token | " |
Token ID | 1 |
Feature activation | -2.56e-01 |
Pos logit contributions | |
He | +1.326 |
Did | +1.259 |
It | +1.258 |
They | +1.233 |
She | +1.230 |
Neg logit contributions | |
ishing | -1.434 |
ater | -1.430 |
ies | -1.402 |
veil | -1.373 |
eking | -1.371 |
Token | I |
Token ID | 40 |
Feature activation | -4.45e-01 |
Pos logit contributions | |
He | +0.412 |
From | +0.397 |
| +0.390 |
They | +0.377 |
It | +0.374 |
Neg logit contributions | |
prett | -0.582 |
veil | -0.573 |
cream | -0.570 |
ater | -0.550 |
instrument | -0.535 |
Token | know |
Token ID | 760 |
Feature activation | -1.15e-01 |
Pos logit contributions | |
| +0.994 |
She | +0.948 |
He | +0.937 |
They | +0.927 |
It | +0.910 |
Neg logit contributions | |
ishing | -0.931 |
breath | -0.929 |
crown | -0.916 |
luck | -0.916 |
ies | -0.905 |
Token | , |
Token ID | 11 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
| +0.280 |
She | +0.261 |
They | +0.254 |
It | +0.253 |
He | +0.250 |
Neg logit contributions | |
ishing | -0.247 |
breath | -0.240 |
eking | -0.234 |
grip | -0.226 |
crown | -0.225 |
Token | ," |
Token ID | 553 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
eking | +0.112 |
ishing | +0.111 |
ier | +0.103 |
ts | +0.093 |
ies | +0.091 |
Neg logit contributions | |
| -0.167 |
She | -0.150 |
| -0.147 |
They | -0.146 |
It | -0.146 |
Token | she |
Token ID | 673 |
Feature activation | -2.22e-01 |
Pos logit contributions | |
crown | +0.054 |
breath | +0.053 |
ies | +0.053 |
veil | +0.052 |
luck | +0.050 |
Neg logit contributions | |
It | -0.041 |
| -0.040 |
From | -0.040 |
Even | -0.039 |
They | -0.039 |
Token | replied |
Token ID | 8712 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
| +0.451 |
She | +0.419 |
He | +0.412 |
They | +0.407 |
It | +0.401 |
Neg logit contributions | |
ishing | -0.529 |
breath | -0.520 |
eking | -0.514 |
crown | -0.513 |
luck | -0.507 |
Token | . |
Token ID | 13 |
Feature activation | +5.83e-02 |
Pos logit contributions | |
| +0.389 |
She | +0.365 |
He | +0.360 |
They | +0.356 |
It | +0.353 |
Neg logit contributions | |
ishing | -0.282 |
breath | -0.275 |
eking | -0.271 |
crown | -0.270 |
luck | -0.267 |
Token |
|
Token ID | 198 |
Feature activation | -6.90e-01 |
Pos logit contributions | |
crown | +0.128 |
veil | +0.124 |
breath | +0.123 |
party | +0.123 |
home | +0.123 |
Neg logit contributions | |
Well | -0.042 |
Did | -0.035 |
Even | -0.033 |
Don | -0.033 |
Even | -0.032 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
Did | +0.667 |
He | +0.518 |
Maybe | +0.485 |
They | +0.447 |
There | +0.436 |
Neg logit contributions | |
ep | -1.802 |
breath | -1.787 |
- | -1.787 |
veil | -1.774 |
trick | -1.731 |
Token | Mom |
Token ID | 29252 |
Feature activation | -2.93e-01 |
Pos logit contributions | |
He | +1.279 |
It | +1.224 |
Did | +1.213 |
They | +1.195 |
She | +1.144 |
Neg logit contributions | |
ater | -1.338 |
ault | -1.320 |
veil | -1.319 |
breath | -1.305 |
cream | -1.300 |
Token | my |
Token ID | 1820 |
Feature activation | -4.22e-02 |
Pos logit contributions | |
| +0.678 |
She | +0.636 |
He | +0.630 |
They | +0.626 |
It | +0.619 |
Neg logit contributions | |
ishing | -0.611 |
breath | -0.599 |
crown | -0.593 |
ies | -0.588 |
eking | -0.588 |
Token | and |
Token ID | 290 |
Feature activation | +3.65e-02 |
Pos logit contributions | |
| +0.097 |
She | +0.092 |
He | +0.091 |
They | +0.090 |
It | +0.089 |
Neg logit contributions | |
ishing | -0.087 |
breath | -0.085 |
crown | -0.085 |
eking | -0.084 |
ies | -0.084 |
Token | Daddy |
Token ID | 26926 |
Feature activation | -1.48e-01 |
Pos logit contributions | |
breath | +0.073 |
ishing | +0.073 |
crown | +0.072 |
veil | +0.071 |
luck | +0.070 |
Neg logit contributions | |
| -0.087 |
She | -0.082 |
He | -0.081 |
They | -0.080 |
It | -0.079 |
Token | said |
Token ID | 531 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
| +0.318 |
She | +0.297 |
He | +0.291 |
They | +0.288 |
It | +0.283 |
Neg logit contributions | |
ishing | -0.337 |
breath | -0.331 |
eking | -0.328 |
crown | -0.327 |
luck | -0.323 |
Token | pay |
Token ID | 1414 |
Feature activation | +5.12e-02 |
Pos logit contributions | |
eking | +0.065 |
ishing | +0.064 |
plank | +0.057 |
island | +0.056 |
thieves | +0.056 |
Neg logit contributions | |
| -0.100 |
She | -0.094 |
It | -0.087 |
They | -0.086 |
He | -0.085 |
Token | attention |
Token ID | 3241 |
Feature activation | +4.89e-01 |
Pos logit contributions | |
ishing | +0.085 |
breath | +0.083 |
veil | +0.083 |
crown | +0.082 |
eking | +0.080 |
Neg logit contributions | |
| -0.145 |
He | -0.134 |
She | -0.133 |
They | -0.132 |
It | -0.129 |
Token | to |
Token ID | 284 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
ishing | +1.000 |
breath | +0.972 |
eking | +0.957 |
crown | +0.949 |
veil | +0.935 |
Neg logit contributions | |
| -1.198 |
She | -1.117 |
He | -1.101 |
They | -1.091 |
It | -1.074 |
Token | Tim |
Token ID | 5045 |
Feature activation | -3.03e-01 |
Pos logit contributions | |
| +0.028 |
She | +0.026 |
He | +0.026 |
They | +0.025 |
It | +0.025 |
Neg logit contributions | |
ishing | -0.024 |
breath | -0.022 |
eking | -0.022 |
veil | -0.022 |
crown | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
| +0.695 |
She | +0.643 |
He | +0.634 |
They | +0.624 |
It | +0.594 |
Neg logit contributions | |
breath | -0.661 |
eking | -0.651 |
ishing | -0.650 |
crown | -0.645 |
veil | -0.644 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
mine | +0.141 |
airplane | +0.136 |
crown | +0.133 |
castle | +0.130 |
umbrella | +0.130 |
Neg logit contributions | |
Well | -0.039 |
Don | -0.038 |
Even | -0.034 |
Would | -0.030 |
It | -0.029 |
Token |
|
Token ID | 198 |
Feature activation | -6.87e-01 |
Pos logit contributions | |
Did | +1.206 |
She | +1.097 |
There | +1.015 |
It | +1.008 |
They | +1.002 |
Neg logit contributions | |
mess | -1.420 |
mistake | -1.405 |
adventure | -1.394 |
ep | -1.375 |
home | -1.341 |
Token | Tim |
Token ID | 14967 |
Feature activation | -2.89e-01 |
Pos logit contributions | |
It | +1.257 |
She | +1.229 |
He | +1.217 |
They | +1.202 |
Did | +1.168 |
Neg logit contributions | |
game | -1.427 |
cream | -1.391 |
ep | -1.387 |
plank | -1.372 |
eking | -1.368 |
Token | and |
Token ID | 290 |
Feature activation | +3.51e-02 |
Pos logit contributions | |
| +0.652 |
She | +0.610 |
He | +0.601 |
They | +0.596 |
It | +0.589 |
Neg logit contributions | |
ishing | -0.625 |
breath | -0.614 |
crown | -0.605 |
eking | -0.603 |
luck | -0.597 |
Token | Spot |
Token ID | 15899 |
Feature activation | -2.77e-01 |
Pos logit contributions | |
breath | +0.095 |
crown | +0.094 |
ies | +0.091 |
ishing | +0.090 |
veil | +0.090 |
Neg logit contributions | |
| -0.059 |
She | -0.055 |
They | -0.052 |
It | -0.052 |
He | -0.052 |
Token | ran |
Token ID | 4966 |
Feature activation | -9.12e-03 |
Pos logit contributions | |
| +0.627 |
She | +0.587 |
He | +0.579 |
They | +0.570 |
It | +0.567 |
Neg logit contributions | |
ishing | -0.591 |
breath | -0.587 |
crown | -0.577 |
eking | -0.571 |
luck | -0.571 |
Token | was |
Token ID | 373 |
Feature activation | -2.30e-01 |
Pos logit contributions | |
ishing | +0.579 |
eking | +0.566 |
breath | +0.562 |
crown | +0.559 |
veil | +0.555 |
Neg logit contributions | |
| -0.450 |
He | -0.420 |
She | -0.418 |
They | -0.399 |
It | -0.397 |
Token | very |
Token ID | 845 |
Feature activation | +1.91e-01 |
Pos logit contributions | |
| +0.612 |
She | +0.559 |
He | +0.555 |
It | +0.547 |
They | +0.543 |
Neg logit contributions | |
ishing | -0.443 |
ies | -0.417 |
eking | -0.405 |
veil | -0.404 |
breath | -0.401 |
Token | sorry |
Token ID | 7926 |
Feature activation | +1.17e-01 |
Pos logit contributions | |
ishing | +0.322 |
airplane | +0.303 |
veil | +0.299 |
eking | +0.292 |
breath | +0.291 |
Neg logit contributions | |
| -0.550 |
He | -0.516 |
She | -0.515 |
They | -0.507 |
It | -0.493 |
Token | . |
Token ID | 13 |
Feature activation | +6.54e-02 |
Pos logit contributions | |
ishing | +0.215 |
breath | +0.208 |
eking | +0.207 |
crown | +0.205 |
veil | +0.202 |
Neg logit contributions | |
| -0.309 |
She | -0.292 |
He | -0.289 |
They | -0.286 |
It | -0.282 |
Token |
|
Token ID | 198 |
Feature activation | -6.90e-01 |
Pos logit contributions | |
mine | +0.137 |
crown | +0.132 |
egg | +0.132 |
angel | +0.130 |
castle | +0.130 |
Neg logit contributions | |
Well | -0.046 |
Even | -0.040 |
Would | -0.040 |
Don | -0.038 |
Finally | -0.037 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.794 |
It | +0.604 |
She | +0.558 |
Even | +0.545 |
All | +0.543 |
Neg logit contributions | |
mistake | -1.771 |
breath | -1.720 |
mess | -1.713 |
ep | -1.684 |
dangerous | -1.682 |
Token | But |
Token ID | 1537 |
Feature activation | -2.65e-01 |
Pos logit contributions | |
It | +1.402 |
He | +1.341 |
Did | +1.321 |
She | +1.294 |
There | +1.253 |
Neg logit contributions | |
plank | -1.242 |
breath | -1.234 |
ater | -1.219 |
eking | -1.218 |
ep | -1.182 |
Token | Cur |
Token ID | 4424 |
Feature activation | -2.27e-01 |
Pos logit contributions | |
| +0.576 |
She | +0.536 |
He | +0.527 |
They | +0.521 |
It | +0.514 |
Neg logit contributions | |
ishing | -0.604 |
breath | -0.590 |
eking | -0.585 |
crown | -0.581 |
luck | -0.575 |
Token | ly |
Token ID | 306 |
Feature activation | +3.17e-01 |
Pos logit contributions | |
| +0.681 |
He | +0.657 |
She | +0.651 |
It | +0.636 |
They | +0.635 |
Neg logit contributions | |
veil | -0.302 |
breath | -0.288 |
crown | -0.286 |
mess | -0.276 |
ishing | -0.274 |
Token | wouldn |
Token ID | 3636 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
ishing | +0.701 |
breath | +0.685 |
eking | +0.682 |
crown | +0.667 |
luck | +0.663 |
Neg logit contributions | |
| -0.717 |
She | -0.666 |
He | -0.657 |
They | -0.646 |
It | -0.641 |
Token | 't |
Token ID | 470 |
Feature activation | -3.22e-01 |
Pos logit contributions | |
breath | +0.195 |
balance | +0.185 |
grip | +0.180 |
falling | +0.179 |
mess | +0.170 |
Neg logit contributions | |
They | -0.154 |
He | -0.143 |
She | -0.141 |
From | -0.140 |
Did | -0.137 |
Token | sorry |
Token ID | 7926 |
Feature activation | +9.67e-02 |
Pos logit contributions | |
| +0.468 |
She | +0.440 |
They | +0.432 |
He | +0.429 |
It | +0.421 |
Neg logit contributions | |
ishing | -0.445 |
eking | -0.440 |
breath | -0.438 |
crown | -0.426 |
veil | -0.421 |
Token | for |
Token ID | 329 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
ishing | +0.167 |
breath | +0.165 |
crown | +0.162 |
luck | +0.160 |
ies | +0.160 |
Neg logit contributions | |
| -0.255 |
She | -0.243 |
He | -0.239 |
They | -0.237 |
It | -0.236 |
Token | him |
Token ID | 683 |
Feature activation | -1.64e-01 |
Pos logit contributions | |
ishing | +0.606 |
crown | +0.578 |
veil | +0.575 |
ies | +0.571 |
breath | +0.565 |
Neg logit contributions | |
| -0.791 |
She | -0.719 |
He | -0.714 |
They | -0.711 |
It | -0.695 |
Token | . |
Token ID | 13 |
Feature activation | +6.00e-02 |
Pos logit contributions | |
| +0.405 |
He | +0.364 |
She | +0.363 |
They | +0.359 |
It | +0.343 |
Neg logit contributions | |
ishing | -0.344 |
veil | -0.332 |
eking | -0.331 |
breath | -0.330 |
crown | -0.330 |
Token |
|
Token ID | 198 |
Feature activation | -6.91e-01 |
Pos logit contributions | |
mine | +0.127 |
crown | +0.120 |
egg | +0.120 |
mess | +0.119 |
mistake | +0.118 |
Neg logit contributions | |
Well | -0.042 |
Over | -0.033 |
Even | -0.032 |
Would | -0.032 |
In | -0.032 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.972 |
She | +0.809 |
It | +0.786 |
All | +0.770 |
There | +0.762 |
Neg logit contributions | |
mistake | -1.591 |
mess | -1.573 |
ep | -1.511 |
dangerous | -1.495 |
breath | -1.490 |
Token | " |
Token ID | 1 |
Feature activation | -2.61e-01 |
Pos logit contributions | |
It | +1.317 |
Did | +1.276 |
She | +1.245 |
He | +1.240 |
There | +1.231 |
Neg logit contributions | |
ep | -1.256 |
plank | -1.254 |
ater | -1.250 |
eking | -1.219 |
breath | -1.204 |
Token | Okay |
Token ID | 16454 |
Feature activation | -1.94e-01 |
Pos logit contributions | |
She | +0.457 |
It | +0.438 |
They | +0.436 |
| +0.436 |
He | +0.430 |
Neg logit contributions | |
cream | -0.630 |
ishing | -0.615 |
veil | -0.613 |
plank | -0.604 |
crown | -0.601 |
Token | , |
Token ID | 11 |
Feature activation | +1.12e-01 |
Pos logit contributions | |
| +0.505 |
She | +0.466 |
He | +0.460 |
They | +0.455 |
It | +0.450 |
Neg logit contributions | |
ishing | -0.383 |
eking | -0.367 |
crown | -0.359 |
breath | -0.357 |
veil | -0.354 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.84e-01 |
Pos logit contributions | |
ishing | +0.278 |
breath | +0.273 |
eking | +0.270 |
crown | +0.270 |
luck | +0.267 |
Neg logit contributions | |
| -0.218 |
She | -0.201 |
He | -0.198 |
They | -0.196 |
It | -0.193 |
Token | . |
Token ID | 13 |
Feature activation | +7.51e-02 |
Pos logit contributions | |
| +0.741 |
She | +0.688 |
He | +0.684 |
They | +0.673 |
It | +0.659 |
Neg logit contributions | |
