Token | have |
Token ID | 423 |
Feature activation | +8.33e-03 |
Pos logit contributions | |
robe | +3.923 |
Spoiler | +3.084 |
ンジ | +2.794 |
taboola | +2.787 |
hib | +2.780 |
Neg logit contributions | |
bia | -2.564 |
rele | -2.540 |
constitu | -2.537 |
ibaba | -2.479 |
iland | -2.406 |
Token | morphed |
Token ID | 49976 |
Feature activation | +1.10e-01 |
Pos logit contributions | |
bia | +0.031 |
rele | +0.031 |
constitu | +0.030 |
ibaba | +0.030 |
iland | +0.029 |
Neg logit contributions | |
robe | -0.053 |
Spoiler | -0.043 |
ンジ | -0.040 |
taboola | -0.040 |
hib | -0.039 |
Token | into |
Token ID | 656 |
Feature activation | +4.66e-01 |
Pos logit contributions | |
rele | +0.427 |
bia | +0.426 |
constitu | +0.421 |
ibaba | +0.406 |
iland | +0.405 |
Neg logit contributions | |
robe | -0.689 |
Spoiler | -0.546 |
ンジ | -0.513 |
taboola | -0.505 |
Doors | -0.501 |
Token | nerves |
Token ID | 25377 |
Feature activation | +3.00e-01 |
Pos logit contributions | |
rele | +1.759 |
bia | +1.720 |
constitu | +1.709 |
ibaba | +1.633 |
iland | +1.633 |
Neg logit contributions | |
robe | -2.997 |
Spoiler | -2.373 |
ンジ | -2.256 |
taboola | -2.225 |
hib | -2.195 |
Token | , |
Token ID | 11 |
Feature activation | +3.98e-01 |
Pos logit contributions | |
rele | +1.065 |
bia | +1.051 |
constitu | +1.039 |
ibaba | +0.997 |
iland | +0.994 |
Neg logit contributions | |
robe | -1.994 |
Spoiler | -1.594 |
ンジ | -1.512 |
taboola | -1.494 |
hib | -1.477 |
Token | dread |
Token ID | 15157 |
Feature activation | +2.07e+00 |
Pos logit contributions | |
rele | +1.476 |
bia | +1.456 |
constitu | +1.439 |
iland | +1.385 |
ibaba | +1.383 |
Neg logit contributions | |
robe | -2.580 |
Spoiler | -2.052 |
ンジ | -1.941 |
taboola | -1.916 |
hib | -1.895 |
Token | and |
Token ID | 290 |
Feature activation | +5.18e-01 |
Pos logit contributions | |
rele | +7.040 |
bia | +6.732 |
constitu | +6.600 |
owing | +6.438 |
ibaba | +6.432 |
Neg logit contributions | |
robe | -13.027 |
Spoiler | -10.347 |
ンジ | -9.997 |
taboola | -9.787 |
hib | -9.742 |
Token | bewild |
Token ID | 34672 |
Feature activation | +2.89e-01 |
Pos logit contributions | |
rele | +1.917 |
bia | +1.868 |
constitu | +1.862 |
ibaba | +1.776 |
iland | +1.773 |
Neg logit contributions | |
robe | -3.374 |
Spoiler | -2.680 |
ンジ | -2.557 |
taboola | -2.519 |
hib | -2.487 |
Token | er |
Token ID | 263 |
Feature activation | +6.23e-01 |
Pos logit contributions | |
rele | +1.621 |
bia | +1.604 |
constitu | +1.585 |
ibaba | +1.551 |
iland | +1.549 |
Neg logit contributions | |
robe | -1.322 |
Spoiler | -0.946 |
ンジ | -0.874 |
taboola | -0.844 |
ヤ | -0.833 |
Token | ment |
Token ID | 434 |
Feature activation | +8.89e-01 |
Pos logit contributions | |
bia | +2.382 |
rele | +2.381 |
constitu | +2.268 |
iland | +2.254 |
ibaba | +2.224 |
Neg logit contributions | |
robe | -3.939 |
Spoiler | -3.124 |
ンジ | -2.973 |
hib | -2.885 |
ヤ | -2.860 |
Token | . |
Token ID | 13 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
rele | +3.413 |
bia | +3.279 |
constitu | +3.250 |
iland | +3.141 |
ibaba | +3.113 |
Neg logit contributions | |
robe | -5.623 |
Spoiler | -4.415 |
ンジ | -4.244 |
taboola | -4.133 |
hib | -4.124 |
Token | raids |
Token ID | 20371 |
Feature activation | +8.66e-01 |
Pos logit contributions | |
rele | +0.294 |
bia | +0.290 |
constitu | +0.287 |
ibaba | +0.278 |
iland | +0.278 |
Neg logit contributions | |
robe | -0.444 |
Spoiler | -0.349 |
ンジ | -0.328 |
taboola | -0.323 |
Doors | -0.321 |
Token | . |
Token ID | 13 |
Feature activation | +1.73e-01 |
Pos logit contributions | |
rele | +3.295 |
bia | +3.207 |
constitu | +3.147 |
iland | +3.067 |
edi | +3.008 |
Neg logit contributions | |
robe | -5.498 |
Spoiler | -4.347 |
ンジ | -4.152 |
taboola | -4.064 |
ヤ | -4.062 |
Token |
|
Token ID | 198 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
rele | +0.686 |
bia | +0.686 |
constitu | +0.675 |
iland | +0.657 |
ibaba | +0.645 |
Neg logit contributions | |
robe | -1.071 |
Spoiler | -0.828 |
ンジ | -0.789 |
taboola | -0.778 |
ヤ | -0.767 |
Token |
|
Token ID | 198 |
Feature activation | +8.93e-01 |
Pos logit contributions | |
bia | +3.164 |
rele | +3.138 |
edi | +2.952 |
iland | +2.940 |
owing | +2.864 |
Neg logit contributions | |
robe | -6.754 |
Spoiler | -5.291 |
ンジ | -5.249 |
taboola | -5.086 |
ヤ | -5.075 |
Token | " |
Token ID | 1 |
Feature activation | +6.19e-01 |
Pos logit contributions | |
bia | +3.561 |
rele | +3.472 |
constitu | +3.412 |
iland | +3.390 |
edi | +3.338 |
Neg logit contributions | |
robe | -5.412 |
Spoiler | -4.280 |
ンジ | -4.002 |
Doors | -3.918 |
taboola | -3.906 |
Token | Air |
Token ID | 16170 |
Feature activation | +1.99e+00 |
Pos logit contributions | |
rele | +2.536 |
bia | +2.527 |
constitu | +2.476 |
iland | +2.410 |
ibaba | +2.374 |
Neg logit contributions | |
robe | -3.737 |
Spoiler | -2.906 |
ンジ | -2.735 |
taboola | -2.694 |
hib | -2.655 |
Token | strikes |
Token ID | 8956 |
Feature activation | +2.38e-01 |
Pos logit contributions | |
bia | +5.959 |
rele | +5.924 |
numbers | +5.417 |
iland | +5.370 |
edi | +5.300 |
Neg logit contributions | |
robe | -13.543 |
Spoiler | -10.523 |
ンジ | -10.201 |
Crate | -10.044 |
taboola | -10.017 |
Token | were |
Token ID | 547 |
Feature activation | +2.45e-01 |
Pos logit contributions | |
rele | +0.921 |
bia | +0.909 |
constitu | +0.901 |
ibaba | +0.873 |
iland | +0.865 |
Neg logit contributions | |
robe | -1.504 |
Spoiler | -1.189 |
ンジ | -1.122 |
taboola | -1.110 |
hib | -1.098 |
Token | carried |
Token ID | 5281 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
rele | +1.064 |
bia | +1.054 |
constitu | +1.044 |
ibaba | +1.016 |
iland | +1.011 |
Neg logit contributions | |
robe | -1.427 |
Spoiler | -1.100 |
ンジ | -1.036 |
taboola | -1.018 |
hib | -1.010 |
Token | out |
Token ID | 503 |
Feature activation | -3.41e-02 |
Pos logit contributions | |
robe | +3.507 |
Spoiler | +2.726 |
ンジ | +2.589 |
taboola | +2.535 |
ヤ | +2.495 |
Neg logit contributions | |
bia | -2.498 |
rele | -2.470 |
constitu | -2.454 |
ibaba | -2.425 |
ilial | -2.348 |
Token | to |
Token ID | 284 |
Feature activation | +3.16e-01 |
Pos logit contributions | |
robe | +0.210 |
Spoiler | +0.166 |
ンジ | +0.155 |
taboola | +0.154 |
hib | +0.153 |
Neg logit contributions | |
bia | -0.135 |
rele | -0.135 |
constitu | -0.134 |
ibaba | -0.131 |
iland | -0.128 |
Token | develops |
Token ID | 21126 |
Feature activation | +5.28e-01 |
Pos logit contributions | |
rele | +1.721 |
bia | +1.699 |
constitu | +1.691 |
ibaba | +1.635 |
iland | +1.635 |
Neg logit contributions | |
robe | -1.595 |
Spoiler | -1.163 |
ンジ | -1.082 |
taboola | -1.055 |
ヤ | -1.035 |
Token | in |
Token ID | 287 |
Feature activation | +1.42e+00 |
Pos logit contributions | |
rele | +1.795 |
bia | +1.733 |
constitu | +1.669 |
iland | +1.632 |
ibaba | +1.625 |
Neg logit contributions | |
robe | -3.594 |
Spoiler | -2.869 |
ンジ | -2.783 |
taboola | -2.728 |
ヤ | -2.697 |
Token | patients |
Token ID | 3871 |
Feature activation | +6.86e-01 |
Pos logit contributions | |
rele | +3.844 |
numbers | +3.459 |
bia | +3.416 |
edi | +3.336 |
constitu | +3.300 |
Neg logit contributions | |
robe | -10.527 |
Spoiler | -8.731 |
ンジ | -8.423 |
ヤ | -8.248 |
taboola | -8.165 |
Token | aged |
Token ID | 9722 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
rele | +1.867 |
bia | +1.764 |
constitu | +1.715 |
edi | +1.627 |
iland | +1.618 |
Neg logit contributions | |
robe | -5.182 |
Spoiler | -4.271 |
ンジ | -4.148 |
taboola | -4.032 |
ヤ | -4.007 |
Token | 12 |
Token ID | 1105 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
rele | +1.114 |
bia | +1.077 |
constitu | +1.064 |
iland | +1.027 |
ibaba | +1.022 |
Neg logit contributions | |
robe | -2.026 |
Spoiler | -1.613 |
ンジ | -1.540 |
taboola | -1.512 |
ヤ | -1.497 |
Token | to |
Token ID | 284 |
Feature activation | +1.96e+00 |
Pos logit contributions | |
rele | +3.448 |
bia | +3.176 |
numbers | +3.077 |
iland | +3.063 |
edi | +3.059 |
Neg logit contributions | |
robe | -7.451 |
Spoiler | -6.012 |
ンジ | -5.863 |
ヤ | -5.644 |
taboola | -5.568 |
Token | 50 |
Token ID | 2026 |
Feature activation | +6.38e-01 |
Pos logit contributions | |
rele | +3.675 |
numbers | +3.272 |
number | +3.148 |
bia | +3.045 |
learning | +3.043 |
Neg logit contributions | |
robe | -15.524 |
Spoiler | -12.770 |
ンジ | -12.738 |
taboola | -12.367 |
ヤ | -12.275 |
Token | , |
Token ID | 11 |
Feature activation | +6.89e-01 |
Pos logit contributions | |
rele | +2.361 |
bia | +2.325 |
edi | +2.200 |
iland | +2.183 |
constitu | +2.163 |
Neg logit contributions | |
robe | -4.072 |
Spoiler | -3.217 |
ンジ | -3.100 |
taboola | -3.042 |
ヤ | -2.986 |
Token | characterized |
Token ID | 16264 |
Feature activation | +1.39e-01 |
Pos logit contributions | |
rele | +2.872 |
bia | +2.818 |
constitu | +2.774 |
iland | +2.691 |
ibaba | +2.668 |
Neg logit contributions | |
robe | -4.155 |
Spoiler | -3.222 |
ンジ | -3.060 |
taboola | -2.994 |
Doors | -2.968 |
Token | by |
Token ID | 416 |
Feature activation | +5.31e-01 |
Pos logit contributions | |
rele | +0.558 |
bia | +0.547 |
constitu | +0.538 |
iland | +0.521 |
ibaba | +0.520 |
Neg logit contributions | |
robe | -0.864 |
Spoiler | -0.678 |
ンジ | -0.642 |
taboola | -0.632 |
hib | -0.624 |
Token | multiple |
Token ID | 3294 |
Feature activation | +3.40e-01 |
Pos logit contributions | |
rele | +2.033 |
bia | +1.956 |
constitu | +1.936 |
ibaba | +1.866 |
iland | +1.843 |
Neg logit contributions | |
robe | -3.364 |
Spoiler | -2.685 |
ンジ | -2.558 |
taboola | -2.507 |
ヤ | -2.498 |
Token | animal |
Token ID | 5044 |
Feature activation | +7.89e-02 |
Pos logit contributions | |
robe | +1.997 |
Spoiler | +1.543 |
ンジ | +1.467 |
taboola | +1.462 |
hib | +1.442 |
Neg logit contributions | |
rele | -1.429 |
constitu | -1.426 |
bia | -1.425 |
ibaba | -1.380 |
iland | -1.380 |
Token | in |
Token ID | 287 |
Feature activation | -2.92e-01 |
Pos logit contributions | |
rele | +0.310 |
bia | +0.308 |
constitu | +0.305 |
ibaba | +0.295 |
iland | +0.294 |
Neg logit contributions | |
robe | -0.492 |
Spoiler | -0.387 |
ンジ | -0.365 |
taboola | -0.361 |
hib | -0.359 |
Token | the |
Token ID | 262 |
Feature activation | -1.26e-01 |
Pos logit contributions | |
robe | +1.734 |
Spoiler | +1.365 |
ンジ | +1.279 |
hib | +1.275 |
taboola | +1.266 |
Neg logit contributions | |
rele | -1.199 |
bia | -1.199 |
constitu | -1.195 |
ibaba | -1.152 |
iland | -1.145 |
Token | house |
Token ID | 2156 |
Feature activation | -1.60e-01 |
Pos logit contributions | |
robe | +0.779 |
Spoiler | +0.622 |
ンジ | +0.578 |
Doors | +0.576 |
taboola | +0.573 |
Neg logit contributions | |
rele | -0.497 |
bia | -0.497 |
constitu | -0.489 |
ibaba | -0.477 |
iland | -0.476 |
Token | I |
Token ID | 314 |
Feature activation | +3.56e-01 |
Pos logit contributions | |
robe | +0.961 |
Spoiler | +0.750 |
ンジ | +0.700 |
taboola | +0.694 |
hib | +0.691 |
Neg logit contributions | |
rele | -0.661 |
bia | -0.658 |
constitu | -0.656 |
ibaba | -0.640 |
iland | -0.631 |
Token | listened |
Token ID | 16399 |
Feature activation | +1.94e+00 |
Pos logit contributions | |
rele | +1.419 |
bia | +1.395 |
constitu | +1.383 |
iland | +1.332 |
ibaba | +1.329 |
Neg logit contributions | |
robe | -2.215 |
Spoiler | -1.739 |
ンジ | -1.646 |
taboola | -1.622 |
hib | -1.603 |
Token | to |
Token ID | 284 |
Feature activation | +9.16e-01 |
Pos logit contributions | |
rele | +6.160 |
constitu | +5.525 |
iland | +5.452 |
bia | +5.370 |
edi | +5.306 |
Neg logit contributions | |
robe | -12.761 |
Spoiler | -10.037 |
ンジ | -9.835 |
taboola | -9.619 |
ヤ | -9.563 |
Token | Barbara |
Token ID | 16685 |
Feature activation | +4.29e-01 |
Pos logit contributions | |
rele | +3.505 |
constitu | +3.336 |
bia | +3.322 |
iland | +3.190 |
edi | +3.153 |
Neg logit contributions | |
robe | -5.824 |
Spoiler | -4.541 |
ンジ | -4.395 |
taboola | -4.321 |
ヤ | -4.240 |
Token | � |
Token ID | 447 |
Feature activation | -5.01e-01 |
Pos logit contributions | |
rele | +1.829 |
bia | +1.801 |
constitu | +1.787 |
iland | +1.724 |
ibaba | +1.722 |
Neg logit contributions | |
robe | -2.552 |
Spoiler | -1.978 |
ンジ | -1.865 |
taboola | -1.837 |
hib | -1.813 |
Token | � |
Token ID | 247 |
Feature activation | +2.03e-02 |
Pos logit contributions | |
robe | +2.177 |
Spoiler | +1.503 |
taboola | +1.389 |
ンジ | +1.347 |
roo | +1.324 |
Neg logit contributions | |
bia | -2.902 |
rele | -2.891 |
constitu | -2.867 |
iland | -2.815 |
ibaba | -2.789 |
Token | s |
Token ID | 82 |
Feature activation | +9.48e-02 |
Pos logit contributions | |
rele | +0.082 |
constitu | +0.081 |
bia | +0.081 |
ibaba | +0.078 |
iland | +0.077 |
Neg logit contributions | |
robe | -0.124 |
Spoiler | -0.097 |
ンジ | -0.091 |
taboola | -0.091 |
hib | -0.090 |
Token | : |
Token ID | 25 |
Feature activation | -4.21e-01 |
Pos logit contributions | |
robe | +3.470 |
Spoiler | +2.722 |
hib | +2.423 |
ンジ | +2.420 |
Doors | +2.412 |
Neg logit contributions | |
constitu | -2.793 |
bia | -2.750 |
rele | -2.706 |
ibaba | -2.646 |
iland | -2.634 |
Token |
|
Token ID | 198 |
Feature activation | -6.36e-01 |
Pos logit contributions | |
robe | +2.379 |
Spoiler | +1.890 |
ンジ | +1.711 |
taboola | +1.693 |
Doors | +1.688 |
Neg logit contributions | |
bia | -1.848 |
constitu | -1.846 |
rele | -1.821 |
iland | -1.773 |
ibaba | -1.755 |
Token |
|
Token ID | 198 |
Feature activation | -2.31e-01 |
Pos logit contributions | |
robe | +3.411 |
Spoiler | +2.642 |
taboola | +2.387 |
hib | +2.371 |
Doors | +2.371 |
Neg logit contributions | |
constitu | -3.095 |
rele | -2.994 |
bia | -2.950 |
ibaba | -2.931 |
iland | -2.862 |
Token | < |
Token ID | 27 |
Feature activation | +1.06e+00 |
Pos logit contributions | |
robe | +1.345 |
Spoiler | +1.024 |
taboola | +0.961 |
ンジ | +0.954 |
hib | +0.952 |
Neg logit contributions | |
constitu | -1.024 |
rele | -1.018 |
bia | -0.993 |
ibaba | -0.971 |
ilial | -0.964 |
Token | body |
Token ID | 2618 |
Feature activation | -2.55e-02 |
Pos logit contributions | |
rele | +4.439 |
bia | +4.339 |
constitu | +4.263 |
ibaba | +4.112 |
iland | +4.111 |
Neg logit contributions | |
robe | -6.264 |
Spoiler | -4.807 |
ンジ | -4.554 |
taboola | -4.458 |
hib | -4.417 |
Token | ng |
Token ID | 23370 |
Feature activation | +1.93e+00 |
Pos logit contributions | |
robe | +0.154 |
Spoiler | +0.121 |
ンジ | +0.113 |
taboola | +0.113 |
hib | +0.111 |
Neg logit contributions | |
rele | -0.104 |
bia | -0.104 |
constitu | -0.103 |
ibaba | -0.100 |
iland | -0.100 |
Token | - |
Token ID | 12 |
Feature activation | +9.70e-01 |
Pos logit contributions | |
rele | +6.092 |
iland | +5.645 |
bia | +5.626 |
shooter | +5.614 |
numbers | +5.533 |
Neg logit contributions | |
robe | -12.455 |
Spoiler | -10.027 |
ンジ | -9.447 |
taboola | -9.363 |
ヤ | -9.316 |
Token | controller |
Token ID | 36500 |
Feature activation | +4.72e-02 |
Pos logit contributions | |
rele | +3.840 |
bia | +3.754 |
constitu | +3.661 |
iland | +3.589 |
edi | +3.526 |
Neg logit contributions | |
robe | -6.026 |
Spoiler | -4.749 |
ンジ | -4.519 |
taboola | -4.396 |
ヤ | -4.370 |
Token | =" |
Token ID | 2625 |
Feature activation | +6.58e-01 |
Pos logit contributions | |
bia | +0.206 |
rele | +0.205 |
constitu | +0.204 |
ibaba | +0.199 |
iland | +0.198 |
Neg logit contributions | |
robe | -0.272 |
Spoiler | -0.211 |
ンジ | -0.197 |
taboola | -0.193 |
hib | -0.193 |
Token | Phone |
Token ID | 6132 |
Feature activation | +5.16e-01 |
Pos logit contributions | |
rele | +3.019 |
bia | +2.977 |
constitu | +2.929 |
ibaba | +2.847 |
iland | +2.846 |
Neg logit contributions | |
robe | -3.644 |
Spoiler | -2.785 |
ンジ | -2.620 |
taboola | -2.539 |
Doors | -2.528 |
Token | List |
Token ID | 8053 |
Feature activation | +3.54e-02 |
Pos logit contributions | |
rele | +2.169 |
bia | +2.135 |
constitu | +2.110 |
iland | +2.047 |
ibaba | +2.038 |
Neg logit contributions | |
robe | -3.080 |
Spoiler | -2.397 |
ンジ | -2.265 |
taboola | -2.220 |
hib | -2.199 |
Token | regain |
Token ID | 21509 |
Feature activation | +5.50e-01 |
Pos logit contributions | |
rele | +2.310 |
bia | +2.249 |
constitu | +2.231 |
ibaba | +2.147 |
iland | +2.142 |
Neg logit contributions | |
robe | -3.713 |
Spoiler | -2.913 |
ンジ | -2.780 |
taboola | -2.743 |
ヤ | -2.695 |
Token | the |
Token ID | 262 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
rele | +2.387 |
bia | +2.333 |
constitu | +2.320 |
iland | +2.232 |
ibaba | +2.231 |
Neg logit contributions | |
robe | -3.233 |
Spoiler | -2.489 |
ンジ | -2.356 |
taboola | -2.326 |
hib | -2.285 |
Token | honor |
Token ID | 7522 |
Feature activation | +8.29e-01 |
Pos logit contributions | |
bia | +0.902 |
rele | +0.902 |
constitu | +0.892 |
ibaba | +0.871 |
iland | +0.868 |
Neg logit contributions | |
robe | -1.436 |
Spoiler | -1.143 |
ンジ | -1.069 |
Doors | -1.050 |
taboola | -1.047 |
Token | he |
Token ID | 339 |
Feature activation | +9.44e-01 |
Pos logit contributions | |
rele | +2.629 |
bia | +2.556 |
constitu | +2.516 |
iland | +2.389 |
ibaba | +2.388 |
Neg logit contributions | |
robe | -5.791 |
Spoiler | -4.673 |
ンジ | -4.483 |
taboola | -4.448 |
ヤ | -4.379 |
Token | had |
Token ID | 550 |
Feature activation | +1.22e+00 |
Pos logit contributions | |
rele | +3.481 |
bia | +3.328 |
constitu | +3.241 |
iland | +3.168 |
edi | +3.105 |
Neg logit contributions | |
robe | -6.138 |
Spoiler | -4.893 |
ンジ | -4.687 |
taboola | -4.651 |
ヤ | -4.598 |
Token | lost |
Token ID | 2626 |
Feature activation | +1.92e+00 |
Pos logit contributions | |
rele | +4.330 |
bia | +3.977 |
constitu | +3.907 |
iland | +3.758 |
edi | +3.719 |
Neg logit contributions | |
robe | -8.110 |
Spoiler | -6.442 |
taboola | -6.269 |
ンジ | -6.258 |
ヤ | -6.151 |
Token | when |
Token ID | 618 |
Feature activation | +3.70e-01 |
Pos logit contributions | |
rele | +6.569 |
constitu | +5.845 |
bia | +5.809 |
learning | +5.775 |
edi | +5.614 |
Neg logit contributions | |
robe | -12.103 |
taboola | -9.736 |
Spoiler | -9.587 |
ンジ | -9.564 |
ヤ | -9.470 |
Token | he |
Token ID | 339 |
Feature activation | -9.22e-02 |
Pos logit contributions | |
rele | +1.383 |
bia | +1.365 |
constitu | +1.352 |
iland | +1.298 |
ibaba | +1.296 |
Neg logit contributions | |
robe | -2.387 |
Spoiler | -1.898 |
ンジ | -1.795 |
taboola | -1.766 |
hib | -1.755 |
Token | murdered |
Token ID | 12864 |
Feature activation | +7.65e-03 |
Pos logit contributions | |
robe | +0.578 |
Spoiler | +0.458 |
ンジ | +0.429 |
taboola | +0.417 |
Doors | +0.416 |
Neg logit contributions | |
constitu | -0.354 |
bia | -0.354 |
rele | -0.352 |
ibaba | -0.344 |
iland | -0.339 |
Token | Yoshi |
Token ID | 38936 |
Feature activation | -1.76e-01 |
Pos logit contributions | |
bia | +0.035 |
rele | +0.035 |
constitu | +0.034 |
ibaba | +0.033 |
iland | +0.033 |
Neg logit contributions | |
robe | -0.043 |
Spoiler | -0.033 |
ンジ | -0.031 |
hib | -0.030 |
Doors | -0.030 |
Token | . |
Token ID | 13 |
Feature activation | -4.68e-02 |
Pos logit contributions | |
robe | +1.030 |
Spoiler | +0.807 |
ンジ | +0.746 |
hib | +0.733 |
Doors | +0.729 |
Neg logit contributions | |
constitu | -0.756 |
bia | -0.755 |
rele | -0.754 |
ibaba | -0.730 |
iland | -0.721 |
Token | : |
Token ID | 25 |
Feature activation | -4.69e-02 |
Pos logit contributions | |
rele | +2.015 |
bia | +1.951 |
constitu | +1.939 |
iland | +1.874 |
ibaba | +1.871 |
Neg logit contributions | |
robe | -3.306 |
Spoiler | -2.600 |
ンジ | -2.476 |
taboola | -2.429 |
ヤ | -2.410 |
Token | a |
Token ID | 257 |
Feature activation | -1.68e-02 |
Pos logit contributions | |
robe | +0.278 |
Spoiler | +0.217 |
ンジ | +0.201 |
Doors | +0.198 |
taboola | +0.198 |
Neg logit contributions | |
bia | -0.199 |
rele | -0.199 |
constitu | -0.196 |
iland | -0.190 |
ibaba | -0.188 |
Token | young |
Token ID | 1862 |
Feature activation | -1.85e-01 |
Pos logit contributions | |
robe | +0.105 |
Spoiler | +0.083 |
ンジ | +0.078 |
taboola | +0.077 |
Doors | +0.076 |
Neg logit contributions | |
bia | -0.066 |
rele | -0.066 |
constitu | -0.065 |
iland | -0.062 |
ibaba | -0.062 |
Token | professional |
Token ID | 4708 |
Feature activation | +6.60e-01 |
Pos logit contributions | |
robe | +1.116 |
Spoiler | +0.871 |
ンジ | +0.810 |
Doors | +0.796 |
hib | +0.788 |
Neg logit contributions | |
bia | -0.767 |
rele | -0.760 |
constitu | -0.754 |
ibaba | -0.734 |
iland | -0.728 |
Token | qu |
Token ID | 627 |
Feature activation | +1.64e-01 |
Pos logit contributions | |
rele | +2.688 |
bia | +2.593 |
constitu | +2.584 |
ibaba | +2.493 |
iland | +2.475 |
Neg logit contributions | |
robe | -4.035 |
Spoiler | -3.139 |
ンジ | -3.010 |
taboola | -2.963 |
hib | -2.928 |
Token | its |
Token ID | 896 |
Feature activation | +1.87e+00 |
Pos logit contributions | |
rele | +0.642 |
bia | +0.632 |
constitu | +0.630 |
ibaba | +0.601 |
iland | +0.600 |
Neg logit contributions | |
robe | -1.037 |
Spoiler | -0.813 |
ンジ | -0.767 |
taboola | -0.758 |
hib | -0.754 |
Token | his |
Token ID | 465 |
Feature activation | +4.02e-01 |
Pos logit contributions | |
rele | +7.153 |
constitu | +6.501 |
bia | +6.465 |
ibaba | +6.441 |
learning | +6.396 |
Neg logit contributions | |
robe | -10.994 |
Spoiler | -8.572 |
taboola | -8.447 |
ヤ | -8.357 |
ンジ | -8.333 |
Token | job |
Token ID | 1693 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
rele | +1.186 |
bia | +1.163 |
constitu | +1.150 |
iland | +1.090 |
ibaba | +1.089 |
Neg logit contributions | |
robe | -2.911 |
Spoiler | -2.375 |
ンジ | -2.267 |
taboola | -2.241 |
hib | -2.220 |
Token | , |
Token ID | 11 |
Feature activation | +3.46e-01 |
Pos logit contributions | |
rele | +4.415 |
bia | +4.173 |
constitu | +4.081 |
ibaba | +3.965 |
edi | +3.914 |
Neg logit contributions | |
robe | -7.999 |
Spoiler | -6.353 |
ンジ | -6.223 |
taboola | -6.147 |
ヤ | -6.040 |
Token | becomes |
Token ID | 4329 |
Feature activation | -1.95e-01 |
Pos logit contributions | |
rele | +1.358 |
bia | +1.351 |
constitu | +1.337 |
iland | +1.286 |
ibaba | +1.283 |
Neg logit contributions | |
robe | -2.166 |
Spoiler | -1.704 |
ンジ | -1.608 |
taboola | -1.586 |
hib | -1.568 |
Token | a |
Token ID | 257 |
Feature activation | -6.31e-02 |
Pos logit contributions | |
robe | +1.188 |
Spoiler | +0.938 |
ンジ | +0.876 |
taboola | +0.861 |
hib | +0.857 |
Neg logit contributions | |
bia | -0.793 |
rele | -0.788 |
constitu | -0.780 |
iland | -0.756 |
ibaba | -0.751 |
Token | zones |
Token ID | 14123 |
Feature activation | +4.72e-01 |
Pos logit contributions | |
rele | +0.923 |
bia | +0.915 |
constitu | +0.903 |
iland | +0.876 |
ibaba | +0.874 |
Neg logit contributions | |
robe | -1.454 |
Spoiler | -1.146 |
ンジ | -1.081 |
taboola | -1.064 |
Doors | -1.056 |
Token | in |
Token ID | 287 |
Feature activation | -6.44e-02 |
Pos logit contributions | |
rele | +1.970 |
bia | +1.938 |
constitu | +1.921 |
iland | +1.851 |
ibaba | +1.849 |
Neg logit contributions | |
robe | -2.845 |
Spoiler | -2.210 |
ンジ | -2.094 |
taboola | -2.061 |
hib | -2.033 |
Token | passport |
Token ID | 17981 |
Feature activation | +2.03e-02 |
Pos logit contributions | |
robe | +0.390 |
Spoiler | +0.306 |
ンジ | +0.286 |
taboola | +0.284 |
Doors | +0.282 |
Neg logit contributions | |
rele | -0.263 |
bia | -0.261 |
constitu | -0.259 |
iland | -0.251 |
ibaba | -0.249 |
Token | images |
Token ID | 4263 |
Feature activation | +4.49e-01 |
Pos logit contributions | |
rele | +0.077 |
bia | +0.077 |
constitu | +0.076 |
iland | +0.073 |
ibaba | +0.073 |
Neg logit contributions | |
robe | -0.129 |
Spoiler | -0.102 |
ンジ | -0.096 |
taboola | -0.095 |
Doors | -0.094 |
Token | # |
Token ID | 1303 |
Feature activation | +1.20e+00 |
Pos logit contributions | |
rele | +1.911 |
bia | +1.875 |
constitu | +1.861 |
ibaba | +1.794 |
iland | +1.794 |
Neg logit contributions | |
robe | -2.667 |
Spoiler | -2.059 |
ンジ | -1.958 |
taboola | -1.920 |
hib | -1.894 |
Token | import |
Token ID | 1330 |
Feature activation | +1.87e+00 |
Pos logit contributions | |
rele | +5.471 |
bia | +5.319 |
constitu | +5.234 |
ibaba | +5.142 |
edi | +5.094 |
Neg logit contributions | |
robe | -6.523 |
Spoiler | -4.912 |
ンジ | -4.686 |
taboola | -4.557 |
ヤ | -4.536 |
Token | the |
Token ID | 262 |
Feature activation | +3.92e-01 |
Pos logit contributions | |
rele | +7.723 |
constitu | +7.308 |
bia | +7.178 |
numbers | +7.161 |
ibaba | +7.108 |
Neg logit contributions | |
robe | -10.698 |
Spoiler | -8.037 |
ヤ | -7.625 |
ンジ | -7.604 |
hib | -7.491 |
Token | necessary |
Token ID | 3306 |
Feature activation | +7.51e-01 |
Pos logit contributions | |
rele | +1.552 |
bia | +1.533 |
constitu | +1.517 |
iland | +1.462 |
ibaba | +1.460 |
Neg logit contributions | |
robe | -2.443 |
Spoiler | -1.922 |
ンジ | -1.816 |
taboola | -1.793 |
hib | -1.770 |
Token | packages |
Token ID | 10392 |
Feature activation | +8.05e-01 |
Pos logit contributions | |
rele | +2.927 |
constitu | +2.831 |
bia | +2.820 |
iland | +2.680 |
ibaba | +2.677 |
Neg logit contributions | |
robe | -4.752 |
Spoiler | -3.713 |
ンジ | -3.551 |
taboola | -3.491 |
ヤ | -3.457 |
Token | from |
Token ID | 422 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
rele | +3.206 |
bia | +3.101 |
constitu | +3.100 |
ibaba | +2.962 |
iland | +2.952 |
Neg logit contributions | |
robe | -5.014 |
Spoiler | -3.890 |
ンジ | -3.740 |
taboola | -3.652 |
ヤ | -3.634 |
Token | im |
Token ID | 545 |
Feature activation | +8.40e-03 |
Pos logit contributions | |
rele | +5.995 |
constitu | +5.735 |
bia | +5.715 |
ibaba | +5.529 |
shooter | +5.353 |
Neg logit contributions | |
robe | -7.821 |
Spoiler | -5.893 |
ンジ | -5.657 |
ヤ | -5.489 |
hib | -5.472 |
Token | the |
Token ID | 262 |
Feature activation | +2.77e-01 |
Pos logit contributions | |
rele | +1.375 |
bia | +1.349 |
constitu | +1.341 |
iland | +1.286 |
ibaba | +1.279 |
Neg logit contributions | |
robe | -2.219 |
Spoiler | -1.739 |
ンジ | -1.659 |
taboola | -1.632 |
ヤ | -1.616 |
Token | BBC |
Token ID | 7823 |
Feature activation | +3.68e-01 |
Pos logit contributions | |
rele | +1.115 |
bia | +1.098 |
constitu | +1.090 |
ibaba | +1.050 |
iland | +1.049 |
Neg logit contributions | |
robe | -1.712 |
Spoiler | -1.342 |
ンジ | -1.266 |
taboola | -1.250 |
hib | -1.236 |
Token | that |
Token ID | 326 |
Feature activation | +2.71e-01 |
Pos logit contributions | |
rele | +1.360 |
bia | +1.336 |
constitu | +1.327 |
ibaba | +1.272 |
iland | +1.270 |
Neg logit contributions | |
robe | -2.392 |
Spoiler | -1.901 |
ンジ | -1.800 |
taboola | -1.776 |
hib | -1.761 |
Token | the |
Token ID | 262 |
Feature activation | +4.39e-01 |
Pos logit contributions | |
rele | +1.065 |
bia | +1.051 |
constitu | +1.040 |
ibaba | +1.008 |
iland | +1.001 |
Neg logit contributions | |
robe | -1.691 |
Spoiler | -1.337 |
ンジ | -1.259 |
taboola | -1.240 |
hib | -1.233 |
Token | overnight |
Token ID | 13417 |
Feature activation | +6.99e-01 |
Pos logit contributions | |
rele | +1.743 |
bia | +1.717 |
constitu | +1.703 |
ibaba | +1.642 |
iland | +1.640 |
Neg logit contributions | |
robe | -2.733 |
Spoiler | -2.150 |
ンジ | -2.032 |
taboola | -2.002 |
hib | -1.981 |
Token | air |
Token ID | 1633 |
Feature activation | +1.86e+00 |
Pos logit contributions | |
rele | +2.730 |
bia | +2.666 |
constitu | +2.650 |
iland | +2.530 |
ibaba | +2.520 |
Neg logit contributions | |
robe | -4.404 |
Spoiler | -3.461 |
ンジ | -3.282 |
taboola | -3.251 |
hib | -3.196 |
Token | strikes |
Token ID | 8956 |
Feature activation | +6.15e-01 |
Pos logit contributions | |
bia | +5.447 |
rele | +5.395 |
numbers | +4.964 |
edi | +4.845 |
constitu | +4.838 |
Neg logit contributions | |
robe | -12.737 |
Spoiler | -10.201 |
ンジ | -9.753 |
taboola | -9.621 |
Doors | -9.586 |
Token | had |
Token ID | 550 |
Feature activation | +2.72e-01 |
Pos logit contributions | |
rele | +2.370 |
bia | +2.326 |
constitu | +2.297 |
iland | +2.217 |
ibaba | +2.189 |
Neg logit contributions | |
robe | -3.912 |
Spoiler | -3.079 |
ンジ | -2.926 |
taboola | -2.883 |
ヤ | -2.854 |
Token | been |
Token ID | 587 |
Feature activation | +3.47e-01 |
Pos logit contributions | |
rele | +1.003 |
bia | +0.992 |
constitu | +0.982 |
ibaba | +0.944 |
iland | +0.941 |
Neg logit contributions | |
robe | -1.763 |
Spoiler | -1.402 |
ンジ | -1.329 |
taboola | -1.311 |
hib | -1.300 |
Token | extended |
Token ID | 7083 |
Feature activation | -8.58e-02 |
Pos logit contributions | |
rele | +1.524 |
bia | +1.503 |
constitu | +1.491 |
ibaba | +1.440 |
iland | +1.440 |
Neg logit contributions | |
robe | -2.016 |
Spoiler | -1.553 |
ンジ | -1.461 |
taboola | -1.437 |
hib | -1.420 |
Token | to |
Token ID | 284 |
Feature activation | +5.06e-01 |
Pos logit contributions | |
robe | +0.521 |
Spoiler | +0.409 |
ンジ | +0.380 |
taboola | +0.377 |
Doors | +0.375 |
Neg logit contributions | |
rele | -0.349 |
constitu | -0.348 |
bia | -0.348 |
ibaba | -0.336 |
iland | -0.330 |
Token | ( |
Token ID | 357 |
Feature activation | +1.68e+00 |
Pos logit contributions | |
rele | +4.469 |
bia | +4.208 |
edi | +4.184 |
iland | +4.105 |
numbers | +4.004 |
Neg logit contributions | |
robe | -8.843 |
Spoiler | -6.906 |
ンジ | -6.876 |
taboola | -6.724 |
ヤ | -6.601 |
Token | args |
Token ID | 26498 |
Feature activation | -5.56e-02 |
Pos logit contributions | |
rele | +4.053 |
numbers | +3.634 |
constitu | +3.612 |
edi | +3.609 |
bia | +3.447 |
Neg logit contributions | |
robe | -12.935 |
Spoiler | -10.358 |
ンジ | -10.220 |
taboola | -10.084 |
hib | -10.060 |
Token | [ |
Token ID | 685 |
Feature activation | +9.15e-01 |
Pos logit contributions | |
robe | +0.296 |
Spoiler | +0.226 |
ンジ | +0.206 |
taboola | +0.205 |
Doors | +0.205 |
Neg logit contributions | |
constitu | -0.271 |
rele | -0.269 |
bia | -0.267 |
ibaba | -0.259 |
iland | -0.256 |
Token | " |
Token ID | 366 |
Feature activation | +5.56e-01 |
Pos logit contributions | |
rele | +2.999 |
constitu | +2.904 |
bia | +2.899 |
edi | +2.759 |
iland | +2.744 |
Neg logit contributions | |
robe | -6.274 |
Spoiler | -5.052 |
ンジ | -4.841 |
taboola | -4.807 |
ヤ | -4.720 |
Token | images |
Token ID | 17566 |
Feature activation | +7.77e-01 |
Pos logit contributions | |
rele | +1.187 |
bia | +1.156 |
constitu | +1.121 |
iland | +1.054 |
ibaba | +1.052 |
Neg logit contributions | |
robe | -4.474 |
Spoiler | -3.742 |
ンジ | -3.599 |
taboola | -3.545 |
ヤ | -3.518 |
Token | " |
Token ID | 1 |
Feature activation | +1.83e+00 |
Pos logit contributions | |
rele | +2.713 |
bia | +2.633 |
iland | +2.559 |
edi | +2.550 |
ibaba | +2.509 |
Neg logit contributions | |
robe | -5.157 |
Spoiler | -4.050 |
ンジ | -3.983 |
taboola | -3.799 |
ヤ | -3.777 |
Token | ] |
Token ID | 2361 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
rele | +3.646 |
edi | +3.609 |
iland | +3.424 |
bia | +3.394 |
aning | +3.352 |
Neg logit contributions | |
robe | -14.111 |
Spoiler | -11.536 |
ンジ | -11.416 |
taboola | -11.192 |
hib | -10.932 |
Token | ) |
Token ID | 1267 |
Feature activation | +1.29e+00 |
Pos logit contributions | |
rele | +2.286 |
constitu | +2.141 |
bia | +2.082 |
edi | +1.960 |
iland | +1.936 |
Neg logit contributions | |
robe | -8.980 |
Spoiler | -7.438 |
ンジ | -7.399 |
taboola | -7.208 |
ヤ | -7.083 |
Token | : |
Token ID | 1058 |
Feature activation | +4.18e-01 |
Pos logit contributions | |
rele | +4.353 |
constitu | +4.155 |
bia | +4.128 |
edi | +3.951 |
ibaba | +3.937 |
Neg logit contributions | |
robe | -8.619 |
Spoiler | -6.839 |
ンジ | -6.750 |
ヤ | -6.586 |
taboola | -6.576 |
Token | # |
Token ID | 1303 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
rele | +1.514 |
bia | +1.470 |
constitu | +1.454 |
iland | +1.391 |
ibaba | +1.390 |
Neg logit contributions | |
robe | -2.759 |
Spoiler | -2.188 |
ンジ | -2.095 |
taboola | -2.059 |
ヤ | -2.053 |
Token | load |
Token ID | 3440 |
Feature activation | +5.31e-01 |
Pos logit contributions | |
rele | +4.001 |
constitu | +3.628 |
bia | +3.587 |
ibaba | +3.555 |
edi | +3.443 |
Neg logit contributions | |
robe | -7.843 |
Spoiler | -6.292 |
ンジ | -6.092 |
ヤ | -5.985 |
taboola | -5.979 |
Token | medium |
Token ID | 7090 |
Feature activation | +7.18e-01 |
Pos logit contributions | |
rele | +4.014 |
bia | +3.793 |
constitu | +3.782 |
shooter | +3.634 |
numbers | +3.618 |
Neg logit contributions | |
robe | -5.564 |
Spoiler | -4.229 |
ンジ | -4.038 |
roo | -3.909 |
taboola | -3.897 |
Token | - |
Token ID | 12 |
Feature activation | +7.09e-01 |
Pos logit contributions | |
rele | +2.629 |
bia | +2.501 |
constitu | +2.436 |
iland | +2.371 |
edi | +2.365 |
Neg logit contributions | |
robe | -4.695 |
Spoiler | -3.668 |
ンジ | -3.575 |
taboola | -3.440 |
ヤ | -3.414 |
Token | high |
Token ID | 8929 |
Feature activation | +9.84e-01 |
Pos logit contributions | |
bia | +2.113 |
rele | +2.102 |
iland | +1.963 |
edi | +1.949 |
constitu | +1.897 |
Neg logit contributions | |
robe | -5.024 |
Spoiler | -4.098 |
ンジ | -3.977 |
Doors | -3.878 |
ヤ | -3.855 |
Token | heat |
Token ID | 4894 |
Feature activation | +6.15e-01 |
Pos logit contributions | |
rele | +2.081 |
bia | +1.873 |
constitu | +1.730 |
shooter | +1.717 |
numbers | +1.644 |
Neg logit contributions | |
robe | -7.868 |
Spoiler | -6.499 |
ンジ | -6.452 |
ヤ | -6.208 |
taboola | -6.187 |
Token | . |
Token ID | 13 |
Feature activation | +5.92e-01 |
Pos logit contributions | |
rele | +2.310 |
bia | +2.246 |
iland | +2.147 |
constitu | +2.128 |
edi | +2.100 |
Neg logit contributions | |
robe | -3.909 |
Spoiler | -3.068 |
ンジ | -2.984 |
taboola | -2.904 |
ヤ | -2.881 |
Token | Place |
Token ID | 8474 |
Feature activation | +1.83e+00 |
Pos logit contributions | |
rele | +3.172 |
bia | +3.101 |
constitu | +3.062 |
iland | +3.002 |
ibaba | +2.988 |
Neg logit contributions | |
robe | -2.823 |
Spoiler | -1.972 |
ンジ | -1.905 |
taboola | -1.870 |
ヤ | -1.809 |
Token | f |
Token ID | 277 |
Feature activation | +3.98e-01 |
Pos logit contributions | |
rele | +4.686 |
numbers | +3.971 |
shooter | +3.909 |
learning | +3.864 |
number | +3.808 |
Neg logit contributions | |
robe | -13.447 |
Spoiler | -10.932 |
taboola | -10.774 |
ンジ | -10.768 |
ヤ | -10.675 |
Token | rit |
Token ID | 799 |
Feature activation | -1.39e-01 |
Pos logit contributions | |
rele | +1.231 |
bia | +1.208 |
constitu | +1.162 |
iland | +1.144 |
ibaba | +1.141 |
Neg logit contributions | |
robe | -2.816 |
Spoiler | -2.293 |
ンジ | -2.198 |
taboola | -2.158 |
ヤ | -2.148 |
Token | ters |
Token ID | 1010 |
Feature activation | +4.30e-01 |
Pos logit contributions | |
robe | +0.759 |
Spoiler | +0.572 |
ンジ | +0.533 |
hib | +0.528 |
taboola | +0.527 |
Neg logit contributions | |
rele | -0.659 |
constitu | -0.656 |
bia | -0.645 |
ibaba | -0.625 |
iland | -0.623 |
Token | in |
Token ID | 287 |
Feature activation | +9.90e-01 |
Pos logit contributions | |
rele | +1.518 |
bia | +1.471 |
constitu | +1.436 |
iland | +1.401 |
ibaba | +1.387 |
Neg logit contributions | |
robe | -2.868 |
Spoiler | -2.286 |
ンジ | -2.195 |
taboola | -2.161 |
ヤ | -2.137 |
Token | the |
Token ID | 262 |
Feature activation | +7.20e-01 |
Pos logit contributions | |
rele | +3.168 |
bia | +2.935 |
numbers | +2.846 |
iland | +2.828 |
edi | +2.798 |
Neg logit contributions | |
robe | -6.741 |
Spoiler | -5.379 |
ンジ | -5.217 |
ヤ | -5.182 |
taboola | -5.149 |
Token | paths |
Token ID | 13532 |
Feature activation | +6.59e-01 |
Pos logit contributions | |
rele | +3.998 |
constitu | +3.855 |
bia | +3.825 |
ibaba | +3.652 |
edi | +3.644 |
Neg logit contributions | |
robe | -6.623 |
Spoiler | -5.208 |
ンジ | -4.957 |
ヤ | -4.842 |
taboola | -4.829 |
Token | . |
Token ID | 13 |
Feature activation | +1.38e+00 |
Pos logit contributions | |
rele | +2.965 |
bia | +2.906 |
constitu | +2.892 |
iland | +2.804 |
ibaba | +2.788 |
Neg logit contributions | |
robe | -3.744 |
Spoiler | -2.866 |
ンジ | -2.719 |
taboola | -2.652 |
ヤ | -2.620 |
Token | list |
Token ID | 4868 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
rele | +5.014 |
bia | +4.848 |
iland | +4.631 |
constitu | +4.590 |
edi | +4.555 |
Neg logit contributions | |
robe | -8.530 |
Spoiler | -7.094 |
ンジ | -6.615 |
ヤ | -6.527 |
taboola | -6.471 |
Token | _ |
Token ID | 62 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
rele | +3.618 |
bia | +3.489 |
constitu | +3.442 |
iland | +3.413 |
edi | +3.363 |
Neg logit contributions | |
robe | -8.077 |
Spoiler | -6.639 |
ンジ | -6.403 |
taboola | -6.276 |
ヤ | -6.133 |
Token | images |
Token ID | 17566 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
rele | +1.237 |
constitu | +1.236 |
bia | +1.206 |
ibaba | +1.167 |
iland | +1.154 |
Neg logit contributions | |
robe | -1.768 |
Spoiler | -1.342 |
taboola | -1.274 |
ンジ | -1.273 |
hib | -1.266 |
Token | ( |
Token ID | 7 |
Feature activation | +1.82e+00 |
Pos logit contributions | |
rele | +4.104 |
bia | +3.998 |
constitu | +3.920 |
iland | +3.842 |
edi | +3.809 |
Neg logit contributions | |
robe | -8.273 |
Spoiler | -6.687 |
ンジ | -6.508 |
taboola | -6.371 |
ヤ | -6.225 |
Token | args |
Token ID | 22046 |
Feature activation | +4.49e-01 |
Pos logit contributions | |
rele | +7.259 |
bia | +7.258 |
edi | +7.051 |
constitu | +6.909 |
iland | +6.870 |
Neg logit contributions | |
robe | -10.297 |
Spoiler | -8.350 |
ンジ | -7.624 |
ヤ | -7.409 |
Crate | -7.379 |
Token | [" |
Token ID | 14692 |
Feature activation | +6.78e-01 |
Pos logit contributions | |
rele | +1.745 |
bia | +1.711 |
constitu | +1.697 |
iland | +1.637 |
ibaba | +1.628 |
Neg logit contributions | |
robe | -2.829 |
Spoiler | -2.231 |
ンジ | -2.124 |
taboola | -2.084 |
ヤ | -2.055 |
Token | images |
Token ID | 17566 |
Feature activation | +5.38e-01 |
Pos logit contributions | |
rele | +3.016 |
bia | +2.987 |
constitu | +2.932 |
ibaba | +2.848 |
iland | +2.847 |
Neg logit contributions | |
robe | -3.860 |
Spoiler | -2.995 |
ンジ | -2.800 |
taboola | -2.713 |
ヤ | -2.698 |
Token | "] |
Token ID | 8973 |
Feature activation | +1.36e+00 |
Pos logit contributions | |
rele | +2.469 |
bia | +2.419 |
constitu | +2.415 |
ibaba | +2.327 |
iland | +2.322 |
Neg logit contributions | |
robe | -2.979 |
Spoiler | -2.275 |
ンジ | -2.170 |
taboola | -2.087 |
Doors | -2.065 |
Token | ): |
Token ID | 2599 |
Feature activation | +3.93e-01 |
Pos logit contributions | |
rele | +4.347 |
constitu | +4.155 |
bia | +4.153 |
iland | +4.004 |
edi | +3.990 |
Neg logit contributions | |
robe | -9.244 |
Spoiler | -7.343 |
ンジ | -7.231 |
taboola | -6.986 |
roo | -6.870 |
Token | the |
Token ID | 262 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
rele | +2.477 |
bia | +2.398 |
constitu | +2.395 |
ibaba | +2.287 |
iland | +2.285 |
Neg logit contributions | |
robe | -4.023 |
Spoiler | -3.148 |
ンジ | -3.025 |
taboola | -2.980 |
ヤ | -2.945 |
Token | terminal |
Token ID | 12094 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
bia | +0.848 |
rele | +0.846 |
constitu | +0.830 |
ibaba | +0.805 |
iland | +0.803 |
Neg logit contributions | |
robe | -1.499 |
Spoiler | -1.200 |
ンジ | -1.127 |
taboola | -1.114 |
Doors | -1.104 |
Token | , |
Token ID | 11 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
rele | +1.116 |
bia | +1.101 |
constitu | +1.092 |
iland | +1.053 |
ibaba | +1.052 |
Neg logit contributions | |
robe | -1.556 |
Spoiler | -1.206 |
ンジ | -1.136 |
taboola | -1.119 |
hib | -1.105 |
Token | you |
Token ID | 345 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
bia | +0.572 |
constitu | +0.564 |
rele | +0.563 |
iland | +0.542 |
ibaba | +0.539 |
Neg logit contributions | |
robe | -0.981 |
Spoiler | -0.788 |
ンジ | -0.732 |
taboola | -0.726 |
hib | -0.721 |
Token | will |
Token ID | 481 |
Feature activation | +5.22e-01 |
Pos logit contributions | |
rele | +0.803 |
bia | +0.799 |
constitu | +0.796 |
ibaba | +0.764 |
iland | +0.763 |
Neg logit contributions | |
robe | -1.118 |
Spoiler | -0.866 |
ンジ | -0.812 |
taboola | -0.803 |
hib | -0.795 |
Token | paste |
Token ID | 17008 |
Feature activation | +1.82e+00 |
Pos logit contributions | |
rele | +1.837 |
bia | +1.785 |
constitu | +1.760 |
iland | +1.693 |
ibaba | +1.687 |
Neg logit contributions | |
robe | -3.492 |
Spoiler | -2.798 |
ンジ | -2.666 |
taboola | -2.630 |
hib | -2.597 |
Token | or |
Token ID | 393 |
Feature activation | +3.96e-01 |
Pos logit contributions | |
rele | +6.217 |
constitu | +5.787 |
edi | +5.691 |
bia | +5.572 |
numbers | +5.498 |
Neg logit contributions | |
robe | -11.575 |
Spoiler | -8.905 |
taboola | -8.824 |
ンジ | -8.727 |
hib | -8.621 |
Token | type |
Token ID | 2099 |
Feature activation | +7.13e-01 |
Pos logit contributions | |
bia | +1.541 |
rele | +1.538 |
constitu | +1.534 |
iland | +1.473 |
ibaba | +1.472 |
Neg logit contributions | |
robe | -2.494 |
Spoiler | -1.970 |
ンジ | -1.845 |
taboola | -1.829 |
hib | -1.814 |
Token | in |
Token ID | 287 |
Feature activation | +5.85e-01 |
Pos logit contributions | |
rele | +2.890 |
constitu | +2.793 |
bia | +2.780 |
edi | +2.657 |
ibaba | +2.655 |
Neg logit contributions | |
robe | -4.379 |
Spoiler | -3.398 |
ンジ | -3.239 |
taboola | -3.212 |
ヤ | -3.160 |
Token | this |
Token ID | 428 |
Feature activation | -3.35e-02 |
Pos logit contributions | |
rele | +2.400 |
constitu | +2.320 |
bia | +2.318 |
ibaba | +2.224 |
iland | +2.212 |
Neg logit contributions | |
robe | -3.571 |
Spoiler | -2.770 |
ンジ | -2.635 |
taboola | -2.608 |
hib | -2.568 |
Token | as |
Token ID | 355 |
Feature activation | +5.30e-01 |
Pos logit contributions | |
robe | +0.213 |
Spoiler | +0.171 |
ンジ | +0.159 |
taboola | +0.156 |
Doors | +0.156 |
Neg logit contributions | |
bia | -0.127 |
rele | -0.126 |
constitu | -0.125 |
iland | -0.121 |
ibaba | -0.120 |
Token | ous |
Token ID | 516 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
rele | +2.767 |
bia | +2.730 |
iland | +2.587 |
edi | +2.586 |
constitu | +2.579 |
Neg logit contributions | |
robe | -5.387 |
Spoiler | -4.370 |
ンジ | -4.138 |
ヤ | -4.060 |
taboola | -4.053 |
Token | labor |
Token ID | 4827 |
Feature activation | +7.41e-01 |
Pos logit contributions | |
rele | +4.326 |
constitu | +3.958 |
bia | +3.954 |
edi | +3.749 |
numbers | +3.748 |
Neg logit contributions | |
robe | -8.422 |
Spoiler | -6.775 |
ンジ | -6.423 |
taboola | -6.370 |
ヤ | -6.324 |
Token | they |
Token ID | 484 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
rele | +2.870 |
bia | +2.779 |
constitu | +2.736 |
iland | +2.661 |
edi | +2.645 |
Neg logit contributions | |
robe | -4.675 |
Spoiler | -3.721 |
ンジ | -3.525 |
taboola | -3.451 |
ヤ | -3.429 |
Token | were |
Token ID | 547 |
Feature activation | +1.01e+00 |
Pos logit contributions | |
rele | +3.463 |
bia | +3.192 |
iland | +3.112 |
constitu | +3.036 |
edi | +3.030 |
Neg logit contributions | |
robe | -8.395 |
Spoiler | -7.011 |
ンジ | -6.689 |
taboola | -6.640 |
ヤ | -6.627 |
Token | forced |
Token ID | 4137 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
rele | +3.676 |
bia | +3.426 |
constitu | +3.419 |
iland | +3.291 |
edi | +3.243 |
Neg logit contributions | |
robe | -6.608 |
Spoiler | -5.356 |
ンジ | -5.095 |
taboola | -5.041 |
ヤ | -4.992 |
Token | to |
Token ID | 284 |
Feature activation | +1.82e+00 |
Pos logit contributions | |
rele | +3.670 |
constitu | +3.401 |
bia | +3.344 |
iland | +3.307 |
ilial | +3.192 |
Neg logit contributions | |
robe | -7.572 |
Spoiler | -6.206 |
taboola | -5.851 |
ンジ | -5.706 |
Doors | -5.682 |
Token | do |
Token ID | 466 |
Feature activation | +1.75e+00 |
Pos logit contributions | |
rele | +4.402 |
bia | +3.401 |
constitu | +3.396 |
iland | +3.368 |
edi | +3.278 |
Neg logit contributions | |
robe | -13.975 |
Spoiler | -11.933 |
taboola | -11.505 |
ンジ | -11.451 |
ヤ | -11.394 |
Token | ." |
Token ID | 526 |
Feature activation | +2.29e-01 |
Pos logit contributions | |
rele | +6.016 |
bia | +5.448 |
iland | +5.438 |
numbers | +5.343 |
constitu | +5.315 |
Neg logit contributions | |
robe | -11.008 |
Spoiler | -9.124 |
ンジ | -8.719 |
taboola | -8.612 |
ヤ | -8.588 |
Token | I |
Token ID | 40 |
Feature activation | -3.44e-01 |
Pos logit contributions | |
rele | +0.917 |
bia | +0.901 |
constitu | +0.899 |
ibaba | +0.865 |
iland | +0.857 |
Neg logit contributions | |
robe | -1.420 |
Spoiler | -1.116 |
ンジ | -1.058 |
taboola | -1.042 |
hib | -1.032 |
Token | ke |
Token ID | 365 |
Feature activation | -6.82e-04 |
Pos logit contributions | |
robe | +2.190 |
Spoiler | +1.686 |
hib | +1.610 |
ンジ | +1.599 |
taboola | +1.588 |
Neg logit contributions | |
constitu | -1.303 |
bia | -1.286 |
rele | -1.284 |
ibaba | -1.224 |
ilial | -1.210 |
Token | a |
Token ID | 64 |
Feature activation | +3.26e-01 |
Pos logit contributions | |
robe | +0.004 |
Spoiler | +0.003 |
ンジ | +0.003 |
taboola | +0.003 |
hib | +0.003 |
Neg logit contributions | |
rele | -0.003 |
constitu | -0.003 |
bia | -0.003 |
ibaba | -0.003 |
iland | -0.003 |
Token | - |
Token ID | 12 |
Feature activation | -3.02e-01 |
Pos logit contributions | |
robe | +0.977 |
Spoiler | +0.760 |
ンジ | +0.710 |
taboola | +0.696 |
Doors | +0.691 |
Neg logit contributions | |
bia | -0.727 |
rele | -0.723 |
constitu | -0.715 |
iland | -0.691 |
ibaba | -0.685 |
Token | for |
Token ID | 1640 |
Feature activation | -2.71e-01 |
Pos logit contributions | |
robe | +1.534 |
Spoiler | +1.111 |
ンジ | +1.033 |
hib | +1.031 |
taboola | +1.026 |
Neg logit contributions | |
constitu | -1.568 |
rele | -1.553 |
bia | -1.528 |
ibaba | -1.487 |
ilial | -1.478 |
Token | sp |
Token ID | 599 |
Feature activation | +1.28e-01 |
Pos logit contributions | |
robe | +1.653 |
Spoiler | +1.306 |
ンジ | +1.204 |
Doors | +1.197 |
hib | +1.194 |
Neg logit contributions | |
bia | -1.088 |
constitu | -1.062 |
rele | -1.057 |
iland | -1.034 |
ilial | -1.032 |
Token | urt |
Token ID | 3325 |
Feature activation | +3.51e-01 |
Pos logit contributions | |
rele | +0.571 |
constitu | +0.561 |
bia | +0.560 |
ibaba | +0.533 |
iland | +0.532 |
Neg logit contributions | |
robe | -0.746 |
Spoiler | -0.567 |
taboola | -0.527 |
ンジ | -0.525 |
hib | -0.523 |
Token | in |
Token ID | 287 |
Feature activation | +9.92e-02 |
Pos logit contributions | |
rele | +1.163 |
bia | +1.150 |
constitu | +1.139 |
iland | +1.089 |
ibaba | +1.082 |
Neg logit contributions | |
robe | -2.416 |
Spoiler | -1.945 |
ンジ | -1.847 |
taboola | -1.833 |
hib | -1.810 |
Token | air |
Token ID | 1633 |
Feature activation | +1.82e+00 |
Pos logit contributions | |
bia | +0.375 |
rele | +0.373 |
constitu | +0.371 |
iland | +0.360 |
ibaba | +0.356 |
Neg logit contributions | |
robe | -0.630 |
Spoiler | -0.499 |
ンジ | -0.470 |
taboola | -0.466 |
Doors | -0.462 |
Token | traffic |
Token ID | 4979 |
Feature activation | +4.36e-01 |
Pos logit contributions | |
rele | +6.970 |
bia | +6.816 |
numbers | +6.519 |
constitu | +6.499 |
ibaba | +6.376 |
Neg logit contributions | |
robe | -10.943 |
Spoiler | -8.478 |
ンジ | -8.100 |
taboola | -7.929 |
ヤ | -7.912 |
Token | fails |
Token ID | 10143 |
Feature activation | +6.13e-01 |
Pos logit contributions | |
rele | +1.723 |
bia | +1.689 |
constitu | +1.677 |
iland | +1.611 |
ibaba | +1.608 |
Neg logit contributions | |
robe | -2.727 |
Spoiler | -2.141 |
ンジ | -2.031 |
taboola | -1.997 |
hib | -1.973 |
Token | to |
Token ID | 284 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
rele | +2.342 |
bia | +2.280 |
constitu | +2.263 |
iland | +2.202 |
ibaba | +2.198 |
Neg logit contributions | |
robe | -3.905 |
Spoiler | -3.064 |
ンジ | -2.931 |
taboola | -2.865 |
ヤ | -2.833 |
Token | material |
Token ID | 2587 |
Feature activation | +2.30e-01 |
Pos logit contributions | |
robe | +1.718 |
Spoiler | +1.324 |
taboola | +1.240 |
ンジ | +1.232 |
Doors | +1.224 |
Neg logit contributions | |
constitu | -1.266 |
bia | -1.264 |
rele | -1.232 |
ibaba | -1.200 |
ilial | -1.192 |
Token | ize |
Token ID | 1096 |
Feature activation | +6.00e-02 |
Pos logit contributions | |
rele | +0.613 |
bia | +0.596 |
constitu | +0.584 |
iland | +0.555 |
ibaba | +0.553 |
Neg logit contributions | |
robe | -1.741 |
Spoiler | -1.429 |
ンジ | -1.370 |
taboola | -1.355 |
hib | -1.340 |
Token | North |
Token ID | 2258 |
Feature activation | +3.49e-02 |
Pos logit contributions | |
robe | +0.870 |
Spoiler | +0.685 |
ンジ | +0.647 |
Doors | +0.636 |
hib | +0.628 |