ishing | -0.541 |
breath | -0.523 |
eking | -0.522 |
crown | -0.516 |
veil | -0.516 |
Token | and |
Token ID | 290 |
Feature activation | +2.40e-02 |
Pos logit contributions | |
| +0.672 |
She | +0.631 |
He | +0.621 |
They | +0.615 |
It | +0.610 |
Neg logit contributions | |
ishing | -0.611 |
breath | -0.601 |
crown | -0.594 |
eking | -0.591 |
luck | -0.585 |
Token | Sam |
Token ID | 3409 |
Feature activation | -2.42e-01 |
Pos logit contributions | |
crown | +0.063 |
breath | +0.063 |
ishing | +0.061 |
ies | +0.061 |
luck | +0.061 |
Neg logit contributions | |
| -0.041 |
She | -0.037 |
He | -0.036 |
It | -0.036 |
They | -0.036 |
Token | are |
Token ID | 389 |
Feature activation | -4.50e-02 |
Pos logit contributions | |
| +0.521 |
She | +0.485 |
He | +0.477 |
They | +0.474 |
It | +0.463 |
Neg logit contributions | |
ishing | -0.559 |
breath | -0.545 |
eking | -0.539 |
crown | -0.535 |
veil | -0.530 |
Token | sad |
Token ID | 6507 |
Feature activation | +2.33e-01 |
Pos logit contributions | |
| +0.124 |
He | +0.114 |
They | +0.114 |
She | +0.113 |
It | +0.108 |
Neg logit contributions | |
ishing | -0.085 |
breath | -0.080 |
veil | -0.079 |
plank | -0.078 |
eking | -0.076 |
Token | . |
Token ID | 13 |
Feature activation | +5.89e-02 |
Pos logit contributions | |
ishing | +0.460 |
breath | +0.447 |
eking | +0.445 |
crown | +0.443 |
veil | +0.439 |
Neg logit contributions | |
| -0.588 |
She | -0.549 |
He | -0.545 |
They | -0.542 |
It | -0.527 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
red | +0.112 |
black | +0.107 |
ugly | +0.103 |
blue | +0.102 |
track | +0.102 |
Neg logit contributions | |
Even | -0.055 |
enny | -0.054 |
Well | -0.053 |
Would | -0.049 |
Although | -0.047 |
Token |
|
Token ID | 198 |
Feature activation | -6.86e-01 |
Pos logit contributions | |
Did | +0.870 |
She | +0.776 |
He | +0.660 |
Her | +0.567 |
There | +0.565 |
Neg logit contributions | |
mistake | -1.798 |
ep | -1.761 |
mess | -1.742 |
breath | -1.694 |
ck | -1.672 |
Token | They |
Token ID | 2990 |
Feature activation | -4.02e-01 |
Pos logit contributions | |
She | +1.413 |
He | +1.391 |
It | +1.359 |
Did | +1.338 |
They | +1.247 |
Neg logit contributions | |
ep | -1.349 |
eking | -1.295 |
breath | -1.289 |
plank | -1.276 |
eping | -1.263 |
Token | run |
Token ID | 1057 |
Feature activation | +1.69e-01 |
Pos logit contributions | |
| +1.034 |
She | +0.986 |
They | +0.961 |
He | +0.959 |
It | +0.946 |
Neg logit contributions | |
ishing | -0.800 |
eking | -0.735 |
ies | -0.723 |
breath | -0.711 |
luck | -0.701 |
Token | to |
Token ID | 284 |
Feature activation | -7.96e-03 |
Pos logit contributions | |
ishing | +0.380 |
eking | +0.379 |
veil | +0.371 |
crown | +0.367 |
ies | +0.355 |
Neg logit contributions | |
| -0.400 |
She | -0.364 |
They | -0.361 |
He | -0.358 |
It | -0.346 |
Token | their |
Token ID | 511 |
Feature activation | -1.19e-03 |
Pos logit contributions | |
| +0.019 |
She | +0.018 |
He | +0.018 |
They | +0.018 |
It | +0.017 |
Neg logit contributions | |
breath | -0.016 |
ishing | -0.016 |
crown | -0.015 |
eking | -0.015 |
luck | -0.015 |
Token | and |
Token ID | 290 |
Feature activation | -2.54e-03 |
Pos logit contributions | |
ishing | +0.316 |
eking | +0.301 |
breath | +0.299 |
crown | +0.297 |
veil | +0.291 |
Neg logit contributions | |
| -0.306 |
She | -0.286 |
They | -0.282 |
He | -0.281 |
It | -0.277 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.38e-01 |
Pos logit contributions | |
| +0.006 |
She | +0.006 |
He | +0.006 |
They | +0.006 |
It | +0.005 |
Neg logit contributions | |
ishing | -0.005 |
eking | -0.005 |
breath | -0.005 |
luck | -0.005 |
crown | -0.005 |
Token | said |
Token ID | 531 |
Feature activation | -1.36e-01 |
Pos logit contributions | |
| +0.542 |
She | +0.509 |
He | +0.502 |
They | +0.496 |
It | +0.491 |
Neg logit contributions | |
ishing | -0.499 |
breath | -0.494 |
crown | -0.487 |
eking | -0.484 |
luck | -0.481 |
Token | . |
Token ID | 13 |
Feature activation | +5.81e-02 |
Pos logit contributions | |
| +0.326 |
She | +0.306 |
He | +0.302 |
They | +0.299 |
It | +0.296 |
Neg logit contributions | |
ishing | -0.276 |
breath | -0.270 |
eking | -0.265 |
crown | -0.265 |
veil | -0.262 |
Token |
|
Token ID | 198 |
Feature activation | -6.91e-01 |
Pos logit contributions | |
boat | +0.118 |
red | +0.118 |
crown | +0.117 |
older | +0.116 |
hat | +0.116 |
Neg logit contributions | |
Well | -0.044 |
Don | -0.039 |
Even | -0.039 |
Ell | -0.034 |
Would | -0.033 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.881 |
She | +0.658 |
There | +0.553 |
Her | +0.550 |
Well | +0.535 |
Neg logit contributions | |
mess | -1.670 |
mistake | -1.668 |
ep | -1.666 |
breath | -1.631 |
sand | -1.591 |
Token | They |
Token ID | 2990 |
Feature activation | -4.03e-01 |
Pos logit contributions | |
It | +1.351 |
Did | +1.329 |
She | +1.311 |
He | +1.247 |
There | +1.232 |
Neg logit contributions | |
ater | -1.303 |
plank | -1.288 |
ep | -1.284 |
cream | -1.280 |
breath | -1.222 |
Token | went |
Token ID | 1816 |
Feature activation | -6.37e-02 |
Pos logit contributions | |
| +0.901 |
She | +0.832 |
They | +0.819 |
He | +0.817 |
It | +0.798 |
Neg logit contributions | |
ishing | -0.910 |
eking | -0.876 |
breath | -0.864 |
luck | -0.858 |
crown | -0.847 |
Token | to |
Token ID | 284 |
Feature activation | -1.12e-02 |
Pos logit contributions | |
| +0.148 |
She | +0.133 |
They | +0.132 |
He | +0.130 |
It | +0.121 |
Neg logit contributions | |
ishing | -0.149 |
eking | -0.145 |
veil | -0.139 |
mess | -0.138 |
egg | -0.138 |
Token | their |
Token ID | 511 |
Feature activation | -5.49e-03 |
Pos logit contributions | |
| +0.027 |
She | +0.026 |
He | +0.026 |
They | +0.025 |
It | +0.025 |
Neg logit contributions | |
breath | -0.022 |
ishing | -0.021 |
crown | -0.021 |
veil | -0.021 |
luck | -0.021 |
Token | room |
Token ID | 2119 |
Feature activation | +3.04e-01 |
Pos logit contributions | |
| +0.014 |
They | +0.013 |
She | +0.012 |
He | +0.012 |
The | +0.012 |
Neg logit contributions | |
eking | -0.012 |
ier | -0.012 |
ishing | -0.011 |
iest | -0.011 |
eping | -0.010 |
Token | to |
Token ID | 284 |
Feature activation | -4.52e-02 |
Pos logit contributions | |
eking | +0.553 |
veil | +0.536 |
ishing | +0.527 |
exam | +0.518 |
crown | +0.513 |
Neg logit contributions | |
| -0.594 |
They | -0.550 |
He | -0.528 |
It | -0.502 |
She | -0.497 |
Token | fight |
Token ID | 1907 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
| +0.118 |
She | +0.110 |
He | +0.108 |
They | +0.108 |
It | +0.107 |
Neg logit contributions | |
ishing | -0.088 |
eking | -0.083 |
luck | -0.080 |
ies | -0.080 |
breath | -0.079 |
Token | with |
Token ID | 351 |
Feature activation | +8.75e-03 |
Pos logit contributions | |
ishing | +0.592 |
breath | +0.577 |
eking | +0.571 |
crown | +0.568 |
veil | +0.565 |
Neg logit contributions | |
| -0.761 |
She | -0.711 |
He | -0.702 |
They | -0.698 |
It | -0.686 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.37e-01 |
Pos logit contributions | |
ishing | +0.019 |
eking | +0.018 |
breath | +0.018 |
veil | +0.018 |
crown | +0.018 |
Neg logit contributions | |
| -0.021 |
She | -0.019 |
He | -0.019 |
They | -0.019 |
It | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | +6.05e-02 |
Pos logit contributions | |
| +0.572 |
She | +0.519 |
He | +0.518 |
They | +0.516 |
It | +0.495 |
Neg logit contributions | |
ishing | -0.498 |
eking | -0.492 |
breath | -0.491 |
veil | -0.483 |
crown | -0.481 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
mess | +0.130 |
mine | +0.129 |
crown | +0.127 |
castle | +0.126 |
mistake | +0.121 |
Neg logit contributions | |
Well | -0.047 |
Even | -0.039 |
Don | -0.038 |
Over | -0.034 |
There | -0.033 |
Token |
|
Token ID | 198 |
Feature activation | -6.85e-01 |
Pos logit contributions | |
Did | +0.812 |
She | +0.559 |
There | +0.545 |
It | +0.529 |
Her | +0.525 |
Neg logit contributions | |
mess | -1.859 |
mistake | -1.834 |
home | -1.717 |
ep | -1.696 |
breath | -1.689 |
Token | " |
Token ID | 1 |
Feature activation | -2.62e-01 |
Pos logit contributions | |
It | +1.381 |
Did | +1.310 |
She | +1.302 |
He | +1.296 |
They | +1.250 |
Neg logit contributions | |
plank | -1.282 |
ater | -1.262 |
ep | -1.233 |
cream | -1.220 |
mess | -1.219 |
Token | Please |
Token ID | 5492 |
Feature activation | -2.35e-01 |
Pos logit contributions | |
She | +0.517 |
| +0.507 |
From | +0.499 |
It | +0.496 |
They | +0.480 |
Neg logit contributions | |
cream | -0.577 |
crown | -0.560 |
ishing | -0.554 |
veil | -0.553 |
mess | -0.552 |
Token | , |
Token ID | 11 |
Feature activation | +1.12e-01 |
Pos logit contributions | |
| +0.622 |
She | +0.577 |
He | +0.572 |
They | +0.568 |
It | +0.562 |
Neg logit contributions | |
ishing | -0.442 |
eking | -0.423 |
veil | -0.418 |
breath | -0.417 |
crown | -0.415 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.29e-01 |
Pos logit contributions | |
breath | +0.268 |
ishing | +0.268 |
crown | +0.267 |
luck | +0.263 |
veil | +0.263 |
Neg logit contributions | |
| -0.218 |
She | -0.204 |
He | -0.199 |
They | -0.198 |
It | -0.194 |
Token | ," |
Token ID | 553 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
He | +0.543 |
| +0.538 |
They | +0.531 |
| +0.505 |
It | +0.491 |
Neg logit contributions | |
veil | -1.032 |
ies | -1.017 |
mess | -1.013 |
ishing | -1.009 |
size | -1.005 |
Token | she |
Token ID | 673 |
Feature activation | -2.22e-01 |
Pos logit contributions | |
crown | +0.054 |
breath | +0.053 |
ies | +0.053 |
veil | +0.052 |
luck | +0.050 |
Neg logit contributions | |
It | -0.041 |
| -0.040 |
From | -0.040 |
Even | -0.039 |
They | -0.039 |
Token | explained |
Token ID | 4893 |
Feature activation | -4.43e-02 |
Pos logit contributions | |
| +0.451 |
She | +0.419 |
He | +0.412 |
They | +0.407 |
It | +0.401 |
Neg logit contributions | |
ishing | -0.529 |
breath | -0.520 |
eking | -0.514 |
crown | -0.513 |
luck | -0.507 |
Token | . |
Token ID | 13 |
Feature activation | +5.92e-02 |
Pos logit contributions | |
| +0.108 |
She | +0.101 |
He | +0.099 |
They | +0.098 |
It | +0.096 |
Neg logit contributions | |
ishing | -0.093 |
eking | -0.088 |
breath | -0.088 |
crown | -0.087 |
mess | -0.085 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-01 |
Pos logit contributions | |
crown | +0.135 |
veil | +0.132 |
airplane | +0.128 |
phone | +0.128 |
home | +0.127 |
Neg logit contributions | |
Well | -0.037 |
Even | -0.031 |
Did | -0.030 |
Don | -0.029 |
Even | -0.028 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.635 |
He | +0.491 |
Maybe | +0.418 |
There | +0.407 |
It | +0.405 |
Neg logit contributions | |
ep | -1.835 |
- | -1.810 |
veil | -1.794 |
mistake | -1.788 |
breath | -1.783 |
Token | " |
Token ID | 1 |
Feature activation | -2.57e-01 |
Pos logit contributions | |
He | +1.278 |
It | +1.244 |
Did | +1.209 |
They | +1.165 |
There | +1.144 |
Neg logit contributions | |
ater | -1.323 |
veil | -1.313 |
breath | -1.299 |
ep | -1.293 |
eking | -1.287 |
Token | Oh |
Token ID | 5812 |
Feature activation | -2.97e-01 |
Pos logit contributions | |
He | +0.407 |
It | +0.396 |
| +0.388 |
From | +0.379 |
They | +0.373 |
Neg logit contributions | |
cream | -0.608 |
veil | -0.606 |
ater | -0.596 |
ishing | -0.586 |
eking | -0.584 |
Token | , |
Token ID | 11 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
| +0.794 |
She | +0.751 |
He | +0.741 |
They | +0.735 |
It | +0.728 |
Neg logit contributions | |
ishing | -0.528 |
breath | -0.516 |
eking | -0.507 |
crown | -0.505 |
veil | -0.498 |
Token | I |
Token ID | 314 |
Feature activation | -2.74e-01 |
Pos logit contributions | |
ishing | +0.254 |
breath | +0.244 |
eking | +0.242 |
ies | +0.233 |
crown | +0.233 |
Neg logit contributions | |
| -0.237 |
She | -0.224 |
He | -0.216 |
It | -0.216 |
They | -0.215 |
Token | see |
Token ID | 766 |
Feature activation | -1.10e-01 |
Pos logit contributions | |
| +0.641 |
She | +0.628 |
He | +0.625 |
They | +0.615 |
It | +0.602 |
Neg logit contributions | |
breath | -0.522 |
luck | -0.518 |
mess | -0.517 |
ishing | -0.513 |
crown | -0.511 |
Token | ," |
Token ID | 553 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
ishing | +0.318 |
eking | +0.309 |
breath | +0.306 |
veil | +0.305 |
crown | +0.304 |
Neg logit contributions | |
| -0.364 |
She | -0.335 |
He | -0.331 |
It | -0.329 |
They | -0.328 |
Token | replied |
Token ID | 8712 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
crown | +0.054 |
breath | +0.053 |
ies | +0.053 |
veil | +0.052 |
luck | +0.050 |
Neg logit contributions | |
It | -0.041 |
| -0.040 |
From | -0.040 |
Even | -0.039 |
They | -0.039 |
Token | Nut |
Token ID | 11959 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
| +0.359 |
She | +0.345 |
He | +0.341 |
They | +0.341 |
It | +0.332 |
Neg logit contributions | |
breath | -0.289 |
ishing | -0.286 |
crown | -0.286 |
veil | -0.280 |
luck | -0.278 |
Token | ty |
Token ID | 774 |
Feature activation | +2.99e-01 |
Pos logit contributions | |