Neg logit contributions | |
rele | -0.595 |
bia | -0.590 |
constitu | -0.584 |
ibaba | -0.578 |
iland | -0.570 |
Token | W |
Token ID | 370 |
Feature activation | -7.82e-02 |
Pos logit contributions | |
rele | +0.162 |
bia | +0.160 |
constitu | +0.160 |
ibaba | +0.154 |
iland | +0.153 |
Neg logit contributions | |
robe | -0.194 |
Spoiler | -0.148 |
ンジ | -0.138 |
taboola | -0.136 |
hib | -0.134 |
Token | az |
Token ID | 1031 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
robe | +0.351 |
Spoiler | +0.243 |
taboola | +0.221 |
ンジ | +0.220 |
hib | +0.220 |
Neg logit contributions | |
rele | -0.447 |
constitu | -0.447 |
bia | -0.440 |
iland | -0.425 |
ibaba | -0.424 |
Token | ir |
Token ID | 343 |
Feature activation | +5.94e-01 |
Pos logit contributions | |
rele | +0.416 |
constitu | +0.411 |
bia | +0.408 |
ibaba | +0.389 |
iland | +0.389 |
Neg logit contributions | |
robe | -0.621 |
Spoiler | -0.483 |
ンジ | -0.456 |
taboola | -0.451 |
hib | -0.449 |
Token | istan |
Token ID | 4103 |
Feature activation | +2.23e-01 |
Pos logit contributions | |
rele | +2.054 |
bia | +2.050 |
constitu | +2.030 |
iland | +1.958 |
ibaba | +1.917 |
Neg logit contributions | |
robe | -3.984 |
Spoiler | -3.187 |
ンジ | -3.028 |
taboola | -3.021 |
Doors | -2.959 |
Token | air |
Token ID | 1633 |
Feature activation | +1.81e+00 |
Pos logit contributions | |
rele | +0.895 |
bia | +0.881 |
constitu | +0.873 |
ibaba | +0.846 |
iland | +0.839 |
Neg logit contributions | |
robe | -1.373 |
Spoiler | -1.078 |
ンジ | -1.018 |
taboola | -0.996 |
hib | -0.992 |
Token | raids |
Token ID | 20371 |
Feature activation | +3.45e-01 |
Pos logit contributions | |
bia | +5.474 |
rele | +5.245 |
iland | +4.981 |
edi | +4.868 |
numbers | +4.767 |
Neg logit contributions | |
robe | -12.469 |
Spoiler | -10.028 |
ンジ | -9.676 |
taboola | -9.563 |
Doors | -9.476 |
Token | . |
Token ID | 13 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
rele | +1.428 |
bia | +1.405 |
constitu | +1.387 |
iland | +1.344 |
ibaba | +1.331 |
Neg logit contributions | |
robe | -2.094 |
Spoiler | -1.635 |
ンジ | -1.547 |
taboola | -1.520 |
ヤ | -1.501 |
Token |
|
Token ID | 198 |
Feature activation | +9.20e-01 |
Pos logit contributions | |
rele | +0.540 |
bia | +0.533 |
constitu | +0.528 |
iland | +0.511 |
ibaba | +0.508 |
Neg logit contributions | |
robe | -0.781 |
Spoiler | -0.605 |
ンジ | -0.572 |
taboola | -0.564 |
hib | -0.557 |
Token |
|
Token ID | 198 |
Feature activation | +8.60e-01 |
Pos logit contributions | |
bia | +2.961 |
rele | +2.955 |
edi | +2.775 |
iland | +2.755 |
owing | +2.688 |
Neg logit contributions | |
robe | -6.183 |
Spoiler | -4.846 |
ンジ | -4.800 |
taboola | -4.654 |
ヤ | -4.637 |
Token | K |
Token ID | 42 |
Feature activation | +3.81e-01 |
Pos logit contributions | |
bia | +3.403 |
rele | +3.330 |
constitu | +3.267 |
iland | +3.233 |
edi | +3.192 |
Neg logit contributions | |
robe | -5.246 |
Spoiler | -4.152 |
ンジ | -3.885 |
Doors | -3.809 |
taboola | -3.800 |
Token | point |
Token ID | 966 |
Feature activation | -6.07e-01 |
Pos logit contributions | |
robe | +2.565 |
Spoiler | +1.997 |
hib | +1.808 |
Doors | +1.802 |
taboola | +1.802 |
Neg logit contributions | |
bia | -2.125 |
rele | -2.101 |
constitu | -2.091 |
iland | -2.067 |
ibaba | -2.051 |
Token | : |
Token ID | 25 |
Feature activation | -1.39e-01 |
Pos logit contributions | |
robe | +3.214 |
Spoiler | +2.405 |
ンジ | +2.195 |
taboola | +2.172 |
ヤ | +2.162 |
Neg logit contributions | |
rele | -2.928 |
bia | -2.905 |
constitu | -2.902 |
ibaba | -2.820 |
iland | -2.818 |
Token | This |
Token ID | 770 |
Feature activation | -3.99e-01 |
Pos logit contributions | |
robe | +0.734 |
Spoiler | +0.564 |
Doors | +0.507 |
ンジ | +0.503 |
taboola | +0.499 |
Neg logit contributions | |
rele | -0.671 |
bia | -0.670 |
constitu | -0.665 |
iland | -0.646 |
ibaba | -0.645 |
Token | # |
Token ID | 1303 |
Feature activation | +4.46e-01 |
Pos logit contributions | |
robe | +2.310 |
Spoiler | +1.834 |
Doors | +1.655 |
taboola | +1.645 |
ンジ | +1.642 |
Neg logit contributions | |
bia | -1.725 |
rele | -1.686 |
constitu | -1.679 |
ibaba | -1.646 |
iland | -1.625 |
Token | J |
Token ID | 41 |
Feature activation | +8.15e-01 |
Pos logit contributions | |
rele | +1.975 |
bia | +1.947 |
constitu | +1.931 |
iland | +1.867 |
ibaba | +1.864 |
Neg logit contributions | |
robe | -2.569 |
Spoiler | -1.974 |
ンジ | -1.858 |
taboola | -1.826 |
hib | -1.803 |
Token | eb |
Token ID | 1765 |
Feature activation | +1.80e+00 |
Pos logit contributions | |
rele | +3.453 |
bia | +3.370 |
constitu | +3.326 |
edi | +3.249 |
iland | +3.245 |
Neg logit contributions | |
robe | -4.793 |
Spoiler | -3.749 |
ンジ | -3.605 |
taboola | -3.489 |
ヤ | -3.452 |
Token | No |
Token ID | 2949 |
Feature activation | +6.61e-01 |
Pos logit contributions | |
bia | +7.369 |
rele | +7.296 |
constitu | +7.006 |
edi | +6.799 |
iland | +6.730 |
Neg logit contributions | |
robe | -10.303 |
Spoiler | -8.199 |
ンジ | -7.869 |
taboola | -7.523 |
Minotaur | -7.487 |
Token | Filter |
Token ID | 22417 |
Feature activation | +6.98e-01 |
Pos logit contributions | |
rele | +2.864 |
bia | +2.831 |
constitu | +2.792 |
iland | +2.702 |
ibaba | +2.689 |
Neg logit contributions | |
robe | -3.854 |
Spoiler | -2.976 |
ンジ | -2.818 |
taboola | -2.768 |
ヤ | -2.725 |
Token | video |
Token ID | 2008 |
Feature activation | +4.64e-01 |
Pos logit contributions | |
rele | +2.823 |
bia | +2.742 |
constitu | +2.712 |
iland | +2.628 |
ibaba | +2.612 |
Neg logit contributions | |
robe | -4.264 |
Spoiler | -3.332 |
ンジ | -3.186 |
taboola | -3.132 |
Doors | -3.085 |
Token | that |
Token ID | 326 |
Feature activation | -1.86e-01 |
Pos logit contributions | |
rele | +1.880 |
bia | +1.841 |
constitu | +1.821 |
iland | +1.753 |
ibaba | +1.747 |
Neg logit contributions | |
robe | -2.858 |
Spoiler | -2.229 |
ンジ | -2.126 |
taboola | -2.092 |
hib | -2.066 |
Token | shows |
Token ID | 2523 |
Feature activation | -4.94e-01 |
Pos logit contributions | |
robe | +1.096 |
Spoiler | +0.865 |
ンジ | +0.794 |
taboola | +0.788 |
Doors | +0.787 |
Neg logit contributions | |
rele | -0.784 |
constitu | -0.781 |
bia | -0.780 |
ibaba | -0.754 |
iland | -0.751 |
Token | istani |
Token ID | 16688 |
Feature activation | +3.48e-02 |
Pos logit contributions | |
robe | +0.962 |
Spoiler | +0.564 |
ンジ | +0.513 |
taboola | +0.480 |
hib | +0.470 |
Neg logit contributions | |
rele | -2.095 |
bia | -2.062 |
constitu | -2.062 |
ibaba | -2.030 |
ilial | -1.996 |
Token | fighter |
Token ID | 10543 |
Feature activation | +4.37e-02 |
Pos logit contributions | |
rele | +0.144 |
bia | +0.143 |
constitu | +0.141 |
ibaba | +0.138 |
iland | +0.136 |
Neg logit contributions | |
robe | -0.209 |
Spoiler | -0.164 |
ンジ | -0.153 |
taboola | -0.151 |
Doors | -0.151 |
Token | jets |
Token ID | 20792 |
Feature activation | +7.82e-01 |
Pos logit contributions | |
rele | +0.230 |
bia | +0.227 |
constitu | +0.227 |
ibaba | +0.224 |
iland | +0.222 |
Neg logit contributions | |
robe | -0.215 |
Spoiler | -0.158 |
ンジ | -0.143 |
hib | -0.140 |
taboola | -0.140 |
Token | have |
Token ID | 423 |
Feature activation | +3.29e-01 |
Pos logit contributions | |
rele | +2.636 |
bia | +2.568 |
constitu | +2.538 |
iland | +2.423 |
edi | +2.393 |
Neg logit contributions | |
robe | -5.340 |
Spoiler | -4.263 |
ンジ | -4.110 |
taboola | -4.059 |
ヤ | -4.028 |
Token | launched |
Token ID | 5611 |
Feature activation | +6.51e-01 |
Pos logit contributions | |
rele | +1.305 |
bia | +1.284 |
constitu | +1.274 |
iland | +1.226 |
ibaba | +1.225 |
Neg logit contributions | |
robe | -2.050 |
Spoiler | -1.610 |
ンジ | -1.523 |
taboola | -1.502 |
hib | -1.484 |
Token | air |
Token ID | 1633 |
Feature activation | +1.79e+00 |
Pos logit contributions | |
rele | +2.478 |
bia | +2.408 |
constitu | +2.364 |
iland | +2.263 |
ibaba | +2.260 |
Neg logit contributions | |
robe | -4.200 |
Spoiler | -3.297 |
ンジ | -3.152 |
taboola | -3.118 |
ヤ | -3.062 |
Token | strikes |
Token ID | 8956 |
Feature activation | +7.18e-01 |
Pos logit contributions | |
bia | +4.808 |
rele | +4.683 |
iland | +4.190 |
numbers | +4.183 |
edi | +4.116 |
Neg logit contributions | |
robe | -12.983 |
Spoiler | -10.492 |
ンジ | -10.305 |
ヤ | -10.142 |
taboola | -10.054 |
Token | on |
Token ID | 319 |
Feature activation | +6.05e-01 |
Pos logit contributions | |
rele | +2.634 |
bia | +2.572 |
constitu | +2.523 |
iland | +2.446 |
edi | +2.406 |
Neg logit contributions | |
robe | -4.695 |
Spoiler | -3.733 |
ンジ | -3.543 |
taboola | -3.499 |
ヤ | -3.482 |
Token | suspected |
Token ID | 9885 |
Feature activation | +9.80e-01 |
Pos logit contributions | |
rele | +2.210 |
bia | +2.161 |
constitu | +2.141 |
iland | +2.046 |
ibaba | +2.035 |
Neg logit contributions | |
robe | -3.979 |
Spoiler | -3.154 |
ンジ | -3.001 |
taboola | -2.969 |
ヤ | -2.931 |
Token | militant |
Token ID | 18614 |
Feature activation | +2.60e-01 |
Pos logit contributions | |
rele | +2.958 |
bia | +2.872 |
constitu | +2.834 |
iland | +2.656 |
edi | +2.641 |
Neg logit contributions | |
robe | -7.033 |
Spoiler | -5.679 |
taboola | -5.458 |
ンジ | -5.451 |
ヤ | -5.368 |
Token | hide |
Token ID | 7808 |
Feature activation | -2.39e-01 |
Pos logit contributions | |
rele | +0.918 |
bia | +0.914 |
constitu | +0.900 |
ibaba | +0.884 |
ilial | +0.864 |
Neg logit contributions | |
robe | -1.704 |
Spoiler | -1.378 |
ンジ | -1.294 |
Doors | -1.276 |
hib | -1.273 |
Token | � |
Token ID | 247 |
Feature activation | +2.46e-01 |
Pos logit contributions | |
rele | +2.178 |
bia | +2.132 |
constitu | +2.126 |
ibaba | +2.077 |
iland | +2.076 |
Neg logit contributions | |
robe | -1.893 |
Spoiler | -1.371 |
ンジ | -1.263 |
taboola | -1.223 |
hib | -1.216 |
Token | t |
Token ID | 83 |
Feature activation | +2.06e-01 |
Pos logit contributions | |
rele | +1.017 |
bia | +1.005 |
constitu | +0.999 |
ibaba | +0.961 |
iland | +0.958 |
Neg logit contributions | |
robe | -1.488 |
Spoiler | -1.158 |
ンジ | -1.093 |
taboola | -1.082 |
hib | -1.068 |
Token | happened |
Token ID | 3022 |
Feature activation | +1.86e-01 |
Pos logit contributions | |
rele | +0.782 |
bia | +0.770 |
constitu | +0.763 |
iland | +0.732 |
ibaba | +0.732 |
Neg logit contributions | |
robe | -1.321 |
Spoiler | -1.046 |
ンジ | -0.991 |
taboola | -0.978 |
hib | -0.967 |
Token | yet |
Token ID | 1865 |
Feature activation | +3.48e-01 |
Pos logit contributions | |
rele | +0.737 |
bia | +0.727 |
constitu | +0.721 |
ibaba | +0.694 |
iland | +0.694 |
Neg logit contributions | |
robe | -1.155 |
Spoiler | -0.910 |
ンジ | -0.858 |
taboola | -0.847 |
hib | -0.838 |
Token | ! |
Token ID | 0 |
Feature activation | +1.46e-02 |
Pos logit contributions | |
rele | +1.464 |
bia | +1.439 |
constitu | +1.427 |
iland | +1.379 |
ibaba | +1.375 |
Neg logit contributions | |
robe | -2.083 |
Spoiler | -1.611 |
ンジ | -1.531 |
taboola | -1.506 |
hib | -1.481 |
Token | David |
Token ID | 3271 |
Feature activation | +1.79e+00 |
Pos logit contributions | |
rele | +0.071 |
bia | +0.070 |
constitu | +0.070 |
ibaba | +0.068 |
iland | +0.068 |
Neg logit contributions | |
robe | -0.077 |
Spoiler | -0.059 |
ンジ | -0.054 |
taboola | -0.053 |
Doors | -0.053 |
Token | : |
Token ID | 25 |
Feature activation | +1.73e-01 |
Pos logit contributions | |
rele | +6.421 |
bia | +6.013 |
constitu | +5.927 |
iland | +5.780 |
numbers | +5.732 |
Neg logit contributions | |
robe | -10.654 |
ンジ | -8.028 |
taboola | -7.998 |
Spoiler | -7.988 |
hib | -7.747 |
Token | Our |
Token ID | 3954 |
Feature activation | +2.20e-01 |
Pos logit contributions | |
rele | +0.886 |
bia | +0.876 |
constitu | +0.872 |
ibaba | +0.847 |
iland | +0.843 |
Neg logit contributions | |
robe | -0.880 |
Spoiler | -0.658 |
ンジ | -0.603 |
taboola | -0.594 |
Doors | -0.588 |
Token | daughter |
Token ID | 4957 |
Feature activation | +1.41e-01 |
Pos logit contributions | |
rele | +0.830 |
bia | +0.817 |
constitu | +0.809 |
ibaba | +0.782 |
iland | +0.779 |
Neg logit contributions | |
robe | -1.414 |
Spoiler | -1.126 |
ンジ | -1.062 |
taboola | -1.050 |
Doors | -1.037 |
Token | is |
Token ID | 318 |
Feature activation | +3.68e-01 |
Pos logit contributions | |
rele | +0.567 |
bia | +0.561 |
constitu | +0.559 |
ibaba | +0.538 |
iland | +0.536 |
Neg logit contributions | |
robe | -0.867 |
Spoiler | -0.684 |
ンジ | -0.642 |
taboola | -0.633 |
hib | -0.626 |
Token | a |
Token ID | 257 |
Feature activation | +6.30e-02 |
Pos logit contributions | |
rele | +1.343 |
bia | +1.329 |
constitu | +1.318 |
ibaba | +1.266 |
iland | +1.262 |
Neg logit contributions | |
robe | -2.400 |
Spoiler | -1.913 |
ンジ | -1.812 |
taboola | -1.785 |
hib | -1.775 |
Token | s |
Token ID | 82 |
Feature activation | -1.61e-01 |
Pos logit contributions | |
rele | +1.588 |
bia | +1.568 |
constitu | +1.396 |
iland | +1.387 |
edi | +1.384 |
Neg logit contributions | |
robe | -6.175 |
Spoiler | -5.149 |
ンジ | -5.073 |
taboola | -4.933 |
ヤ | -4.910 |
Token | . |
Token ID | 13 |
Feature activation | +2.77e-01 |
Pos logit contributions | |
robe | +0.937 |
Spoiler | +0.727 |
ンジ | +0.677 |
taboola | +0.670 |
hib | +0.662 |
Neg logit contributions | |
rele | -0.704 |
constitu | -0.696 |
bia | -0.695 |
ibaba | -0.670 |
iland | -0.666 |
Token | Good |
Token ID | 4599 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
rele | +1.236 |
bia | +1.215 |
constitu | +1.202 |
iland | +1.164 |
ibaba | +1.157 |
Neg logit contributions | |
robe | -1.598 |
Spoiler | -1.210 |
ンジ | -1.153 |
taboola | -1.131 |
ヤ | -1.113 |
Token | updates |
Token ID | 5992 |
Feature activation | +6.57e-01 |
Pos logit contributions | |
rele | +3.479 |
numbers | +3.245 |
constitu | +3.199 |
bia | +3.068 |
owing | +3.047 |
Neg logit contributions | |
robe | -6.684 |
Spoiler | -5.347 |
ンジ | -5.217 |
ヤ | -5.210 |
taboola | -5.054 |
Token | ? |
Token ID | 30 |
Feature activation | +9.08e-01 |
Pos logit contributions | |
rele | +1.821 |
bia | +1.731 |
constitu | +1.617 |
iland | +1.583 |
edi | +1.574 |
Neg logit contributions | |
robe | -4.897 |
Spoiler | -3.933 |
ンジ | -3.875 |
taboola | -3.775 |
ヤ | -3.739 |
Token | No |
Token ID | 1400 |
Feature activation | +1.79e+00 |
Pos logit contributions | |
rele | +2.741 |
bia | +2.679 |
edi | +2.440 |
constitu | +2.432 |
iland | +2.426 |
Neg logit contributions | |
robe | -6.454 |
taboola | -4.887 |
ンジ | -4.885 |
Spoiler | -4.830 |
hib | -4.776 |
Token | more |
Token ID | 517 |
Feature activation | +6.03e-01 |
Pos logit contributions | |
rele | +3.815 |
numbers | +3.558 |
number | +3.505 |
learning | +3.333 |
owing | +3.144 |
Neg logit contributions | |
robe | -13.989 |
ンジ | -11.683 |
ヤ | -11.547 |
Spoiler | -11.409 |
Pengu | -10.974 |
Token | bad |
Token ID | 2089 |
Feature activation | +5.96e-01 |
Pos logit contributions | |
rele | +2.008 |
bia | +1.906 |
constitu | +1.831 |
iland | +1.751 |
ibaba | +1.747 |
Neg logit contributions | |
robe | -4.176 |
Spoiler | -3.338 |
ンジ | -3.222 |
taboola | -3.177 |
ヤ | -3.172 |
Token | news |
Token ID | 1705 |
Feature activation | +8.12e-01 |
Pos logit contributions | |
rele | +1.412 |
bia | +1.318 |
constitu | +1.306 |
ibaba | +1.204 |
iland | +1.183 |
Neg logit contributions | |
robe | -4.715 |
Spoiler | -3.877 |
ンジ | -3.770 |
taboola | -3.708 |
ヤ | -3.690 |
Token | pl |
Token ID | 458 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
rele | +2.569 |
bia | +2.413 |
constitu | +2.412 |
ibaba | +2.233 |
edi | +2.204 |
Neg logit contributions | |
robe | -5.795 |
Spoiler | -4.577 |
ンジ | -4.562 |
taboola | -4.458 |
ヤ | -4.387 |
Token | s |
Token ID | 82 |
Feature activation | +2.45e-01 |
Pos logit contributions | |
bia | +1.897 |
rele | +1.848 |
edi | +1.651 |
iland | +1.600 |
ibaba | +1.501 |
Neg logit contributions | |
robe | -8.944 |
Spoiler | -7.500 |
ンジ | -7.491 |
Minotaur | -7.223 |
ヤ | -7.202 |
Token | Glen |
Token ID | 20716 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
robe | +1.872 |
Spoiler | +1.411 |
Doors | +1.269 |
ンジ | +1.266 |
Crate | +1.244 |
Neg logit contributions | |
rele | -1.840 |
bia | -1.821 |
constitu | -1.814 |
ibaba | -1.781 |
iland | -1.754 |
Token | Gab |
Token ID | 12300 |
Feature activation | -1.52e-01 |
Pos logit contributions | |
robe | +0.978 |
Spoiler | +0.741 |
ンジ | +0.686 |
Doors | +0.674 |
hib | +0.669 |
Neg logit contributions | |
rele | -0.836 |
constitu | -0.821 |
bia | -0.817 |
ibaba | -0.791 |
iland | -0.779 |
Token | ert |
Token ID | 861 |
Feature activation | -3.42e-01 |
Pos logit contributions | |
robe | +0.877 |
Spoiler | +0.657 |
hib | +0.614 |
ンジ | +0.610 |
roo | +0.610 |
Neg logit contributions | |
rele | -0.686 |
constitu | -0.685 |
bia | -0.660 |
ibaba | -0.647 |
iland | -0.637 |
Token | was |
Token ID | 373 |
Feature activation | -7.13e-01 |
Pos logit contributions | |
robe | +2.130 |
Spoiler | +1.628 |
taboola | +1.533 |
hib | +1.530 |
Doors | +1.523 |
Neg logit contributions | |
rele | -1.373 |
constitu | -1.371 |
bia | -1.344 |
ibaba | -1.329 |
iland | -1.270 |
Token | not |
Token ID | 407 |
Feature activation | -1.07e+00 |
Pos logit contributions | |
robe | +4.709 |
Spoiler | +3.703 |
taboola | +3.603 |
hib | +3.574 |
Doors | +3.526 |
Neg logit contributions | |
bia | -2.495 |
ibaba | -2.425 |
constitu | -2.373 |
ilial | -2.294 |
rele | -2.269 |
Token | returned |
Token ID | 4504 |
Feature activation | -2.39e+00 |
Pos logit contributions | |
robe | +6.179 |
Spoiler | +4.625 |
taboola | +4.532 |
ンジ | +4.334 |
hib | +4.333 |
Neg logit contributions | |
bia | -4.540 |
ibaba | -4.460 |
constitu | -4.254 |
ilial | -4.246 |
iland | -4.137 |
Token | . |
Token ID | 13 |
Feature activation | -3.50e-01 |
Pos logit contributions | |
robe | +10.134 |
Spoiler | +7.250 |
taboola | +7.051 |
Doors | +6.860 |
Crate | +6.680 |
Neg logit contributions | |
bia | -11.544 |
constitu | -11.539 |
ibaba | -11.418 |
iland | -11.080 |
rele | -11.000 |
Token | Hudson |
Token ID | 19995 |
Feature activation | +3.82e-01 |
Pos logit contributions | |
robe | +1.944 |
Spoiler | +1.511 |
ンジ | +1.386 |
taboola | +1.377 |
Doors | +1.376 |
Neg logit contributions | |
rele | -1.636 |
bia | -1.585 |
constitu | -1.583 |
ibaba | -1.541 |
iland | -1.514 |
Token | County |
Token ID | 3418 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
rele | +1.243 |
bia | +1.214 |
constitu | +1.204 |
iland | +1.158 |
ibaba | +1.132 |
Neg logit contributions | |
robe | -2.643 |
Spoiler | -2.133 |
ンジ | -2.036 |
taboola | -2.010 |
hib | -1.991 |
Token | Executive |
Token ID | 10390 |
Feature activation | -1.20e-01 |
Pos logit contributions | |
rele | +0.488 |
bia | +0.484 |
constitu | +0.480 |
ibaba | +0.467 |
iland | +0.463 |
Neg logit contributions | |
robe | -0.608 |
Spoiler | -0.466 |
ンジ | -0.435 |
taboola | -0.429 |
Doors | -0.424 |
Token | Tom |
Token ID | 4186 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
robe | +0.602 |
Spoiler | +0.445 |
ンジ | +0.407 |
Doors | +0.398 |
taboola | +0.393 |
Neg logit contributions | |
rele | -0.626 |
bia | -0.619 |
constitu | -0.617 |
ibaba | -0.600 |
iland | -0.599 |
Token | the |
Token ID | 262 |
Feature activation | -2.39e-02 |
Pos logit contributions | |
rele | +0.015 |
bia | +0.015 |
constitu | +0.015 |
ibaba | +0.015 |
iland | +0.015 |
Neg logit contributions | |
robe | -0.026 |
Spoiler | -0.021 |
ンジ | -0.020 |
Doors | -0.019 |
taboola | -0.019 |
Token | phrase |
Token ID | 9546 |
Feature activation | -2.19e-01 |
Pos logit contributions | |
robe | +0.144 |
Spoiler | +0.113 |
ンジ | +0.106 |
Doors | +0.105 |
taboola | +0.104 |
Neg logit contributions | |
rele | -0.098 |
bia | -0.098 |
constitu | -0.097 |
ibaba | -0.095 |
iland | -0.094 |
Token | started |
Token ID | 2067 |
Feature activation | -2.16e-01 |
Pos logit contributions | |
robe | +1.317 |
Spoiler | +1.039 |
ンジ | +0.969 |
taboola | +0.957 |
hib | +0.948 |
Neg logit contributions | |
bia | -0.894 |
rele | -0.886 |
constitu | -0.882 |
ibaba | -0.865 |
iland | -0.856 |
Token | on |
Token ID | 319 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
robe | +1.333 |
Spoiler | +1.065 |
ンジ | +0.991 |
taboola | +0.974 |
ヤ | +0.973 |
Neg logit contributions | |
bia | -0.845 |
rele | -0.838 |
constitu | -0.831 |
ibaba | -0.823 |
iland | -0.810 |
Token | B |
Token ID | 347 |
Feature activation | -6.82e-01 |
Pos logit contributions | |
robe | +0.694 |
Spoiler | +0.537 |
ンジ | +0.500 |
taboola | +0.491 |
Doors | +0.491 |
Neg logit contributions | |
rele | -0.542 |
bia | -0.538 |
constitu | -0.534 |
iland | -0.519 |
ibaba | -0.517 |
Token | flat |
Token ID | 6228 |
Feature activation | -2.19e+00 |
Pos logit contributions | |
robe | +3.897 |
Spoiler | +3.036 |
ンジ | +2.861 |
hib | +2.810 |
Doors | +2.795 |
Neg logit contributions | |
rele | -2.893 |
constitu | -2.892 |
bia | -2.884 |
ibaba | -2.791 |
iland | -2.773 |
Token | , |
Token ID | 11 |
Feature activation | +2.13e-02 |
Pos logit contributions | |
robe | +11.198 |
Spoiler | +8.641 |
taboola | +8.225 |
hib | +8.091 |
ヤ | +8.010 |
Neg logit contributions | |
constitu | -8.486 |
bia | -8.391 |
rele | -8.372 |
ibaba | -8.297 |
iland | -8.197 |
Token | or |
Token ID | 393 |
Feature activation | +2.85e-01 |
Pos logit contributions | |
bia | +0.076 |
rele | +0.076 |
constitu | +0.075 |
ibaba | +0.073 |
iland | +0.072 |
Neg logit contributions | |
robe | -0.139 |
Spoiler | -0.113 |
ンジ | -0.106 |
taboola | -0.105 |
hib | -0.104 |
Token | that |
Token ID | 326 |
Feature activation | -6.39e-02 |
Pos logit contributions | |
rele | +1.076 |
bia | +1.059 |
constitu | +1.050 |
ibaba | +1.009 |
iland | +1.009 |
Neg logit contributions | |
robe | -1.833 |
Spoiler | -1.454 |
ンジ | -1.377 |
taboola | -1.359 |
hib | -1.345 |
Token | it |
Token ID | 340 |
Feature activation | -5.34e-01 |
Pos logit contributions | |
robe | +0.405 |
Spoiler | +0.323 |
ンジ | +0.303 |
Doors | +0.299 |
taboola | +0.299 |
Neg logit contributions | |
rele | -0.242 |
bia | -0.240 |
constitu | -0.238 |
ibaba | -0.231 |
iland | -0.230 |
Token | happened |
Token ID | 3022 |
Feature activation | -1.37e-01 |
Pos logit contributions | |
robe | +3.182 |
Spoiler | +2.534 |
taboola | +2.356 |
ンジ | +2.341 |
hib | +2.306 |
Neg logit contributions | |
bia | -2.151 |
constitu | -2.123 |
ibaba | -2.113 |
rele | -2.080 |
iland | -2.041 |
Token | London |
Token ID | 3576 |
Feature activation | -5.25e-02 |
Pos logit contributions | |
robe | +1.853 |
Spoiler | +1.427 |
Crate | +1.333 |
Doors | +1.317 |
hib | +1.303 |
Neg logit contributions | |
constitu | -1.450 |
rele | -1.430 |
bia | -1.388 |
ilial | -1.365 |
ibaba | -1.345 |
Token | . |
Token ID | 13 |
Feature activation | +2.16e-01 |
Pos logit contributions | |
robe | +0.308 |
Spoiler | +0.239 |
ンジ | +0.223 |
taboola | +0.220 |
Doors | +0.219 |
Neg logit contributions | |
rele | -0.227 |
constitu | -0.225 |
bia | -0.224 |
ibaba | -0.216 |
iland | -0.215 |
Token |
|
Token ID | 198 |
Feature activation | -2.66e-01 |
Pos logit contributions | |
rele | +0.956 |
bia | +0.943 |
constitu | +0.935 |
iland | +0.904 |
ibaba | +0.903 |
Neg logit contributions | |
robe | -1.243 |
Spoiler | -0.954 |
ンジ | -0.897 |
taboola | -0.883 |
hib | -0.872 |
Token |
|
Token ID | 198 |
Feature activation | -2.29e-01 |
Pos logit contributions | |
robe | +1.538 |
Spoiler | +1.195 |
ンジ | +1.107 |
taboola | +1.096 |
hib | +1.088 |
Neg logit contributions | |
constitu | -1.172 |
rele | -1.171 |
bia | -1.154 |
ibaba | -1.123 |
iland | -1.108 |
Token | The |
Token ID | 464 |
Feature activation | +2.42e-01 |
Pos logit contributions | |
robe | +1.304 |
Spoiler | +1.000 |
ンジ | +0.933 |
taboola | +0.929 |
hib | +0.917 |
Neg logit contributions | |
rele | -1.058 |
constitu | -1.047 |
ibaba | -1.021 |
bia | -1.019 |
ilial | -0.986 |
Token | Walking |
Token ID | 21276 |
Feature activation | -2.14e+00 |
Pos logit contributions | |
rele | +1.090 |
bia | +1.080 |
constitu | +1.071 |
ibaba | +1.040 |
iland | +1.035 |
Neg logit contributions | |
robe | -1.381 |
Spoiler | -1.063 |
ンジ | -0.990 |
taboola | -0.977 |
Doors | -0.969 |
Token | Dead |
Token ID | 5542 |
Feature activation | -4.31e-01 |
Pos logit contributions | |
robe | +5.621 |
Spoiler | +3.917 |
Doors | +3.150 |
Crate | +3.050 |
Door | +2.904 |
Neg logit contributions | |
constitu | -14.900 |
ilial | -14.416 |
rele | -14.360 |
bia | -14.299 |
ibaba | -14.123 |
Token | is |
Token ID | 318 |
Feature activation | -3.53e-01 |
Pos logit contributions | |
robe | +2.510 |
Spoiler | +1.983 |
Doors | +1.808 |
ンジ | +1.793 |
hib | +1.782 |
Neg logit contributions | |
bia | -1.846 |
rele | -1.829 |
constitu | -1.819 |
ibaba | -1.800 |
iland | -1.758 |
Token | coming |
Token ID | 2406 |
Feature activation | -8.43e-01 |
Pos logit contributions | |
robe | +2.177 |
Spoiler | +1.735 |
hib | +1.609 |
Doors | +1.606 |
ンジ | +1.604 |
Neg logit contributions | |
bia | -1.386 |
constitu | -1.364 |
rele | -1.364 |
ibaba | -1.341 |
iland | -1.307 |
Token | back |
Token ID | 736 |
Feature activation | -1.07e+00 |
Pos logit contributions | |
robe | +4.662 |
Spoiler | +3.694 |
Doors | +3.375 |
hib | +3.318 |
ンジ | +3.286 |
Neg logit contributions | |
bia | -3.766 |
constitu | -3.751 |
ibaba | -3.705 |
rele | -3.651 |
ilial | -3.539 |
Token | to |
Token ID | 284 |
Feature activation | -5.87e-01 |
Pos logit contributions | |
robe | +5.911 |
Spoiler | +4.706 |
Doors | +4.295 |
hib | +4.248 |
ヤ | +4.161 |
Neg logit contributions | |
constitu | -4.658 |
bia | -4.626 |
ibaba | -4.597 |
rele | -4.421 |
iland | -4.342 |
Token | keep |
Token ID | 1394 |
Feature activation | +9.32e-02 |
Pos logit contributions | |
robe | +2.346 |
Spoiler | +1.852 |
ンジ | +1.748 |
taboola | +1.726 |
hib | +1.714 |
Neg logit contributions | |
bia | -1.493 |
constitu | -1.490 |
rele | -1.458 |
ibaba | -1.430 |
iland | -1.395 |
Token | things |
Token ID | 1243 |
Feature activation | +1.63e-01 |
Pos logit contributions | |
rele | +0.366 |
bia | +0.364 |
constitu | +0.359 |
ibaba | +0.347 |
iland | +0.345 |
Neg logit contributions | |
robe | -0.583 |
Spoiler | -0.461 |
ンジ | -0.434 |
taboola | -0.428 |
hib | -0.424 |
Token | fresh |
Token ID | 4713 |
Feature activation | +1.07e-01 |
Pos logit contributions | |
bia | +0.614 |
rele | +0.605 |
constitu | +0.601 |
ibaba | +0.585 |
iland | +0.576 |
Neg logit contributions | |
robe | -1.046 |
Spoiler | -0.836 |
ンジ | -0.784 |
Doors | -0.770 |
taboola | -0.770 |
Token | with |
Token ID | 351 |
Feature activation | +4.88e-01 |
Pos logit contributions | |
rele | +0.408 |
bia | +0.403 |
constitu | +0.399 |
ibaba | +0.384 |
iland | +0.383 |
Neg logit contributions | |
robe | -0.680 |
Spoiler | -0.538 |
ンジ | -0.509 |
taboola | -0.502 |
hib | -0.497 |
Token | The |
Token ID | 383 |
Feature activation | +3.82e-02 |
Pos logit contributions | |
rele | +1.922 |
bia | +1.871 |
constitu | +1.864 |
iland | +1.793 |
ibaba | +1.773 |
Neg logit contributions | |
robe | -3.051 |
Spoiler | -2.388 |
ンジ | -2.272 |
taboola | -2.247 |
ヤ | -2.210 |
Token | Walking |
Token ID | 21276 |
Feature activation | -2.11e+00 |
Pos logit contributions | |
rele | +0.183 |
bia | +0.180 |
constitu | +0.179 |
ibaba | +0.174 |
iland | +0.173 |
Neg logit contributions | |
robe | -0.207 |
Spoiler | -0.157 |
ンジ | -0.146 |
taboola | -0.143 |
Doors | -0.143 |
Token | Dead |
Token ID | 5542 |
Feature activation | -4.05e-01 |
Pos logit contributions | |
robe | +5.621 |
Spoiler | +4.033 |
Doors | +3.305 |
Crate | +3.225 |
Door | +3.077 |
Neg logit contributions | |
constitu | -14.616 |
rele | -14.247 |
ibaba | -14.092 |
bia | -14.045 |
ilial | -14.033 |
Token | in |
Token ID | 287 |
Feature activation | -8.15e-02 |
Pos logit contributions | |
robe | +2.495 |
Spoiler | +1.994 |
ンジ | +1.843 |
Doors | +1.836 |
hib | +1.831 |
Neg logit contributions | |
bia | -1.596 |
constitu | -1.584 |
rele | -1.572 |
ibaba | -1.543 |
iland | -1.514 |
Token | each |
Token ID | 1123 |
Feature activation | +1.47e-01 |
Pos logit contributions | |
robe | +0.505 |
Spoiler | +0.398 |
ンジ | +0.373 |
taboola | +0.369 |
hib | +0.368 |
Neg logit contributions | |
rele | -0.325 |
bia | -0.324 |
constitu | -0.319 |
ibaba | -0.311 |
iland | -0.305 |
Token | incarnation |
Token ID | 29185 |
Feature activation | -4.74e-01 |
Pos logit contributions | |
bia | +0.566 |
rele | +0.561 |
constitu | +0.550 |
ibaba | +0.545 |
iland | +0.532 |
Neg logit contributions | |
robe | -0.935 |
Spoiler | -0.744 |
ンジ | -0.697 |
taboola | -0.689 |
Doors | -0.687 |
Token | , |
Token ID | 11 |
Feature activation | +2.53e-01 |
Pos logit contributions | |
robe | +2.975 |
Spoiler | +2.401 |
ンジ | +2.229 |
taboola | +2.207 |
hib | +2.197 |
Neg logit contributions | |
bia | -1.808 |
constitu | -1.778 |
rele | -1.760 |
ibaba | -1.737 |
iland | -1.711 |
Token | Boston |
Token ID | 6182 |
Feature activation | -2.92e-01 |
Pos logit contributions | |
rele | +0.341 |
bia | +0.336 |
constitu | +0.334 |
iland | +0.322 |
ibaba | +0.322 |
Neg logit contributions | |
robe | -0.452 |
Spoiler | -0.348 |
ンジ | -0.328 |
taboola | -0.323 |
hib | -0.319 |
Token | . |
Token ID | 13 |
Feature activation | -3.81e-01 |
Pos logit contributions | |
robe | +1.922 |
Spoiler | +1.533 |
ンジ | +1.441 |
hib | +1.419 |
taboola | +1.418 |
Neg logit contributions | |
rele | -1.042 |
constitu | -1.037 |
bia | -1.025 |
ibaba | -0.988 |
iland | -0.977 |
Token | ( |
Token ID | 357 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
robe | +1.952 |
Spoiler | +1.508 |
Doors | +1.378 |
ンジ | +1.348 |
Crate | +1.327 |
Neg logit contributions | |
rele | -1.896 |
constitu | -1.871 |
bia | -1.864 |
ibaba | -1.826 |
iland | -1.777 |
Token | Wa |
Token ID | 33484 |
Feature activation | +4.71e-02 |
Pos logit contributions | |
rele | +1.214 |
bia | +1.193 |
constitu | +1.184 |
ibaba | +1.138 |
iland | +1.137 |
Neg logit contributions | |
robe | -1.917 |
Spoiler | -1.506 |
ンジ | -1.426 |
taboola | -1.405 |
hib | -1.389 |
Token | ub |
Token ID | 549 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
rele | +0.211 |
constitu | +0.210 |
bia | +0.208 |
iland | +0.199 |
ibaba | +0.197 |
Neg logit contributions | |
robe | -0.271 |
Spoiler | -0.206 |
ンジ | -0.192 |
hib | -0.191 |
taboola | -0.190 |
Token | ges |
Token ID | 3212 |
Feature activation | -2.07e+00 |
Pos logit contributions | |
robe | +0.142 |
Spoiler | +0.109 |
ンジ | +0.102 |
taboola | +0.101 |
hib | +0.100 |
Neg logit contributions | |
rele | -0.110 |
constitu | -0.108 |
bia | -0.107 |
iland | -0.103 |
ibaba | -0.103 |
Token | hig |
Token ID | 25196 |
Feature activation | -1.46e+00 |
Pos logit contributions | |
robe | +7.172 |
hib | +4.944 |
Spoiler | +4.439 |
taboola | +4.062 |
erald | +3.898 |
Neg logit contributions | |
constitu | -13.278 |
rele | -12.050 |
bia | -11.780 |
ilial | -11.743 |
shooter | -11.688 |
Token | Rice |
Token ID | 13823 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
robe | +5.214 |
Spoiler | +3.542 |
Crate | +3.173 |
Doors | +3.009 |
roo | +3.005 |
Neg logit contributions | |
constitu | -9.406 |
bia | -8.989 |
rele | -8.908 |
ibaba | -8.811 |
ilial | -8.781 |
Token | / |
Token ID | 14 |
Feature activation | -5.78e-01 |
Pos logit contributions | |
robe | +0.648 |
Spoiler | +0.513 |
ンジ | +0.480 |
taboola | +0.477 |
hib | +0.472 |
Neg logit contributions | |
rele | -0.408 |
constitu | -0.402 |
bia | -0.401 |
ibaba | -0.387 |
iland | -0.383 |
Token | CBC |
Token ID | 29208 |
Feature activation | +2.47e-01 |
Pos logit contributions | |
robe | +2.324 |
Spoiler | +1.545 |
ンジ | +1.422 |
taboola | +1.412 |
ヤ | +1.342 |
Neg logit contributions | |
rele | -3.580 |
constitu | -3.531 |
bia | -3.468 |
ibaba | -3.453 |
iland | -3.370 |
Token | ) |
Token ID | 8 |
Feature activation | +4.26e-01 |
Pos logit contributions | |
rele | +0.888 |
bia | +0.877 |
constitu | +0.868 |
iland | +0.829 |
edi | +0.812 |
Neg logit contributions | |
robe | -1.619 |
Spoiler | -1.292 |
ンジ | -1.234 |
taboola | -1.220 |
ヤ | -1.199 |
Token | . |
Token ID | 13 |
Feature activation | -1.70e-01 |
Pos logit contributions | |
robe | +2.527 |
Spoiler | +1.938 |
taboola | +1.778 |
ンジ | +1.775 |
hib | +1.760 |
Neg logit contributions | |
constitu | -2.046 |
rele | -2.032 |
bia | -1.998 |
ibaba | -1.953 |
iland | -1.924 |
Token | You |
Token ID | 921 |
Feature activation | +4.36e-01 |
Pos logit contributions | |
robe | +0.999 |
Spoiler | +0.779 |
ンジ | +0.727 |
taboola | +0.720 |
hib | +0.710 |
Neg logit contributions | |
rele | -0.730 |
constitu | -0.720 |
bia | -0.719 |
ibaba | -0.694 |
iland | -0.689 |
Token | can |
Token ID | 460 |
Feature activation | -6.28e-01 |
Pos logit contributions | |
rele | +1.827 |
bia | +1.763 |
constitu | +1.705 |
iland | +1.672 |
edi | +1.661 |
Neg logit contributions | |
robe | -2.608 |
Spoiler | -2.042 |
ンジ | -1.986 |
taboola | -1.875 |
Doors | -1.874 |
Token | read |
Token ID | 1100 |
Feature activation | -7.83e-01 |
Pos logit contributions | |
robe | +4.209 |
Spoiler | +3.368 |
taboola | +3.205 |
hib | +3.159 |
ンジ | +3.119 |
Neg logit contributions | |
constitu | -2.194 |
bia | -2.108 |
rele | -2.065 |
ilial | -2.014 |
iland | -2.012 |
Token | our |
Token ID | 674 |
Feature activation | -6.41e-01 |
Pos logit contributions | |
robe | +4.236 |
Spoiler | +3.236 |
ンジ | +3.013 |
taboola | +2.983 |
hib | +2.933 |
Neg logit contributions | |
constitu | -3.683 |
rele | -3.676 |
bia | -3.591 |
ibaba | -3.546 |
iland | -3.447 |
Token | Privacy |
Token ID | 16777 |
Feature activation | -2.07e+00 |
Pos logit contributions | |
robe | +3.412 |
Spoiler | +2.683 |
ンジ | +2.463 |
taboola | +2.416 |
Doors | +2.356 |
Neg logit contributions | |
rele | -3.034 |
constitu | -2.987 |
bia | -2.963 |
ibaba | -2.938 |
iland | -2.886 |
Token | Policy |
Token ID | 7820 |
Feature activation | -1.19e+00 |
Pos logit contributions | |
robe | +3.825 |
Spoiler | +2.808 |
Crate | +1.743 |
Doors | +1.540 |
ンジ | +1.349 |
Neg logit contributions | |
constitu | -15.972 |
ibaba | -15.480 |
rele | -15.452 |
iland | -15.039 |
bia | -14.934 |
Token | here |
Token ID | 994 |
Feature activation | -1.05e-01 |
Pos logit contributions | |
robe | +5.389 |
Spoiler | +4.133 |
taboola | +3.674 |
hib | +3.598 |
Doors | +3.503 |
Neg logit contributions | |
bia | -6.327 |
constitu | -6.316 |
rele | -6.195 |
ilial | -6.122 |
iland | -6.087 |
Token | . |
Token ID | 13 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
robe | +0.611 |
Spoiler | +0.474 |
ンジ | +0.442 |
taboola | +0.439 |
hib | +0.433 |
Neg logit contributions | |
rele | -0.455 |
bia | -0.449 |
constitu | -0.448 |
ibaba | -0.431 |
iland | -0.429 |
Token | Sign |
Token ID | 5865 |
Feature activation | -5.20e-01 |
Pos logit contributions | |
rele | +0.604 |
bia | +0.596 |
constitu | +0.588 |
iland | +0.577 |
ibaba | +0.576 |
Neg logit contributions | |
robe | -0.434 |
Spoiler | -0.290 |
ンジ | -0.270 |
hib | -0.255 |
ヤ | -0.254 |
Token | up |
Token ID | 510 |
Feature activation | -2.87e-01 |
Pos logit contributions | |
robe | +2.990 |
Spoiler | +2.311 |
taboola | +2.136 |
ンジ | +2.135 |
hib | +2.110 |
Neg logit contributions | |
constitu | -2.300 |
bia | -2.285 |
rele | -2.277 |
ibaba | -2.204 |
iland | -2.195 |
Token | the |
Token ID | 262 |
Feature activation | -2.40e-01 |
Pos logit contributions | |
robe | +2.905 |
Spoiler | +2.301 |
Doors | +2.108 |
taboola | +2.080 |
Crate | +2.080 |
Neg logit contributions | |
bia | -2.065 |
constitu | -2.055 |
rele | -2.018 |
ibaba | -2.005 |
iland | -1.973 |
Token | Fred |
Token ID | 8559 |
Feature activation | -2.82e-01 |
Pos logit contributions | |
robe | +1.454 |
Spoiler | +1.137 |
ンジ | +1.055 |
Doors | +1.054 |
Crate | +1.047 |
Neg logit contributions | |
bia | -0.978 |
rele | -0.977 |
constitu | -0.976 |
ibaba | -0.936 |
iland | -0.935 |
Token | Meyer |
Token ID | 23975 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
robe | +1.478 |
Spoiler | +1.100 |
ンジ | +1.014 |
taboola | +1.009 |
Doors | +1.006 |
Neg logit contributions | |
rele | -1.404 |
constitu | -1.376 |
bia | -1.364 |
ibaba | -1.318 |
iland | -1.311 |
Token | at |
Token ID | 379 |
Feature activation | -5.53e-01 |
Pos logit contributions | |
rele | +0.476 |
bia | +0.470 |
constitu | +0.469 |
ibaba | +0.452 |
iland | +0.450 |
Neg logit contributions | |
robe | -0.623 |
Spoiler | -0.481 |
ンジ | -0.446 |
Doors | -0.441 |
taboola | -0.441 |
Token | 3 |
Token ID | 513 |
Feature activation | -6.32e-01 |
Pos logit contributions | |
robe | +3.457 |
Spoiler | +2.691 |
Doors | +2.505 |
ンジ | +2.482 |
hib | +2.476 |
Neg logit contributions | |
rele | -2.162 |
constitu | -2.159 |
bia | -2.148 |
iland | -2.081 |
ibaba | -2.064 |
Token | 184 |
Token ID | 22883 |
Feature activation | -2.06e+00 |
Pos logit contributions | |
robe | +3.254 |
Spoiler | +2.400 |
Doors | +2.266 |
ンジ | +2.229 |
ヤ | +2.212 |
Neg logit contributions | |
rele | -3.105 |
bia | -3.093 |
constitu | -3.092 |
iland | -3.026 |
ibaba | -2.976 |
Token | Ocean |
Token ID | 10692 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
robe | +9.431 |
Spoiler | +7.695 |
Doors | +7.166 |
Crate | +6.966 |
Door | +6.900 |
Neg logit contributions | |
constitu | -9.602 |
ibaba | -9.362 |
rele | -9.354 |
bia | -8.963 |
iland | -8.891 |
Token | Beach |
Token ID | 8511 |
Feature activation | -1.87e-01 |
Pos logit contributions | |
robe | +0.066 |
Spoiler | +0.049 |
ンジ | +0.045 |
Doors | +0.044 |
hib | +0.044 |
Neg logit contributions | |
rele | -0.066 |
constitu | -0.066 |
bia | -0.065 |
ibaba | -0.063 |
iland | -0.062 |
Token | Highway |
Token ID | 14914 |
Feature activation | +2.87e-01 |
Pos logit contributions | |
robe | +1.015 |
Spoiler | +0.767 |
ンジ | +0.703 |
Doors | +0.700 |
taboola | +0.691 |
Neg logit contributions | |
constitu | -0.883 |
rele | -0.882 |
bia | -0.880 |
ibaba | -0.860 |
iland | -0.841 |
Token | with |
Token ID | 351 |
Feature activation | +5.11e-01 |
Pos logit contributions | |
rele | +1.147 |
bia | +1.131 |
constitu | +1.123 |
ibaba | +1.082 |
iland | +1.079 |
Neg logit contributions | |
robe | -1.783 |
Spoiler | -1.399 |
ンジ | -1.318 |
taboola | -1.301 |
hib | -1.287 |
Token | the |
Token ID | 262 |
Feature activation | +3.77e-01 |
Pos logit contributions | |
rele | +1.957 |
bia | +1.919 |
constitu | +1.900 |
iland | +1.825 |
ibaba | +1.822 |
Neg logit contributions | |
robe | -3.257 |
Spoiler | -2.574 |
ンジ | -2.449 |
taboola | -2.411 |
hib | -2.381 |
Token |
|
Token ID | 198 |
Feature activation | -7.55e-01 |
Pos logit contributions | |
robe | +3.469 |
Spoiler | +2.669 |
taboola | +2.419 |
ンジ | +2.415 |
Doors | +2.415 |
Neg logit contributions | |
constitu | -3.037 |
rele | -2.969 |
bia | -2.932 |
ibaba | -2.906 |
iland | -2.824 |
Token | Ab |
Token ID | 4826 |
Feature activation | -4.05e-01 |
Pos logit contributions | |
robe | +4.185 |
Spoiler | +3.116 |
taboola | +2.950 |
ンジ | +2.921 |
ヤ | +2.905 |
Neg logit contributions | |
rele | -3.495 |
constitu | -3.451 |
bia | -3.402 |
ibaba | -3.393 |
iland | -3.298 |
Token | ig |
Token ID | 328 |
Feature activation | -2.64e-01 |
Pos logit contributions | |
robe | +2.098 |
Spoiler | +1.510 |
hib | +1.414 |
ンジ | +1.389 |
Doors | +1.375 |
Neg logit contributions | |
rele | -2.070 |
constitu | -2.040 |
bia | -1.973 |
iland | -1.952 |
ibaba | -1.927 |
Token | ail |
Token ID | 603 |
Feature activation | +5.83e-01 |
Pos logit contributions | |
robe | +1.522 |
Spoiler | +1.155 |
roo | +1.088 |
hib | +1.085 |
erald | +1.082 |
Neg logit contributions | |
constitu | -1.152 |
rele | -1.151 |
bia | -1.117 |
iland | -1.110 |
ibaba | -1.109 |
Token | van |
Token ID | 5719 |
Feature activation | -6.91e-01 |
Pos logit contributions | |
rele | +2.675 |
bia | +2.647 |
constitu | +2.606 |
iland | +2.540 |
ibaba | +2.515 |
Neg logit contributions | |
robe | -3.254 |
Spoiler | -2.450 |
ンジ | -2.342 |
taboola | -2.314 |
hib | -2.253 |
Token | den |
Token ID | 2853 |
Feature activation | -2.03e+00 |
Pos logit contributions | |
robe | +3.801 |
Spoiler | +2.915 |
Doors | +2.741 |
hib | +2.634 |
Door | +2.606 |
Neg logit contributions | |
rele | -3.211 |
constitu | -3.139 |
bia | -3.133 |
ilial | -3.083 |
ibaba | -3.064 |
Token | No |
Token ID | 1400 |
Feature activation | -4.33e-01 |
Pos logit contributions | |
robe | +7.825 |
Spoiler | +6.456 |
Doors | +6.250 |
Door | +5.902 |
Vik | +5.735 |
Neg logit contributions | |
rele | -11.116 |
constitu | -10.746 |
ilial | -10.647 |
ibaba | -10.609 |
iland | -10.303 |
Token | ort |
Token ID | 419 |
Feature activation | +2.44e-01 |
Pos logit contributions | |
robe | +3.210 |
Spoiler | +2.599 |
hib | +2.496 |
Doors | +2.478 |
roo | +2.463 |
Neg logit contributions | |
rele | -1.203 |
constitu | -1.192 |
bia | -1.166 |
ibaba | -1.103 |
ilial | -1.076 |
Token | , |
Token ID | 11 |
Feature activation | -5.69e-01 |
Pos logit contributions | |
rele | +1.045 |
bia | +1.029 |
constitu | +1.020 |
iland | +0.986 |
ibaba | +0.984 |
Neg logit contributions | |
robe | -1.446 |
Spoiler | -1.121 |
ンジ | -1.057 |
taboola | -1.040 |
hib | -1.025 |
Token | Morris |
Token ID | 14433 |
Feature activation | +3.85e-02 |
Pos logit contributions | |
robe | +3.122 |
Spoiler | +2.443 |
Doors | +2.232 |
ンジ | +2.184 |
taboola | +2.177 |
Neg logit contributions | |
rele | -2.569 |
bia | -2.562 |
constitu | -2.539 |
iland | -2.500 |
ibaba | -2.492 |
Token | ville |
Token ID | 4244 |
Feature activation | -7.30e-02 |
Pos logit contributions | |
rele | +0.179 |
bia | +0.176 |
constitu | +0.175 |
ibaba | +0.170 |
iland | +0.168 |
Neg logit contributions | |
robe | -0.214 |
Spoiler | -0.164 |
ンジ | -0.152 |
hib | -0.149 |
Doors | -0.149 |
Token | ar |
Token ID | 610 |
Feature activation | -1.04e+00 |
Pos logit contributions | |
robe | +2.743 |
Spoiler | +1.915 |
ンジ | +1.690 |
taboola | +1.681 |
Doors | +1.656 |
Neg logit contributions | |
bia | -3.960 |
constitu | -3.915 |
rele | -3.884 |
iland | -3.837 |
ibaba | -3.832 |
Token | X |
Token ID | 55 |
Feature activation | -7.57e-02 |
Pos logit contributions | |
robe | +4.592 |
Spoiler | +3.211 |
taboola | +3.171 |
erald | +3.013 |
ンジ | +3.007 |
Neg logit contributions | |
constitu | -5.660 |
rele | -5.653 |
bia | -5.443 |
ibaba | -5.416 |
edi | -5.285 |
Token | iv |
Token ID | 452 |
Feature activation | -4.03e-01 |
Pos logit contributions | |
robe | +0.451 |
Spoiler | +0.355 |
ンジ | +0.331 |
hib | +0.328 |
taboola | +0.327 |
Neg logit contributions | |
rele | -0.315 |
constitu | -0.315 |
bia | -0.309 |
ibaba | -0.299 |
iland | -0.293 |
Token | : |
Token ID | 25 |
Feature activation | -9.49e-01 |
Pos logit contributions | |
robe | +2.026 |
Spoiler | +1.562 |
hib | +1.397 |
Doors | +1.391 |
ンジ | +1.367 |
Neg logit contributions | |
bia | -1.997 |
constitu | -1.983 |
rele | -1.966 |
ibaba | -1.945 |
iland | -1.913 |
Token | 14 |
Token ID | 1478 |
Feature activation | -1.00e+00 |
Pos logit contributions | |
robe | +4.939 |
Spoiler | +3.706 |
hib | +3.509 |
ヤ | +3.422 |
taboola | +3.413 |
Neg logit contributions | |
constitu | -4.409 |
bia | -4.344 |
rele | -4.267 |
ibaba | -4.234 |
iland | -4.172 |
Token | 05 |
Token ID | 2713 |
Feature activation | -2.02e+00 |
Pos logit contributions | |
robe | +4.823 |
Spoiler | +3.620 |
hib | +3.409 |
taboola | +3.343 |
Doors | +3.319 |
Neg logit contributions | |
constitu | -4.929 |
ibaba | -4.909 |
bia | -4.851 |
rele | -4.834 |
iland | -4.780 |
Token | . |
Token ID | 13 |
Feature activation | -8.76e-01 |
Pos logit contributions | |
robe | +8.042 |
Spoiler | +6.029 |
hib | +5.575 |
Doors | +5.494 |
ヤ | +5.485 |
Neg logit contributions | |
constitu | -10.225 |
ibaba | -10.089 |
rele | -10.038 |
iland | -9.890 |
ilial | -9.782 |
Token | 7 |
Token ID | 22 |
Feature activation | -4.01e-01 |
Pos logit contributions | |
robe | +4.426 |
Spoiler | +3.445 |
ンジ | +3.154 |
Doors | +3.126 |
taboola | +3.112 |
Neg logit contributions | |
constitu | -4.283 |
rele | -4.199 |
bia | -4.098 |
ibaba | -4.044 |
iland | -3.973 |
Token | 88 |
Token ID | 3459 |
Feature activation | -6.95e-01 |
Pos logit contributions | |
robe | +2.083 |
Spoiler | +1.591 |
ンジ | +1.456 |
hib | +1.440 |
taboola | +1.434 |
Neg logit contributions | |
constitu | -1.967 |
rele | -1.944 |
bia | -1.928 |
ibaba | -1.879 |
iland | -1.869 |
Token | 1 |
Token ID | 16 |
Feature activation | -7.65e-01 |
Pos logit contributions | |
robe | +3.766 |
Spoiler | +2.895 |
hib | +2.706 |
ンジ | +2.655 |
taboola | +2.651 |
Neg logit contributions | |
constitu | -3.172 |
rele | -3.161 |
bia | -3.105 |
ibaba | -3.077 |
iland | -3.060 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -9.17e-02 |
Pos logit contributions | |
robe | +4.081 |
Spoiler | +3.128 |
hib | +2.943 |
taboola | +2.886 |
Doors | +2.848 |
Neg logit contributions | |
constitu | -3.569 |
rele | -3.498 |
bia | -3.473 |
ibaba | -3.461 |
iland | -3.356 |
Token | year |
Token ID | 614 |
Feature activation | +4.71e-01 |
Pos logit contributions | |
robe | +0.654 |
Spoiler | +0.520 |
ンジ | +0.483 |
hib | +0.480 |
Doors | +0.478 |
Neg logit contributions | |
bia | -0.418 |
rele | -0.416 |
constitu | -0.411 |
ibaba | -0.402 |
iland | -0.395 |
Token | 's |
Token ID | 338 |
Feature activation | +4.16e-02 |
Pos logit contributions | |
rele | +1.702 |
bia | +1.668 |
constitu | +1.619 |
iland | +1.579 |
edi | +1.548 |
Neg logit contributions | |
robe | -3.086 |
Spoiler | -2.427 |
ンジ | -2.352 |
taboola | -2.310 |
ヤ | -2.272 |
Token | event |
Token ID | 1785 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
rele | +0.172 |
bia | +0.171 |
constitu | +0.169 |
ibaba | +0.164 |
iland | +0.163 |
Neg logit contributions | |
robe | -0.251 |
Spoiler | -0.197 |
ンジ | -0.185 |
taboola | -0.182 |
hib | -0.180 |
Token | - |
Token ID | 532 |
Feature activation | +1.57e-01 |
Pos logit contributions | |
robe | +0.710 |
Spoiler | +0.549 |
ンジ | +0.513 |
taboola | +0.504 |
hib | +0.502 |
Neg logit contributions | |
rele | -0.530 |
bia | -0.524 |
constitu | -0.522 |
ibaba | -0.503 |
iland | -0.501 |
Token | The |
Token ID | 383 |
Feature activation | -2.69e-02 |
Pos logit contributions | |
rele | +0.729 |
bia | +0.722 |
constitu | +0.719 |
ibaba | +0.695 |
iland | +0.692 |
Neg logit contributions | |
robe | -0.871 |
Spoiler | -0.668 |
ンジ | -0.620 |
taboola | -0.611 |
Doors | -0.606 |
Token | Walking |
Token ID | 21276 |
Feature activation | -2.00e+00 |
Pos logit contributions | |
robe | +0.139 |
Spoiler | +0.105 |
ンジ | +0.096 |
Doors | +0.095 |
taboola | +0.094 |
Neg logit contributions | |
rele | -0.136 |
bia | -0.134 |
constitu | -0.134 |
ibaba | -0.130 |
iland | -0.129 |
Token | Dead |
Token ID | 5542 |
Feature activation | -4.64e-01 |
Pos logit contributions | |
robe | +5.043 |
Spoiler | +3.559 |
Doors | +2.711 |
Minotaur | +2.393 |
Door | +2.376 |
Neg logit contributions | |
constitu | -13.829 |
bia | -13.766 |
ilial | -13.755 |
rele | -13.753 |
ibaba | -13.524 |
Token | and |
Token ID | 290 |
Feature activation | -3.54e-02 |
Pos logit contributions | |
robe | +2.995 |
Spoiler | +2.438 |
Doors | +2.234 |
ンジ | +2.233 |
hib | +2.210 |
Neg logit contributions | |
rele | -1.699 |
constitu | -1.677 |
bia | -1.672 |
ibaba | -1.626 |
iland | -1.578 |
Token | An |
Token ID | 1052 |
Feature activation | -2.67e-01 |
Pos logit contributions | |
robe | +0.181 |
Spoiler | +0.136 |
ンジ | +0.124 |
Doors | +0.123 |
taboola | +0.122 |
Neg logit contributions | |
rele | -0.180 |
bia | -0.178 |
constitu | -0.177 |
ibaba | -0.173 |
iland | -0.171 |
Token | American |
Token ID | 1605 |
Feature activation | -8.25e-01 |
Pos logit contributions | |
robe | +1.453 |
Spoiler | +1.121 |
Doors | +1.023 |
ンジ | +1.021 |
taboola | +1.016 |
Neg logit contributions | |
rele | -1.264 |
constitu | -1.256 |
bia | -1.220 |
ibaba | -1.184 |
ilial | -1.184 |
Token | Werewolf |
Token ID | 40261 |
Feature activation | -3.02e-01 |
Pos logit contributions | |
robe | +3.488 |
Spoiler | +2.674 |
Doors | +2.258 |
Crate | +2.257 |
Minotaur | +2.254 |
Neg logit contributions | |
rele | -4.764 |
constitu | -4.655 |
ibaba | -4.533 |
bia | -4.476 |
iland | -4.427 |
Token | . |
Token ID | 13 |
Feature activation | -4.49e-01 |
Pos logit contributions | |
robe | +5.111 |
Spoiler | +3.842 |
ンジ | +3.568 |
hib | +3.503 |