| +0.439 |
She | +0.421 |
They | +0.417 |
He | +0.413 |
It | +0.411 |
Neg logit contributions | |
ishing | -0.301 |
breath | -0.293 |
crown | -0.289 |
eking | -0.288 |
ies | -0.287 |
Token | . |
Token ID | 13 |
Feature activation | +6.34e-02 |
Pos logit contributions | |
ishing | +0.511 |
breath | +0.504 |
crown | +0.503 |
ies | +0.495 |
luck | +0.490 |
Neg logit contributions | |
| -0.780 |
She | -0.747 |
They | -0.737 |
He | -0.736 |
It | -0.730 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
crown | +0.153 |
breath | +0.153 |
airplane | +0.151 |
grip | +0.149 |
home | +0.149 |
Neg logit contributions | |
They | -0.042 |
Did | -0.041 |
It | -0.037 |
She | -0.036 |
Well | -0.034 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.726 |
They | +0.639 |
There | +0.636 |
She | +0.619 |
It | +0.580 |
Neg logit contributions | |
- | -1.869 |
ep | -1.816 |
ig | -1.731 |
is | -1.729 |
eping | -1.711 |
Token | L |
Token ID | 43 |
Feature activation | -2.75e-01 |
Pos logit contributions | |
They | +1.381 |
It | +1.367 |
She | +1.344 |
Did | +1.332 |
There | +1.330 |
Neg logit contributions | |
ishing | -1.247 |
eking | -1.226 |
- | -1.225 |
ising | -1.215 |
ep | -1.204 |
Token | ily |
Token ID | 813 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
| +0.663 |
They | +0.633 |
She | +0.628 |
He | +0.619 |
It | +0.615 |
Neg logit contributions | |
ishing | -0.566 |
eking | -0.537 |
ies | -0.525 |
luck | -0.523 |
crown | -0.520 |
Token | and |
Token ID | 290 |
Feature activation | +3.37e-02 |
Pos logit contributions | |
| +0.226 |
She | +0.214 |
He | +0.213 |
They | +0.212 |
It | +0.211 |
Neg logit contributions | |
ishing | -0.212 |
crown | -0.209 |
breath | -0.209 |
ies | -0.207 |
veil | -0.206 |
Token | Nut |
Token ID | 11959 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
breath | +0.077 |
ishing | +0.076 |
crown | +0.075 |
ies | +0.075 |
luck | +0.075 |
Neg logit contributions | |
| -0.067 |
They | -0.064 |
She | -0.063 |
It | -0.063 |
He | -0.063 |
Token | Tom |
Token ID | 4186 |
Feature activation | -3.13e-01 |
Pos logit contributions | |
ishing | +0.129 |
eking | +0.124 |
breath | +0.121 |
ies | +0.117 |
mess | +0.116 |
Neg logit contributions | |
| -0.135 |
He | -0.126 |
She | -0.125 |
They | -0.125 |
It | -0.123 |
Token | are |
Token ID | 389 |
Feature activation | -4.36e-02 |
Pos logit contributions | |
| +0.754 |
She | +0.714 |
He | +0.702 |
They | +0.696 |
It | +0.691 |
Neg logit contributions | |
ishing | -0.614 |
breath | -0.608 |
crown | -0.599 |
eking | -0.596 |
ies | -0.589 |
Token | sad |
Token ID | 6507 |
Feature activation | +2.20e-01 |
Pos logit contributions | |
| +0.124 |
He | +0.113 |
They | +0.112 |
She | +0.111 |
It | +0.107 |
Neg logit contributions | |
ishing | -0.078 |
veil | -0.072 |
eking | -0.069 |
airplane | -0.069 |
island | -0.068 |
Token | . |
Token ID | 13 |
Feature activation | +4.92e-02 |
Pos logit contributions | |
ishing | +0.456 |
eking | +0.445 |
breath | +0.439 |
veil | +0.439 |
crown | +0.438 |
Neg logit contributions | |
| -0.544 |
She | -0.504 |
He | -0.504 |
They | -0.501 |
It | -0.474 |
Token |
|
Token ID | 198 |
Feature activation | -6.91e-01 |
Pos logit contributions | |
mine | +0.109 |
mess | +0.105 |
earthquake | +0.104 |
ugly | +0.103 |
castle | +0.101 |
Neg logit contributions | |
Well | -0.034 |
Would | -0.031 |
Even | -0.031 |
Don | -0.027 |
All | -0.025 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.834 |
She | +0.613 |
There | +0.547 |
Well | +0.537 |
It | +0.531 |
Neg logit contributions | |
mistake | -1.806 |
mess | -1.753 |
home | -1.683 |
ep | -1.660 |
breath | -1.624 |
Token | They |
Token ID | 2990 |
Feature activation | -4.01e-01 |
Pos logit contributions | |
Did | +1.370 |
It | +1.356 |
She | +1.305 |
He | +1.268 |
There | +1.251 |
Neg logit contributions | |
ep | -1.239 |
eking | -1.165 |
ind | -1.157 |
ater | -1.149 |
plank | -1.142 |
Token | have |
Token ID | 423 |
Feature activation | -1.49e-01 |
Pos logit contributions | |
| +1.037 |
She | +0.988 |
He | +0.966 |
They | +0.965 |
It | +0.950 |
Neg logit contributions | |
ishing | -0.775 |
eking | -0.722 |
ies | -0.709 |
breath | -0.706 |
luck | -0.696 |
Token | an |
Token ID | 281 |
Feature activation | +2.38e-03 |
Pos logit contributions | |
| +0.347 |
They | +0.315 |
She | +0.312 |
He | +0.312 |
It | +0.307 |
Neg logit contributions | |
ishing | -0.341 |
eking | -0.329 |
ies | -0.325 |
veil | -0.319 |
breath | -0.316 |
Token | idea |
Token ID | 2126 |
Feature activation | +3.84e-01 |
Pos logit contributions | |
ishing | +0.004 |
ies | +0.004 |
eking | +0.004 |
plank | +0.003 |
breath | +0.003 |
Neg logit contributions | |
He | -0.007 |
| -0.006 |
It | -0.006 |
She | -0.006 |
They | -0.006 |
Token | . |
Token ID | 13 |
Feature activation | +7.60e-02 |
Pos logit contributions | |
ishing | +0.700 |
breath | +0.687 |
crown | +0.675 |
eking | +0.674 |
luck | +0.665 |
Neg logit contributions | |
| -0.997 |
She | -0.942 |
He | -0.930 |
They | -0.922 |
It | -0.912 |
Token | high |
Token ID | 1029 |
Feature activation | +2.46e-01 |
Pos logit contributions | |
ishing | +0.228 |
breath | +0.206 |
eking | +0.206 |
veil | +0.200 |
crown | +0.200 |
Neg logit contributions | |
| -0.463 |
She | -0.437 |
He | -0.437 |
They | -0.428 |
It | -0.421 |
Token | ." |
Token ID | 526 |
Feature activation | +4.51e-02 |
Pos logit contributions | |
ishing | +0.476 |
crown | +0.467 |
veil | +0.460 |
luck | +0.455 |
eking | +0.453 |
Neg logit contributions | |
| -0.627 |
He | -0.586 |
She | -0.581 |
They | -0.557 |
It | -0.539 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.43e-01 |
Pos logit contributions | |
crown | +0.122 |
ies | +0.121 |
breath | +0.121 |
home | +0.117 |
veil | +0.117 |
Neg logit contributions | |
It | -0.020 |
Did | -0.019 |
Well | -0.018 |
Even | -0.016 |
Even | -0.015 |
Token | says |
Token ID | 1139 |
Feature activation | -1.36e-01 |
Pos logit contributions | |
| +0.531 |
She | +0.498 |
He | +0.489 |
It | +0.487 |
They | +0.486 |
Neg logit contributions | |
breath | -0.508 |
crown | -0.505 |
ishing | -0.499 |
veil | -0.497 |
mess | -0.497 |
Token | . |
Token ID | 13 |
Feature activation | +5.78e-02 |
Pos logit contributions | |
| +0.347 |
She | +0.324 |
He | +0.321 |
They | +0.317 |
It | +0.314 |
Neg logit contributions | |
ishing | -0.264 |
breath | -0.253 |
veil | -0.250 |
eking | -0.250 |
crown | -0.249 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
crown | +0.136 |
breath | +0.135 |
hat | +0.132 |
bigger | +0.128 |
red | +0.128 |
Neg logit contributions | |
Well | -0.039 |
Did | -0.039 |
From | -0.034 |
Even | -0.033 |
Even | -0.031 |
Token |
|
Token ID | 198 |
Feature activation | -6.85e-01 |
Pos logit contributions | |
Did | +0.988 |
There | +0.673 |
She | +0.654 |
Well | +0.645 |
Her | +0.590 |
Neg logit contributions | |
breath | -1.731 |
mistake | -1.704 |
mess | -1.678 |
- | -1.630 |
trick | -1.593 |
Token | They |
Token ID | 2990 |
Feature activation | -3.99e-01 |
Pos logit contributions | |
It | +1.378 |
Did | +1.368 |
She | +1.288 |
There | +1.269 |
From | +1.243 |
Neg logit contributions | |
breath | -1.343 |
ater | -1.320 |
plank | -1.305 |
cream | -1.302 |
ep | -1.272 |
Token | hold |
Token ID | 1745 |
Feature activation | +7.83e-02 |
Pos logit contributions | |
| +1.075 |
She | +1.023 |
They | +1.003 |
He | +0.999 |
It | +0.978 |
Neg logit contributions | |
ishing | -0.744 |
eking | -0.696 |
ies | -0.669 |
luck | -0.646 |
problem | -0.634 |
Token | hands |
Token ID | 2832 |
Feature activation | +1.74e-01 |
Pos logit contributions | |
ishing | +0.127 |
eking | +0.123 |
breath | +0.122 |
crown | +0.122 |
veil | +0.120 |
Neg logit contributions | |
| -0.222 |
She | -0.210 |
He | -0.208 |
They | -0.207 |
It | -0.205 |
Token | and |
Token ID | 290 |
Feature activation | +3.77e-02 |
Pos logit contributions | |
ishing | +0.398 |
crown | +0.396 |
eking | +0.391 |
veil | +0.390 |
plank | +0.386 |
Neg logit contributions | |
| -0.395 |
She | -0.354 |
They | -0.354 |
He | -0.348 |
It | -0.334 |
Token | ack |
Token ID | 441 |
Feature activation | +5.96e-02 |
Pos logit contributions | |
ishing | +0.276 |
eking | +0.259 |
ies | +0.258 |
breath | +0.255 |
crown | +0.250 |
Neg logit contributions | |
| -0.461 |
She | -0.445 |
He | -0.442 |
They | -0.430 |
It | -0.427 |
Token | angrily |
Token ID | 34425 |
Feature activation | -5.74e-02 |
Pos logit contributions | |
ishing | +0.112 |
breath | +0.110 |
eking | +0.108 |
crown | +0.107 |
luck | +0.107 |
Neg logit contributions | |
| -0.154 |
She | -0.143 |
He | -0.141 |
They | -0.141 |
It | -0.139 |
Token | at |
Token ID | 379 |
Feature activation | +1.97e-01 |
Pos logit contributions | |
| +0.144 |
He | +0.132 |
They | +0.132 |
She | +0.131 |
It | +0.128 |
Neg logit contributions | |
ishing | -0.114 |
eking | -0.111 |
breath | -0.109 |
veil | -0.109 |
crown | -0.108 |
Token | him |
Token ID | 683 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ishing | +0.418 |
eking | +0.401 |
crown | +0.400 |
breath | +0.399 |
veil | +0.398 |
Neg logit contributions | |
| -0.467 |
She | -0.430 |
He | -0.426 |
They | -0.423 |
It | -0.413 |
Token | . |
Token ID | 13 |
Feature activation | +6.38e-02 |
Pos logit contributions | |
| +0.391 |
She | +0.360 |
He | +0.358 |
They | +0.355 |
It | +0.346 |
Neg logit contributions | |
ishing | -0.308 |
breath | -0.301 |
eking | -0.295 |
veil | -0.293 |
crown | -0.293 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
crown | +0.148 |
mine | +0.146 |
airplane | +0.142 |
mess | +0.142 |
mistake | +0.140 |
Neg logit contributions | |
Well | -0.038 |
Don | -0.036 |
Even | -0.033 |
Did | -0.032 |
Would | -0.030 |
Token |
|
Token ID | 198 |
Feature activation | -6.87e-01 |
Pos logit contributions | |
Did | +0.718 |
She | +0.553 |
It | +0.514 |
They | +0.504 |
There | +0.493 |
Neg logit contributions | |
mistake | -1.856 |
breath | -1.774 |
mess | -1.772 |
ep | -1.763 |
home | -1.736 |
Token | Gary |
Token ID | 33820 |
Feature activation | -1.86e-01 |
Pos logit contributions | |
It | +1.250 |
She | +1.196 |
They | +1.192 |
Did | +1.188 |
He | +1.167 |
Neg logit contributions | |
ater | -1.381 |
ault | -1.348 |
plank | -1.329 |
breath | -1.323 |
ep | -1.312 |
Token | didn |
Token ID | 1422 |
Feature activation | -7.65e-02 |
Pos logit contributions | |
| +0.391 |
She | +0.366 |
He | +0.359 |
They | +0.355 |
It | +0.351 |
Neg logit contributions | |
ishing | -0.421 |
breath | -0.418 |
crown | -0.411 |
eking | -0.410 |
luck | -0.407 |
Token | 't |
Token ID | 470 |
Feature activation | -3.19e-01 |
Pos logit contributions | |
Did | +0.128 |
They | +0.124 |
| +0.123 |
She | +0.122 |
He | +0.117 |
Neg logit contributions | |
breath | -0.178 |
balance | -0.173 |
luck | -0.172 |
grip | -0.169 |
job | -0.167 |
Token | understand |
Token ID | 1833 |
Feature activation | -7.40e-02 |
Pos logit contributions | |
| +0.775 |
He | +0.736 |
She | +0.727 |
They | +0.710 |
It | +0.681 |
Neg logit contributions | |
ishing | -0.698 |
airplane | -0.646 |
veil | -0.646 |
eking | -0.645 |
island | -0.626 |
Token | trying |
Token ID | 2111 |
Feature activation | +1.95e-01 |
Pos logit contributions | |
breath | +0.157 |
ishing | +0.156 |
crown | +0.156 |
veil | +0.154 |
eking | +0.153 |
Neg logit contributions | |
| -0.140 |
She | -0.130 |
He | -0.128 |
They | -0.127 |
It | -0.125 |
Token | to |
Token ID | 284 |
Feature activation | -2.26e-02 |
Pos logit contributions | |
ishing | +0.463 |
eking | +0.439 |
breath | +0.434 |
veil | +0.428 |
crown | +0.422 |
Neg logit contributions | |
| -0.442 |
She | -0.393 |
He | -0.390 |
They | -0.389 |
It | -0.380 |
Token | keep |
Token ID | 1394 |
Feature activation | +5.84e-02 |
Pos logit contributions | |
| +0.061 |
She | +0.054 |
He | +0.054 |
It | +0.053 |
They | +0.053 |
Neg logit contributions | |
ishing | -0.046 |
ies | -0.041 |
eking | -0.041 |
luck | -0.039 |
veil | -0.038 |
Token | up |
Token ID | 510 |
Feature activation | +2.78e-01 |
Pos logit contributions | |
ies | +0.125 |
island | +0.125 |
ishing | +0.124 |
crown | +0.124 |
veil | +0.123 |
Neg logit contributions | |
| -0.134 |
He | -0.117 |
She | -0.117 |
It | -0.116 |
| -0.115 |
Token | . |
Token ID | 13 |
Feature activation | +6.31e-02 |
Pos logit contributions | |
ishing | +0.559 |
eking | +0.532 |
veil | +0.527 |
crown | +0.526 |
breath | +0.522 |
Neg logit contributions | |
| -0.704 |
He | -0.645 |
She | -0.639 |
They | -0.629 |
It | -0.608 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
home | +0.147 |
airplane | +0.144 |
mine | +0.144 |
job | +0.143 |
crown | +0.143 |
Neg logit contributions | |
Well | -0.029 |
Don | -0.027 |
Even | -0.024 |
Did | -0.024 |
That | -0.023 |
Token |
|
Token ID | 198 |
Feature activation | -6.86e-01 |
Pos logit contributions | |
Did | +0.741 |
She | +0.650 |
They | +0.632 |
It | +0.606 |
He | +0.574 |
Neg logit contributions | |
mistake | -1.748 |
breath | -1.714 |
adventure | -1.705 |
mess | -1.683 |
ep | -1.680 |
Token | The |