taboola | +3.477 |
Neg logit contributions | |
bia | -4.291 |
constitu | -4.231 |
rele | -4.198 |
ibaba | -4.145 |
iland | -4.057 |
Token | Like |
Token ID | 4525 |
Feature activation | +2.74e-01 |
Pos logit contributions | |
robe | +2.361 |
Spoiler | +1.898 |
ンジ | +1.712 |
Doors | +1.697 |
taboola | +1.660 |
Neg logit contributions | |
rele | -2.170 |
constitu | -2.129 |
bia | -2.123 |
ibaba | -2.082 |
iland | -2.049 |
Token | a |
Token ID | 257 |
Feature activation | -5.54e-01 |
Pos logit contributions | |
rele | +1.083 |
bia | +1.071 |
constitu | +1.059 |
ibaba | +1.023 |
iland | +1.022 |
Neg logit contributions | |
robe | -1.709 |
Spoiler | -1.348 |
ンジ | -1.271 |
taboola | -1.251 |
Doors | -1.239 |
Token | breakup |
Token ID | 41584 |
Feature activation | -1.11e+00 |
Pos logit contributions | |
robe | +3.367 |
Spoiler | +2.685 |
ンジ | +2.499 |
Doors | +2.477 |
roo | +2.441 |
Neg logit contributions | |
bia | -2.227 |
rele | -2.144 |
constitu | -2.131 |
ibaba | -2.129 |
ilial | -2.118 |
Token | on |
Token ID | 319 |
Feature activation | +1.14e-01 |
Pos logit contributions | |
robe | +6.578 |
Spoiler | +5.151 |
roo | +4.757 |
ンジ | +4.721 |
Doors | +4.720 |
Neg logit contributions | |
bia | -4.479 |
ibaba | -4.453 |
constitu | -4.453 |
iland | -4.316 |
rele | -4.275 |
Token | Valentine |
Token ID | 23078 |
Feature activation | -1.99e+00 |
Pos logit contributions | |
rele | +0.458 |
bia | +0.453 |
constitu | +0.449 |
iland | +0.432 |
ibaba | +0.431 |
Neg logit contributions | |
robe | -0.700 |
Spoiler | -0.549 |
ンジ | -0.518 |
taboola | -0.510 |
hib | -0.504 |
Token | � |
Token ID | 447 |
Feature activation | -3.53e-01 |
Pos logit contributions | |
robe | +9.739 |
Spoiler | +7.317 |
Crate | +6.741 |
Vik | +6.612 |
Doors | +6.573 |
Neg logit contributions | |
bia | -8.978 |
constitu | -8.938 |
iland | -8.796 |
ibaba | -8.774 |
ilial | -8.699 |
Token | � |
Token ID | 247 |
Feature activation | +3.90e-02 |
Pos logit contributions | |
robe | +1.520 |
Spoiler | +1.043 |
ンジ | +0.947 |
taboola | +0.940 |
hib | +0.909 |
Neg logit contributions | |
rele | -2.074 |
bia | -2.066 |
constitu | -2.053 |
iland | -1.996 |
ibaba | -1.993 |
Token | s |
Token ID | 82 |
Feature activation | -2.24e-01 |
Pos logit contributions | |
rele | +0.142 |
constitu | +0.139 |
bia | +0.139 |
ibaba | +0.133 |
iland | +0.133 |
Neg logit contributions | |
robe | -0.256 |
Spoiler | -0.204 |
ンジ | -0.193 |
taboola | -0.191 |
hib | -0.189 |
Token | Day |
Token ID | 3596 |
Feature activation | -8.65e-02 |
Pos logit contributions | |
robe | +1.177 |
Spoiler | +0.881 |
ンジ | +0.813 |
taboola | +0.807 |
Doors | +0.794 |
Neg logit contributions | |
rele | -1.097 |
bia | -1.096 |
constitu | -1.083 |
iland | -1.055 |
ibaba | -1.053 |
Token | . |
Token ID | 13 |
Feature activation | -7.00e-02 |
Pos logit contributions | |
robe | +0.512 |
Spoiler | +0.397 |
ンジ | +0.372 |
taboola | +0.367 |
hib | +0.363 |
Neg logit contributions | |
rele | -0.371 |
bia | -0.366 |
constitu | -0.363 |
iland | -0.350 |
ibaba | -0.350 |
Token | All |
Token ID | 1439 |
Feature activation | +5.31e-01 |
Pos logit contributions | |
robe | +2.348 |
Spoiler | +1.862 |
Doors | +1.644 |
ンジ | +1.619 |
Crate | +1.604 |
Neg logit contributions | |
rele | -2.440 |
constitu | -2.426 |
bia | -2.414 |
ibaba | -2.360 |
iland | -2.333 |
Token | three |
Token ID | 1115 |
Feature activation | +3.13e-01 |
Pos logit contributions | |
rele | +1.846 |
bia | +1.791 |
constitu | +1.770 |
ibaba | +1.695 |
iland | +1.690 |
Neg logit contributions | |
robe | -3.580 |
Spoiler | -2.853 |
ンジ | -2.726 |
taboola | -2.695 |
hib | -2.665 |
Token | could |
Token ID | 714 |
Feature activation | -2.17e-01 |
Pos logit contributions | |
rele | +1.006 |
bia | +0.985 |
constitu | +0.975 |
iland | +0.929 |
ibaba | +0.928 |
Neg logit contributions | |
robe | -2.185 |
Spoiler | -1.767 |
ンジ | -1.684 |
taboola | -1.664 |
hib | -1.647 |
Token | not |
Token ID | 407 |
Feature activation | -1.74e-01 |
Pos logit contributions | |
robe | +1.397 |
Spoiler | +1.111 |
taboola | +1.034 |
ンジ | +1.029 |
hib | +1.021 |
Neg logit contributions | |
constitu | -0.790 |
bia | -0.786 |
rele | -0.774 |
ibaba | -0.768 |
ilial | -0.745 |
Token | be |
Token ID | 307 |
Feature activation | -7.70e-01 |
Pos logit contributions | |
robe | +1.135 |
Spoiler | +0.910 |
taboola | +0.847 |
ンジ | +0.844 |
Doors | +0.839 |
Neg logit contributions | |
constitu | -0.626 |
bia | -0.622 |
ibaba | -0.608 |
rele | -0.607 |
iland | -0.587 |
Token | further |
Token ID | 2252 |
Feature activation | -1.97e+00 |
Pos logit contributions | |
robe | +4.675 |
Spoiler | +3.816 |
taboola | +3.527 |
ンジ | +3.485 |
Doors | +3.478 |
Neg logit contributions | |
bia | -2.940 |
ibaba | -2.850 |
constitu | -2.844 |
rele | -2.720 |
iland | -2.696 |
Token | from |
Token ID | 422 |
Feature activation | +5.09e-02 |
Pos logit contributions | |
robe | +10.025 |
Spoiler | +8.395 |
roo | +7.498 |
Doors | +7.419 |
taboola | +7.416 |
Neg logit contributions | |
ibaba | -8.192 |
bia | -8.105 |
constitu | -7.666 |
rele | -7.566 |
iland | -7.396 |
Token | the |
Token ID | 262 |
Feature activation | +5.66e-02 |
Pos logit contributions | |
bia | +0.199 |
rele | +0.199 |
constitu | +0.197 |
iland | +0.191 |
ibaba | +0.189 |
Neg logit contributions | |
robe | -0.319 |
Spoiler | -0.252 |
ンジ | -0.235 |
taboola | -0.231 |
ヤ | -0.230 |
Token | truth |
Token ID | 3872 |
Feature activation | -1.30e-01 |
Pos logit contributions | |
bia | +0.178 |
rele | +0.175 |
constitu | +0.173 |
iland | +0.168 |
ibaba | +0.166 |
Neg logit contributions | |
robe | -0.397 |
Spoiler | -0.325 |
ンジ | -0.305 |
taboola | -0.302 |
Doors | -0.301 |
Token | . |
Token ID | 13 |
Feature activation | -4.64e-02 |
Pos logit contributions | |
robe | +0.759 |
Spoiler | +0.586 |
ンジ | +0.544 |
taboola | +0.536 |
hib | +0.533 |
Neg logit contributions | |
rele | -0.566 |
bia | -0.563 |
constitu | -0.560 |
iland | -0.541 |
ibaba | -0.540 |
Token | Zed |
Token ID | 46159 |
Feature activation | -6.03e-02 |
Pos logit contributions | |
robe | +0.244 |
Spoiler | +0.188 |
ンジ | +0.171 |
Doors | +0.168 |
taboola | +0.167 |
Neg logit contributions | |
rele | -0.228 |
bia | -0.225 |
constitu | -0.224 |
ibaba | -0.218 |
iland | -0.217 |
Token | a |
Token ID | 64 |
Feature activation | +2.74e-01 |
Pos logit contributions | |
rele | +0.351 |
bia | +0.345 |
constitu | +0.345 |
iland | +0.332 |
ibaba | +0.331 |
Neg logit contributions | |
robe | -0.440 |
Spoiler | -0.336 |
ンジ | -0.314 |
taboola | -0.310 |
hib | -0.306 |
Token | is |
Token ID | 318 |
Feature activation | -1.98e-01 |
Pos logit contributions | |
rele | +1.078 |
bia | +1.063 |
constitu | +1.053 |
ibaba | +1.012 |
iland | +1.012 |
Neg logit contributions | |
robe | -1.713 |
Spoiler | -1.347 |
ンジ | -1.275 |
taboola | -1.257 |
hib | -1.243 |
Token | only |
Token ID | 691 |
Feature activation | +3.69e-01 |
Pos logit contributions | |
robe | +1.259 |
Spoiler | +0.985 |
hib | +0.932 |
ンジ | +0.930 |
Crate | +0.915 |
Neg logit contributions | |
bia | -0.748 |
constitu | -0.730 |
rele | -0.726 |
ibaba | -0.716 |
iland | -0.707 |
Token | the |
Token ID | 262 |
Feature activation | +1.10e-01 |
Pos logit contributions | |
rele | +1.215 |
bia | +1.190 |
constitu | +1.178 |
iland | +1.124 |
ibaba | +1.122 |
Neg logit contributions | |
robe | -2.555 |
Spoiler | -2.061 |
ンジ | -1.963 |
taboola | -1.940 |
hib | -1.919 |
Token | tip |
Token ID | 8171 |
Feature activation | -6.60e-01 |
Pos logit contributions | |
rele | +0.395 |
bia | +0.393 |
constitu | +0.388 |
ibaba | +0.375 |
iland | +0.375 |
Neg logit contributions | |
robe | -0.728 |
Spoiler | -0.581 |
ンジ | -0.549 |
hib | -0.539 |
taboola | -0.538 |
Token | of |
Token ID | 286 |
Feature activation | -1.96e+00 |
Pos logit contributions | |
robe | +4.984 |
Spoiler | +4.102 |
ンジ | +3.932 |
taboola | +3.890 |
hib | +3.846 |
Neg logit contributions | |
rele | -1.721 |
bia | -1.714 |
constitu | -1.651 |
iland | -1.578 |
ibaba | -1.565 |
Token | the |
Token ID | 262 |
Feature activation | -1.69e+00 |
Pos logit contributions | |
robe | +9.538 |
Spoiler | +7.299 |
Doors | +6.830 |
ンジ | +6.817 |
hib | +6.704 |
Neg logit contributions | |
bia | -9.164 |
iland | -8.631 |
constitu | -8.598 |
rele | -8.381 |
ilial | -8.338 |
Token | iceberg |
Token ID | 34849 |
Feature activation | -6.20e-01 |
Pos logit contributions | |
robe | +5.593 |
Spoiler | +3.568 |
ンジ | +3.093 |
roo | +2.978 |
Doors | +2.975 |
Neg logit contributions | |
bia | -10.803 |
ilial | -10.450 |
ibaba | -10.391 |
iland | -10.387 |
constitu | -10.341 |
Token | ," |
Token ID | 553 |
Feature activation | -1.98e-01 |
Pos logit contributions | |
robe | +3.612 |
Spoiler | +2.767 |
taboola | +2.588 |
hib | +2.562 |
ンジ | +2.550 |
Neg logit contributions | |
bia | -2.664 |
constitu | -2.588 |
rele | -2.558 |
ibaba | -2.514 |
iland | -2.495 |
Token | he |
Token ID | 339 |
Feature activation | -2.25e-01 |
Pos logit contributions | |
robe | +1.246 |
Spoiler | +0.991 |
ンジ | +0.930 |
taboola | +0.914 |
hib | +0.907 |
Neg logit contributions | |
rele | -0.769 |
bia | -0.765 |
constitu | -0.758 |
ibaba | -0.721 |
iland | -0.718 |
Token | told |
Token ID | 1297 |
Feature activation | -3.08e-01 |
Pos logit contributions | |
robe | +1.707 |
Spoiler | +1.399 |
ンジ | +1.349 |
taboola | +1.332 |
hib | +1.316 |
Neg logit contributions | |
rele | -0.580 |
constitu | -0.571 |
bia | -0.569 |
ibaba | -0.540 |
iland | -0.524 |
Token | а |
Token ID | 16142 |
Feature activation | -1.23e-02 |
Pos logit contributions | |
robe | +1.312 |
Spoiler | +0.983 |
ンジ | +0.926 |
ヤ | +0.908 |
taboola | +0.901 |
Neg logit contributions | |
rele | -1.148 |
bia | -1.139 |
constitu | -1.132 |
ibaba | -1.100 |
iland | -1.092 |
Token | р |
Token ID | 21169 |
Feature activation | -3.25e-01 |
Pos logit contributions | |
robe | +0.065 |
Spoiler | +0.049 |
ンジ | +0.046 |
taboola | +0.045 |
hib | +0.044 |
Neg logit contributions | |
rele | -0.060 |
bia | -0.059 |
constitu | -0.059 |
ibaba | -0.057 |
iland | -0.057 |
Token | а |
Token ID | 16142 |
Feature activation | +1.85e-02 |
Pos logit contributions | |
robe | +1.683 |
Spoiler | +1.237 |
ンジ | +1.172 |
ヤ | +1.150 |
hib | +1.127 |
Neg logit contributions | |
rele | -1.607 |
bia | -1.596 |
constitu | -1.591 |
ibaba | -1.560 |
iland | -1.531 |
Token | к |
Token ID | 31583 |
Feature activation | -2.75e-01 |
Pos logit contributions | |
rele | +0.092 |
bia | +0.091 |
constitu | +0.091 |
ibaba | +0.088 |
iland | +0.088 |
Neg logit contributions | |
robe | -0.096 |
Spoiler | -0.071 |
ンジ | -0.067 |
taboola | -0.065 |
ヤ | -0.065 |
Token | т |
Token ID | 20375 |
Feature activation | -1.38e+00 |
Pos logit contributions | |
robe | +1.433 |
Spoiler | +1.059 |
ンジ | +0.991 |
ヤ | +0.968 |
taboola | +0.963 |
Neg logit contributions | |
bia | -1.351 |
rele | -1.350 |
constitu | -1.341 |
ibaba | -1.315 |
iland | -1.298 |
Token | е |
Token ID | 16843 |
Feature activation | -1.93e+00 |
Pos logit contributions | |
robe | +6.681 |
Spoiler | +4.953 |
ンジ | +4.627 |
ヤ | +4.562 |
Doors | +4.507 |
Neg logit contributions | |
constitu | -6.540 |
bia | -6.539 |
rele | -6.449 |
ibaba | -6.444 |
ilial | -6.329 |
Token | р |
Token ID | 21169 |
Feature activation | -8.03e-01 |
Pos logit contributions | |
robe | +8.528 |
Spoiler | +6.082 |
roo | +5.795 |
ヤ | +5.741 |
ンジ | +5.702 |
Neg logit contributions | |
constitu | -9.362 |
ibaba | -9.360 |
bia | -9.355 |
iland | -9.151 |
rele | -9.130 |
Token | н |
Token ID | 22177 |
Feature activation | -4.83e-01 |
Pos logit contributions | |
robe | +3.932 |
Spoiler | +2.867 |
ンジ | +2.693 |
ヤ | +2.647 |
Doors | +2.586 |
Neg logit contributions | |
bia | -4.030 |
constitu | -4.028 |
rele | -4.020 |
ibaba | -3.959 |
iland | -3.895 |
Token | о� |
Token ID | 25443 |
Feature activation | -9.36e-01 |
Pos logit contributions | |
robe | +2.411 |
Spoiler | +1.759 |
ンジ | +1.670 |
ヤ | +1.617 |
taboola | +1.602 |
Neg logit contributions | |
rele | -2.451 |
bia | -2.451 |
constitu | -2.436 |
ibaba | -2.392 |
iland | -2.358 |
Token | � |
Token ID | 117 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
robe | +3.767 |
Spoiler | +2.470 |
ンジ | +2.318 |
taboola | +2.267 |
roo | +2.249 |
Neg logit contributions | |
rele | -5.602 |
constitu | -5.564 |
iland | -5.421 |
ilial | -5.406 |
bia | -5.401 |
Token | � |
Token ID | 12466 |
Feature activation | -3.65e-01 |
Pos logit contributions | |
robe | +0.760 |
Spoiler | +0.556 |
ンジ | +0.514 |
ヤ | +0.498 |
taboola | +0.497 |
Neg logit contributions | |
rele | -0.813 |
bia | -0.808 |
constitu | -0.801 |
ibaba | -0.787 |
iland | -0.783 |
Token | met |
Token ID | 4164 |
Feature activation | -3.17e-01 |
Pos logit contributions | |
robe | +0.175 |
Spoiler | +0.128 |
ンジ | +0.118 |
taboola | +0.115 |
ヤ | +0.114 |
Neg logit contributions | |
rele | -0.185 |
bia | -0.183 |
constitu | -0.181 |
iland | -0.176 |
ibaba | -0.176 |
Token | re |
Token ID | 260 |
Feature activation | -6.64e-01 |
Pos logit contributions | |
robe | +1.852 |
Spoiler | +1.426 |
ンジ | +1.337 |
taboola | +1.323 |
hib | +1.305 |
Neg logit contributions | |
rele | -1.384 |
bia | -1.366 |
constitu | -1.363 |
ibaba | -1.313 |
iland | -1.310 |
Token | distance |
Token ID | 5253 |
Feature activation | -8.58e-02 |
Pos logit contributions | |
robe | +4.032 |
Spoiler | +3.165 |
ンジ | +2.973 |
taboola | +2.949 |
Doors | +2.908 |
Neg logit contributions | |
bia | -2.636 |
rele | -2.603 |
constitu | -2.590 |
ibaba | -2.509 |
iland | -2.494 |
Token | from |
Token ID | 422 |
Feature activation | -6.67e-01 |
Pos logit contributions | |
robe | +0.587 |
Spoiler | +0.473 |
ンジ | +0.450 |
taboola | +0.445 |
hib | +0.440 |
Neg logit contributions | |
rele | -0.287 |
bia | -0.283 |
constitu | -0.279 |
ibaba | -0.267 |
iland | -0.267 |
Token | 2 |
Token ID | 362 |
Feature activation | -4.85e-01 |
Pos logit contributions | |
robe | +3.647 |
Spoiler | +2.820 |
Doors | +2.598 |
ンジ | +2.582 |
Crate | +2.563 |
Neg logit contributions | |
bia | -3.034 |
rele | -3.027 |
constitu | -2.999 |
iland | -2.940 |
ibaba | -2.927 |
Token | 221 |
Token ID | 26115 |
Feature activation | -1.91e+00 |
Pos logit contributions | |
robe | +2.668 |
Spoiler | +2.025 |
Doors | +1.896 |
ンジ | +1.865 |
hib | +1.861 |
Neg logit contributions | |
rele | -2.227 |
constitu | -2.213 |
bia | -2.209 |
ibaba | -2.154 |
iland | -2.141 |
Token | Y |
Token ID | 575 |
Feature activation | -8.47e-01 |
Pos logit contributions | |
robe | +8.750 |
Spoiler | +6.841 |
Doors | +6.554 |
Crate | +6.389 |
Door | +6.199 |
Neg logit contributions | |
ibaba | -8.778 |
constitu | -8.710 |
rele | -8.452 |
bia | -8.437 |
iland | -8.381 |
Token | onge |
Token ID | 14220 |
Feature activation | -1.34e+00 |
Pos logit contributions | |
robe | +2.497 |
Spoiler | +1.433 |
Crate | +1.109 |
Doors | +1.092 |
hib | +1.085 |
Neg logit contributions | |
constitu | -6.080 |
rele | -5.919 |
ibaba | -5.885 |
bia | -5.871 |
ilial | -5.790 |
Token | Cond |
Token ID | 9724 |
Feature activation | -1.08e+00 |
Pos logit contributions | |
robe | +6.989 |
Spoiler | +5.254 |
Doors | +5.010 |
Crate | +4.854 |
taboola | +4.798 |
Neg logit contributions | |
bia | -6.032 |
constitu | -6.029 |
ibaba | -5.943 |
rele | -5.796 |
iland | -5.767 |
Token | os |
Token ID | 418 |
Feature activation | -7.42e-01 |
Pos logit contributions | |
robe | +5.985 |
Spoiler | +4.445 |
Doors | +4.204 |
roo | +4.165 |
erald | +4.098 |
Neg logit contributions | |
constitu | -4.895 |
rele | -4.782 |
bia | -4.607 |
ibaba | -4.565 |
iland | -4.476 |
Token | to |
Token ID | 284 |
Feature activation | -3.46e-01 |
Pos logit contributions | |
robe | +4.339 |
Spoiler | +3.339 |
ンジ | +3.133 |
Doors | +3.112 |
taboola | +3.112 |
Neg logit contributions | |
constitu | -3.120 |
bia | -3.119 |
rele | -3.119 |
ibaba | -3.002 |
iland | -2.968 |
Token | of |
Token ID | 286 |
Feature activation | +3.56e-02 |
Pos logit contributions | |
rele | +1.708 |
bia | +1.659 |
constitu | +1.648 |
iland | +1.569 |
ibaba | +1.566 |
Neg logit contributions | |
robe | -3.541 |
Spoiler | -2.849 |
ンジ | -2.723 |
taboola | -2.688 |
hib | -2.655 |
Token | factors |
Token ID | 5087 |
Feature activation | -3.98e-01 |
Pos logit contributions | |
bia | +0.131 |
rele | +0.129 |
constitu | +0.128 |
ibaba | +0.124 |
iland | +0.124 |
Neg logit contributions | |
robe | -0.231 |
Spoiler | -0.184 |
ンジ | -0.173 |
Doors | -0.172 |
taboola | -0.171 |
Token | determines |
Token ID | 15947 |
Feature activation | -3.32e-01 |
Pos logit contributions | |
robe | +2.377 |
Spoiler | +1.885 |
Doors | +1.725 |
ンジ | +1.722 |
hib | +1.714 |
Neg logit contributions | |
bia | -1.643 |
constitu | -1.605 |
rele | -1.599 |
ibaba | -1.575 |
iland | -1.566 |
Token | whether |
Token ID | 1771 |
Feature activation | +5.07e-02 |
Pos logit contributions | |
robe | +2.009 |
Spoiler | +1.585 |
Doors | +1.456 |
taboola | +1.452 |
ンジ | +1.451 |
Neg logit contributions | |
bia | -1.365 |
constitu | -1.318 |
rele | -1.311 |
iland | -1.296 |
ibaba | -1.279 |
Token | a |
Token ID | 257 |
Feature activation | -3.84e-01 |
Pos logit contributions | |
rele | +0.179 |
bia | +0.178 |
constitu | +0.175 |
iland | +0.168 |
ibaba | +0.168 |
Neg logit contributions | |
robe | -0.337 |
Spoiler | -0.270 |
ンジ | -0.255 |
taboola | -0.253 |
Doors | -0.251 |
Token | locker |
Token ID | 17355 |
Feature activation | -1.90e+00 |
Pos logit contributions | |
robe | +2.060 |
Spoiler | +1.542 |
Doors | +1.418 |
Crate | +1.413 |
taboola | +1.412 |
Neg logit contributions | |
bia | -1.869 |
rele | -1.801 |
constitu | -1.799 |
ibaba | -1.783 |
edi | -1.753 |
Token | is |
Token ID | 318 |
Feature activation | -2.71e-01 |
Pos logit contributions | |
robe | +10.105 |
Spoiler | +7.687 |
Crate | +7.356 |
Doors | +7.344 |
taboola | +7.186 |
Neg logit contributions | |
bia | -8.121 |
constitu | -8.021 |
ibaba | -7.862 |
iland | -7.675 |
ilial | -7.528 |
Token | open |
Token ID | 1280 |
Feature activation | -2.32e-01 |
Pos logit contributions | |
robe | +1.492 |
Spoiler | +1.136 |
Doors | +1.044 |
taboola | +1.041 |
hib | +1.033 |
Neg logit contributions | |
bia | -1.269 |
constitu | -1.235 |
rele | -1.227 |
ibaba | -1.213 |
iland | -1.205 |
Token | or |
Token ID | 393 |
Feature activation | -2.39e-01 |
Pos logit contributions | |
robe | +1.491 |
Spoiler | +1.182 |
ンジ | +1.106 |
taboola | +1.098 |
Doors | +1.095 |
Neg logit contributions | |
bia | -0.870 |
rele | -0.867 |
constitu | -0.863 |
ibaba | -0.824 |
iland | -0.820 |
Token | closed |
Token ID | 4838 |
Feature activation | -4.02e-01 |
Pos logit contributions | |
robe | +1.317 |
Spoiler | +1.008 |
taboola | +0.924 |
ンジ | +0.919 |
hib | +0.919 |
Neg logit contributions | |
bia | -1.106 |
constitu | -1.096 |
rele | -1.091 |
ibaba | -1.068 |
iland | -1.053 |
Token | because |
Token ID | 780 |
Feature activation | +4.82e-01 |
Pos logit contributions | |
robe | +2.331 |
Spoiler | +1.797 |
ンジ | +1.658 |
taboola | +1.651 |
Doors | +1.643 |
Neg logit contributions | |
bia | -1.753 |
rele | -1.737 |
constitu | -1.736 |
ibaba | -1.676 |
iland | -1.668 |
Token | Sign |
Token ID | 5865 |
Feature activation | -6.08e-01 |
Pos logit contributions | |
robe | +1.766 |
Spoiler | +1.176 |
taboola | +0.993 |
ンジ | +0.992 |
Doors | +0.957 |
Neg logit contributions | |
rele | -3.043 |
bia | -3.032 |
constitu | -3.019 |
ibaba | -2.957 |
iland | -2.950 |
Token | Up |
Token ID | 3205 |
Feature activation | -2.06e-01 |
Pos logit contributions | |
robe | +1.962 |
Spoiler | +1.206 |
taboola | +0.993 |
Doors | +0.991 |
ンジ | +0.964 |
Neg logit contributions | |
rele | -4.232 |
constitu | -4.217 |
bia | -4.199 |
ibaba | -4.098 |
iland | -4.076 |
Token | Continue |
Token ID | 10054 |
Feature activation | -8.52e-01 |
Pos logit contributions | |
robe | +1.031 |
Spoiler | +0.751 |
ンジ | +0.709 |
taboola | +0.680 |
hib | +0.680 |
Neg logit contributions | |
rele | -1.069 |
bia | -1.051 |
constitu | -1.044 |
iland | -1.012 |
ibaba | -1.009 |
Token | reading |
Token ID | 3555 |
Feature activation | -9.63e-01 |
Pos logit contributions | |
robe | +5.518 |
Spoiler | +4.490 |
taboola | +4.190 |
Crate | +4.148 |
Doors | +4.121 |
Neg logit contributions | |
constitu | -3.100 |
bia | -2.988 |
rele | -2.939 |
ibaba | -2.916 |
iland | -2.913 |
Token | the |
Token ID | 262 |
Feature activation | -9.02e-01 |
Pos logit contributions | |
robe | +5.213 |
Spoiler | +4.086 |
taboola | +3.668 |
Crate | +3.628 |
Doors | +3.626 |
Neg logit contributions | |
constitu | -4.461 |
bia | -4.363 |
rele | -4.277 |
iland | -4.272 |
ibaba | -4.256 |
Token | main |
Token ID | 1388 |
Feature activation | -1.89e+00 |
Pos logit contributions | |
robe | +5.035 |
Spoiler | +4.027 |
Doors | +3.673 |
taboola | +3.672 |
ンジ | +3.565 |
Neg logit contributions | |
bia | -3.912 |
constitu | -3.895 |
iland | -3.890 |
ibaba | -3.730 |
edi | -3.700 |
Token | story |
Token ID | 1621 |
Feature activation | -6.49e-01 |
Pos logit contributions | |
robe | +7.751 |
Spoiler | +6.131 |
taboola | +5.400 |
Doors | +5.395 |
roo | +5.158 |
Neg logit contributions | |
bia | -10.186 |
constitu | -10.128 |
ibaba | -9.614 |
iland | -9.568 |
ilial | -9.369 |
Token | Please |
Token ID | 4222 |
Feature activation | +2.16e-01 |
Pos logit contributions | |
robe | +2.903 |
Spoiler | +2.163 |
taboola | +1.945 |
ンジ | +1.926 |
Doors | +1.889 |
Neg logit contributions | |
constitu | -3.578 |
bia | -3.569 |
rele | -3.567 |
edi | -3.483 |
iland | -3.482 |
Token | verify |
Token ID | 11767 |
Feature activation | -2.67e-01 |
Pos logit contributions | |
rele | +0.738 |
bia | +0.706 |
constitu | +0.699 |
iland | +0.669 |
ibaba | +0.663 |
Neg logit contributions | |
robe | -1.473 |
Spoiler | -1.183 |
ンジ | -1.134 |
Doors | -1.106 |
taboola | -1.105 |
Token | you |
Token ID | 345 |
Feature activation | -3.27e-01 |
Pos logit contributions | |
robe | +1.936 |
Spoiler | +1.576 |
taboola | +1.510 |
ンジ | +1.493 |
roo | +1.480 |
Neg logit contributions | |
constitu | -0.774 |
rele | -0.770 |
bia | -0.761 |
ibaba | -0.734 |
iland | -0.723 |
Token | 're |
Token ID | 821 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
robe | +1.890 |
Spoiler | +1.432 |
taboola | +1.353 |
ンジ | +1.322 |
ヤ | +1.310 |
Neg logit contributions | |
constitu | -1.425 |
rele | -1.424 |
bia | -1.412 |
ibaba | -1.385 |
iland | -1.371 |
Token | base |
Token ID | 2779 |
Feature activation | -3.10e-01 |
Pos logit contributions | |
robe | +1.746 |
Spoiler | +1.384 |
ンジ | +1.292 |
hib | +1.278 |
ヤ | +1.269 |
Neg logit contributions | |
rele | -1.094 |
bia | -1.091 |
constitu | -1.089 |
iland | -1.059 |
ibaba | -1.058 |
Token | , |
Token ID | 11 |
Feature activation | -3.81e-02 |
Pos logit contributions | |
robe | +1.841 |
Spoiler | +1.439 |
ンジ | +1.328 |
taboola | +1.313 |
hib | +1.312 |
Neg logit contributions | |
bia | -1.302 |
rele | -1.288 |
constitu | -1.274 |
ibaba | -1.258 |
iland | -1.249 |
Token | nothing |
Token ID | 2147 |
Feature activation | -4.77e-01 |
Pos logit contributions | |
robe | +0.227 |
Spoiler | +0.181 |
ンジ | +0.165 |
Doors | +0.164 |
taboola | +0.164 |
Neg logit contributions | |
bia | -0.158 |
rele | -0.157 |
constitu | -0.155 |
iland | -0.151 |
ibaba | -0.151 |
Token | could |
Token ID | 714 |
Feature activation | -2.90e-01 |
Pos logit contributions | |
robe | +2.863 |
Spoiler | +2.298 |
Doors | +2.093 |
ンジ | +2.091 |
taboola | +2.089 |
Neg logit contributions | |
bia | -1.927 |
rele | -1.884 |
iland | -1.867 |
ibaba | -1.855 |
constitu | -1.847 |
Token | be |
Token ID | 307 |
Feature activation | -9.89e-02 |
Pos logit contributions | |
robe | +1.774 |
Spoiler | +1.417 |
taboola | +1.312 |
ンジ | +1.308 |
Doors | +1.298 |
Neg logit contributions | |
bia | -1.150 |
constitu | -1.139 |
rele | -1.135 |
ibaba | -1.120 |
iland | -1.090 |
Token | further |
Token ID | 2252 |
Feature activation | -1.89e+00 |
Pos logit contributions | |
robe | +0.653 |
Spoiler | +0.526 |
ンジ | +0.495 |
taboola | +0.494 |
roo | +0.490 |
Neg logit contributions | |