Token ID | 464 |
Feature activation | -3.76e-01 |
Pos logit contributions | |
It | +1.203 |
They | +1.167 |
She | +1.162 |
He | +1.158 |
Did | +1.116 |
Neg logit contributions | |
game | -1.400 |
ater | -1.380 |
eking | -1.377 |
plank | -1.369 |
breath | -1.364 |
Token | two |
Token ID | 734 |
Feature activation | +3.74e-02 |
Pos logit contributions | |
| +1.033 |
He | +0.949 |
She | +0.925 |
They | +0.908 |
It | +0.888 |
Neg logit contributions | |
ishing | -0.707 |
eking | -0.631 |
ies | -0.580 |
breath | -0.570 |
et | -0.531 |
Token | of |
Token ID | 286 |
Feature activation | -4.33e-06 |
Pos logit contributions | |
ishing | +0.063 |
eking | +0.060 |
veil | +0.057 |
ater | +0.056 |
breath | +0.055 |
Neg logit contributions | |
| -0.104 |
He | -0.100 |
She | -0.099 |
They | -0.098 |
It | -0.094 |
Token | them |
Token ID | 606 |
Feature activation | -3.41e-02 |
Token | ," |
Token ID | 553 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
attention | +0.228 |
job | +0.220 |
toys | +0.220 |
food | +0.219 |
clothes | +0.217 |
Neg logit contributions | |
Well | -0.098 |
Did | -0.096 |
Int | -0.092 |
Exc | -0.090 |
Would | -0.086 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.90e-01 |
Pos logit contributions | |
crown | +0.054 |
breath | +0.053 |
ies | +0.053 |
veil | +0.052 |
luck | +0.050 |
Neg logit contributions | |
It | -0.041 |
| -0.040 |
From | -0.040 |
Even | -0.039 |
They | -0.039 |
Token | said |
Token ID | 531 |
Feature activation | -1.29e-01 |
Pos logit contributions | |
| +0.627 |
She | +0.587 |
He | +0.575 |
It | +0.574 |
They | +0.573 |
Neg logit contributions | |
breath | -0.626 |
ishing | -0.624 |
crown | -0.622 |
ies | -0.614 |
luck | -0.612 |
Token | . |
Token ID | 13 |
Feature activation | +5.68e-02 |
Pos logit contributions | |
| +0.332 |
She | +0.311 |
He | +0.308 |
They | +0.305 |
It | +0.301 |
Neg logit contributions | |
ishing | -0.245 |
breath | -0.238 |
eking | -0.235 |
crown | -0.234 |
veil | -0.232 |
Token |
|
Token ID | 198 |
Feature activation | -6.90e-01 |
Pos logit contributions | |
crown | +0.117 |
bigger | +0.109 |
hat | +0.109 |
breath | +0.109 |
red | +0.108 |
Neg logit contributions | |
Did | -0.058 |
Well | -0.058 |
From | -0.054 |
Even | -0.050 |
Eating | -0.048 |
Token |
|
Token ID | 198 |
Feature activation | -6.88e-01 |
Pos logit contributions | |
Did | +0.915 |
There | +0.620 |
Well | +0.578 |
She | +0.564 |
However | +0.547 |
Neg logit contributions | |
ep | -1.884 |
- | -1.763 |
breath | -1.744 |
mistake | -1.742 |
ies | -1.670 |
Token | " |
Token ID | 1 |
Feature activation | -2.60e-01 |
Pos logit contributions | |
Did | +1.440 |
It | +1.418 |
She | +1.326 |
There | +1.324 |
He | +1.287 |
Neg logit contributions | |
ep | -1.321 |
breath | -1.287 |
cream | -1.241 |
ater | -1.237 |
plank | -1.236 |
Token | But |
Token ID | 1537 |
Feature activation | -2.68e-01 |
Pos logit contributions | |
From | +0.451 |
| +0.425 |
| +0.421 |
It | +0.415 |
Even | +0.411 |
Neg logit contributions | |
cream | -0.598 |
crown | -0.550 |
veil | -0.547 |
ater | -0.542 |
plank | -0.532 |
Token | it |
Token ID | 340 |
Feature activation | -7.82e-02 |
Pos logit contributions | |
| +0.536 |
She | +0.495 |
He | +0.489 |
They | +0.481 |
It | +0.477 |
Neg logit contributions | |
ishing | -0.664 |
breath | -0.646 |
eking | -0.640 |
crown | -0.632 |
veil | -0.628 |
Token | is |
Token ID | 318 |
Feature activation | -6.52e-02 |
Pos logit contributions | |
| +0.156 |
She | +0.144 |
He | +0.144 |
They | +0.142 |
It | +0.141 |
Neg logit contributions | |
ishing | -0.199 |
breath | -0.193 |
eking | -0.190 |
veil | -0.187 |
crown | -0.186 |
Token | too |
Token ID | 1165 |
Feature activation | +1.51e-01 |
Pos logit contributions | |
| +0.161 |
He | +0.148 |
They | +0.146 |
She | +0.141 |
It | +0.141 |
Neg logit contributions | |
ishing | -0.150 |
eking | -0.134 |
ies | -0.133 |
veil | -0.130 |
breath | -0.130 |
Token | He |
Token ID | 679 |
Feature activation | -6.10e-01 |
Pos logit contributions | |
mess | +0.206 |
new | +0.206 |
toys | +0.203 |
home | +0.202 |
job | +0.202 |
Neg logit contributions | |
Don | -0.084 |
From | -0.082 |
Well | -0.080 |
Since | -0.075 |
Boo | -0.070 |
Token | would |
Token ID | 561 |
Feature activation | -3.25e-02 |
Pos logit contributions | |
| +1.371 |
She | +1.273 |
He | +1.258 |
They | +1.244 |
It | +1.224 |
Neg logit contributions | |
ishing | -1.342 |
breath | -1.325 |
eking | -1.312 |
crown | -1.295 |
luck | -1.286 |
Token | clean |
Token ID | 3424 |
Feature activation | +3.10e-01 |
Pos logit contributions | |
| +0.089 |
She | +0.086 |
He | +0.085 |
They | +0.084 |
It | +0.083 |
Neg logit contributions | |
ishing | -0.059 |
eking | -0.055 |
ies | -0.054 |
luck | -0.053 |
breath | -0.053 |
Token | up |
Token ID | 510 |
Feature activation | +3.04e-01 |
Pos logit contributions | |
eking | +0.572 |
veil | +0.568 |
ishing | +0.548 |
crown | +0.545 |
breath | +0.545 |
Neg logit contributions | |
| -0.822 |
He | -0.780 |
They | -0.775 |
She | -0.766 |
It | -0.762 |
Token | his |
Token ID | 465 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
eking | +0.739 |
ishing | +0.737 |
veil | +0.733 |
breath | +0.709 |
crown | +0.699 |
Neg logit contributions | |
| -0.650 |
He | -0.601 |
They | -0.592 |
She | -0.581 |
It | -0.550 |
Token | toys |
Token ID | 14958 |
Feature activation | +4.04e-01 |
Pos logit contributions | |
| +0.034 |
He | +0.031 |
They | +0.031 |
She | +0.030 |
The | +0.029 |
Neg logit contributions | |
eking | -0.016 |
veil | -0.014 |
ier | -0.014 |
island | -0.014 |
crown | -0.014 |
Token | and |
Token ID | 290 |
Feature activation | +5.13e-02 |
Pos logit contributions | |
eking | +0.942 |
ishing | +0.934 |
veil | +0.934 |
crown | +0.926 |
breath | +0.921 |
Neg logit contributions | |
| -0.915 |
She | -0.835 |
He | -0.822 |
They | -0.806 |
It | -0.756 |
Token | put |
Token ID | 1234 |
Feature activation | +1.69e-01 |
Pos logit contributions | |
ishing | +0.097 |
eking | +0.094 |
breath | +0.090 |
luck | +0.089 |
veil | +0.088 |
Neg logit contributions | |
| -0.136 |
He | -0.128 |
She | -0.128 |
They | -0.127 |
It | -0.124 |
Token | them |
Token ID | 606 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
eking | +0.389 |
ishing | +0.381 |
veil | +0.376 |
ies | +0.375 |
decision | +0.373 |
Neg logit contributions | |
| -0.378 |
He | -0.342 |
It | -0.340 |
They | -0.339 |
She | -0.332 |
Token | away |
Token ID | 1497 |
Feature activation | +4.25e-01 |
Pos logit contributions | |
| +0.038 |
He | +0.034 |
They | +0.034 |
She | +0.033 |
| +0.031 |
Neg logit contributions | |
decision | -0.032 |
mistake | -0.032 |
pearl | -0.032 |
veil | -0.031 |
island | -0.031 |
Token | . |
Token ID | 13 |
Feature activation | +9.07e-02 |
Pos logit contributions | |
veil | +0.989 |
eking | +0.977 |
ishing | +0.962 |
breath | +0.954 |
crown | +0.934 |
Neg logit contributions | |
| -0.979 |
He | -0.900 |
They | -0.884 |
She | -0.851 |
The | -0.788 |
Token | find |
Token ID | 1064 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
ishing | +0.105 |
eking | +0.100 |
breath | +0.099 |
crown | +0.098 |
luck | +0.097 |
Neg logit contributions | |
| -0.153 |
She | -0.144 |
He | -0.143 |
They | -0.142 |
It | -0.140 |
Token | out |
Token ID | 503 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
ishing | +0.282 |
eking | +0.274 |
breath | +0.265 |
mess | +0.264 |
ies | +0.261 |
Neg logit contributions | |
| -0.399 |
She | -0.371 |
It | -0.366 |
He | -0.366 |
They | -0.361 |
Token | where |
Token ID | 810 |
Feature activation | +8.86e-02 |
Pos logit contributions | |
ishing | +0.538 |
breath | +0.517 |
eking | +0.507 |
crown | +0.505 |
veil | +0.504 |
Neg logit contributions | |
| -0.657 |
She | -0.623 |
He | -0.604 |
They | -0.600 |
It | -0.595 |
Token | it |
Token ID | 340 |
Feature activation | -5.19e-02 |
Pos logit contributions | |
ishing | +0.214 |
breath | +0.211 |
eking | +0.208 |
crown | +0.208 |
luck | +0.206 |
Neg logit contributions | |
| -0.178 |
She | -0.165 |
He | -0.162 |
They | -0.160 |
It | -0.158 |
Token | was |
Token ID | 373 |
Feature activation | -1.85e-01 |
Pos logit contributions | |
| +0.091 |
She | +0.083 |
He | +0.082 |
They | +0.081 |
It | +0.079 |
Neg logit contributions | |
ishing | -0.139 |
breath | -0.137 |
eking | -0.136 |
crown | -0.136 |
veil | -0.134 |
Token | going |
Token ID | 1016 |
Feature activation | +2.68e-01 |
Pos logit contributions | |
| +0.511 |
He | +0.445 |
They | +0.433 |
She | +0.427 |
It | +0.417 |
Neg logit contributions | |
ishing | -0.366 |
breath | -0.338 |
crown | -0.335 |
veil | -0.334 |
isher | -0.327 |
Token | . |
Token ID | 13 |
Feature activation | +9.88e-02 |
Pos logit contributions | |
ishing | +0.470 |
breath | +0.463 |
eking | +0.454 |
crown | +0.452 |
luck | +0.449 |
Neg logit contributions | |
| -0.703 |
She | -0.667 |
He | -0.658 |
They | -0.653 |
It | -0.650 |
Token | She |
Token ID | 1375 |
Feature activation | -6.11e-01 |
Pos logit contributions | |
job | +0.253 |
adventure | +0.244 |
attention | +0.243 |
calendar | +0.241 |
mess | +0.241 |
Neg logit contributions | |
He | -0.070 |
Well | -0.069 |
| -0.066 |
Would | -0.058 |
He | -0.053 |
Token | ran |
Token ID | 4966 |
Feature activation | +1.40e-02 |
Pos logit contributions | |
| +1.372 |
He | +1.248 |
She | +1.238 |
They | +1.227 |
It | +1.204 |
Neg logit contributions | |
ishing | -1.355 |
eking | -1.314 |
luck | -1.302 |
breath | -1.300 |
attention | -1.287 |
Token | off |
Token ID | 572 |
Feature activation | +2.43e-01 |
Pos logit contributions | |
ishing | +0.025 |
eking | +0.025 |
veil | +0.024 |
crown | +0.024 |
breath | +0.024 |
Neg logit contributions | |
| -0.038 |
She | -0.036 |
He | -0.035 |
They | -0.035 |
It | -0.034 |
Token | into |
Token ID | 656 |
Feature activation | +2.27e-01 |
Pos logit contributions | |
ishing | +0.536 |
eking | +0.523 |
veil | +0.509 |
luck | +0.508 |
crown | +0.508 |
Neg logit contributions | |
| -0.572 |
She | -0.529 |
He | -0.515 |
They | -0.511 |
It | -0.497 |
Token | wanted |
Token ID | 2227 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
ies | +0.601 |
crown | +0.598 |
ishing | +0.595 |
veil | +0.590 |
breath | +0.587 |
Neg logit contributions | |
| -0.646 |
He | -0.602 |
She | -0.601 |
It | -0.595 |
They | -0.591 |
Token | to |
Token ID | 284 |
Feature activation | +4.88e-03 |
Pos logit contributions | |
| +0.257 |
He | +0.232 |
She | +0.231 |
It | +0.224 |
They | +0.224 |
Neg logit contributions | |
ishing | -0.311 |
eking | -0.302 |
breath | -0.293 |
ies | -0.291 |
veil | -0.284 |
Token | perform |
Token ID | 1620 |
Feature activation | +1.80e-01 |
Pos logit contributions | |
ishing | +0.009 |
eking | +0.009 |
breath | +0.008 |
ies | +0.008 |
luck | +0.008 |
Neg logit contributions | |
| -0.013 |
She | -0.012 |
He | -0.012 |
They | -0.012 |
It | -0.012 |
Token | a |
Token ID | 257 |
Feature activation | +4.95e-02 |
Pos logit contributions | |
ishing | +0.397 |
eking | +0.378 |
breath | +0.365 |
egg | +0.365 |
island | +0.362 |
Neg logit contributions | |
| -0.441 |
He | -0.399 |
She | -0.394 |
They | -0.383 |
It | -0.383 |
Token | magic |
Token ID | 5536 |
Feature activation | +4.65e-01 |
Pos logit contributions | |
ishing | +0.094 |
eking | +0.089 |
thieves | +0.078 |
wy | +0.077 |
et | +0.077 |
Neg logit contributions | |
| -0.145 |
| -0.126 |
It | -0.117 |
They | -0.116 |
He | -0.114 |
Token | trick |
Token ID | 6908 |
Feature activation | +4.67e-01 |
Pos logit contributions | |
eking | +0.706 |
ishing | +0.688 |
cking | +0.610 |
wy | +0.585 |
thieves | +0.585 |
Neg logit contributions | |
| -1.502 |
It | -1.335 |
He | -1.331 |
She | -1.309 |
They | -1.290 |
Token | too |
Token ID | 1165 |
Feature activation | +1.65e-01 |
Pos logit contributions | |
ishing | +0.957 |
breath | +0.939 |
eking | +0.919 |
mess | +0.899 |
veil | +0.884 |
Neg logit contributions | |
| -1.172 |
She | -1.089 |
He | -1.088 |
They | -1.060 |
It | -1.053 |
Token | . |
Token ID | 13 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
ishing | +0.352 |
eking | +0.337 |
breath | +0.336 |
crown | +0.330 |
veil | +0.326 |
Neg logit contributions | |
| -0.409 |
She | -0.381 |
He | -0.379 |
They | -0.373 |
It | -0.365 |
Token | She |
Token ID | 1375 |
Feature activation | -6.19e-01 |
Pos logit contributions | |
job | +0.223 |
mess | +0.220 |
mine | +0.214 |
problem | +0.213 |
argument | +0.211 |
Neg logit contributions | |
| -0.077 |
Would | -0.071 |
Well | -0.070 |
Even | -0.066 |
Since | -0.061 |
Token | was |
Token ID | 373 |
Feature activation | -1.93e-01 |
Pos logit contributions | |
| +1.223 |
She | +1.105 |
He | +1.090 |
They | +1.086 |
It | +1.060 |
Neg logit contributions | |
ishing | -1.532 |
eking | -1.514 |
breath | -1.487 |
luck | -1.478 |
attention | -1.470 |
Token | so |
Token ID | 523 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
| +0.488 |
She | +0.454 |
He | +0.442 |
It | +0.439 |
They | +0.428 |
Neg logit contributions | |
ishing | -0.403 |
ies | -0.388 |
island | -0.371 |
breath | -0.371 |
plank | -0.368 |
Token | my |
Token ID | 616 |
Feature activation | +2.04e-01 |
Pos logit contributions | |
ishing | +0.023 |
breath | +0.022 |
eking | +0.022 |
crown | +0.022 |
veil | +0.022 |
Neg logit contributions | |
| -0.034 |
She | -0.031 |
They | -0.031 |
He | -0.031 |
It | -0.031 |
Token | recommendation |
Token ID | 15602 |
Feature activation | +1.62e-01 |
Pos logit contributions | |
eking | +0.337 |
ishing | +0.318 |
plank | +0.299 |
veil | +0.298 |
breath | +0.285 |
Neg logit contributions | |
| -0.601 |
They | -0.536 |
She | -0.524 |
It | -0.521 |
He | -0.520 |
Token | . |
Token ID | 13 |
Feature activation | +9.47e-02 |
Pos logit contributions | |
ishing | +0.302 |
breath | +0.297 |
eking | +0.292 |
crown | +0.289 |
veil | +0.286 |
Neg logit contributions | |
| -0.423 |
She | -0.398 |
He | -0.392 |
They | -0.389 |
It | -0.383 |
Token | Your |
Token ID | 3406 |
Feature activation | -2.04e-01 |
Pos logit contributions | |
job | +0.198 |
luck | +0.186 |
helper | +0.183 |
operation | +0.176 |
artist | +0.174 |
Neg logit contributions | |
| -0.095 |
He | -0.090 |
She | -0.087 |
His | -0.085 |
He | -0.085 |
Token | lemon |
Token ID | 18873 |
Feature activation | +2.54e-01 |
Pos logit contributions | |
| +0.585 |
It | +0.512 |
They | +0.508 |
She | +0.500 |
| +0.500 |
Neg logit contributions | |
eking | -0.341 |
ishing | -0.335 |
veil | -0.308 |
plank | -0.306 |
breath | -0.305 |
Token | ade |
Token ID | 671 |
Feature activation | +2.95e-01 |
Pos logit contributions | |
ishing | +0.560 |
eking | +0.554 |
breath | +0.549 |
veil | +0.530 |
airplane | +0.526 |
Neg logit contributions | |
| -0.587 |
She | -0.546 |
He | -0.535 |
They | -0.531 |
It | -0.525 |
Token | is |
Token ID | 318 |
Feature activation | -4.36e-02 |
Pos logit contributions | |
ishing | +0.705 |
breath | +0.703 |
eking | +0.698 |
veil | +0.684 |
crown | +0.677 |
Neg logit contributions | |
| -0.608 |
She | -0.568 |
He | -0.554 |
They | -0.552 |
It | -0.543 |
Token | perfect |
Token ID | 2818 |
Feature activation | +3.40e-01 |
Pos logit contributions | |
| +0.119 |
It | +0.104 |
She | +0.103 |
He | +0.102 |
They | +0.102 |
Neg logit contributions | |
ishing | -0.082 |
eking | -0.078 |
plank | -0.078 |
breath | -0.077 |
ies | -0.076 |
Token | now |
Token ID | 783 |
Feature activation | +1.77e-01 |
Pos logit contributions | |
eking | +0.733 |
breath | +0.692 |
ishing | +0.691 |
veil | +0.688 |
crown | +0.680 |
Neg logit contributions | |
| -0.852 |
She | -0.772 |
He | -0.769 |
They | -0.760 |
It | -0.738 |
Token | !" |
Token ID | 2474 |
Feature activation | +2.18e-01 |
Pos logit contributions | |
ishing | +0.375 |
breath | +0.373 |
eking | +0.369 |
crown | +0.364 |
veil | +0.360 |
Neg logit contributions | |
| -0.439 |
She | -0.398 |
They | -0.395 |
He | -0.389 |
It | -0.387 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.09e-01 |
Pos logit contributions | |
tea | +0.557 |
mess | +0.557 |
ies | +0.550 |
cake | +0.529 |
- | +0.521 |
Neg logit contributions | |
He | -0.159 |
Where | -0.149 |
He | -0.138 |
Did | -0.137 |
Who | -0.133 |
Token | . |
Token ID | 13 |
Feature activation | +9.43e-02 |
Pos logit contributions | |
ishing | +0.570 |
eking | +0.558 |
breath | +0.557 |
crown | +0.552 |
veil | +0.550 |
Neg logit contributions | |
| -0.747 |
She | -0.694 |
He | -0.688 |
They | -0.684 |
It | -0.667 |
Token | They |
Token ID | 1119 |
Feature activation | -6.02e-01 |
Pos logit contributions | |
mess | +0.227 |
problem | +0.222 |
job | +0.221 |
mistake | +0.219 |
attention | +0.213 |
Neg logit contributions | |
Would | -0.086 |
Did | -0.080 |
Well | -0.073 |
Would | -0.071 |
Ell | -0.068 |
Token | start |
Token ID | 923 |
Feature activation | +2.47e-01 |
Pos logit contributions | |
| +1.581 |
She | +1.501 |
He | +1.475 |
They | +1.474 |
It | +1.440 |
Neg logit contributions | |
ishing | -1.175 |
eking | -1.085 |
ies | -1.065 |
breath | -1.029 |
veil | -1.024 |
Token | to |
Token ID | 284 |
Feature activation | +1.40e-02 |
Pos logit contributions | |
ishing | +0.574 |
veil | +0.573 |
crown | +0.568 |
eking | +0.566 |
luck | +0.559 |
Neg logit contributions | |
| -0.544 |
She | -0.488 |
They | -0.486 |
He | -0.485 |
It | -0.468 |
Token | pick |
Token ID | 2298 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
ishing | +0.022 |
eking | +0.021 |
veil | +0.021 |
luck | +0.020 |
crown | +0.020 |
Neg logit contributions | |
| -0.043 |
They | -0.040 |
She | -0.040 |
He | -0.039 |
It | -0.038 |
Token | up |
Token ID | 510 |
Feature activation | +3.08e-01 |
Pos logit contributions | |
eking | +0.190 |
veil | +0.180 |
crown | +0.176 |
breath | +0.175 |
ishing | +0.173 |
Neg logit contributions | |
| -0.338 |
They | -0.316 |
He | -0.314 |
She | -0.312 |
It | -0.309 |
Token | the |
Token ID | 262 |
Feature activation | +3.72e-02 |
Pos logit contributions | |
eking | +0.763 |
ishing | +0.750 |
veil | +0.745 |
crown | +0.725 |
breath | +0.718 |
Neg logit contributions | |
| -0.663 |
They | -0.608 |
He | -0.598 |
She | -0.583 |
It | -0.561 |
Token | bricks |
Token ID | 28902 |
Feature activation | +2.20e-01 |
Pos logit contributions | |
ishing | +0.057 |
veil | +0.055 |
eking | +0.055 |
breath | +0.054 |
crown | +0.053 |
Neg logit contributions | |
| -0.114 |
She | -0.106 |
They | -0.105 |
He | -0.105 |
It | -0.103 |
Token | and |
Token ID | 290 |
Feature activation | +5.67e-02 |
Pos logit contributions | |
ishing | +0.486 |
eking | +0.483 |
veil | +0.471 |
breath | +0.466 |
crown | +0.464 |
Neg logit contributions | |
| -0.524 |
She | -0.479 |
He | -0.478 |
They | -0.471 |
It | -0.446 |
Token | build |
Token ID | 1382 |
Feature activation | +1.10e-01 |
Pos logit contributions | |
ishing | +0.114 |
eking | +0.109 |
luck | +0.105 |
ies | +0.105 |
veil | +0.105 |
Neg logit contributions | |
| -0.145 |
She | -0.136 |
They | -0.136 |
He | -0.136 |
It | -0.133 |
Token | the |
Token ID | 262 |
Feature activation | +3.67e-02 |
Pos logit contributions | |
ishing | +0.235 |
veil | +0.228 |
eking | +0.226 |
breath | +0.216 |
ies | +0.215 |
Neg logit contributions | |
| -0.278 |
They | -0.258 |
He | -0.254 |
It | -0.252 |
She | -0.245 |
Token | . |
Token ID | 13 |
Feature activation | +9.49e-02 |
Pos logit contributions | |
eking | +0.818 |
ishing | +0.804 |
attention | +0.786 |
breath | +0.785 |
luck | +0.781 |
Neg logit contributions | |
| -1.170 |
He | -1.085 |
They | -1.067 |
She | -1.046 |
It | -0.981 |
Token | They |
Token ID | 1119 |
Feature activation | -6.03e-01 |
Pos logit contributions | |
island | +0.221 |
breath | +0.212 |
eye | +0.206 |
sand | +0.205 |
clothes | +0.205 |
Neg logit contributions | |
He | -0.091 |
His | -0.091 |
Well | -0.090 |
Even | -0.080 |
John | -0.077 |
Token | see |
Token ID | 766 |
Feature activation | -8.44e-02 |
Pos logit contributions | |
| +1.525 |
She | +1.434 |
He | +1.420 |
They | +1.409 |
It | +1.384 |
Neg logit contributions | |
ishing | -1.192 |
eking | -1.137 |
breath | -1.115 |
ies | -1.109 |
crown | -1.097 |
Token | the |
Token ID | 262 |
Feature activation | +3.19e-02 |
Pos logit contributions | |
| +0.199 |
She | +0.183 |
They | +0.181 |
He | +0.181 |
It | +0.179 |
Neg logit contributions | |
ishing | -0.192 |
eking | -0.183 |
breath | -0.180 |
ies | -0.175 |
amount | -0.171 |
Token | shell |
Token ID | 7582 |
Feature activation | +4.52e-01 |
Pos logit contributions | |
ishing | +0.045 |
eking | +0.040 |
ist | +0.036 |
exam | +0.035 |
et | +0.034 |
Neg logit contributions | |
| -0.099 |
He | -0.090 |
It | -0.085 |
She | -0.083 |
They | -0.082 |
Token | and |
Token ID | 290 |
Feature activation | +5.63e-02 |
Pos logit contributions | |
ishing | +1.036 |
breath | +0.992 |
job | +0.962 |
exam | +0.960 |
eking | +0.953 |
Neg logit contributions | |
| -1.040 |
He | -0.967 |
She | -0.963 |
They | -0.939 |
It | -0.910 |
Token | the |
Token ID | 262 |
Feature activation | +3.17e-02 |
Pos logit contributions | |
ishing | +0.112 |
eking | +0.108 |
breath | +0.107 |
crown | +0.105 |
luck | +0.105 |
Neg logit contributions | |
| -0.140 |
She | -0.132 |
He | -0.131 |
They | -0.130 |
It | -0.128 |
Token | pearl |
Token ID | 43836 |
Feature activation | +4.64e-01 |
Pos logit contributions | |
ishing | +0.048 |
eking | +0.042 |
et | +0.036 |
airplane | +0.035 |
ier | +0.035 |
Neg logit contributions | |
| -0.096 |
He | -0.089 |
She | -0.085 |
They | -0.084 |
It | -0.084 |
Token | on |
Token ID | 319 |
Feature activation | +2.47e-01 |
Pos logit contributions | |
ishing | +0.975 |
exam | +0.945 |
breath | +0.927 |
job | +0.908 |
eking | +0.900 |
Neg logit contributions | |
| -1.120 |
He | -1.034 |
She | -1.033 |
They | -1.029 |
It | -0.998 |
Token | the |
Token ID | 262 |
Feature activation | +3.14e-02 |
Pos logit contributions | |
breath | +0.483 |
crown | +0.470 |
ishing | +0.470 |
veil | +0.465 |
luck | +0.459 |
Neg logit contributions | |
| -0.593 |
She | -0.566 |
He | -0.562 |
They | -0.548 |
It | -0.542 |
Token | ground |
Token ID | 2323 |
Feature activation | +3.13e-01 |
Pos logit contributions | |
ishing | +0.059 |
ies | +0.049 |
eking | +0.048 |
problem | +0.046 |
breath | +0.045 |
Neg logit contributions | |
| -0.088 |
He | -0.082 |
They | -0.077 |
It | -0.076 |
She | -0.076 |
Token | away |
Token ID | 1497 |
Feature activation | +4.25e-01 |
Pos logit contributions | |
ishing | +0.153 |
eking | +0.144 |
crown | +0.137 |
veil | +0.137 |
breath | +0.137 |
Neg logit contributions | |
| -0.611 |
She | -0.586 |
He | -0.579 |
They | -0.573 |
It | -0.569 |
Token | . |
Token ID | 13 |
Feature activation | +9.49e-02 |
Pos logit contributions | |
ishing | +0.936 |
eking | +0.887 |
veil | +0.886 |
airplane | +0.876 |
exam | +0.868 |
Neg logit contributions | |
| -1.043 |
She | -0.943 |
He | -0.922 |
They | -0.905 |
It | -0.860 |
Token | Sam |
Token ID | 3409 |
Feature activation | -2.21e-01 |
Pos logit contributions | |
attention | +0.215 |
luck | +0.204 |
accomplishment | +0.197 |
way | +0.191 |
decision | +0.191 |
Neg logit contributions | |
Did | -0.090 |
Would | -0.077 |
E | -0.075 |
He | -0.069 |
His | -0.066 |
Token | was |
Token ID | 373 |
Feature activation | -1.89e-01 |
Pos logit contributions | |
| +0.434 |
She | +0.398 |
He | +0.391 |
They | +0.385 |
It | +0.381 |
Neg logit contributions | |
ishing | -0.549 |
breath | -0.538 |
eking | -0.533 |
crown | -0.530 |
luck | -0.527 |
Token | so |
Token ID | 523 |
Feature activation | +1.09e-01 |
Pos logit contributions | |
| +0.498 |
She | +0.458 |
He | +0.450 |
It | +0.445 |
They | +0.437 |
Neg logit contributions | |
ishing | -0.368 |
ies | -0.345 |
breath | -0.341 |
eking | -0.336 |
veil | -0.333 |
Token | proud |
Token ID | 6613 |
Feature activation | +1.40e-01 |
Pos logit contributions | |
breath | +0.157 |
ishing | +0.154 |
crown | +0.153 |
luck | +0.151 |
eking | +0.151 |
Neg logit contributions | |
| -0.316 |
She | -0.304 |
He | -0.301 |
They | -0.300 |
It | -0.295 |
Token | - |
Token ID | 532 |
Feature activation | +5.68e-02 |
Pos logit contributions | |
ishing | +0.288 |
breath | +0.283 |
eking | +0.279 |
crown | +0.278 |
veil | +0.274 |
Neg logit contributions | |
| -0.333 |
She | -0.312 |
He | -0.308 |
They | -0.304 |
It | -0.301 |
Token | she |
Token ID | 673 |
Feature activation | -1.96e-01 |
Pos logit contributions | |
breath | +0.137 |
crown | +0.136 |
ishing | +0.136 |
luck | +0.135 |
veil | +0.134 |
Neg logit contributions | |
| -0.108 |
She | -0.100 |
He | -0.099 |
They | -0.099 |
It | -0.096 |
Token | had |
Token ID | 550 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
| +0.385 |
She | +0.354 |
He | +0.349 |
They | +0.343 |
It | +0.340 |
Neg logit contributions | |
ishing | -0.487 |
breath | -0.478 |
eking | -0.473 |
crown | -0.470 |
luck | -0.466 |
Token | been |
Token ID | 587 |
Feature activation | +3.02e-02 |
Pos logit contributions | |
| +0.253 |
She | +0.236 |
He | +0.233 |
They | +0.230 |
It | +0.229 |
Neg logit contributions | |
ishing | -0.202 |
breath | -0.197 |
eking | -0.194 |
crown | -0.193 |
ies | -0.192 |
Token | brave |
Token ID | 14802 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
ies | +0.053 |
ishing | +0.053 |
mess | +0.052 |
breath | +0.050 |
veil | +0.050 |
Neg logit contributions | |
| -0.085 |
She | -0.074 |
| -0.073 |
He | -0.072 |
It | -0.072 |
Token | so |
Token ID | 523 |
Feature activation | +1.00e-01 |
Pos logit contributions | |
| +0.519 |
She | +0.475 |
He | +0.466 |
It | +0.463 |
| +0.448 |
Neg logit contributions | |
ishing | -0.405 |
ies | -0.391 |
breath | -0.374 |
island | -0.374 |
plank | -0.370 |
Token | happy |
Token ID | 3772 |
Feature activation | +1.87e-01 |
Pos logit contributions | |
ishing | +0.176 |
breath | +0.169 |
veil | +0.167 |
eking | +0.165 |
island | +0.165 |
Neg logit contributions | |
| -0.279 |
She | -0.261 |
He | -0.258 |
It | -0.252 |
They | -0.252 |
Token | ! |
Token ID | 0 |
Feature activation | +1.97e-01 |
Pos logit contributions | |
ishing | +0.342 |
breath | +0.337 |
crown | +0.332 |