bia | -0.343 |
constitu | -0.342 |
rele | -0.337 |
ibaba | -0.329 |
iland | -0.327 |
Token | from |
Token ID | 422 |
Feature activation | +1.85e-01 |
Pos logit contributions | |
robe | +9.897 |
Spoiler | +8.039 |
roo | +7.318 |
Crate | +7.148 |
Doors | +7.116 |
Neg logit contributions | |
bia | -7.849 |
ibaba | -7.846 |
constitu | -7.402 |
rele | -7.341 |
iland | -7.338 |
Token | the |
Token ID | 262 |
Feature activation | +2.60e-01 |
Pos logit contributions | |
rele | +0.725 |
bia | +0.711 |
constitu | +0.706 |
ibaba | +0.678 |
iland | +0.677 |
Neg logit contributions | |
robe | -1.160 |
Spoiler | -0.912 |
ンジ | -0.865 |
taboola | -0.854 |
hib | -0.843 |
Token | truth |
Token ID | 3872 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
rele | +0.789 |
bia | +0.789 |
constitu | +0.771 |
iland | +0.741 |
ibaba | +0.738 |
Neg logit contributions | |
robe | -1.857 |
Spoiler | -1.514 |
ンジ | -1.438 |
taboola | -1.416 |
Doors | -1.411 |
Token | . |
Token ID | 13 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
robe | +1.050 |
Spoiler | +0.810 |
ンジ | +0.751 |
taboola | +0.740 |
hib | +0.737 |
Neg logit contributions | |
rele | -0.806 |
bia | -0.800 |
constitu | -0.796 |
iland | -0.772 |
ibaba | -0.771 |
Token | He |
Token ID | 679 |
Feature activation | -8.23e-02 |
Pos logit contributions | |
rele | +0.093 |
bia | +0.092 |
constitu | +0.091 |
ibaba | +0.089 |
iland | +0.088 |
Neg logit contributions | |
robe | -0.104 |
Spoiler | -0.080 |
ンジ | -0.074 |
Doors | -0.072 |
taboola | -0.072 |
Token | comes |
Token ID | 2058 |
Feature activation | -2.00e-01 |
Pos logit contributions | |
robe | +1.658 |
Spoiler | +1.258 |
Doors | +1.147 |
taboola | +1.144 |
ンジ | +1.138 |
Neg logit contributions | |
bia | -1.444 |
constitu | -1.421 |
rele | -1.402 |
iland | -1.393 |
ibaba | -1.393 |
Token | by |
Token ID | 416 |
Feature activation | -2.63e-01 |
Pos logit contributions | |
robe | +1.211 |
Spoiler | +0.953 |
ンジ | +0.884 |
taboola | +0.873 |
Doors | +0.867 |
Neg logit contributions | |
bia | -0.828 |
rele | -0.817 |
constitu | -0.814 |
ibaba | -0.785 |
iland | -0.782 |
Token | , |
Token ID | 11 |
Feature activation | +2.09e-01 |
Pos logit contributions | |
robe | +1.570 |
Spoiler | +1.220 |
ンジ | +1.126 |
Doors | +1.113 |
hib | +1.112 |
Neg logit contributions | |
bia | -1.111 |
rele | -1.096 |
constitu | -1.094 |
ibaba | -1.057 |
iland | -1.050 |
Token | every |
Token ID | 790 |
Feature activation | -1.79e-01 |
Pos logit contributions | |
bia | +0.819 |
rele | +0.813 |
constitu | +0.804 |
iland | +0.778 |
ibaba | +0.773 |
Neg logit contributions | |
robe | -1.314 |
Spoiler | -1.041 |
ンジ | -0.972 |
taboola | -0.962 |
Doors | -0.957 |
Token | third |
Token ID | 2368 |
Feature activation | -1.84e-01 |
Pos logit contributions | |
robe | +1.131 |
Spoiler | +0.887 |
ンジ | +0.833 |
Doors | +0.826 |
taboola | +0.823 |
Neg logit contributions | |
bia | -0.704 |
rele | -0.683 |
constitu | -0.681 |
ibaba | -0.664 |
iland | -0.654 |
Token | locker |
Token ID | 17355 |
Feature activation | -1.88e+00 |
Pos logit contributions | |
robe | +1.055 |
Spoiler | +0.807 |
ンジ | +0.751 |
taboola | +0.748 |
Doors | +0.739 |
Neg logit contributions | |
bia | -0.824 |
rele | -0.814 |
constitu | -0.811 |
ibaba | -0.785 |
iland | -0.777 |
Token | . |
Token ID | 13 |
Feature activation | +2.34e-01 |
Pos logit contributions | |
robe | +9.152 |
Spoiler | +6.838 |
Doors | +6.329 |
Crate | +6.308 |
taboola | +6.164 |
Neg logit contributions | |
bia | -8.969 |
constitu | -8.621 |
ibaba | -8.601 |
iland | -8.416 |
ilial | -8.264 |
Token | If |
Token ID | 1002 |
Feature activation | +4.94e-01 |
Pos logit contributions | |
rele | +1.135 |
bia | +1.120 |
constitu | +1.112 |
iland | +1.078 |
ibaba | +1.077 |
Neg logit contributions | |
robe | -1.247 |
Spoiler | -0.933 |
ンジ | -0.873 |
taboola | -0.858 |
hib | -0.845 |
Token | it |
Token ID | 340 |
Feature activation | -3.94e-01 |
Pos logit contributions | |
rele | +1.709 |
bia | +1.657 |
constitu | +1.649 |
ibaba | +1.577 |
iland | +1.571 |
Neg logit contributions | |
robe | -3.333 |
Spoiler | -2.666 |
ンジ | -2.558 |
taboola | -2.519 |
hib | -2.483 |
Token | 's |
Token ID | 338 |
Feature activation | +2.66e-01 |
Pos logit contributions | |
robe | +2.412 |
Spoiler | +1.920 |
taboola | +1.811 |
ンジ | +1.781 |
hib | +1.766 |
Neg logit contributions | |
bia | -1.561 |
constitu | -1.543 |
rele | -1.524 |
ibaba | -1.485 |
iland | -1.480 |
Token | open |
Token ID | 1280 |
Feature activation | -2.21e-01 |
Pos logit contributions | |
rele | +0.959 |
bia | +0.943 |
constitu | +0.934 |
iland | +0.896 |
ibaba | +0.895 |
Neg logit contributions | |
robe | -1.757 |
Spoiler | -1.402 |
ンジ | -1.330 |
taboola | -1.313 |
hib | -1.299 |
Token | goes |
Token ID | 2925 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
robe | +2.613 |
Spoiler | +2.020 |
Doors | +1.862 |
taboola | +1.847 |
hib | +1.825 |
Neg logit contributions | |
bia | -2.044 |
constitu | -2.007 |
ibaba | -1.961 |
iland | -1.953 |
rele | -1.941 |
Token | by |
Token ID | 416 |
Feature activation | -6.26e-01 |
Pos logit contributions | |
robe | +0.075 |
Spoiler | +0.059 |
ンジ | +0.055 |
taboola | +0.055 |
hib | +0.054 |
Neg logit contributions | |
rele | -0.052 |
bia | -0.052 |
constitu | -0.051 |
iland | -0.049 |
ibaba | -0.049 |
Token | . |
Token ID | 13 |
Feature activation | +7.12e-02 |
Pos logit contributions | |
robe | +3.577 |
Spoiler | +2.755 |
Doors | +2.496 |
ンジ | +2.488 |
hib | +2.480 |
Neg logit contributions | |
bia | -2.783 |
constitu | -2.742 |
rele | -2.689 |
ibaba | -2.661 |
iland | -2.622 |
Token | Every |
Token ID | 3887 |
Feature activation | +2.33e-02 |
Pos logit contributions | |
rele | +0.345 |
bia | +0.341 |
constitu | +0.339 |
ibaba | +0.328 |
iland | +0.327 |
Neg logit contributions | |
robe | -0.383 |
Spoiler | -0.293 |
ンジ | -0.269 |
Doors | -0.264 |
taboola | -0.263 |
Token | fourth |
Token ID | 5544 |
Feature activation | -3.78e-01 |
Pos logit contributions | |
bia | +0.088 |
rele | +0.087 |
constitu | +0.086 |
ibaba | +0.084 |
iland | +0.083 |
Neg logit contributions | |
robe | -0.149 |
Spoiler | -0.118 |
ンジ | -0.111 |
taboola | -0.110 |
Doors | -0.110 |
Token | locker |
Token ID | 17355 |
Feature activation | -1.88e+00 |
Pos logit contributions | |
robe | +2.135 |
Spoiler | +1.629 |
taboola | +1.513 |
ンジ | +1.508 |
Crate | +1.491 |
Neg logit contributions | |
bia | -1.730 |
constitu | -1.694 |
rele | -1.677 |
ibaba | -1.641 |
iland | -1.613 |
Token | , |
Token ID | 11 |
Feature activation | +2.65e-01 |
Pos logit contributions | |
robe | +9.373 |
Spoiler | +7.113 |
Doors | +6.723 |
Crate | +6.681 |
Minotaur | +6.393 |
Neg logit contributions | |
bia | -8.720 |
ibaba | -8.572 |
constitu | -8.440 |
iland | -8.307 |
edi | -7.945 |
Token | he |
Token ID | 339 |
Feature activation | -2.60e-01 |
Pos logit contributions | |
bia | +1.070 |
rele | +1.069 |
constitu | +1.053 |
iland | +1.018 |
ibaba | +1.012 |
Neg logit contributions | |
robe | -1.624 |
Spoiler | -1.276 |
ンジ | -1.198 |
taboola | -1.182 |
Doors | -1.172 |
Token | changes |
Token ID | 2458 |
Feature activation | -4.67e-02 |
Pos logit contributions | |
robe | +1.279 |
Spoiler | +0.935 |
taboola | +0.862 |
Doors | +0.849 |
hib | +0.848 |
Neg logit contributions | |
bia | -1.360 |
constitu | -1.356 |
rele | -1.331 |
ibaba | -1.325 |
iland | -1.314 |
Token | the |
Token ID | 262 |
Feature activation | -6.33e-02 |
Pos logit contributions | |
robe | +0.290 |
Spoiler | +0.227 |
ンジ | +0.212 |
taboola | +0.210 |
Doors | +0.209 |
Neg logit contributions | |
bia | -0.188 |
rele | -0.185 |
constitu | -0.183 |
iland | -0.179 |
ibaba | -0.179 |
Token | state |
Token ID | 1181 |
Feature activation | +1.18e-01 |
Pos logit contributions | |
robe | +0.343 |
Spoiler | +0.259 |
ンジ | +0.239 |
Doors | +0.237 |
taboola | +0.236 |
Neg logit contributions | |
bia | -0.303 |
rele | -0.298 |
constitu | -0.296 |
iland | -0.291 |
ibaba | -0.290 |
Token | ization |
Token ID | 1634 |
Feature activation | +2.42e-01 |
Pos logit contributions | |
rele | +3.184 |
bia | +3.148 |
constitu | +3.022 |
iland | +2.965 |
edi | +2.930 |
Neg logit contributions | |
robe | -6.642 |
Spoiler | -5.406 |
ンジ | -5.154 |
taboola | -5.052 |
hib | -5.016 |
Token | for |
Token ID | 329 |
Feature activation | +3.33e-01 |
Pos logit contributions | |
rele | +0.999 |
bia | +0.990 |
constitu | +0.981 |
ibaba | +0.949 |
iland | +0.944 |
Neg logit contributions | |
robe | -1.464 |
Spoiler | -1.143 |
ンジ | -1.071 |
taboola | -1.057 |
Doors | -1.047 |
Token | CPU |
Token ID | 9135 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
rele | +1.390 |
bia | +1.372 |
constitu | +1.359 |
iland | +1.315 |
ibaba | +1.312 |
Neg logit contributions | |
robe | -2.001 |
Spoiler | -1.558 |
ンジ | -1.466 |
taboola | -1.446 |
Doors | -1.429 |
Token | , |
Token ID | 11 |
Feature activation | +8.09e-01 |
Pos logit contributions | |
rele | +0.834 |
bia | +0.825 |
constitu | +0.819 |
ibaba | +0.787 |
iland | +0.786 |
Neg logit contributions | |
robe | -1.579 |
Spoiler | -1.263 |
ンジ | -1.200 |
taboola | -1.184 |
Doors | -1.177 |
Token | memory |
Token ID | 4088 |
Feature activation | -7.24e-01 |
Pos logit contributions | |
rele | +2.540 |
bia | +2.449 |
constitu | +2.428 |
ibaba | +2.297 |
iland | +2.296 |
Neg logit contributions | |
robe | -5.705 |
Spoiler | -4.605 |
ンジ | -4.417 |
taboola | -4.370 |
hib | -4.322 |
Token | , |
Token ID | 11 |
Feature activation | +1.06e+00 |
Pos logit contributions | |
robe | +4.309 |
Spoiler | +3.353 |
Doors | +3.122 |
ンジ | +3.111 |
taboola | +3.057 |
Neg logit contributions | |
constitu | -2.913 |
bia | -2.903 |
rele | -2.882 |
ibaba | -2.848 |
iland | -2.821 |
Token | network |
Token ID | 3127 |
Feature activation | +1.45e-01 |
Pos logit contributions | |
rele | +3.564 |
constitu | +3.417 |
bia | +3.392 |
ibaba | +3.230 |
edi | +3.195 |
Neg logit contributions | |
robe | -7.137 |
Spoiler | -5.693 |
ンジ | -5.470 |
taboola | -5.405 |
hib | -5.352 |
Token | , |
Token ID | 11 |
Feature activation | +1.14e+00 |
Pos logit contributions | |
rele | +0.573 |
bia | +0.566 |
constitu | +0.560 |
ibaba | +0.541 |
iland | +0.541 |
Neg logit contributions | |
robe | -0.908 |
Spoiler | -0.714 |
ンジ | -0.673 |
taboola | -0.664 |
Doors | -0.657 |
Token | and |
Token ID | 290 |
Feature activation | +8.29e-01 |
Pos logit contributions | |
rele | +3.719 |
constitu | +3.539 |
bia | +3.504 |
ibaba | +3.358 |
iland | +3.297 |
Neg logit contributions | |
robe | -7.852 |
Spoiler | -6.303 |
ンジ | -6.056 |
taboola | -5.971 |
hib | -5.933 |
Token | storage |
Token ID | 6143 |
Feature activation | -7.04e-01 |
Pos logit contributions | |
rele | +2.877 |
constitu | +2.758 |
bia | +2.754 |
ibaba | +2.621 |
iland | +2.604 |
Neg logit contributions | |
robe | -5.576 |
Spoiler | -4.467 |
ンジ | -4.283 |
taboola | -4.209 |
hib | -4.173 |
Token | . |
Token ID | 13 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
robe | +3.923 |
Spoiler | +2.995 |
Doors | +2.832 |
ンジ | +2.748 |
taboola | +2.720 |
Neg logit contributions | |
constitu | -3.079 |
bia | -3.077 |
rele | -3.039 |
iland | -3.029 |
ibaba | -3.023 |
Token | after |
Token ID | 706 |
Feature activation | +3.60e-01 |
Pos logit contributions | |
rele | +1.283 |
bia | +1.279 |
constitu | +1.272 |
ibaba | +1.225 |
iland | +1.222 |
Neg logit contributions | |
robe | -2.111 |
Spoiler | -1.681 |
ンジ | -1.570 |
taboola | -1.544 |
Doors | -1.539 |
Token | leaving |
Token ID | 4305 |
Feature activation | -9.75e-02 |
Pos logit contributions | |
rele | +1.345 |
bia | +1.342 |
constitu | +1.340 |
iland | +1.286 |
ibaba | +1.285 |
Neg logit contributions | |
robe | -2.317 |
Spoiler | -1.839 |
ンジ | -1.726 |
taboola | -1.701 |
Doors | -1.698 |
Token | Kentucky |
Token ID | 11867 |
Feature activation | +1.10e-01 |
Pos logit contributions | |
robe | +0.604 |
Spoiler | +0.480 |
ンジ | +0.446 |
Doors | +0.442 |
taboola | +0.437 |
Neg logit contributions | |
constitu | -0.383 |
bia | -0.382 |
rele | -0.382 |
ibaba | -0.369 |
iland | -0.367 |
Token | and |
Token ID | 290 |
Feature activation | +1.86e-01 |
Pos logit contributions | |
rele | +0.462 |
constitu | +0.459 |
bia | +0.458 |
ibaba | +0.442 |
iland | +0.441 |
Neg logit contributions | |
robe | -0.661 |
Spoiler | -0.517 |
ンジ | -0.483 |
taboola | -0.473 |
hib | -0.471 |
Token | transferring |
Token ID | 26140 |
Feature activation | +3.03e-01 |
Pos logit contributions | |
bia | +0.684 |
constitu | +0.678 |
rele | +0.671 |
ibaba | +0.660 |
iland | +0.649 |
Neg logit contributions | |
robe | -1.209 |
Spoiler | -0.961 |
ンジ | -0.900 |
taboola | -0.894 |
Doors | -0.891 |
Token | to |
Token ID | 284 |
Feature activation | +1.25e+00 |
Pos logit contributions | |
rele | +1.203 |
bia | +1.184 |
constitu | +1.174 |
iland | +1.129 |
ibaba | +1.128 |
Neg logit contributions | |
robe | -1.887 |
Spoiler | -1.482 |
ンジ | -1.402 |
taboola | -1.382 |
hib | -1.366 |
Token | Louisville |
Token ID | 21525 |
Feature activation | +2.05e-01 |
Pos logit contributions | |
rele | +4.956 |
bia | +4.711 |
constitu | +4.636 |
iland | +4.490 |
ibaba | +4.464 |
Neg logit contributions | |
robe | -7.627 |
Spoiler | -5.896 |
ンジ | -5.710 |
taboola | -5.650 |
hib | -5.527 |
Token | as |
Token ID | 355 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
rele | +0.874 |
bia | +0.862 |
constitu | +0.859 |
ibaba | +0.827 |
iland | +0.827 |
Neg logit contributions | |
robe | -1.212 |
Spoiler | -0.942 |
ンジ | -0.885 |
taboola | -0.870 |
hib | -0.860 |
Token | a |
Token ID | 257 |
Feature activation | -2.44e-01 |
Pos logit contributions | |
rele | +0.048 |
bia | +0.048 |
constitu | +0.048 |
iland | +0.046 |
ibaba | +0.046 |
Neg logit contributions | |
robe | -0.071 |
Spoiler | -0.056 |
ンジ | -0.052 |
Doors | -0.051 |
hib | -0.051 |
Token | soccer |
Token ID | 11783 |
Feature activation | -2.49e-01 |
Pos logit contributions | |
robe | +1.684 |
Spoiler | +1.377 |
ンジ | +1.300 |
Doors | +1.286 |
hib | +1.266 |
Neg logit contributions | |
bia | -0.791 |
rele | -0.775 |
constitu | -0.770 |
ibaba | -0.742 |
iland | -0.741 |
Token | player |
Token ID | 2137 |
Feature activation | +3.72e-01 |
Pos logit contributions | |
robe | +1.606 |
Spoiler | +1.270 |
ンジ | +1.211 |
hib | +1.176 |
roo | +1.160 |
Neg logit contributions | |
bia | -0.915 |
constitu | -0.911 |
rele | -0.896 |
ibaba | -0.891 |
iland | -0.871 |
Token | time |
Token ID | 640 |
Feature activation | -5.85e-01 |
Pos logit contributions | |
robe | +4.734 |
Spoiler | +3.820 |
Doors | +3.392 |
ンジ | +3.382 |
Crate | +3.363 |
Neg logit contributions | |
bia | -3.593 |
rele | -3.440 |
ibaba | -3.430 |
iland | -3.406 |
ilial | -3.386 |
Token | . |
Token ID | 13 |
Feature activation | -2.29e-01 |
Pos logit contributions | |
robe | +3.285 |
Spoiler | +2.583 |
ンジ | +2.352 |
taboola | +2.325 |
hib | +2.315 |
Neg logit contributions | |
bia | -2.580 |
rele | -2.571 |
constitu | -2.569 |
ibaba | -2.513 |
iland | -2.482 |
Token | Maybe |
Token ID | 6674 |
Feature activation | +8.37e-02 |
Pos logit contributions | |
robe | +1.274 |
Spoiler | +1.007 |
ンジ | +0.917 |
taboola | +0.903 |
Doors | +0.903 |
Neg logit contributions | |
rele | -1.060 |
bia | -1.038 |
constitu | -1.035 |
ibaba | -1.012 |
iland | -0.996 |
Token | the |
Token ID | 262 |
Feature activation | +9.03e-02 |
Pos logit contributions | |
rele | +0.325 |
bia | +0.322 |
constitu | +0.318 |
ibaba | +0.309 |
iland | +0.307 |
Neg logit contributions | |
robe | -0.527 |
Spoiler | -0.418 |
ンジ | -0.390 |
taboola | -0.387 |
hib | -0.384 |
Token | true |
Token ID | 2081 |
Feature activation | +2.61e-01 |
Pos logit contributions | |
rele | +0.390 |
bia | +0.387 |
constitu | +0.382 |
ibaba | +0.372 |
iland | +0.371 |
Neg logit contributions | |
robe | -0.530 |
Spoiler | -0.412 |
ンジ | -0.384 |
taboola | -0.379 |
hib | -0.375 |
Token | voice |
Token ID | 3809 |
Feature activation | +1.26e+00 |
Pos logit contributions | |
rele | +1.081 |
bia | +1.073 |
constitu | +1.058 |
iland | +1.026 |
ibaba | +1.026 |
Neg logit contributions | |
robe | -1.572 |
Spoiler | -1.229 |
ンジ | -1.153 |
taboola | -1.137 |
hib | -1.126 |
Token | of |
Token ID | 286 |
Feature activation | -2.16e-01 |
Pos logit contributions | |
rele | +3.724 |
constitu | +3.581 |
bia | +3.532 |
ibaba | +3.277 |
edi | +3.259 |
Neg logit contributions | |
robe | -9.000 |
Spoiler | -7.187 |
ンジ | -7.031 |
taboola | -6.846 |
Doors | -6.828 |
Token | fandom |
Token ID | 35387 |
Feature activation | -1.05e+00 |
Pos logit contributions | |
robe | +1.271 |
Spoiler | +1.000 |
ンジ | +0.920 |
taboola | +0.908 |
Doors | +0.908 |
Neg logit contributions | |
bia | -0.912 |
rele | -0.909 |
constitu | -0.891 |
iland | -0.879 |
ibaba | -0.879 |
Token | has |
Token ID | 468 |
Feature activation | -4.08e-01 |
Pos logit contributions | |
robe | +5.898 |
Spoiler | +4.832 |
hib | +4.315 |
taboola | +4.287 |
ヤ | +4.280 |
Neg logit contributions | |
bia | -4.308 |
ibaba | -4.263 |
iland | -4.177 |
constitu | -4.143 |
rele | -4.107 |
Token | spoken |
Token ID | 9635 |
Feature activation | +6.25e-01 |
Pos logit contributions | |
robe | +2.741 |
Spoiler | +2.220 |
ンジ | +2.087 |
taboola | +2.072 |
hib | +2.055 |
Neg logit contributions | |
bia | -1.365 |
constitu | -1.344 |
ibaba | -1.336 |
rele | -1.307 |
iland | -1.273 |
Token | and |
Token ID | 290 |
Feature activation | -9.68e-02 |
Pos logit contributions | |
rele | +2.581 |
bia | +2.491 |
constitu | +2.482 |
iland | +2.380 |
ibaba | +2.346 |
Neg logit contributions | |
robe | -3.809 |
Spoiler | -2.937 |
ンジ | -2.831 |
taboola | -2.782 |
hib | -2.735 |
Token | . |
Token ID | 13 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
rele | +1.372 |
bia | +1.354 |
constitu | +1.304 |
iland | +1.290 |
edi | +1.273 |
Neg logit contributions | |
robe | -2.160 |
Spoiler | -1.680 |
ンジ | -1.623 |
taboola | -1.571 |
ヤ | -1.563 |
Token | R |
Token ID | 371 |
Feature activation | -1.37e-01 |
Pos logit contributions | |
robe | +0.656 |
Spoiler | +0.506 |
ンジ | +0.460 |
taboola | +0.454 |
hib | +0.451 |
Neg logit contributions | |
rele | -0.578 |
bia | -0.573 |
constitu | -0.571 |
iland | -0.559 |
ibaba | -0.556 |
Token | f |
Token ID | 69 |
Feature activation | +7.13e-01 |
Pos logit contributions | |
robe | +0.832 |
Spoiler | +0.643 |
hib | +0.602 |
taboola | +0.601 |
ンジ | +0.601 |
Neg logit contributions | |
rele | -0.571 |
constitu | -0.562 |
bia | -0.547 |
ibaba | -0.535 |
iland | -0.534 |
Token | 2 |
Token ID | 17 |
Feature activation | +4.59e-01 |
Pos logit contributions | |
rele | +2.873 |
bia | +2.868 |
constitu | +2.785 |
iland | +2.718 |
ibaba | +2.684 |
Neg logit contributions | |
robe | -4.346 |
Spoiler | -3.402 |
ンジ | -3.217 |
taboola | -3.162 |
ヤ | -3.119 |
Token | { |
Token ID | 1391 |
Feature activation | +5.30e-01 |
Pos logit contributions | |
rele | +1.876 |
bia | +1.864 |
constitu | +1.825 |
ibaba | +1.751 |
iland | +1.748 |
Neg logit contributions | |
robe | -2.801 |
Spoiler | -2.162 |
ンジ | -2.060 |
taboola | -2.040 |
ヤ | -2.020 |
Token | ( |
Token ID | 7 |
Feature activation | +9.94e-01 |
Pos logit contributions | |
rele | +1.909 |
bia | +1.899 |
constitu | +1.797 |
iland | +1.793 |
edi | +1.775 |
Neg logit contributions | |
robe | -3.443 |
Spoiler | -2.725 |
ンジ | -2.609 |
taboola | -2.546 |
Doors | -2.533 |
Token | 418 |
Token ID | 39667 |
Feature activation | -3.61e-01 |
Pos logit contributions | |
rele | +4.272 |
bia | +4.248 |
constitu | +4.128 |
iland | +4.043 |
edi | +4.012 |
Neg logit contributions | |
robe | -5.699 |
Spoiler | -4.338 |
ンジ | -4.101 |
taboola | -4.083 |
hib | -3.969 |
Token | s |
Token ID | 82 |
Feature activation | +5.77e-02 |
Pos logit contributions | |
robe | +2.630 |
Spoiler | +2.164 |
taboola | +2.045 |
hib | +2.029 |
ンジ | +2.028 |
Neg logit contributions | |
constitu | -1.057 |
rele | -1.046 |
bia | -0.991 |
ibaba | -0.972 |
iland | -0.968 |
Token | )} |
Token ID | 38165 |
Feature activation | +4.87e-01 |
Pos logit contributions | |
rele | +0.320 |
bia | +0.316 |
constitu | +0.315 |
iland | +0.306 |
ibaba | +0.306 |
Neg logit contributions | |
robe | -0.270 |
Spoiler | -0.193 |
ンジ | -0.177 |
taboola | -0.173 |
hib | -0.170 |
Token | R |
Token ID | 371 |
Feature activation | -4.33e-02 |
Pos logit contributions | |
rele | +2.190 |
bia | +2.157 |
constitu | +2.139 |
ibaba | +2.061 |
iland | +2.056 |
Neg logit contributions | |
robe | -2.779 |
Spoiler | -2.108 |
ンジ | -2.004 |
taboola | -1.970 |
ヤ | -1.945 |
Token | xc |
Token ID | 25306 |
Feature activation | +6.87e-01 |
Pos logit contributions | |
robe | +0.271 |
Spoiler | +0.212 |
ンジ | +0.199 |
taboola | +0.199 |
hib | +0.198 |
Neg logit contributions | |
rele | -0.173 |
constitu | -0.170 |
bia | -0.166 |
iland | -0.161 |
ibaba | -0.161 |
Token | giving |
Token ID | 3501 |
Feature activation | -2.20e-01 |
Pos logit contributions | |
rele | +1.003 |
bia | +0.996 |
constitu | +0.984 |
ibaba | +0.950 |
iland | +0.943 |
Neg logit contributions | |
robe | -1.999 |
Spoiler | -1.609 |
ンジ | -1.519 |
taboola | -1.504 |
Doors | -1.494 |
Token | up |
Token ID | 510 |
Feature activation | +7.59e-01 |
Pos logit contributions | |
robe | +1.286 |
Spoiler | +1.008 |
Doors | +0.922 |
ンジ | +0.916 |
taboola | +0.904 |
Neg logit contributions | |
bia | -0.951 |
constitu | -0.936 |
rele | -0.930 |
ibaba | -0.913 |
iland | -0.907 |
Token | , |
Token ID | 11 |
Feature activation | -4.53e-02 |
Pos logit contributions | |
rele | +2.977 |
bia | +2.881 |
constitu | +2.866 |
iland | +2.747 |
ibaba | +2.729 |
Neg logit contributions | |
robe | -4.779 |
Spoiler | -3.738 |
ンジ | -3.591 |
taboola | -3.545 |
ヤ | -3.472 |
Token | what |
Token ID | 644 |
Feature activation | +5.78e-01 |
Pos logit contributions | |
robe | +0.280 |
Spoiler | +0.224 |
Doors | +0.206 |
ンジ | +0.205 |
taboola | +0.204 |
Neg logit contributions | |
bia | -0.178 |
constitu | -0.173 |
rele | -0.173 |
ibaba | -0.170 |
iland | -0.169 |
Token | she |
Token ID | 673 |
Feature activation | +9.99e-01 |
Pos logit contributions | |
rele | +2.032 |
bia | +1.974 |
constitu | +1.967 |
iland | +1.874 |
ibaba | +1.864 |
Neg logit contributions | |
robe | -3.884 |
Spoiler | -3.098 |
ンジ | -2.965 |
taboola | -2.918 |
hib | -2.884 |
Token | said |
Token ID | 531 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
rele | +3.347 |
bia | +3.208 |
constitu | +3.151 |
iland | +3.042 |
ibaba | +2.983 |
Neg logit contributions | |
robe | -6.866 |
Spoiler | -5.528 |
ンジ | -5.295 |
taboola | -5.204 |
hib | -5.166 |
Token | she |
Token ID | 673 |
Feature activation | +6.04e-01 |
Pos logit contributions | |
rele | +4.630 |
constitu | +4.406 |
bia | +4.347 |
iland | +4.148 |
edi | +4.116 |
Neg logit contributions | |
robe | -7.696 |
Spoiler | -5.991 |
ンジ | -5.915 |
taboola | -5.757 |
ヤ | -5.653 |
Token | and |
Token ID | 290 |
Feature activation | +2.09e-01 |
Pos logit contributions | |
rele | +2.225 |
bia | +2.207 |
constitu | +2.197 |
ibaba | +2.112 |
iland | +2.099 |
Neg logit contributions | |
robe | -3.931 |
Spoiler | -3.120 |
ンジ | -2.952 |
taboola | -2.918 |
hib | -2.884 |
Token | President |
Token ID | 1992 |
Feature activation | +4.51e-01 |
Pos logit contributions | |
rele | +0.843 |
bia | +0.839 |
constitu | +0.831 |
ibaba | +0.805 |
iland | +0.803 |
Neg logit contributions | |
robe | -1.292 |
Spoiler | -1.015 |
ンジ | -0.944 |
taboola | -0.932 |
Doors | -0.928 |
Token | Barack |
Token ID | 8732 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
rele | +1.808 |
bia | +1.780 |
constitu | +1.755 |
iland | +1.694 |
ibaba | +1.692 |
Neg logit contributions | |
robe | -2.784 |
Spoiler | -2.180 |
ンジ | -2.075 |
taboola | -2.035 |
ヤ | -2.009 |
Token | Obama |
Token ID | 2486 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
rele | +3.919 |
bia | +3.785 |
constitu | +3.606 |
iland | +3.535 |
edi | +3.448 |
Neg logit contributions | |
robe | -8.241 |
Spoiler | -6.597 |
ンジ | -6.461 |
taboola | -6.349 |
ヤ | -6.266 |
Token | the |
Token ID | 262 |
Feature activation | +6.84e-01 |