eking | +0.328 |
ies | +0.327 |
Neg logit contributions | |
| -0.476 |
She | -0.451 |
He | -0.445 |
They | -0.443 |
It | -0.440 |
Token | He |
Token ID | 679 |
Feature activation | -6.13e-01 |
Pos logit contributions | |
home | +0.389 |
friends | +0.372 |
job | +0.357 |
old | +0.357 |
toy | +0.352 |
Neg logit contributions | |
Would | -0.159 |
Did | -0.151 |
Don | -0.136 |
Is | -0.133 |
Do | -0.125 |
Token | promised |
Token ID | 8072 |
Feature activation | -8.87e-02 |
Pos logit contributions | |
| +1.322 |
She | +1.236 |
He | +1.216 |
They | +1.205 |
It | +1.187 |
Neg logit contributions | |
ishing | -1.386 |
breath | -1.368 |
crown | -1.353 |
eking | -1.344 |
veil | -1.333 |
Token | to |
Token ID | 284 |
Feature activation | +3.63e-03 |
Pos logit contributions | |
| +0.217 |
It | +0.211 |
They | +0.211 |
She | +0.209 |
Did | +0.207 |
Neg logit contributions | |
breath | -0.149 |
luck | -0.148 |
attention | -0.146 |
crown | -0.141 |
job | -0.137 |
Token | never |
Token ID | 1239 |
Feature activation | -4.14e-02 |
Pos logit contributions | |
ishing | +0.007 |
breath | +0.007 |
crown | +0.007 |
eking | +0.007 |
luck | +0.006 |
Neg logit contributions | |
| -0.009 |
She | -0.009 |
He | -0.009 |
They | -0.009 |
It | -0.009 |
Token | leave |
Token ID | 2666 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
| +0.107 |
She | +0.100 |
He | +0.099 |
They | +0.095 |
It | +0.094 |
Neg logit contributions | |
ishing | -0.082 |
veil | -0.078 |
eking | -0.078 |
breath | -0.078 |
ies | -0.075 |
Token | his |
Token ID | 465 |
Feature activation | -1.43e-02 |
Pos logit contributions | |
ishing | +0.623 |
eking | +0.616 |
ies | +0.616 |
veil | +0.606 |
breath | +0.605 |
Neg logit contributions | |
| -0.576 |
He | -0.524 |
She | -0.521 |
It | -0.510 |
They | -0.497 |
Token | toy |
Token ID | 13373 |
Feature activation | +3.82e-01 |
Pos logit contributions | |
| +0.038 |
The | +0.034 |
They | +0.033 |
He | +0.033 |
She | +0.033 |
Neg logit contributions | |
eking | -0.028 |
aches | -0.026 |
ier | -0.026 |
cking | -0.025 |
pearl | -0.024 |
Token | computer |
Token ID | 3644 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
eking | +0.693 |
ishing | +0.627 |
pearl | +0.620 |
amount | +0.617 |
luck | +0.616 |
Neg logit contributions | |
| -1.075 |
She | -1.029 |
He | -1.014 |
They | -0.968 |
It | -0.936 |
Token | likes |
Token ID | 7832 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
| +0.218 |
She | +0.205 |
He | +0.203 |
They | +0.200 |
It | +0.199 |
Neg logit contributions | |
ishing | -0.184 |
breath | -0.179 |
crown | -0.177 |
luck | -0.175 |
veil | -0.175 |
Token | to |
Token ID | 284 |
Feature activation | -6.26e-04 |
Pos logit contributions | |
| +0.037 |
She | +0.035 |
They | +0.033 |
He | +0.033 |
It | +0.032 |
Neg logit contributions | |
ishing | -0.040 |
eking | -0.039 |
breath | -0.037 |
luck | -0.037 |
ies | -0.037 |
Token | walk |
Token ID | 2513 |
Feature activation | +2.48e-01 |
Pos logit contributions | |
| +0.002 |
She | +0.002 |
He | +0.002 |
They | +0.002 |
It | +0.002 |
Neg logit contributions | |
ishing | -0.001 |
eking | -0.001 |
luck | -0.001 |
ies | -0.001 |
problem | -0.001 |
Token | with |
Token ID | 351 |
Feature activation | +2.52e-02 |
Pos logit contributions | |
eking | +0.482 |
ishing | +0.469 |
mess | +0.444 |
breath | +0.444 |
luck | +0.441 |
Neg logit contributions | |
| -0.661 |
She | -0.628 |
He | -0.615 |
They | -0.610 |
It | -0.597 |
Token | her |
Token ID | 607 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
ishing | +0.059 |
eking | +0.057 |
ies | +0.055 |
breath | +0.054 |
luck | +0.054 |
Neg logit contributions | |
| -0.056 |
She | -0.053 |
He | -0.051 |
They | -0.050 |
It | -0.050 |
Token | dad |
Token ID | 9955 |
Feature activation | +1.21e-01 |
Pos logit contributions | |
eking | +0.030 |
ishing | +0.030 |
amount | +0.027 |
luck | +0.027 |
ier | +0.027 |
Neg logit contributions | |
| -0.033 |
She | -0.029 |
They | -0.028 |
He | -0.028 |
It | -0.027 |
Token | to |
Token ID | 284 |
Feature activation | -3.62e-04 |
Pos logit contributions | |
eking | +0.248 |
ishing | +0.247 |
breath | +0.247 |
crown | +0.238 |
mess | +0.236 |
Neg logit contributions | |
| -0.305 |
She | -0.281 |
They | -0.276 |
He | -0.276 |
It | -0.264 |
Token | the |
Token ID | 262 |
Feature activation | +2.26e-02 |
Pos logit contributions | |
| +0.001 |
She | +0.001 |
He | +0.001 |
They | +0.001 |
It | +0.001 |
Neg logit contributions | |
ishing | -0.001 |
eking | -0.001 |
ies | -0.001 |
luck | -0.001 |
breath | -0.001 |
Token | park |
Token ID | 3952 |
Feature activation | +1.30e-01 |
Pos logit contributions | |
ishing | +0.044 |
eking | +0.043 |
breath | +0.041 |
ies | +0.040 |
luck | +0.038 |
Neg logit contributions | |
| -0.062 |
She | -0.057 |
They | -0.055 |
He | -0.055 |
It | -0.053 |
Token | . |
Token ID | 13 |
Feature activation | +7.69e-02 |
Pos logit contributions | |
ishing | +0.268 |
breath | +0.262 |
eking | +0.260 |
luck | +0.254 |
crown | +0.253 |
Neg logit contributions | |
| -0.325 |
She | -0.304 |
They | -0.300 |
He | -0.298 |
It | -0.288 |
Token | She |
Token ID | 1375 |
Feature activation | -6.24e-01 |
Pos logit contributions | |
airplane | +0.227 |
exam | +0.224 |
size | +0.216 |
crown | +0.215 |
castle | +0.213 |
Neg logit contributions | |
Well | -0.059 |
| -0.054 |
From | -0.048 |
Don | -0.037 |
From | -0.036 |
Token | sun |
Token ID | 4252 |
Feature activation | +1.99e-01 |
Pos logit contributions | |
| +1.452 |
He | +1.349 |
She | +1.304 |
They | +1.299 |
It | +1.267 |
Neg logit contributions | |
ishing | -1.050 |
eking | -0.959 |
ies | -0.877 |
breath | -0.867 |
luck | -0.840 |
Token | would |
Token ID | 561 |
Feature activation | -3.22e-02 |
Pos logit contributions | |
ishing | +0.547 |
eking | +0.535 |
breath | +0.528 |
luck | +0.525 |
crown | +0.522 |
Neg logit contributions | |
| -0.348 |
She | -0.311 |
He | -0.310 |
They | -0.304 |
It | -0.299 |
Token | come |
Token ID | 1282 |
Feature activation | +1.26e-01 |
Pos logit contributions | |
| +0.089 |
She | +0.083 |
He | +0.083 |
They | +0.081 |
It | +0.079 |
Neg logit contributions | |
ishing | -0.063 |
eking | -0.055 |
ies | -0.054 |
luck | -0.053 |
veil | -0.052 |
Token | back |
Token ID | 736 |
Feature activation | +2.89e-01 |
Pos logit contributions | |
ishing | +0.214 |
eking | +0.207 |
veil | +0.206 |
crown | +0.202 |
mess | +0.198 |
Neg logit contributions | |
| -0.363 |
She | -0.340 |
They | -0.336 |
He | -0.336 |
It | -0.326 |
Token | soon |
Token ID | 2582 |
Feature activation | +7.77e-02 |
Pos logit contributions | |
ishing | +0.597 |
eking | +0.585 |
breath | +0.582 |
crown | +0.578 |
veil | +0.575 |
Neg logit contributions | |
| -0.705 |
She | -0.654 |
He | -0.644 |
They | -0.639 |
It | -0.620 |
Token | . |
Token ID | 13 |
Feature activation | +9.04e-02 |
Pos logit contributions | |
ishing | +0.184 |
eking | +0.173 |
breath | +0.171 |
crown | +0.170 |
veil | +0.170 |
Neg logit contributions | |
| -0.177 |
She | -0.158 |
He | -0.157 |
They | -0.156 |
It | -0.149 |
Token | And |
Token ID | 843 |
Feature activation | -4.38e-01 |
Pos logit contributions | |
home | +0.164 |
friends | +0.164 |
job | +0.160 |
food | +0.158 |
breath | +0.155 |
Neg logit contributions | |
Don | -0.087 |
Would | -0.075 |
She | -0.074 |
Come | -0.065 |
Is | -0.065 |
Token | it |
Token ID | 340 |
Feature activation | -5.88e-02 |
Pos logit contributions | |
| +1.003 |
She | +0.930 |
He | +0.924 |
They | +0.915 |
It | +0.899 |
Neg logit contributions | |
ishing | -0.957 |
breath | -0.926 |
eking | -0.926 |
crown | -0.914 |
veil | -0.911 |
Token | did |
Token ID | 750 |
Feature activation | -2.29e-02 |
Pos logit contributions | |
| +0.116 |
She | +0.105 |
He | +0.105 |
They | +0.103 |
It | +0.101 |
Neg logit contributions | |
ishing | -0.149 |
eking | -0.145 |
breath | -0.144 |
crown | -0.141 |
luck | -0.140 |
Token | ! |
Token ID | 0 |
Feature activation | +2.06e-01 |
Pos logit contributions | |
| +0.055 |
He | +0.050 |
She | +0.050 |
They | +0.050 |
It | +0.049 |
Neg logit contributions | |
ishing | -0.051 |
veil | -0.048 |
eking | -0.048 |
breath | -0.047 |
crown | -0.047 |
Token | The |
Token ID | 383 |
Feature activation | -5.27e-01 |
Pos logit contributions | |
home | +0.389 |
friends | +0.389 |
job | +0.389 |
victory | +0.377 |
breath | +0.375 |
Neg logit contributions | |
Don | -0.155 |
Would | -0.153 |
She | -0.140 |
Did | -0.134 |
She | -0.133 |
Token |
|
Token ID | 198 |
Feature activation | -6.59e-01 |
Pos logit contributions | |
Did | +0.911 |
Well | +0.682 |
What | +0.658 |
Would | +0.649 |
Where | +0.644 |
Neg logit contributions | |
mess | -1.495 |
trick | -1.449 |
- | -1.449 |
balance | -1.406 |
cream | -1.404 |
Token | They |
Token ID | 2990 |
Feature activation | -3.76e-01 |
Pos logit contributions | |
| +1.119 |
It | +1.081 |
He | +1.080 |
She | +1.051 |
They | +1.014 |
Neg logit contributions | |
eking | -1.630 |
cream | -1.622 |
ishing | -1.618 |
breath | -1.611 |
mess | -1.606 |
Token | work |
Token ID | 670 |
Feature activation | +4.47e-01 |
Pos logit contributions | |
| +0.960 |
They | +0.930 |
She | +0.923 |
It | +0.908 |
He | +0.906 |
Neg logit contributions | |
ishing | -0.728 |
eking | -0.702 |
ies | -0.661 |
luck | -0.647 |
problem | -0.643 |
Token | hard |
Token ID | 1327 |
Feature activation | +3.14e-01 |
Pos logit contributions | |
veil | +0.918 |
ishing | +0.892 |
eking | +0.883 |
island | +0.876 |
crown | +0.868 |
Neg logit contributions | |
| -1.150 |
They | -1.071 |
She | -1.046 |
He | -1.023 |
It | -1.017 |
Token | and |
Token ID | 290 |
Feature activation | +5.72e-02 |
Pos logit contributions | |
ishing | +0.609 |
crown | +0.598 |
eking | +0.596 |
breath | +0.595 |
veil | +0.593 |
Neg logit contributions | |
| -0.795 |
She | -0.743 |
He | -0.734 |
They | -0.731 |
It | -0.715 |
Token | have |
Token ID | 423 |
Feature activation | -1.30e-01 |
Pos logit contributions | |
ishing | +0.117 |
eking | +0.107 |
veil | +0.105 |
ies | +0.104 |
egg | +0.102 |
Neg logit contributions | |
| -0.147 |
They | -0.142 |
He | -0.138 |
It | -0.137 |
She | -0.137 |
Token | fun |
Token ID | 1257 |
Feature activation | +3.09e-01 |
Pos logit contributions | |
| +0.397 |
She | +0.369 |
They | +0.369 |
He | +0.366 |
It | +0.364 |
Neg logit contributions | |
ishing | -0.202 |
eking | -0.188 |
ies | -0.187 |
breath | -0.183 |
veil | -0.179 |
Token | . |
Token ID | 13 |
Feature activation | +9.60e-02 |
Pos logit contributions | |
veil | +0.766 |
crown | +0.731 |
ishing | +0.719 |
eking | +0.718 |
luck | +0.718 |
Neg logit contributions | |
| -0.680 |
They | -0.606 |
He | -0.566 |
She | -0.559 |
It | -0.539 |
Token | They |
Token ID | 1119 |
Feature activation | -6.00e-01 |
Pos logit contributions | |
attention | +0.250 |
sand | +0.244 |
mess | +0.243 |
job | +0.243 |
size | +0.235 |
Neg logit contributions | |
He | -0.073 |
Did | -0.071 |
He | -0.066 |
His | -0.062 |
His | -0.060 |
Token | make |
Token ID | 787 |
Feature activation | +1.19e-01 |
Pos logit contributions | |
| +1.406 |
They | +1.338 |
She | +1.330 |
He | +1.299 |
It | +1.298 |
Neg logit contributions | |
ishing | -1.332 |
eking | -1.248 |
ies | -1.190 |
luck | -1.161 |
problem | -1.149 |
Token | a |
Token ID | 257 |
Feature activation | +5.98e-02 |
Pos logit contributions | |
ishing | +0.264 |
exam | +0.251 |
eking | +0.248 |
veil | +0.243 |
breath | +0.237 |
Neg logit contributions | |
| -0.290 |
They | -0.261 |
It | -0.254 |
He | -0.247 |
She | -0.246 |
Token | been |
Token ID | 587 |
Feature activation | +3.12e-02 |
Pos logit contributions | |
| +0.258 |
She | +0.238 |
He | +0.236 |
They | +0.232 |
It | +0.232 |
Neg logit contributions | |
ishing | -0.193 |
breath | -0.182 |
ies | -0.180 |
eking | -0.179 |
crown | -0.177 |
Token | saving |
Token ID | 8914 |
Feature activation | +1.59e-01 |
Pos logit contributions | |
ies | +0.054 |
decision | +0.053 |
ishing | +0.053 |
island | +0.051 |
airplane | +0.051 |
Neg logit contributions | |
| -0.086 |
She | -0.076 |
He | -0.074 |
It | -0.073 |
| -0.073 |
Token | them |
Token ID | 606 |
Feature activation | -7.42e-03 |
Pos logit contributions | |
ishing | +0.395 |
veil | +0.390 |
eking | +0.387 |
crown | +0.383 |
airplane | +0.376 |
Neg logit contributions | |
| -0.323 |
He | -0.289 |
They | -0.280 |
She | -0.276 |
It | -0.274 |
Token | especially |
Token ID | 2592 |
Feature activation | +7.35e-02 |
Pos logit contributions | |
| +0.019 |
She | +0.017 |
They | +0.017 |
He | +0.017 |
It | +0.015 |
Neg logit contributions | |
ishing | -0.015 |
crown | -0.014 |
veil | -0.014 |
eking | -0.014 |
breath | -0.014 |
Token | for |
Token ID | 329 |
Feature activation | +3.41e-01 |
Pos logit contributions | |
breath | +0.137 |
veil | +0.136 |
crown | +0.135 |
airplane | +0.134 |
ishing | +0.131 |
Neg logit contributions | |
| -0.191 |
They | -0.176 |
He | -0.176 |
She | -0.174 |
It | -0.170 |
Token | Pete |
Token ID | 17542 |
Feature activation | -1.36e-01 |
Pos logit contributions | |
ishing | +0.689 |
breath | +0.662 |
crown | +0.660 |
veil | +0.657 |
eking | +0.657 |
Neg logit contributions | |
| -0.853 |
He | -0.774 |
They | -0.770 |
She | -0.770 |
It | -0.752 |
Token | . |
Token ID | 13 |
Feature activation | +9.85e-02 |
Pos logit contributions | |
| +0.327 |
He | +0.295 |
She | +0.295 |
They | +0.293 |
It | +0.276 |
Neg logit contributions | |
ishing | -0.288 |
eking | -0.282 |
breath | -0.276 |
crown | -0.273 |
veil | -0.270 |
Token | His |
Token ID | 2399 |
Feature activation | -5.19e-01 |
Pos logit contributions | |
shell | +0.172 |
pearl | +0.171 |
luck | +0.171 |
oy | +0.167 |
home | +0.164 |
Neg logit contributions | |
Mark | -0.142 |
Ch | -0.133 |
Ann | -0.128 |
Don | -0.123 |
Kim | -0.123 |
Token | dad |
Token ID | 9955 |
Feature activation | +1.35e-01 |
Pos logit contributions | |
| +1.383 |
He | +1.249 |
They | +1.232 |
It | +1.224 |
She | +1.217 |
Neg logit contributions | |
ishing | -1.004 |
eking | -0.925 |
plank | -0.901 |
breath | -0.887 |
crown | -0.884 |
Token | smiled |
Token ID | 13541 |
Feature activation | +3.91e-02 |
Pos logit contributions | |
ishing | +0.315 |
breath | +0.311 |
crown | +0.308 |
eking | +0.306 |
luck | +0.303 |
Neg logit contributions | |
| -0.276 |
She | -0.258 |
He | -0.253 |
They | -0.251 |
It | -0.247 |
Token | and |
Token ID | 290 |
Feature activation | +5.84e-02 |
Pos logit contributions | |
ishing | +0.076 |
breath | +0.075 |
eking | +0.074 |
crown | +0.074 |
veil | +0.072 |
Neg logit contributions | |
| -0.097 |
She | -0.092 |
He | -0.091 |
They | -0.090 |
It | -0.088 |
Token | ! |
Token ID | 0 |
Feature activation | +1.96e-01 |
Pos logit contributions | |
ishing | +0.096 |
eking | +0.095 |
breath | +0.093 |
crown | +0.090 |
luck | +0.090 |
Neg logit contributions | |
| -0.125 |
She | -0.118 |
He | -0.117 |
It | -0.114 |
They | -0.114 |
Token | Hur |
Token ID | 10077 |
Feature activation | -2.79e-01 |
Pos logit contributions | |
exam | +0.413 |
invitation | +0.399 |
enemy | +0.398 |
luck | +0.391 |
angel | +0.388 |
Neg logit contributions | |
| -0.250 |
Well | -0.232 |
They | -0.213 |
She | -0.209 |
Don | -0.209 |
Token | ry |
Token ID | 563 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
| +0.756 |
She | +0.709 |
He | +0.700 |
They | +0.697 |
It | +0.689 |
Neg logit contributions | |
ishing | -0.470 |
breath | -0.466 |
crown | -0.465 |
veil | -0.460 |
eking | -0.456 |
Token | , |
Token ID | 11 |
Feature activation | +1.21e-01 |
Pos logit contributions | |
ishing | +0.180 |
eking | +0.175 |
crown | +0.174 |
breath | +0.173 |
veil | +0.170 |
Neg logit contributions | |
| -0.334 |
She | -0.315 |
He | -0.309 |
It | -0.307 |
They | -0.305 |
Token | I |
Token ID | 314 |
Feature activation | -2.62e-01 |
Pos logit contributions | |
ishing | +0.284 |
breath | +0.277 |
eking | +0.275 |
crown | +0.273 |
luck | +0.269 |
Neg logit contributions | |
| -0.256 |
She | -0.238 |
He | -0.234 |
They | -0.231 |
It | -0.229 |
Token | have |
Token ID | 423 |
Feature activation | -1.40e-01 |
Pos logit contributions | |
| +0.531 |
She | +0.486 |
He | +0.479 |
They | +0.475 |
It | +0.469 |
Neg logit contributions | |
ishing | -0.642 |
breath | -0.626 |
eking | -0.623 |
crown | -0.619 |
veil | -0.612 |
Token | a |
Token ID | 257 |
Feature activation | +4.41e-02 |
Pos logit contributions | |
| +0.323 |
It | +0.288 |
She | +0.287 |
He | +0.284 |
They | +0.280 |
Neg logit contributions | |
ishing | -0.324 |
ies | -0.307 |
eking | -0.305 |
breath | -0.297 |
decision | -0.291 |
Token | present |
Token ID | 1944 |
Feature activation | +3.87e-01 |
Pos logit contributions | |
ishing | +0.078 |
eking | +0.075 |
ies | +0.065 |
ts | +0.062 |
cking | +0.061 |
Neg logit contributions | |
| -0.134 |
He | -0.121 |
She | -0.119 |
They | -0.119 |
| -0.119 |
Token | for |
Token ID | 329 |
Feature activation | +3.30e-01 |
Pos logit contributions | |
ishing | +0.714 |
eking | +0.692 |
breath | +0.650 |
mess | +0.609 |
luck | +0.603 |
Neg logit contributions | |
| -1.070 |
She | -1.026 |
He | -1.001 |
They | -0.976 |
It | -0.970 |
Token | you |
Token ID | 345 |
Feature activation | +3.14e-02 |
Pos logit contributions | |
ishing | +0.687 |
eking | +0.649 |
crown | +0.647 |
breath | +0.646 |
ies | +0.643 |
Neg logit contributions | |
| -0.811 |
She | -0.749 |
He | -0.733 |
They | -0.731 |
It | -0.722 |
Token | !" |
Token ID | 2474 |
Feature activation | +2.08e-01 |
Pos logit contributions | |
ishing | +0.061 |
eking | +0.058 |
breath | +0.057 |
crown | +0.057 |
veil | +0.056 |
Neg logit contributions | |
| -0.084 |
She | -0.078 |
He | -0.076 |
They | -0.076 |
It | -0.074 |
Token | a |
Token ID | 257 |
Feature activation | +5.63e-02 |
Pos logit contributions | |
ishing | +0.082 |
eking | +0.082 |
ies | +0.079 |
veil | +0.079 |
mess | +0.078 |
Neg logit contributions | |
| -0.090 |
It | -0.082 |
They | -0.082 |
He | -0.081 |
She | -0.081 |
Token | nap |
Token ID | 25422 |
Feature activation | +2.58e-01 |
Pos logit contributions | |
ishing | +0.090 |
eking | +0.088 |
ies | +0.080 |
veil | +0.075 |
luck | +0.075 |
Neg logit contributions | |
| -0.176 |
He | -0.156 |
| -0.156 |
It | -0.156 |
They | -0.154 |
Token | . |
Token ID | 13 |
Feature activation | +9.56e-02 |
Pos logit contributions | |
eking | +0.538 |
ishing | +0.525 |
breath | +0.515 |
mess | +0.511 |
crown | +0.504 |
Neg logit contributions | |
| -0.628 |
He | -0.568 |
They | -0.555 |
She | -0.549 |
It | -0.528 |
Token | Once |
Token ID | 4874 |
Feature activation | -3.38e-01 |
Pos logit contributions | |
airplane | +0.271 |
job | +0.262 |
island | +0.260 |
breath | +0.258 |
breaths | +0.253 |
Neg logit contributions | |
| -0.075 |
Well | -0.057 |
She | -0.055 |
Don | -0.055 |
She | -0.049 |
Token | everyone |
Token ID | 2506 |
Feature activation | +4.61e-02 |
Pos logit contributions | |
| +0.741 |
She | +0.669 |
He | +0.665 |
They | +0.655 |
It | +0.651 |
Neg logit contributions | |
eking | -0.791 |
ishing | -0.779 |
crown | -0.760 |
veil | -0.742 |
breath | -0.739 |
Token | was |
Token ID | 373 |
Feature activation | -1.87e-01 |
Pos logit contributions | |
ishing | +0.121 |
eking | +0.120 |
breath | +0.119 |
crown | +0.117 |
luck | +0.116 |
Neg logit contributions | |
| -0.086 |
She | -0.077 |
He | -0.077 |
They | -0.075 |
It | -0.074 |
Token | asleep |
Token ID | 16039 |
Feature activation | +2.12e-01 |
Pos logit contributions | |
| +0.520 |
He | +0.460 |
She | +0.459 |
| +0.448 |
They | +0.443 |
Neg logit contributions | |
ishing | -0.342 |
veil | -0.329 |
breath | -0.324 |
ies | -0.323 |
airplane | -0.321 |
Token | , |
Token ID | 11 |
Feature activation | +1.33e-01 |
Pos logit contributions | |
ishing | +0.411 |
eking | +0.399 |
breath | +0.395 |
crown | +0.390 |
ies | +0.390 |
Neg logit contributions | |
| -0.549 |
She | -0.509 |
He | -0.508 |
They | -0.499 |
It | -0.492 |
Token | the |
Token ID | 262 |
Feature activation | +3.42e-02 |
Pos logit contributions | |
ishing | +0.302 |
breath | +0.297 |
crown | +0.293 |
eking | +0.293 |
luck | +0.290 |
Neg logit contributions | |
| -0.286 |
She | -0.267 |
He | -0.263 |
They | -0.260 |
It | -0.257 |
Token | golf |
Token ID | 13126 |
Feature activation | +1.98e-01 |
Pos logit contributions | |
eking | +0.062 |
ishing | +0.060 |
ies | +0.058 |
veil | +0.057 |
breath | +0.056 |
Neg logit contributions | |
| -0.099 |
He | -0.091 |
They | -0.090 |
It | -0.090 |
She | -0.088 |
Token | ball |
Token ID | 2613 |
Feature activation | +2.81e-01 |
Pos logit contributions | |
ishing | +0.332 |
breath | +0.325 |
crown | +0.320 |
eking | +0.317 |
luck | +0.314 |
Neg logit contributions | |
| -0.540 |
She | -0.513 |
He | -0.506 |
They | -0.502 |
It | -0.497 |
Token | the |
Token ID | 262 |
Feature activation | +9.89e-03 |
Pos logit contributions | |
crown | +0.253 |
breath | +0.249 |
luck | +0.248 |
veil | +0.247 |
ies | +0.244 |
Neg logit contributions | |
| -0.211 |
She | -0.200 |
They | -0.192 |
It | -0.192 |
He | -0.191 |
Token | soup |
Token ID | 17141 |
Feature activation | +4.12e-01 |
Pos logit contributions | |
ishing | +0.017 |
eking | +0.016 |
ies | +0.014 |
breath | +0.014 |
cking | +0.013 |
Neg logit contributions | |
| -0.029 |
He | -0.027 |
She | -0.026 |
It | -0.026 |
They | -0.025 |
Token | tasted |
Token ID | 29187 |
Feature activation | +7.89e-02 |
Pos logit contributions | |
ishing | +1.090 |
breath | +1.056 |
eking | +1.051 |
luck | +1.025 |
crown | +1.013 |
Neg logit contributions | |
| -0.767 |
She | -0.701 |
He | -0.700 |
They | -0.691 |
It | -0.672 |
Token | delicious |
Token ID | 12625 |
Feature activation | +3.44e-01 |
Pos logit contributions | |
ishing | +0.153 |
breath | +0.145 |
eking | +0.145 |
veil | +0.140 |
crown | +0.138 |
Neg logit contributions | |
| -0.213 |
He | -0.198 |
She | -0.197 |
They | -0.194 |
It | -0.193 |
Token | ! |
Token ID | 0 |
Feature activation | +1.97e-01 |
Pos logit contributions | |
eking | +0.720 |
ishing | +0.710 |
breath | +0.692 |
crown | +0.666 |
veil | +0.663 |
Neg logit contributions | |
| -0.857 |
She | -0.798 |
He | -0.793 |
They | -0.785 |
It | -0.776 |
Token | Tim |
Token ID | 5045 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
airplane | +0.475 |
crown | +0.470 |
invitation | +0.462 |
treasure | +0.456 |
job | +0.454 |
Neg logit contributions | |
Don | -0.098 |
Well | -0.085 |
| -0.083 |
Did | -0.083 |
Even | -0.080 |
Token | my |
Token ID | 1820 |
Feature activation | -3.73e-02 |
Pos logit contributions | |
| +0.632 |
She | +0.586 |
He | +0.579 |
They | +0.562 |
It | +0.555 |
Neg logit contributions | |
ishing | -0.674 |
breath | -0.669 |
eking | -0.663 |
crown | -0.656 |
veil | -0.651 |
Token | realized |
Token ID | 6939 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
| +0.071 |
She | +0.065 |
He | +0.064 |
They | +0.063 |
It | +0.063 |
Neg logit contributions | |
ishing | -0.094 |
breath | -0.093 |
crown | -0.092 |
eking | -0.091 |
luck | -0.090 |
Token | that |
Token ID | 326 |
Feature activation | +5.38e-02 |
Pos logit contributions | |
| +0.063 |
She | +0.058 |
He | +0.058 |
They | +0.057 |
It | +0.056 |
Neg logit contributions | |
ishing | -0.058 |
breath | -0.056 |
eking | -0.055 |
crown | -0.055 |
luck | -0.054 |
Token | sometimes |
Token ID | 3360 |
Feature activation | -6.44e-02 |
Pos logit contributions | |
ishing | +0.132 |
eking | +0.128 |
breath | +0.128 |
crown | +0.125 |
ies | +0.123 |
Neg logit contributions | |
| -0.112 |
She | -0.103 |
He | -0.102 |
They | -0.101 |
It | -0.099 |
Token | things |
Token ID | 1243 |
Feature activation | +3.94e-01 |
Pos logit contributions | |
| +0.150 |
He | +0.140 |
They | +0.139 |
She | +0.139 |
It | +0.135 |
Neg logit contributions | |
ishing | -0.147 |
eking | -0.141 |
breath | -0.137 |
crown | -0.136 |
ater | -0.135 |
Token | take |
Token ID | 1011 |
Feature activation | +3.76e-02 |
Pos logit contributions | |
ishing | +0.020 |
eking | +0.019 |
breath | +0.019 |
luck | +0.019 |
crown | +0.019 |
Neg logit contributions | |
| -0.033 |
She | -0.031 |
He | -0.031 |
They | -0.030 |
It | -0.030 |
Token | it |
Token ID | 340 |
Feature activation | -5.49e-02 |
Pos logit contributions | |
ishing | +0.088 |
eking | +0.086 |
breath | +0.085 |
crown | +0.084 |
ies | +0.084 |
Neg logit contributions | |
| -0.081 |
She | -0.075 |
He | -0.073 |
It | -0.073 |
They | -0.072 |
Token | home |
Token ID | 1363 |
Feature activation | +4.15e-01 |
Pos logit contributions | |
| +0.163 |
She | +0.141 |
He | +0.138 |
| +0.134 |
They | +0.132 |
Neg logit contributions | |
eking | -0.092 |
ishing | -0.091 |
eling | -0.087 |
inches | -0.085 |
pearl | -0.083 |
Token | . |
Token ID | 13 |
Feature activation | +9.09e-02 |
Pos logit contributions | |
ishing | +0.994 |
eking | +0.988 |
problem | +0.976 |
luck | +0.953 |
breath | +0.942 |
Neg logit contributions | |
| -0.949 |
She | -0.820 |
He | -0.796 |
They | -0.779 |
| -0.738 |
Token |
|
Token ID | 198 |
Feature activation | -6.62e-01 |
Pos logit contributions | |
home | +0.220 |
way | +0.201 |
mess | +0.195 |
up | +0.193 |
new | +0.187 |
Neg logit contributions | |
Did | -0.089 |
Well | -0.074 |
He | -0.073 |
His | -0.071 |
Did | -0.070 |
Token |
|
Token ID | 198 |
Feature activation | -6.61e-01 |
Pos logit contributions | |
Did | +0.655 |
He | +0.549 |
They | +0.510 |
She | +0.494 |
Something | +0.480 |
Neg logit contributions | |
home | -1.862 |
ies | -1.857 |
mess | -1.855 |
breath | -1.853 |
- | -1.847 |
Token | As |
Token ID | 1722 |
Feature activation | -3.41e-01 |
Pos logit contributions | |
He | +1.148 |
It | +1.067 |
| +1.067 |
They | +1.053 |
Did | +0.999 |
Neg logit contributions | |
mess | -1.643 |
veil | -1.616 |
game | -1.537 |
ies | -1.537 |
crown | -1.528 |
Token | she |
Token ID | 673 |
Feature activation | -1.93e-01 |
Pos logit contributions | |
| +0.563 |
She | +0.512 |
He | +0.499 |
They | +0.493 |
It | +0.482 |
Neg logit contributions | |
ishing | -0.965 |
breath | -0.949 |
eking | -0.943 |
crown | -0.939 |
luck | -0.931 |
Token | walked |
Token ID | 6807 |