Pos logit contributions | |
rele | +0.973 |
bia | +0.971 |
constitu | +0.961 |
iland | +0.927 |
ibaba | +0.913 |
Neg logit contributions | |
robe | -1.387 |
Spoiler | -1.085 |
ンジ | -1.009 |
taboola | -0.997 |
hib | -0.993 |
Token | left |
Token ID | 1364 |
Feature activation | -1.30e-01 |
Pos logit contributions | |
rele | +2.438 |
bia | +2.389 |
constitu | +2.369 |
iland | +2.266 |
ibaba | +2.264 |
Neg logit contributions | |
robe | -4.541 |
Spoiler | -3.623 |
ンジ | -3.446 |
taboola | -3.400 |
hib | -3.363 |
Token | : |
Token ID | 25 |
Feature activation | -5.30e-01 |
Pos logit contributions | |
robe | +0.765 |
Spoiler | +0.597 |
ンジ | +0.552 |
taboola | +0.541 |
hib | +0.540 |
Neg logit contributions | |
rele | -0.555 |
bia | -0.551 |
constitu | -0.548 |
iland | -0.526 |
ibaba | -0.525 |
Token | Bo |
Token ID | 3248 |
Feature activation | -1.45e-01 |
Pos logit contributions | |
robe | +2.922 |
Spoiler | +2.285 |
Doors | +2.068 |
Crate | +2.031 |
ンジ | +2.029 |
Neg logit contributions | |
bia | -2.422 |
rele | -2.391 |
constitu | -2.379 |
iland | -2.324 |
ibaba | -2.308 |
Token | Scar |
Token ID | 12089 |
Feature activation | +5.59e-02 |
Pos logit contributions | |
robe | +0.804 |
Spoiler | +0.616 |
ンジ | +0.569 |
Doors | +0.555 |
hib | +0.554 |
Neg logit contributions | |
rele | -0.665 |
constitu | -0.664 |
bia | -0.661 |
ibaba | -0.630 |
edi | -0.624 |
Token | b |
Token ID | 65 |
Feature activation | +9.56e-01 |
Pos logit contributions | |
constitu | +0.209 |
rele | +0.205 |
ibaba | +0.197 |
bia | +0.190 |
edi | +0.186 |
Neg logit contributions | |
robe | -0.367 |
Spoiler | -0.285 |
hib | -0.277 |
taboola | -0.275 |
roo | -0.273 |
Token | rough |
Token ID | 740 |
Feature activation | +2.73e-02 |
Pos logit contributions | |
bia | +3.828 |
rele | +3.764 |
constitu | +3.617 |
iland | +3.597 |
ibaba | +3.558 |
Neg logit contributions | |
robe | -5.824 |
Spoiler | -4.653 |
ンジ | -4.384 |
taboola | -4.324 |
hib | -4.252 |
Token | in |
Token ID | 287 |
Feature activation | -4.66e-01 |
Pos logit contributions | |
rele | +0.116 |
bia | +0.115 |
constitu | +0.115 |
iland | +0.111 |
ibaba | +0.110 |
Neg logit contributions | |
robe | -0.163 |
Spoiler | -0.126 |
ンジ | -0.118 |
taboola | -0.116 |
hib | -0.115 |
Token | 10 |
Token ID | 838 |
Feature activation | +3.72e-01 |
Pos logit contributions | |
robe | +2.841 |
Spoiler | +2.233 |
Doors | +2.064 |
hib | +2.043 |
ンジ | +2.041 |
Neg logit contributions | |
bia | -1.894 |
constitu | -1.885 |
rele | -1.845 |
ibaba | -1.811 |
ilial | -1.808 |
Token | th |
Token ID | 400 |
Feature activation | +1.14e+00 |
Pos logit contributions | |
rele | +1.479 |
bia | +1.462 |
constitu | +1.452 |
ibaba | +1.397 |
iland | +1.395 |
Neg logit contributions | |
robe | -2.322 |
Spoiler | -1.823 |
ンジ | -1.722 |
taboola | -1.700 |
Doors | -1.681 |
Token | grade |
Token ID | 9559 |
Feature activation | +6.21e-01 |
Pos logit contributions | |
rele | +3.708 |
bia | +3.538 |
constitu | +3.518 |
iland | +3.392 |
ibaba | +3.318 |
Neg logit contributions | |
robe | -7.829 |
Spoiler | -6.268 |
ンジ | -6.082 |
ヤ | -5.956 |
taboola | -5.943 |
Token |
|
Token ID | 198 |
Feature activation | +2.75e-01 |
Pos logit contributions | |
robe | +0.287 |
Spoiler | +0.223 |
ンジ | +0.210 |
taboola | +0.207 |
hib | +0.205 |
Neg logit contributions | |
rele | -0.203 |
bia | -0.200 |
constitu | -0.200 |
ibaba | -0.192 |
iland | -0.191 |
Token | Our |
Token ID | 5122 |
Feature activation | +5.79e-01 |
Pos logit contributions | |
rele | +1.133 |
bia | +1.112 |
constitu | +1.107 |
ibaba | +1.066 |
iland | +1.063 |
Neg logit contributions | |
robe | -1.673 |
Spoiler | -1.304 |
ンジ | -1.232 |
taboola | -1.215 |
hib | -1.201 |
Token | daughter |
Token ID | 4957 |
Feature activation | +3.23e-01 |
Pos logit contributions | |
rele | +2.162 |
bia | +2.095 |
constitu | +2.063 |
iland | +2.008 |
ibaba | +1.982 |
Neg logit contributions | |
robe | -3.763 |
Spoiler | -2.932 |
ンジ | -2.847 |
taboola | -2.767 |
hib | -2.741 |
Token | is |
Token ID | 318 |
Feature activation | +6.34e-01 |
Pos logit contributions | |
rele | +1.321 |
bia | +1.288 |
constitu | +1.267 |
ibaba | +1.233 |
iland | +1.228 |
Neg logit contributions | |
robe | -1.975 |
Spoiler | -1.539 |
ンジ | -1.468 |
taboola | -1.438 |
hib | -1.421 |
Token | a |
Token ID | 257 |
Feature activation | +6.96e-01 |
Pos logit contributions | |
rele | +2.284 |
bia | +2.203 |
constitu | +2.178 |
ibaba | +2.109 |
iland | +2.076 |
Neg logit contributions | |
robe | -4.180 |
Spoiler | -3.317 |
ンジ | -3.188 |
taboola | -3.135 |
ヤ | -3.081 |
Token | genius |
Token ID | 15632 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
rele | +2.901 |
bia | +2.797 |
constitu | +2.792 |
ibaba | +2.674 |
iland | +2.666 |
Neg logit contributions | |
robe | -4.207 |
Spoiler | -3.245 |
ンジ | -3.118 |
taboola | -3.028 |
ヤ | -2.999 |
Token | . |
Token ID | 13 |
Feature activation | -5.37e-01 |
Pos logit contributions | |
rele | +1.217 |
bia | +1.201 |
constitu | +1.160 |
iland | +1.147 |
edi | +1.128 |
Neg logit contributions | |
robe | -1.903 |
Spoiler | -1.476 |
ンジ | -1.420 |
taboola | -1.388 |
hib | -1.375 |
Token | Ye |
Token ID | 11609 |
Feature activation | +5.51e-01 |
Pos logit contributions | |
robe | +2.806 |
Spoiler | +2.222 |
Doors | +1.987 |
ンジ | +1.978 |
Crate | +1.946 |
Neg logit contributions | |
rele | -2.574 |
bia | -2.538 |
constitu | -2.507 |
ibaba | -2.476 |
iland | -2.436 |
Token | Thu |
Token ID | 26223 |
Feature activation | -7.77e-02 |
Pos logit contributions | |
rele | +2.351 |
bia | +2.321 |
constitu | +2.294 |
ibaba | +2.224 |
iland | +2.219 |
Neg logit contributions | |
robe | -3.250 |
Spoiler | -2.520 |
ンジ | -2.377 |
taboola | -2.340 |
Doors | -2.301 |
Token | za |
Token ID | 4496 |
Feature activation | +5.61e-01 |
Pos logit contributions | |
robe | +0.443 |
Spoiler | +0.339 |
ンジ | +0.320 |
taboola | +0.314 |
hib | +0.314 |
Neg logit contributions | |
rele | -0.349 |
constitu | -0.343 |
bia | -0.342 |
iland | -0.328 |
ibaba | -0.327 |
Token | : |
Token ID | 25 |
Feature activation | -4.04e-01 |
Pos logit contributions | |
rele | +2.517 |
bia | +2.484 |
constitu | +2.452 |
ibaba | +2.388 |
iland | +2.376 |
Neg logit contributions | |
robe | -3.193 |
Spoiler | -2.442 |
ンジ | -2.302 |
taboola | -2.279 |
Doors | -2.222 |
Token | on |
Token ID | 319 |
Feature activation | -2.93e-01 |
Pos logit contributions | |
robe | +1.212 |
Spoiler | +0.944 |
ンジ | +0.881 |
taboola | +0.869 |
Doors | +0.865 |
Neg logit contributions | |
rele | -0.855 |
bia | -0.851 |
constitu | -0.846 |
ibaba | -0.817 |
iland | -0.813 |
Token | 4 |
Token ID | 604 |
Feature activation | -6.09e-01 |
Pos logit contributions | |
robe | +1.645 |
Spoiler | +1.270 |
Doors | +1.169 |
ンジ | +1.152 |
hib | +1.142 |
Neg logit contributions | |
bia | -1.311 |
rele | -1.308 |
constitu | -1.308 |
ibaba | -1.274 |
iland | -1.265 |
Token | June |
Token ID | 2795 |
Feature activation | -6.06e-01 |
Pos logit contributions | |
robe | +3.003 |
Spoiler | +2.153 |
Doors | +2.050 |
ンジ | +2.000 |
hib | +1.998 |
Neg logit contributions | |
constitu | -3.135 |
rele | -3.091 |
bia | -3.045 |
ibaba | -3.037 |
ilial | -3.032 |
Token | by |
Token ID | 416 |
Feature activation | -1.91e-01 |
Pos logit contributions | |
robe | +3.570 |
Spoiler | +2.764 |
Doors | +2.563 |
ンジ | +2.545 |
taboola | +2.543 |
Neg logit contributions | |
constitu | -2.562 |
rele | -2.556 |
bia | -2.537 |
ibaba | -2.477 |
ilial | -2.415 |
Token | Pic |
Token ID | 15085 |
Feature activation | -5.67e-01 |
Pos logit contributions | |
robe | +1.030 |
Spoiler | +0.787 |
ンジ | +0.719 |
Doors | +0.711 |
taboola | +0.709 |
Neg logit contributions | |
rele | -0.908 |
bia | -0.902 |
constitu | -0.895 |
ibaba | -0.877 |
iland | -0.869 |
Token | ador |
Token ID | 7079 |
Feature activation | -9.24e-02 |
Pos logit contributions | |
robe | +3.846 |
Spoiler | +3.071 |
hib | +2.937 |
ンジ | +2.899 |
Doors | +2.879 |
Neg logit contributions | |
constitu | -1.905 |
rele | -1.875 |
ibaba | -1.805 |
bia | -1.803 |
ilial | -1.766 |
Token | at |
Token ID | 379 |
Feature activation | -1.01e+00 |
Pos logit contributions | |
robe | +0.541 |
Spoiler | +0.419 |
ンジ | +0.391 |
taboola | +0.387 |
Doors | +0.385 |
Neg logit contributions | |
rele | -0.399 |
constitu | -0.395 |
bia | -0.395 |
ibaba | -0.380 |
iland | -0.376 |
Token | £ |
Token ID | 4248 |
Feature activation | -9.93e-01 |
Pos logit contributions | |
robe | +5.172 |
Spoiler | +3.983 |
Doors | +3.650 |
Crate | +3.562 |
ンジ | +3.433 |
Neg logit contributions | |
constitu | -4.881 |
bia | -4.800 |
rele | -4.750 |
iland | -4.723 |
ibaba | -4.708 |
Token | 16 |
Token ID | 1433 |
Feature activation | -6.78e-01 |
Pos logit contributions | |
robe | +4.761 |
Spoiler | +3.533 |
Doors | +3.227 |
taboola | +3.195 |
hib | +3.153 |
Neg logit contributions | |
constitu | -5.426 |
rele | -5.293 |
ibaba | -5.224 |
bia | -5.111 |
ilial | -4.979 |
Token | . |
Token ID | 13 |
Feature activation | -5.04e-01 |
Pos logit contributions | |
robe | +3.778 |
Spoiler | +2.923 |
Doors | +2.661 |
ンジ | +2.645 |
taboola | +2.629 |
Neg logit contributions | |
rele | -3.075 |
constitu | -3.069 |
bia | -3.040 |
ibaba | -2.963 |
ilial | -2.930 |
Token | 99 |
Token ID | 2079 |
Feature activation | -5.17e-01 |
Pos logit contributions | |
robe | +3.005 |
Spoiler | +2.385 |
Doors | +2.208 |
ンジ | +2.193 |
taboola | +2.178 |
Neg logit contributions | |
rele | -2.131 |
constitu | -2.095 |
bia | -2.084 |
ibaba | -2.040 |
iland | -1.994 |
Token | spokeswoman |
Token ID | 12262 |
Feature activation | -7.44e-01 |
Pos logit contributions | |
robe | +2.690 |
Spoiler | +2.068 |
ンジ | +1.872 |
Doors | +1.862 |
hib | +1.838 |
Neg logit contributions | |
bia | -2.409 |
rele | -2.405 |
constitu | -2.361 |
ibaba | -2.333 |
iland | -2.330 |
Token | Barbara |
Token ID | 16685 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
robe | +3.513 |
Spoiler | +2.638 |
Doors | +2.353 |
Crate | +2.294 |
ンジ | +2.290 |
Neg logit contributions | |
rele | -3.922 |
bia | -3.903 |
constitu | -3.891 |
ibaba | -3.844 |
iland | -3.792 |
Token | Reynolds |
Token ID | 21995 |
Feature activation | +4.27e-01 |
Pos logit contributions | |
robe | +0.094 |
Spoiler | +0.072 |
ンジ | +0.067 |
taboola | +0.065 |
Doors | +0.065 |
Neg logit contributions | |
rele | -0.081 |
bia | -0.080 |
constitu | -0.079 |
ibaba | -0.077 |
iland | -0.077 |
Token | . |
Token ID | 13 |
Feature activation | +4.79e-01 |
Pos logit contributions | |
rele | +1.775 |
bia | +1.745 |
constitu | +1.722 |
iland | +1.674 |
ibaba | +1.653 |
Neg logit contributions | |
robe | -2.577 |
Spoiler | -1.998 |
ンジ | -1.902 |
taboola | -1.868 |
ヤ | -1.840 |
Token | The |
Token ID | 383 |
Feature activation | +6.65e-01 |
Pos logit contributions | |
rele | +2.177 |
bia | +2.156 |
constitu | +2.128 |
iland | +2.072 |
ibaba | +2.053 |
Neg logit contributions | |
robe | -2.696 |
Spoiler | -2.031 |
ンジ | -1.922 |
taboola | -1.888 |
ヤ | -1.863 |
Token | Environmental |
Token ID | 13272 |
Feature activation | +3.18e-01 |
Pos logit contributions | |
rele | +2.704 |
bia | +2.642 |
constitu | +2.623 |
iland | +2.513 |
ibaba | +2.505 |
Neg logit contributions | |
robe | -4.074 |
Spoiler | -3.177 |
ンジ | -3.018 |
taboola | -2.976 |
ヤ | -2.936 |
Token | Protection |
Token ID | 9985 |
Feature activation | +1.95e-01 |
Pos logit contributions | |
rele | +1.472 |
bia | +1.453 |
constitu | +1.442 |
iland | +1.395 |
ibaba | +1.395 |
Neg logit contributions | |
robe | -1.778 |
Spoiler | -1.353 |
ンジ | -1.266 |
taboola | -1.246 |
hib | -1.230 |
Token | Agency |
Token ID | 7732 |
Feature activation | +3.62e-01 |
Pos logit contributions | |
rele | +1.093 |
bia | +1.081 |
constitu | +1.074 |
ibaba | +1.046 |
iland | +1.046 |
Neg logit contributions | |
robe | -0.892 |
Spoiler | -0.633 |
ンジ | -0.579 |
taboola | -0.566 |
hib | -0.557 |
Token | has |
Token ID | 468 |
Feature activation | -9.99e-02 |
Pos logit contributions | |
rele | +1.440 |
bia | +1.416 |
constitu | +1.404 |
iland | +1.351 |
ibaba | +1.349 |
Neg logit contributions | |
robe | -2.256 |
Spoiler | -1.771 |
ンジ | -1.676 |
taboola | -1.652 |
hib | -1.632 |
Token | retained |
Token ID | 17383 |
Feature activation | +4.55e-01 |
Pos logit contributions | |
robe | +0.654 |
Spoiler | +0.517 |
ンジ | +0.488 |
hib | +0.486 |
taboola | +0.485 |
Neg logit contributions | |
bia | -0.358 |
rele | -0.352 |
constitu | -0.351 |
ibaba | -0.345 |
iland | -0.339 |
Token | some |
Token ID | 617 |
Feature activation | +6.21e-01 |
Pos logit contributions | |
rele | +1.829 |
bia | +1.792 |
constitu | +1.780 |
iland | +1.712 |
ibaba | +1.706 |
Neg logit contributions | |
robe | -2.813 |
Spoiler | -2.204 |
ンジ | -2.089 |
taboola | -2.067 |
ヤ | -2.032 |
Token | and |
Token ID | 290 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
bia | +0.154 |
constitu | +0.151 |
rele | +0.151 |
ibaba | +0.144 |
iland | +0.143 |
Neg logit contributions | |
robe | -0.272 |
Spoiler | -0.217 |
ンジ | -0.204 |
taboola | -0.202 |
hib | -0.201 |
Token | right |
Token ID | 826 |
Feature activation | -3.36e-01 |
Pos logit contributions | |
rele | +0.779 |
bia | +0.779 |
constitu | +0.769 |
ibaba | +0.738 |
iland | +0.737 |
Neg logit contributions | |
robe | -1.398 |
Spoiler | -1.110 |
ンジ | -1.047 |
taboola | -1.035 |
Doors | -1.027 |
Token | now |
Token ID | 783 |
Feature activation | +4.90e-01 |
Pos logit contributions | |
robe | +2.282 |
Spoiler | +1.810 |
ンジ | +1.723 |
Doors | +1.678 |
ヤ | +1.673 |
Neg logit contributions | |
bia | -1.152 |
constitu | -1.124 |
rele | -1.106 |
ibaba | -1.071 |
ilial | -1.049 |
Token | whatever |
Token ID | 4232 |
Feature activation | +2.76e-01 |
Pos logit contributions | |
rele | +1.878 |
bia | +1.819 |
constitu | +1.811 |
ibaba | +1.737 |
iland | +1.734 |
Neg logit contributions | |
robe | -3.134 |
Spoiler | -2.472 |
ンジ | -2.358 |
taboola | -2.333 |
ヤ | -2.294 |
Token | decisions |
Token ID | 5370 |
Feature activation | +3.97e-01 |
Pos logit contributions | |
rele | +1.009 |
bia | +1.002 |
constitu | +0.988 |
iland | +0.949 |
ibaba | +0.948 |
Neg logit contributions | |
robe | -1.798 |
Spoiler | -1.433 |
ンジ | -1.355 |
taboola | -1.336 |
Doors | -1.323 |
Token | have |
Token ID | 423 |
Feature activation | +4.64e-01 |
Pos logit contributions | |
rele | +1.471 |
bia | +1.440 |
constitu | +1.430 |
iland | +1.372 |
ibaba | +1.366 |
Neg logit contributions | |
robe | -2.578 |
Spoiler | -2.045 |
ンジ | -1.948 |
taboola | -1.920 |
hib | -1.896 |
Token | to |
Token ID | 284 |
Feature activation | +8.71e-01 |
Pos logit contributions | |
rele | +1.645 |
bia | +1.588 |
constitu | +1.585 |
iland | +1.514 |
ibaba | +1.506 |
Neg logit contributions | |
robe | -3.102 |
Spoiler | -2.477 |
ンジ | -2.366 |
taboola | -2.339 |
ヤ | -2.303 |
Token | be |
Token ID | 307 |
Feature activation | +8.52e-01 |
Pos logit contributions | |
rele | +3.141 |
bia | +2.948 |
constitu | +2.904 |
iland | +2.831 |
edi | +2.794 |
Neg logit contributions | |
robe | -5.738 |
Spoiler | -4.618 |
ンジ | -4.427 |
taboola | -4.387 |
ヤ | -4.305 |
Token | made |
Token ID | 925 |
Feature activation | +3.89e-01 |
Pos logit contributions | |
rele | +3.177 |
constitu | +3.021 |
bia | +2.989 |
iland | +2.852 |
edi | +2.840 |
Neg logit contributions | |
robe | -5.543 |
Spoiler | -4.425 |
ンジ | -4.243 |
taboola | -4.183 |
ヤ | -4.126 |
Token | going |
Token ID | 1016 |
Feature activation | -2.14e-02 |
Pos logit contributions | |
rele | +1.539 |
bia | +1.500 |
constitu | +1.492 |
iland | +1.436 |
ibaba | +1.427 |
Neg logit contributions | |
robe | -2.437 |
Spoiler | -1.915 |
ンジ | -1.824 |
taboola | -1.793 |
ヤ | -1.770 |
Token | forward |
Token ID | 2651 |
Feature activation | +2.22e-01 |
Pos logit contributions | |
robe | +0.108 |
Spoiler | +0.080 |
ンジ | +0.073 |
hib | +0.071 |
Doors | +0.071 |
Neg logit contributions | |
bia | -0.110 |
constitu | -0.109 |
rele | -0.109 |
ibaba | -0.106 |
iland | -0.105 |
Token | making |
Token ID | 1642 |
Feature activation | -2.08e-01 |
Pos logit contributions | |
robe | +0.568 |
Spoiler | +0.452 |
ンジ | +0.420 |
hib | +0.419 |
Doors | +0.419 |
Neg logit contributions | |
bia | -0.328 |
constitu | -0.320 |
ibaba | -0.314 |
rele | -0.311 |
iland | -0.307 |
Token | the |
Token ID | 262 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
robe | +1.337 |
Spoiler | +1.084 |
ンジ | +1.011 |
hib | +1.004 |
Doors | +1.004 |
Neg logit contributions | |
bia | -0.768 |
constitu | -0.745 |
rele | -0.737 |
ibaba | -0.730 |
iland | -0.726 |
Token | whole |
Token ID | 2187 |
Feature activation | -9.59e-01 |
Pos logit contributions | |
robe | +0.620 |
Spoiler | +0.494 |
ンジ | +0.458 |
Doors | +0.455 |
hib | +0.454 |
Neg logit contributions | |
bia | -0.395 |
constitu | -0.389 |
rele | -0.385 |
ibaba | -0.382 |
iland | -0.381 |
Token | project |
Token ID | 1628 |
Feature activation | -3.54e-01 |
Pos logit contributions | |
robe | +5.740 |
Spoiler | +4.601 |
hib | +4.215 |
ンジ | +4.203 |
Crate | +4.198 |
Neg logit contributions | |
bia | -3.698 |
constitu | -3.580 |
iland | -3.559 |
ibaba | -3.520 |
rele | -3.443 |
Token | work |
Token ID | 670 |
Feature activation | +1.78e-01 |
Pos logit contributions | |
robe | +2.271 |
Spoiler | +1.842 |
ンジ | +1.680 |
taboola | +1.676 |
Doors | +1.674 |
Neg logit contributions | |
bia | -1.299 |
constitu | -1.267 |
ibaba | -1.238 |
rele | -1.214 |
iland | -1.214 |
Token | - |
Token ID | 532 |
Feature activation | -4.34e-01 |
Pos logit contributions | |
rele | +0.690 |
bia | +0.682 |
constitu | +0.678 |
ibaba | +0.651 |
iland | +0.649 |
Neg logit contributions | |
robe | -1.120 |
Spoiler | -0.884 |
ンジ | -0.834 |
taboola | -0.821 |
hib | -0.815 |
Token | and |
Token ID | 290 |
Feature activation | -1.27e-01 |
Pos logit contributions | |
robe | +2.605 |
Spoiler | +2.117 |
hib | +1.938 |
Doors | +1.936 |
ンジ | +1.935 |
Neg logit contributions | |
bia | -1.738 |
constitu | -1.659 |
iland | -1.645 |
rele | -1.637 |
ibaba | -1.626 |
Token | also |
Token ID | 635 |
Feature activation | -2.23e-01 |
Pos logit contributions | |
robe | +0.816 |
Spoiler | +0.654 |
Doors | +0.604 |
ンジ | +0.604 |
hib | +0.603 |
Neg logit contributions | |
bia | -0.475 |
constitu | -0.461 |
rele | -0.453 |
ibaba | -0.451 |
iland | -0.447 |
Token | the |
Token ID | 262 |
Feature activation | -1.73e-01 |
Pos logit contributions | |
robe | +1.425 |
Spoiler | +1.143 |
ンジ | +1.062 |
Doors | +1.053 |
hib | +1.052 |
Neg logit contributions | |
bia | -0.845 |
constitu | -0.812 |
ibaba | -0.803 |
iland | -0.795 |
rele | -0.792 |
Token | management |
Token ID | 4542 |
Feature activation | -4.99e-01 |
Pos logit contributions | |
robe | +1.070 |
Spoiler | +0.850 |
ンジ | +0.786 |
Doors | +0.783 |
hib | +0.782 |
Neg logit contributions | |
bia | -0.690 |
constitu | -0.674 |
rele | -0.669 |
ibaba | -0.664 |
iland | -0.660 |
Token | side |
Token ID | 1735 |
Feature activation | -6.08e-01 |
Pos logit contributions | |
robe | +3.162 |
Spoiler | +2.512 |
hib | +2.319 |
ンジ | +2.315 |
Doors | +2.300 |
Neg logit contributions | |
bia | -1.873 |
ibaba | -1.820 |
constitu | -1.810 |
rele | -1.784 |
iland | -1.756 |
Token | old |
Token ID | 727 |
Feature activation | -7.66e-02 |
Pos logit contributions | |
rele | +0.064 |
bia | +0.062 |
constitu | +0.062 |
ibaba | +0.058 |
iland | +0.058 |
Neg logit contributions | |
robe | -0.169 |
Spoiler | -0.138 |
ンジ | -0.132 |
taboola | -0.131 |
hib | -0.130 |
Token | Hed |
Token ID | 30920 |
Feature activation | -5.20e-01 |
Pos logit contributions | |
robe | +0.439 |
Spoiler | +0.340 |
ンジ | +0.316 |
taboola | +0.310 |
Doors | +0.308 |
Neg logit contributions | |
rele | -0.341 |
bia | -0.338 |
constitu | -0.337 |
ibaba | -0.326 |
iland | -0.324 |
Token | ges |
Token ID | 3212 |
Feature activation | -1.79e-01 |
Pos logit contributions | |
robe | +2.571 |
Spoiler | +1.879 |
hib | +1.751 |
Doors | +1.717 |
ンジ | +1.707 |
Neg logit contributions | |
rele | -2.693 |
constitu | -2.666 |
bia | -2.639 |
ibaba | -2.541 |
ilial | -2.505 |
Token | is |
Token ID | 318 |
Feature activation | +1.95e-01 |
Pos logit contributions | |
robe | +1.075 |
Spoiler | +0.838 |
ンジ | +0.784 |
taboola | +0.772 |
hib | +0.771 |
Neg logit contributions | |
bia | -0.746 |
rele | -0.746 |
constitu | -0.742 |
ibaba | -0.716 |
iland | -0.711 |
Token | sure |
Token ID | 1654 |
Feature activation | +1.37e-01 |
Pos logit contributions | |
rele | +0.794 |
bia | +0.787 |
constitu | +0.778 |
ibaba | +0.752 |
iland | +0.752 |
Neg logit contributions | |
robe | -1.190 |
Spoiler | -0.933 |
ンジ | -0.875 |
taboola | -0.863 |
hib | -0.859 |
Token | to |
Token ID | 284 |
Feature activation | -2.85e-01 |
Pos logit contributions | |
rele | +0.553 |
bia | +0.547 |
constitu | +0.542 |
ibaba | +0.522 |
iland | +0.520 |
Neg logit contributions | |
robe | -0.847 |
Spoiler | -0.667 |
ンジ | -0.627 |
taboola | -0.618 |
hib | -0.614 |
Token | be |
Token ID | 307 |
Feature activation | -4.64e-01 |
Pos logit contributions | |
robe | +1.812 |
Spoiler | +1.442 |
ンジ | +1.348 |
taboola | +1.347 |
Doors | +1.330 |
Neg logit contributions | |
bia | -1.062 |
constitu | -1.058 |
rele | -1.051 |
ibaba | -1.020 |
iland | -1.001 |
Token | a |
Token ID | 257 |
Feature activation | -2.88e-01 |
Pos logit contributions | |
robe | +2.683 |
Spoiler | +2.113 |
hib | +1.951 |
Doors | +1.917 |
ンジ | +1.917 |
Neg logit contributions | |
bia | -1.960 |
constitu | -1.931 |
rele | -1.930 |
ibaba | -1.893 |
iland | -1.881 |
Token | breakout |
Token ID | 31661 |
Feature activation | -7.50e-01 |
Pos logit contributions | |
robe | +1.757 |
Spoiler | +1.412 |
ンジ | +1.296 |
Doors | +1.292 |
taboola | +1.285 |
Neg logit contributions | |
bia | -1.148 |
rele | -1.132 |
constitu | -1.111 |
iland | -1.097 |
ibaba | -1.088 |
Token | star |
Token ID | 3491 |
Feature activation | -4.44e-01 |
Pos logit contributions | |
robe | +4.214 |
Spoiler | +3.314 |
ンジ | +3.044 |
hib | +3.034 |
taboola | +3.019 |
Neg logit contributions | |
bia | -3.265 |
constitu | -3.207 |
rele | -3.153 |
ibaba | -3.150 |
iland | -3.142 |
Token | ), |
Token ID | 828 |
Feature activation | +2.88e-01 |
Pos logit contributions | |
robe | +2.621 |
Spoiler | +2.063 |
ンジ | +1.905 |
hib | +1.894 |
taboola | +1.883 |
Neg logit contributions | |
constitu | -1.867 |
bia | -1.862 |
rele | -1.846 |
ibaba | -1.794 |
iland | -1.777 |
Token | are |
Token ID | 389 |
Feature activation | +7.12e-01 |
Pos logit contributions | |
rele | +1.880 |
bia | +1.844 |
constitu | +1.827 |
iland | +1.760 |
ibaba | +1.754 |
Neg logit contributions | |
robe | -3.059 |
Spoiler | -2.415 |
ンジ | -2.287 |
taboola | -2.256 |
hib | -2.230 |
Token | . |
Token ID | 13 |
Feature activation | +3.99e-01 |
Pos logit contributions | |
rele | +2.633 |
bia | +2.548 |
constitu | +2.526 |
iland | +2.438 |
ibaba | +2.407 |
Neg logit contributions | |
robe | -4.609 |
Spoiler | -3.684 |
ンジ | -3.491 |
taboola | -3.459 |
Doors | -3.412 |
Token | T |
Token ID | 309 |
Feature activation | +3.84e-01 |
Pos logit contributions | |
rele | +1.817 |
bia | +1.789 |
constitu | +1.775 |
iland | +1.718 |
ibaba | +1.713 |
Neg logit contributions | |
robe | -2.251 |
Spoiler | -1.709 |
ンジ | -1.613 |
taboola | -1.588 |
hib | -1.564 |
Token | apes |
Token ID | 7916 |
Feature activation | -8.64e-02 |
Pos logit contributions | |
rele | +1.014 |
bia | +0.988 |
constitu | +0.975 |
iland | +0.918 |
ibaba | +0.915 |
Neg logit contributions | |
robe | -2.901 |
Spoiler | -2.394 |
ンジ | -2.296 |
taboola | -2.267 |
hib | -2.242 |
Token | are |
Token ID | 389 |
Feature activation | +2.26e-01 |
Pos logit contributions | |
robe | +0.542 |
Spoiler | +0.431 |
ンジ | +0.400 |
taboola | +0.398 |
hib | +0.395 |
Neg logit contributions | |
bia | -0.335 |
rele | -0.330 |
constitu | -0.329 |
ibaba | -0.318 |
iland | -0.317 |
Token | encrypted |
Token ID | 19365 |
Feature activation | -4.76e-01 |
Pos logit contributions | |
bia | +0.817 |
rele | +0.814 |
constitu | +0.802 |
ibaba | +0.776 |
iland | +0.775 |