Feature activation | -1.43e-01 |
Pos logit contributions | |
| +0.393 |
She | +0.367 |
He | +0.360 |
They | +0.356 |
It | +0.351 |
Neg logit contributions | |
ishing | -0.462 |
breath | -0.455 |
crown | -0.449 |
eking | -0.447 |
veil | -0.442 |
Token | , |
Token ID | 11 |
Feature activation | +1.32e-01 |
Pos logit contributions | |
| +0.429 |
She | +0.410 |
He | +0.406 |
They | +0.395 |
It | +0.387 |
Neg logit contributions | |
eking | -0.219 |
ishing | -0.214 |
veil | -0.206 |
mess | -0.204 |
breath | -0.202 |
Token | she |
Token ID | 673 |
Feature activation | -1.93e-01 |
Pos logit contributions | |
breath | +0.343 |
ishing | +0.341 |
crown | +0.340 |
veil | +0.336 |
eking | +0.332 |
Neg logit contributions | |
| -0.233 |
She | -0.213 |
He | -0.213 |
They | -0.209 |
It | -0.205 |
Token | a |
Token ID | 257 |
Feature activation | +5.85e-02 |
Pos logit contributions | |
| +0.091 |
He | +0.083 |
It | +0.082 |
They | +0.081 |
She | +0.080 |
Neg logit contributions | |
ishing | -0.079 |
ies | -0.074 |
breath | -0.073 |
eking | -0.073 |
veil | -0.070 |
Token | weighing |
Token ID | 21990 |
Feature activation | +2.69e-01 |
Pos logit contributions | |
ishing | +0.105 |
eking | +0.104 |
ies | +0.099 |
breath | +0.085 |
cking | +0.083 |
Neg logit contributions | |
| -0.174 |
He | -0.160 |
It | -0.154 |
| -0.153 |
She | -0.149 |
Token | scale |
Token ID | 5046 |
Feature activation | +3.51e-01 |
Pos logit contributions | |
eking | +0.395 |
ishing | +0.371 |
breath | +0.361 |
veil | +0.352 |
ies | +0.346 |
Neg logit contributions | |
| -0.822 |
He | -0.770 |
It | -0.748 |
They | -0.718 |
She | -0.702 |
Token | !" |
Token ID | 2474 |
Feature activation | +2.22e-01 |
Pos logit contributions | |
ishing | +0.700 |
eking | +0.691 |
breath | +0.688 |
veil | +0.670 |
crown | +0.646 |
Neg logit contributions | |
| -0.893 |
He | -0.841 |
She | -0.828 |
They | -0.818 |
It | -0.802 |
Token |
|
Token ID | 198 |
Feature activation | -6.56e-01 |
Pos logit contributions | |
breath | +0.664 |
luck | +0.659 |
attention | +0.658 |
anchor | +0.646 |
job | +0.638 |
Neg logit contributions | |
Would | -0.091 |
Did | -0.089 |
She | -0.082 |
They | -0.077 |
It | -0.067 |
Token |
|
Token ID | 198 |
Feature activation | -6.55e-01 |
Pos logit contributions | |
She | +0.599 |
Did | +0.565 |
They | +0.544 |
It | +0.447 |
He | +0.403 |
Neg logit contributions | |
breath | -1.904 |
trick | -1.901 |
home | -1.899 |
anchor | -1.892 |
mistake | -1.869 |
Token | His |
Token ID | 6653 |
Feature activation | -3.67e-01 |
Pos logit contributions | |
She | +1.254 |
They | +1.218 |
It | +1.209 |
| +1.179 |
He | +1.140 |
Neg logit contributions | |
airplane | -1.459 |
anchor | -1.449 |
breath | -1.447 |
luck | -1.446 |
island | -1.443 |
Token | mom |
Token ID | 1995 |
Feature activation | +7.18e-02 |
Pos logit contributions | |
| +0.788 |
He | +0.703 |
She | +0.659 |
They | +0.650 |
| +0.646 |
Neg logit contributions | |
ishing | -0.943 |
eking | -0.932 |
veil | -0.838 |
plank | -0.838 |
island | -0.834 |
Token | was |
Token ID | 373 |
Feature activation | -1.85e-01 |
Pos logit contributions | |
ishing | +0.149 |
breath | +0.146 |
crown | +0.144 |
eking | +0.144 |
luck | +0.142 |
Neg logit contributions | |
| -0.169 |
She | -0.158 |
He | -0.156 |
They | -0.155 |
It | -0.153 |
Token | surprised |
Token ID | 6655 |
Feature activation | +5.94e-02 |
Pos logit contributions | |
| +0.486 |
She | +0.453 |
He | +0.445 |
They | +0.439 |
It | +0.438 |
Neg logit contributions | |
ishing | -0.346 |
breath | -0.333 |
ies | -0.329 |
crown | -0.327 |
eking | -0.324 |
Token | . |
Token ID | 13 |
Feature activation | +9.71e-02 |
Pos logit contributions | |
ishing | +0.109 |
breath | +0.107 |
eking | +0.104 |
crown | +0.104 |
luck | +0.104 |
Neg logit contributions | |
| -0.153 |
She | -0.144 |
He | -0.143 |
They | -0.142 |
It | -0.141 |
Token | fix |
Token ID | 4259 |
Feature activation | +1.42e-01 |
Pos logit contributions | |
ishing | +0.012 |
veil | +0.012 |
eking | +0.012 |
breath | +0.012 |
crown | +0.012 |
Neg logit contributions | |
| -0.020 |
She | -0.019 |
He | -0.019 |
They | -0.018 |
It | -0.018 |
Token | it |
Token ID | 340 |
Feature activation | -6.12e-02 |
Pos logit contributions | |
eking | +0.338 |
ishing | +0.338 |
island | +0.331 |
breath | +0.331 |
crown | +0.330 |
Neg logit contributions | |
| -0.299 |
She | -0.274 |
It | -0.267 |
He | -0.265 |
They | -0.263 |
Token | . |
Token ID | 13 |
Feature activation | +8.87e-02 |
Pos logit contributions | |
| +0.147 |
She | +0.137 |
He | +0.134 |
They | +0.131 |
It | +0.127 |
Neg logit contributions | |
veil | -0.135 |
ishing | -0.133 |
eking | -0.131 |
crown | -0.130 |
breath | -0.129 |
Token | |
Token ID | 220 |
Feature activation | -5.60e-01 |
Pos logit contributions | |
home | +0.197 |
attention | +0.168 |
job | +0.168 |
ish | +0.166 |
way | +0.164 |
Neg logit contributions | |
Did | -0.097 |
Mark | -0.094 |
Steve | -0.087 |
Would | -0.085 |
Did | -0.084 |
Token |
|
Token ID | 198 |
Feature activation | -6.64e-01 |
Pos logit contributions | |
Hi | +0.457 |
Did | +0.435 |
He | +0.389 |
Well | +0.385 |
Would | +0.385 |
Neg logit contributions | |
ies | -1.469 |
attention | -1.447 |
home | -1.432 |
ishing | -1.415 |
eking | -1.410 |
Token |
|
Token ID | 198 |
Feature activation | -6.64e-01 |
Pos logit contributions | |
Did | +0.715 |
There | +0.706 |
It | +0.704 |
He | +0.616 |
They | +0.591 |
Neg logit contributions | |
is | -1.625 |
trick | -1.517 |
home | -1.507 |
ies | -1.490 |
balance | -1.477 |
Token | In |
Token ID | 818 |
Feature activation | -3.70e-01 |
Pos logit contributions | |
He | +1.200 |
It | +1.198 |
They | +1.145 |
She | +1.139 |
| +1.116 |
Neg logit contributions | |
mess | -1.439 |
game | -1.432 |
balance | -1.418 |
breath | -1.417 |
ishing | -1.412 |
Token | the |
Token ID | 262 |
Feature activation | +2.48e-02 |
Pos logit contributions | |
| +0.833 |
She | +0.763 |
He | +0.733 |
They | +0.728 |
It | +0.723 |
Neg logit contributions | |
ishing | -0.844 |
eking | -0.837 |
ies | -0.819 |
crown | -0.811 |
veil | -0.791 |
Token | end |
Token ID | 886 |
Feature activation | +2.91e-01 |
Pos logit contributions | |
ishing | +0.038 |
eking | +0.037 |
ies | +0.037 |
breath | +0.035 |
crown | +0.033 |
Neg logit contributions | |
| -0.077 |
She | -0.073 |
He | -0.071 |
They | -0.070 |
It | -0.069 |
Token | , |
Token ID | 11 |
Feature activation | +1.27e-01 |
Pos logit contributions | |
ies | +0.382 |
breath | +0.369 |
mess | +0.358 |
problem | +0.356 |
ishing | +0.354 |
Neg logit contributions | |
| -0.829 |
He | -0.816 |
It | -0.808 |
She | -0.799 |
They | -0.798 |
Token | the |
Token ID | 262 |
Feature activation | +2.36e-02 |
Pos logit contributions | |
breath | +0.300 |
ishing | +0.293 |
ies | +0.293 |
crown | +0.291 |
veil | +0.290 |
Neg logit contributions | |
| -0.247 |
He | -0.233 |
She | -0.230 |
They | -0.226 |
It | -0.225 |
Token | They |
Token ID | 1119 |
Feature activation | -6.04e-01 |
Pos logit contributions | |
luck | +0.244 |
attention | +0.238 |
job | +0.232 |
enemy | +0.232 |
guys | +0.231 |
Neg logit contributions | |
Would | -0.078 |
| -0.073 |
Well | -0.073 |
| -0.068 |
Did | -0.067 |
Token | liked |
Token ID | 8288 |
Feature activation | -8.53e-02 |
Pos logit contributions | |
| +1.274 |
She | +1.187 |
He | +1.169 |
They | +1.155 |
It | +1.142 |
Neg logit contributions | |
ishing | -1.389 |
breath | -1.372 |
crown | -1.356 |
eking | -1.349 |
luck | -1.338 |
Token | his |
Token ID | 465 |
Feature activation | -6.24e-03 |
Pos logit contributions | |
| +0.189 |
She | +0.174 |
He | +0.173 |
They | +0.173 |
It | +0.169 |
Neg logit contributions | |
ishing | -0.196 |
eking | -0.192 |
breath | -0.190 |
veil | -0.185 |
crown | -0.185 |
Token | workshop |
Token ID | 20243 |
Feature activation | +1.97e-01 |
Pos logit contributions | |
| +0.019 |
He | +0.018 |
They | +0.017 |
She | +0.017 |
The | +0.017 |
Neg logit contributions | |
eking | -0.011 |
ishing | -0.010 |
cking | -0.009 |
ies | -0.009 |
plank | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | +9.43e-02 |
Pos logit contributions | |
eking | +0.443 |
ishing | +0.433 |
veil | +0.409 |
breath | +0.404 |
grip | +0.387 |
Neg logit contributions | |
| -0.484 |
He | -0.464 |
They | -0.456 |
She | -0.443 |
It | -0.415 |
Token | They |
Token ID | 1119 |
Feature activation | -6.03e-01 |
Pos logit contributions | |
luck | +0.216 |
enemy | +0.209 |
idea | +0.209 |
guys | +0.208 |
magic | +0.208 |
Neg logit contributions | |
Would | -0.087 |
Well | -0.086 |
| -0.080 |
oped | -0.075 |
From | -0.073 |
Token | liked |
Token ID | 8288 |
Feature activation | -8.50e-02 |
Pos logit contributions | |
| +1.301 |
She | +1.214 |
He | +1.196 |
They | +1.182 |
It | +1.167 |
Neg logit contributions | |
ishing | -1.365 |
breath | -1.346 |
crown | -1.328 |
eking | -1.324 |
luck | -1.311 |
Token | their |
Token ID | 511 |
Feature activation | +1.81e-02 |
Pos logit contributions | |
| +0.193 |
She | +0.179 |
He | +0.178 |
They | +0.177 |
It | +0.174 |
Neg logit contributions | |
ishing | -0.190 |
eking | -0.187 |
breath | -0.185 |
veil | -0.180 |
crown | -0.179 |
Token | new |
Token ID | 649 |
Feature activation | +3.20e-01 |
Pos logit contributions | |
eking | +0.031 |
ishing | +0.027 |
cking | +0.024 |
ies | +0.024 |
veil | +0.024 |
Neg logit contributions | |
| -0.056 |
He | -0.053 |
They | -0.052 |
She | -0.050 |
It | -0.050 |
Token | watches |
Token ID | 16860 |
Feature activation | +5.34e-02 |
Pos logit contributions | |
eking | +0.570 |
ishing | +0.481 |
cking | +0.460 |
breath | +0.445 |
plank | +0.425 |
Neg logit contributions | |
| -1.001 |
He | -0.934 |
They | -0.917 |
She | -0.916 |
It | -0.879 |
Token | . |
Token ID | 13 |
Feature activation | +9.48e-02 |
Pos logit contributions | |
ishing | +0.112 |
breath | +0.109 |
eking | +0.108 |
veil | +0.105 |
luck | +0.104 |
Neg logit contributions | |
| -0.131 |
She | -0.124 |
He | -0.123 |
They | -0.121 |
It | -0.120 |
Token | happy |
Token ID | 3772 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
| +0.532 |
They | +0.500 |
He | +0.500 |
She | +0.493 |
It | +0.475 |
Neg logit contributions | |
ishing | -0.507 |
veil | -0.488 |
ater | -0.460 |
pearl | -0.454 |
eking | -0.453 |
Token | with |
Token ID | 351 |
Feature activation | +3.00e-02 |
Pos logit contributions | |
eking | +0.422 |
ishing | +0.421 |
veil | +0.410 |
breath | +0.408 |
crown | +0.404 |
Neg logit contributions | |
| -0.449 |
She | -0.405 |
They | -0.404 |
He | -0.402 |
It | -0.381 |
Token | their |
Token ID | 511 |
Feature activation | +1.26e-02 |
Pos logit contributions | |
eking | +0.079 |
ishing | +0.079 |
veil | +0.076 |
ies | +0.074 |
crown | +0.073 |
Neg logit contributions | |
| -0.059 |
They | -0.056 |
It | -0.053 |
She | -0.051 |
He | -0.050 |
Token | path |
Token ID | 3108 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
eking | +0.020 |
ishing | +0.019 |
ier | +0.019 |
le | +0.017 |
wy | +0.017 |
Neg logit contributions | |
| -0.037 |
They | -0.035 |
He | -0.034 |
She | -0.033 |
It | -0.033 |
Token | . |
Token ID | 13 |
Feature activation | +8.27e-02 |
Pos logit contributions | |
ishing | +0.715 |
veil | +0.695 |
eking | +0.692 |
breath | +0.676 |
crown | +0.661 |
Neg logit contributions | |
| -0.686 |
They | -0.656 |
He | -0.647 |
She | -0.647 |
It | -0.607 |
Token | It |
Token ID | 632 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
best | +0.147 |
home | +0.146 |
toys | +0.145 |
track | +0.144 |
tea | +0.141 |
Neg logit contributions | |
Don | -0.072 |
Did | -0.071 |
She | -0.066 |
Int | -0.062 |
Well | -0.060 |
Token | was |
Token ID | 373 |
Feature activation | -1.95e-01 |
Pos logit contributions | |
| +1.021 |
She | +0.927 |
He | +0.916 |
They | +0.899 |
It | +0.878 |
Neg logit contributions | |
ishing | -1.645 |
breath | -1.612 |
eking | -1.600 |
crown | -1.580 |
veil | -1.577 |
Token | fine |
Token ID | 3734 |
Feature activation | +2.83e-01 |
Pos logit contributions | |
| +0.510 |
They | +0.467 |
He | +0.465 |
She | +0.459 |
It | +0.452 |
Neg logit contributions | |
ishing | -0.415 |
eking | -0.386 |
veil | -0.373 |
ies | -0.371 |
ater | -0.363 |
Token | and |
Token ID | 290 |
Feature activation | +4.67e-02 |
Pos logit contributions | |
eking | +0.739 |
veil | +0.703 |
ishing | +0.695 |
breath | +0.685 |
crown | +0.677 |
Neg logit contributions | |
| -0.593 |
They | -0.575 |
He | -0.554 |
She | -0.554 |
It | -0.501 |
Token | long |
Token ID | 890 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
ishing | +0.080 |
veil | +0.076 |
eking | +0.073 |
breath | +0.073 |
exam | +0.071 |
Neg logit contributions | |
| -0.133 |
They | -0.128 |
He | -0.124 |
She | -0.123 |
It | -0.123 |
Token | and |
Token ID | 290 |
Feature activation | +4.64e-02 |
Pos logit contributions | |
ishing | +0.557 |
veil | +0.535 |
breath | +0.534 |
eking | +0.530 |
ies | +0.513 |
Neg logit contributions | |
| -0.529 |
She | -0.491 |
They | -0.490 |
He | -0.489 |
It | -0.457 |