Neg logit contributions | |
robe | -1.480 |
Spoiler | -1.179 |
ンジ | -1.110 |
taboola | -1.098 |
hib | -1.092 |
Token | implying |
Token ID | 30253 |
Feature activation | -3.13e-02 |
Pos logit contributions | |
robe | +2.909 |
Spoiler | +2.286 |
ンジ | +2.130 |
Doors | +2.116 |
taboola | +2.103 |
Neg logit contributions | |
bia | -1.899 |
constitu | -1.869 |
rele | -1.847 |
ibaba | -1.827 |
iland | -1.823 |
Token | that |
Token ID | 326 |
Feature activation | -2.86e-01 |
Pos logit contributions | |
robe | +0.206 |
Spoiler | +0.165 |
ンジ | +0.154 |
taboola | +0.153 |
hib | +0.153 |
Neg logit contributions | |
bia | -0.112 |
constitu | -0.109 |
rele | -0.109 |
iland | -0.107 |
ibaba | -0.105 |
Token | nefarious |
Token ID | 42079 |
Feature activation | -5.50e-01 |
Pos logit contributions | |
robe | +1.889 |
Spoiler | +1.523 |
ンジ | +1.419 |
taboola | +1.416 |
Doors | +1.412 |
Neg logit contributions | |
bia | -1.003 |
rele | -0.977 |
constitu | -0.975 |
iland | -0.953 |
ibaba | -0.947 |
Token | forces |
Token ID | 3386 |
Feature activation | +3.25e-01 |
Pos logit contributions | |
robe | +3.271 |
Spoiler | +2.562 |
ンジ | +2.392 |
Doors | +2.388 |
taboola | +2.378 |
Neg logit contributions | |
bia | -2.231 |
ibaba | -2.194 |
rele | -2.190 |
iland | -2.179 |
constitu | -2.175 |
Token | would |
Token ID | 561 |
Feature activation | -3.90e-01 |
Pos logit contributions | |
rele | +1.246 |
bia | +1.227 |
constitu | +1.216 |
iland | +1.169 |
ibaba | +1.169 |
Neg logit contributions | |
robe | -2.064 |
Spoiler | -1.631 |
ンジ | -1.543 |
taboola | -1.523 |
hib | -1.505 |
Token | s |
Token ID | 82 |
Feature activation | +2.54e-01 |
Pos logit contributions | |
rele | +0.065 |
constitu | +0.064 |
bia | +0.064 |
iland | +0.060 |
ibaba | +0.060 |
Neg logit contributions | |
robe | -0.165 |
Spoiler | -0.134 |
ンジ | -0.129 |
taboola | -0.127 |
hib | -0.127 |
Token | say |
Token ID | 910 |
Feature activation | +3.40e-01 |
Pos logit contributions | |
rele | +0.938 |
bia | +0.933 |
constitu | +0.927 |
iland | +0.886 |
ibaba | +0.885 |
Neg logit contributions | |
robe | -1.650 |
Spoiler | -1.313 |
ンジ | -1.241 |
taboola | -1.224 |
hib | -1.214 |
Token | the |
Token ID | 262 |
Feature activation | +5.78e-02 |
Pos logit contributions | |
rele | +1.219 |
bia | +1.203 |
constitu | +1.191 |
iland | +1.143 |
ibaba | +1.142 |
Neg logit contributions | |
robe | -2.247 |
Spoiler | -1.796 |
ンジ | -1.700 |
taboola | -1.679 |
hib | -1.663 |
Token | rate |
Token ID | 2494 |
Feature activation | +5.67e-02 |
Pos logit contributions | |
bia | +0.212 |
rele | +0.210 |
constitu | +0.206 |
ibaba | +0.202 |
iland | +0.201 |
Neg logit contributions | |
robe | -0.375 |
Spoiler | -0.300 |
ンジ | -0.281 |
taboola | -0.279 |
Doors | -0.276 |
Token | of |
Token ID | 286 |
Feature activation | -6.54e-01 |
Pos logit contributions | |
rele | +0.200 |
bia | +0.200 |
constitu | +0.197 |
iland | +0.190 |
ibaba | +0.189 |
Neg logit contributions | |
robe | -0.376 |
Spoiler | -0.302 |
ンジ | -0.284 |
taboola | -0.282 |
hib | -0.279 |
Token | failure |
Token ID | 5287 |
Feature activation | -2.59e-01 |
Pos logit contributions | |
robe | +3.681 |
Spoiler | +2.892 |
Crate | +2.603 |
Doors | +2.601 |
taboola | +2.590 |
Neg logit contributions | |
bia | -2.821 |
iland | -2.781 |
constitu | -2.744 |
ibaba | -2.717 |
rele | -2.698 |
Token | on |
Token ID | 319 |
Feature activation | -3.77e-02 |
Pos logit contributions | |
robe | +1.503 |
Spoiler | +1.178 |
taboola | +1.088 |
ンジ | +1.081 |
hib | +1.073 |
Neg logit contributions | |
bia | -1.108 |
rele | -1.099 |
constitu | -1.094 |
ibaba | -1.063 |
iland | -1.060 |
Token | the |
Token ID | 262 |
Feature activation | +5.05e-02 |
Pos logit contributions | |
robe | +0.234 |
Spoiler | +0.186 |
ンジ | +0.172 |
taboola | +0.171 |
hib | +0.170 |
Neg logit contributions | |
bia | -0.148 |
rele | -0.147 |
constitu | -0.145 |
iland | -0.142 |
ibaba | -0.140 |
Token | first |
Token ID | 717 |
Feature activation | +4.46e-01 |
Pos logit contributions | |
bia | +0.175 |
rele | +0.172 |
constitu | +0.169 |
ibaba | +0.165 |
iland | +0.164 |
Neg logit contributions | |
robe | -0.338 |
Spoiler | -0.273 |
ンジ | -0.256 |
taboola | -0.254 |
Doors | -0.253 |
Token | attempt |
Token ID | 2230 |
Feature activation | -1.61e-01 |
Pos logit contributions | |
rele | +1.622 |
bia | +1.600 |
constitu | +1.581 |
iland | +1.519 |
ibaba | +1.517 |
Neg logit contributions | |
robe | -2.922 |
Spoiler | -2.327 |
ンジ | -2.205 |
taboola | -2.178 |
hib | -2.155 |
Token | is |
Token ID | 318 |
Feature activation | +7.07e-02 |
Pos logit contributions | |
robe | +0.942 |
Spoiler | +0.735 |
taboola | +0.679 |
ンジ | +0.679 |
hib | +0.671 |
Neg logit contributions | |
bia | -0.691 |
rele | -0.688 |
constitu | -0.684 |
ibaba | -0.662 |
iland | -0.661 |
Token | and |
Token ID | 290 |
Feature activation | -2.46e-01 |
Pos logit contributions | |
robe | +4.531 |
Spoiler | +3.452 |
hib | +3.197 |
Doors | +3.158 |
ヤ | +3.151 |
Neg logit contributions | |
constitu | -3.319 |
bia | -3.308 |
rele | -3.266 |
ibaba | -3.199 |
iland | -3.179 |
Token | cunning |
Token ID | 34218 |
Feature activation | -7.79e-01 |
Pos logit contributions | |
robe | +1.506 |
Spoiler | +1.188 |
ンジ | +1.104 |
taboola | +1.089 |
Doors | +1.086 |
Neg logit contributions | |
bia | -0.992 |
rele | -0.979 |
constitu | -0.977 |
ibaba | -0.958 |
iland | -0.946 |
Token | a |
Token ID | 257 |
Feature activation | -2.15e-01 |
Pos logit contributions | |
robe | +4.519 |
Spoiler | +3.461 |
hib | +3.213 |
ンジ | +3.178 |
Doors | +3.161 |
Neg logit contributions | |
bia | -3.323 |
constitu | -3.294 |
rele | -3.292 |
ibaba | -3.219 |
iland | -3.217 |
Token | poet |
Token ID | 21810 |
Feature activation | +1.85e-01 |
Pos logit contributions | |
robe | +1.241 |
Spoiler | +0.956 |
ンジ | +0.885 |
Doors | +0.879 |
hib | +0.867 |
Neg logit contributions | |
bia | -0.946 |
rele | -0.940 |
constitu | -0.920 |
ibaba | -0.910 |
iland | -0.909 |
Token | ever |
Token ID | 1683 |
Feature activation | -4.13e-01 |
Pos logit contributions | |
rele | +0.754 |
bia | +0.742 |
constitu | +0.736 |
iland | +0.709 |
ibaba | +0.708 |
Neg logit contributions | |
robe | -1.135 |
Spoiler | -0.887 |
ンジ | -0.838 |
taboola | -0.826 |
hib | -0.816 |
Token | to |
Token ID | 284 |
Feature activation | -2.46e-01 |
Pos logit contributions | |
robe | +2.471 |
Spoiler | +1.914 |
taboola | +1.779 |
ンジ | +1.775 |
hib | +1.756 |
Neg logit contributions | |
bia | -1.708 |
constitu | -1.700 |
rele | -1.695 |
ibaba | -1.647 |
iland | -1.618 |
Token | win |
Token ID | 1592 |
Feature activation | +3.93e-01 |
Pos logit contributions | |
robe | +1.557 |
Spoiler | +1.229 |
taboola | +1.145 |
ンジ | +1.140 |
hib | +1.132 |
Neg logit contributions | |
constitu | -0.933 |
bia | -0.926 |
rele | -0.910 |
ibaba | -0.899 |
iland | -0.875 |
Token | too |
Token ID | 1165 |
Feature activation | -2.09e-01 |
Pos logit contributions | |
rele | +1.565 |
bia | +1.530 |
constitu | +1.520 |
iland | +1.461 |
ibaba | +1.456 |
Neg logit contributions | |
robe | -2.444 |
Spoiler | -1.920 |
ンジ | -1.824 |
taboola | -1.798 |
ヤ | -1.773 |
Token | many |
Token ID | 867 |
Feature activation | -5.58e-01 |
Pos logit contributions | |
robe | +1.320 |
Spoiler | +1.045 |
ンジ | +0.978 |
Doors | +0.965 |
hib | +0.964 |
Neg logit contributions | |
bia | -0.799 |
rele | -0.793 |
constitu | -0.778 |
ibaba | -0.770 |
iland | -0.766 |
Token | friends |
Token ID | 2460 |
Feature activation | -7.88e-01 |
Pos logit contributions | |
robe | +3.286 |
Spoiler | +2.582 |
Doors | +2.421 |
ンジ | +2.333 |
Crate | +2.326 |
Neg logit contributions | |
bia | -2.322 |
ibaba | -2.288 |
iland | -2.253 |
constitu | -2.212 |
rele | -2.212 |
Token | — |
Token ID | 851 |
Feature activation | -2.17e-01 |
Pos logit contributions | |
robe | +4.615 |
Spoiler | +3.567 |
hib | +3.263 |
ンジ | +3.206 |
taboola | +3.202 |
Neg logit contributions | |
bia | -3.316 |
constitu | -3.294 |
ibaba | -3.244 |
rele | -3.219 |
iland | -3.194 |
Token | regarding |
Token ID | 5115 |
Feature activation | +1.94e-01 |
Pos logit contributions | |
robe | +0.885 |
Spoiler | +0.698 |
taboola | +0.647 |
ンジ | +0.645 |
hib | +0.643 |
Neg logit contributions | |
rele | -0.548 |
bia | -0.544 |
constitu | -0.541 |
ibaba | -0.525 |
iland | -0.522 |
Token | your |
Token ID | 534 |
Feature activation | +1.52e-01 |
Pos logit contributions | |
rele | +0.805 |
bia | +0.793 |
constitu | +0.786 |
iland | +0.758 |
ibaba | +0.757 |
Neg logit contributions | |
robe | -1.177 |
Spoiler | -0.917 |
ンジ | -0.866 |
taboola | -0.853 |
hib | -0.843 |
Token | vote |
Token ID | 3015 |
Feature activation | +5.56e-01 |
Pos logit contributions | |
rele | +0.590 |
bia | +0.590 |
constitu | +0.574 |
ibaba | +0.566 |
iland | +0.560 |
Neg logit contributions | |
robe | -0.959 |
Spoiler | -0.762 |
ンジ | -0.712 |
taboola | -0.706 |
hib | -0.698 |
Token | , |
Token ID | 11 |
Feature activation | -2.51e-02 |
Pos logit contributions | |
rele | +2.160 |
bia | +2.105 |
constitu | +2.097 |
ibaba | +2.003 |
iland | +2.001 |
Neg logit contributions | |
robe | -3.502 |
Spoiler | -2.758 |
ンジ | -2.643 |
taboola | -2.588 |
ヤ | -2.567 |
Token | let |
Token ID | 1309 |
Feature activation | +6.55e-02 |
Pos logit contributions | |
robe | +0.167 |
Spoiler | +0.135 |
ンジ | +0.126 |
taboola | +0.126 |
hib | +0.124 |
Neg logit contributions | |
rele | -0.088 |
bia | -0.087 |
constitu | -0.086 |
ibaba | -0.083 |
iland | -0.083 |
Token | me |
Token ID | 502 |
Feature activation | -1.34e+00 |
Pos logit contributions | |
rele | +0.248 |
bia | +0.244 |
constitu | +0.242 |
ibaba | +0.232 |
iland | +0.232 |
Neg logit contributions | |
robe | -0.420 |
Spoiler | -0.334 |
ンジ | -0.316 |
taboola | -0.311 |
hib | -0.308 |
Token | know |
Token ID | 760 |
Feature activation | -9.41e-01 |
Pos logit contributions | |
robe | +7.719 |
Spoiler | +6.156 |
Doors | +5.668 |
Crate | +5.639 |
hib | +5.628 |
Neg logit contributions | |
constitu | -5.247 |
ibaba | -5.195 |
bia | -5.150 |
ilial | -5.038 |
rele | -4.810 |
Token | ! |
Token ID | 0 |
Feature activation | -3.69e-01 |
Pos logit contributions | |
robe | +5.365 |
Spoiler | +4.266 |
Doors | +3.832 |
hib | +3.806 |
ンジ | +3.789 |
Neg logit contributions | |
constitu | -3.959 |
bia | -3.889 |
rele | -3.883 |
ibaba | -3.857 |
iland | -3.744 |
Token | Thanks |
Token ID | 9690 |
Feature activation | -3.15e-01 |
Pos logit contributions | |
robe | +2.000 |
Spoiler | +1.569 |
ンジ | +1.419 |
ヤ | +1.400 |
Doors | +1.391 |
Neg logit contributions | |
rele | -1.748 |
bia | -1.719 |
constitu | -1.718 |
ibaba | -1.680 |
iland | -1.644 |
Token | for |
Token ID | 329 |
Feature activation | -5.47e-01 |
Pos logit contributions | |
robe | +1.893 |
Spoiler | +1.490 |
ンジ | +1.370 |
roo | +1.357 |
ヤ | +1.356 |
Neg logit contributions | |
rele | -1.317 |
bia | -1.305 |
constitu | -1.275 |
ibaba | -1.240 |
iland | -1.226 |
Token | voting |
Token ID | 6709 |
Feature activation | -4.90e-01 |
Pos logit contributions | |
robe | +3.527 |
Spoiler | +2.878 |
hib | +2.628 |
Doors | +2.616 |
ンジ | +2.597 |
Neg logit contributions | |
bia | -1.974 |
constitu | -1.924 |
rele | -1.921 |
ibaba | -1.888 |
ilial | -1.855 |
Token | own |
Token ID | 898 |
Feature activation | +4.71e-01 |
Pos logit contributions | |
rele | +0.469 |
bia | +0.463 |
constitu | +0.459 |
iland | +0.444 |
ibaba | +0.444 |
Neg logit contributions | |
robe | -0.649 |
Spoiler | -0.502 |
ンジ | -0.472 |
taboola | -0.465 |
hib | -0.459 |
Token | to |
Token ID | 284 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
rele | +2.120 |
constitu | +2.073 |
bia | +2.067 |
iland | +1.993 |
ibaba | +1.979 |
Neg logit contributions | |
robe | -2.661 |
Spoiler | -2.035 |
ンジ | -1.936 |
taboola | -1.888 |
Doors | -1.864 |
Token | explain |
Token ID | 4727 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
rele | +3.975 |
bia | +3.689 |
constitu | +3.581 |
iland | +3.549 |
edi | +3.473 |
Neg logit contributions | |
robe | -7.259 |
Spoiler | -5.785 |
ンジ | -5.545 |
taboola | -5.442 |
ヤ | -5.437 |
Token | . |
Token ID | 13 |
Feature activation | -8.98e-02 |
Pos logit contributions | |
rele | +0.129 |
bia | +0.128 |
constitu | +0.127 |
ibaba | +0.122 |
iland | +0.122 |
Neg logit contributions | |
robe | -0.183 |
Spoiler | -0.142 |
ンジ | -0.134 |
taboola | -0.132 |
hib | -0.130 |
Token | Ad |
Token ID | 1215 |
Feature activation | -1.33e-01 |
Pos logit contributions | |
robe | +0.490 |
Spoiler | +0.381 |
ンジ | +0.349 |
Doors | +0.343 |
taboola | +0.342 |
Neg logit contributions | |
rele | -0.426 |
bia | -0.418 |
constitu | -0.417 |
ibaba | -0.407 |
iland | -0.403 |
Token | Policy |
Token ID | 7820 |
Feature activation | -1.58e+00 |
Pos logit contributions | |
robe | +0.743 |
Spoiler | +0.564 |
ンジ | +0.522 |
hib | +0.519 |
taboola | +0.513 |
Neg logit contributions | |
rele | -0.615 |
bia | -0.612 |
constitu | -0.609 |
ibaba | -0.586 |
iland | -0.584 |
Token |
|
Token ID | 198 |
Feature activation | -3.42e-01 |
Pos logit contributions | |
robe | +7.713 |
Spoiler | +6.299 |
Doors | +5.337 |
hib | +5.298 |
ンジ | +5.284 |
Neg logit contributions | |
rele | -7.265 |
ibaba | -7.222 |
constitu | -7.179 |
bia | -7.128 |
iland | -7.058 |
Token |
|
Token ID | 198 |
Feature activation | +5.68e-02 |
Pos logit contributions | |
robe | +1.946 |
Spoiler | +1.503 |
ンジ | +1.391 |
taboola | +1.382 |
hib | +1.367 |
Neg logit contributions | |
constitu | -1.547 |
rele | -1.541 |
bia | -1.509 |
ibaba | -1.479 |
iland | -1.456 |
Token | Mod |
Token ID | 5841 |
Feature activation | -7.85e-01 |
Pos logit contributions | |
rele | +0.252 |
constitu | +0.247 |
bia | +0.246 |
ibaba | +0.239 |
iland | +0.236 |
Neg logit contributions | |
robe | -0.332 |
Spoiler | -0.255 |
ンジ | -0.239 |
taboola | -0.237 |
hib | -0.234 |
Token | er |
Token ID | 263 |
Feature activation | -7.96e-01 |
Pos logit contributions | |
robe | +4.118 |
Spoiler | +3.133 |
hib | +2.816 |
taboola | +2.810 |
Crate | +2.791 |
Neg logit contributions | |
constitu | -3.709 |
rele | -3.683 |
ibaba | -3.588 |
bia | -3.573 |
iland | -3.539 |
Token | ator |
Token ID | 1352 |
Feature activation | -3.31e-01 |
Pos logit contributions | |
robe | +3.326 |
Spoiler | +2.322 |
taboola | +2.077 |
hib | +2.029 |
ンジ | +2.028 |
Neg logit contributions | |
rele | -4.606 |
constitu | -4.530 |
ibaba | -4.465 |
bia | -4.463 |
iland | -4.378 |
Token | Championship |
Token ID | 10749 |
Feature activation | +2.68e-01 |
Pos logit contributions | |
robe | +0.433 |
Spoiler | +0.333 |
ンジ | +0.307 |
taboola | +0.302 |
Doors | +0.300 |
Neg logit contributions | |
rele | -0.358 |
bia | -0.356 |
constitu | -0.353 |
iland | -0.344 |
ibaba | -0.342 |
Token | Game |
Token ID | 3776 |
Feature activation | -7.45e-01 |
Pos logit contributions | |
rele | +1.262 |
bia | +1.238 |
constitu | +1.229 |
iland | +1.191 |
ibaba | +1.190 |
Neg logit contributions | |
robe | -1.474 |
Spoiler | -1.109 |
ンジ | -1.048 |
taboola | -1.032 |
hib | -1.017 |
Token | . |
Token ID | 13 |
Feature activation | -3.82e-01 |
Pos logit contributions | |
robe | +4.174 |
Spoiler | +3.282 |
Doors | +2.946 |
ンジ | +2.935 |
ヤ | +2.910 |
Neg logit contributions | |
bia | -3.302 |
constitu | -3.264 |
rele | -3.223 |
iland | -3.171 |
ibaba | -3.159 |
Token |
|
Token ID | 198 |
Feature activation | -6.10e-01 |
Pos logit contributions | |
robe | +2.181 |
Spoiler | +1.707 |
ンジ | +1.577 |
Doors | +1.554 |
taboola | +1.548 |
Neg logit contributions | |
rele | -1.698 |
bia | -1.682 |
constitu | -1.670 |
ibaba | -1.619 |
iland | -1.608 |
Token |
|
Token ID | 198 |
Feature activation | -4.34e-01 |
Pos logit contributions | |
robe | +3.319 |
Spoiler | +2.568 |
ンジ | +2.315 |
taboola | +2.314 |
Doors | +2.313 |
Neg logit contributions | |
constitu | -2.913 |
rele | -2.849 |
bia | -2.805 |
ibaba | -2.797 |
iland | -2.722 |
Token | Setting |
Token ID | 34149 |
Feature activation | -1.59e+00 |
Pos logit contributions | |
robe | +2.511 |
Spoiler | +1.940 |
ンジ | +1.811 |
taboola | +1.796 |
ヤ | +1.771 |
Neg logit contributions | |
rele | -1.917 |
constitu | -1.899 |
bia | -1.871 |
ibaba | -1.853 |
ilial | -1.797 |
Token | the |
Token ID | 262 |
Feature activation | -4.69e-01 |
Pos logit contributions | |
robe | +8.464 |
Spoiler | +6.982 |
Doors | +6.495 |
ンジ | +6.127 |
Crate | +6.073 |
Neg logit contributions | |
bia | -6.889 |
constitu | -6.791 |
iland | -6.735 |
rele | -6.386 |
edi | -6.291 |
Token | Stage |
Token ID | 15371 |
Feature activation | -2.17e-01 |
Pos logit contributions | |
robe | +2.584 |
Spoiler | +2.042 |
Doors | +1.868 |
ンジ | +1.836 |
Crate | +1.814 |
Neg logit contributions | |
bia | -2.148 |
constitu | -2.115 |
rele | -2.101 |
iland | -2.089 |
ibaba | -2.048 |
Token |
|
Token ID | 198 |
Feature activation | -5.88e-01 |
Pos logit contributions | |
robe | +1.278 |
Spoiler | +0.998 |
ンジ | +0.926 |
Doors | +0.910 |
taboola | +0.909 |
Neg logit contributions | |
rele | -0.940 |
bia | -0.927 |
constitu | -0.923 |
iland | -0.884 |
ibaba | -0.883 |
Token |
|
Token ID | 198 |
Feature activation | -4.09e-01 |
Pos logit contributions | |
robe | +3.206 |
Spoiler | +2.480 |
Doors | +2.235 |
hib | +2.234 |
ンジ | +2.232 |
Neg logit contributions | |
constitu | -2.817 |
rele | -2.751 |
bia | -2.709 |
ibaba | -2.692 |
iland | -2.619 |
Token | Like |
Token ID | 7594 |
Feature activation | +2.00e-01 |
Pos logit contributions | |
robe | +2.377 |
Spoiler | +1.819 |
ンジ | +1.716 |
taboola | +1.680 |
hib | +1.670 |
Neg logit contributions | |
rele | -1.812 |
constitu | -1.799 |
bia | -1.767 |
ibaba | -1.747 |
ilial | -1.704 |
Token | , |
Token ID | 11 |
Feature activation | +4.74e-01 |
Pos logit contributions | |
rele | +0.991 |
bia | +0.977 |
constitu | +0.969 |
ibaba | +0.933 |
iland | +0.933 |
Neg logit contributions | |
robe | -1.472 |
Spoiler | -1.150 |
ンジ | -1.084 |
taboola | -1.070 |
hib | -1.057 |
Token | 2016 |
Token ID | 1584 |
Feature activation | -4.89e-01 |
Pos logit contributions | |
rele | +1.598 |
bia | +1.565 |
constitu | +1.549 |
iland | +1.480 |
ibaba | +1.477 |
Neg logit contributions | |
robe | -3.243 |
Spoiler | -2.606 |
ンジ | -2.485 |
taboola | -2.452 |
hib | -2.427 |
Token | , |
Token ID | 11 |
Feature activation | +2.78e-02 |
Pos logit contributions | |
robe | +2.788 |
Spoiler | +2.178 |
Doors | +1.965 |
taboola | +1.960 |
ンジ | +1.940 |
Neg logit contributions | |
bia | -2.144 |
rele | -2.133 |
constitu | -2.131 |
ibaba | -2.083 |
iland | -2.048 |
Token | 8 |
Token ID | 807 |
Feature activation | +1.42e-02 |
Pos logit contributions | |
bia | +0.111 |
rele | +0.110 |
constitu | +0.109 |
ibaba | +0.106 |
iland | +0.105 |
Neg logit contributions | |
robe | -0.171 |
Spoiler | -0.137 |
ンジ | -0.126 |
taboola | -0.125 |
Doors | -0.125 |
Token | : |
Token ID | 25 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
rele | +0.067 |
bia | +0.066 |
constitu | +0.065 |
ibaba | +0.063 |
iland | +0.063 |
Neg logit contributions | |
robe | -0.078 |
Spoiler | -0.059 |
ンジ | -0.055 |
taboola | -0.055 |
hib | -0.054 |
Token | 00 |
Token ID | 405 |
Feature activation | -1.47e+00 |
Pos logit contributions | |
rele | +0.516 |
constitu | +0.510 |
bia | +0.509 |
ibaba | +0.493 |
iland | +0.490 |
Neg logit contributions | |
robe | -0.571 |
Spoiler | -0.432 |
ンジ | -0.399 |
taboola | -0.395 |
hib | -0.389 |
Token | PM |
Token ID | 3122 |
Feature activation | -8.53e-01 |
Pos logit contributions | |
robe | +7.683 |
Spoiler | +5.909 |
Doors | +5.491 |
taboola | +5.402 |
hib | +5.327 |
Neg logit contributions | |
rele | -6.442 |
constitu | -6.425 |
iland | -6.350 |
bia | -6.307 |
ibaba | -6.188 |
Token | PDT |
Token ID | 28288 |
Feature activation | -8.57e-01 |
Pos logit contributions | |
robe | +4.415 |
Spoiler | +3.413 |
Doors | +3.116 |
taboola | +3.052 |
Crate | +3.036 |
Neg logit contributions | |
rele | -4.043 |
bia | -4.019 |
constitu | -4.015 |
ibaba | -3.934 |
iland | -3.885 |
Token |
|
Token ID | 198 |
Feature activation | -5.00e-01 |
Pos logit contributions | |
robe | +4.721 |
Spoiler | +3.641 |
Doors | +3.304 |
taboola | +3.254 |
hib | +3.214 |
Neg logit contributions | |
constitu | -3.906 |
bia | -3.893 |
rele | -3.838 |
ibaba | -3.761 |
iland | -3.712 |
Token |
|
Token ID | 198 |
Feature activation | -2.23e-01 |
Pos logit contributions | |
robe | +2.768 |
Spoiler | +2.140 |
taboola | +1.944 |
ンジ | +1.944 |
Doors | +1.940 |
Neg logit contributions | |
constitu | -2.355 |
rele | -2.298 |
bia | -2.267 |
ibaba | -2.240 |
iland | -2.189 |
Token | The |
Token ID | 464 |
Feature activation | +2.58e-01 |
Pos logit contributions | |
robe | +1.271 |
Spoiler | +0.962 |
taboola | +0.901 |
ンジ | +0.897 |
hib | +0.890 |
Neg logit contributions | |
rele | -1.009 |
constitu | -0.995 |
bia | -0.985 |
ibaba | -0.960 |
iland | -0.947 |
Token | are |
Token ID | 389 |
Feature activation | +3.85e-01 |
Pos logit contributions | |
rele | +0.209 |
bia | +0.207 |
constitu | +0.204 |
ibaba | +0.198 |
iland | +0.197 |
Neg logit contributions | |
robe | -0.329 |
Spoiler | -0.260 |
ンジ | -0.243 |
taboola | -0.241 |
hib | -0.238 |
Token | - |
Token ID | 532 |
Feature activation | -3.58e-01 |
Pos logit contributions | |
rele | +1.504 |
bia | +1.480 |
constitu | +1.467 |
iland | +1.411 |
ibaba | +1.410 |
Neg logit contributions | |
robe | -2.422 |
Spoiler | -1.908 |
ンジ | -1.805 |
taboola | -1.780 |
hib | -1.760 |
Token | that |
Token ID | 326 |
Feature activation | -3.21e-01 |
Pos logit contributions | |
robe | +2.293 |
Spoiler | +1.868 |
ンジ | +1.742 |
taboola | +1.718 |
Doors | +1.715 |
Neg logit contributions | |
bia | -1.292 |
rele | -1.258 |
constitu | -1.251 |
iland | -1.222 |
ibaba | -1.221 |
Token | no |
Token ID | 645 |
Feature activation | +2.65e-01 |
Pos logit contributions | |
robe | +1.958 |
Spoiler | +1.567 |
ンジ | +1.460 |
taboola | +1.450 |
Doors | +1.443 |
Neg logit contributions | |
bia | -1.258 |
rele | -1.251 |
constitu | -1.236 |
ibaba | -1.205 |
iland | -1.199 |
Token | real |
Token ID | 1103 |
Feature activation | -6.04e-01 |
Pos logit contributions | |
rele | +0.926 |
bia | +0.917 |
constitu | +0.905 |
ibaba | +0.875 |
iland | +0.866 |
Neg logit contributions | |
robe | -1.772 |
Spoiler | -1.424 |
ンジ | -1.351 |
taboola | -1.337 |
Doors | -1.316 |
Token | policy |
Token ID | 2450 |
Feature activation | -1.32e+00 |
Pos logit contributions | |
robe | +3.684 |
Spoiler | +2.982 |
ンジ | +2.766 |
taboola | +2.737 |
ヤ | +2.722 |
Neg logit contributions | |
bia | -2.365 |
ibaba | -2.303 |
rele | -2.264 |
iland | -2.263 |
constitu | -2.252 |
Token | changes |
Token ID | 2458 |
Feature activation | -6.37e-01 |
Pos logit contributions | |
robe | +6.956 |
Spoiler | +5.491 |
taboola | +4.894 |
roo | +4.860 |
Doors | +4.829 |
Neg logit contributions | |
bia | -5.958 |
ibaba | -5.788 |
iland | -5.769 |
ilial | -5.743 |
constitu | -5.502 |
Token | are |
Token ID | 389 |
Feature activation | -1.80e-01 |
Pos logit contributions | |
robe | +3.669 |
Spoiler | +2.883 |
hib | +2.645 |
taboola | +2.616 |
Doors | +2.600 |
Neg logit contributions | |
bia | -2.703 |
ibaba | -2.641 |
rele | -2.639 |
ilial | -2.602 |
iland | -2.579 |
Token | needed |
Token ID | 2622 |
Feature activation | -6.24e-02 |
Pos logit contributions | |
robe | +1.258 |
Spoiler | +1.021 |
hib | +0.951 |
Doors | +0.944 |
roo | +0.938 |
Neg logit contributions | |
bia | -0.576 |
ibaba | -0.557 |
constitu | -0.551 |
iland | -0.549 |
ilial | -0.544 |
Token | ( |
Token ID | 357 |
Feature activation | +3.00e-01 |
Pos logit contributions | |
robe | +0.367 |
Spoiler | +0.287 |
ンジ | +0.265 |
taboola | +0.262 |
hib | +0.262 |
Neg logit contributions | |
rele | -0.266 |
bia | -0.265 |
constitu | -0.260 |
ibaba | -0.256 |
iland | -0.253 |
Token | other |
Token ID | 847 |
Feature activation | -5.78e-01 |
Pos logit contributions | |
rele | +1.194 |
bia | +1.170 |
constitu | +1.169 |
ibaba | +1.124 |
iland | +1.118 |
Neg logit contributions | |
robe | -1.870 |
Spoiler | -1.468 |
ンジ | -1.386 |
taboola | -1.367 |
hib | -1.355 |