Token | make |
Token ID | 787 |
Feature activation | +5.57e-02 |
Token | things |
Token ID | 1243 |
Feature activation | +2.06e-02 |
Token | better |
Token ID | 1365 |
Feature activation | +1.43e-01 |
Token | in |
Token ID | 287 |
Feature activation | +6.75e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.88e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.71e-01 |
Token | . |
Token ID | 13 |
Feature activation | -2.51e-03 |
Token | Tom |
Token ID | 4186 |
Feature activation | -4.76e-03 |
Token | learned |
Token ID | 4499 |
Feature activation | -5.23e-03 |
Token | his |
Token ID | 465 |
Feature activation | -4.21e-02 |
Token | lesson |
Token ID | 11483 |
Feature activation | -2.20e-02 |
Token | can |
Token ID | 460 |
Feature activation | +4.08e-02 |
Token | help |
Token ID | 1037 |
Feature activation | +4.87e-02 |
Token | you |
Token ID | 345 |
Feature activation | +7.59e-03 |
Token | enjoy |
Token ID | 2883 |
Feature activation | +8.82e-02 |
Token | nature |
Token ID | 3450 |
Feature activation | +3.90e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.70e-01 |
Token | . |
Token ID | 13 |
Feature activation | +1.02e-03 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.15e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -4.76e-02 |
Token | there |
Token ID | 612 |
Feature activation | -9.12e-02 |
Token | was |
Token ID | 373 |
Feature activation | -2.91e-04 |
Token | doing |
Token ID | 1804 |
Feature activation | -4.60e-02 |
Token | what |
Token ID | 644 |
Feature activation | +5.07e-02 |
Token | you |
Token ID | 345 |
Feature activation | +2.01e-02 |
Token | love |
Token ID | 1842 |
Feature activation | +8.66e-02 |
Token | is |
Token ID | 318 |
Feature activation | +4.84e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.69e-01 |
Token | important |
Token ID | 1593 |
Feature activation | +3.39e-02 |
Token | than |
Token ID | 621 |
Feature activation | +4.51e-02 |
Token | being |
Token ID | 852 |
Feature activation | +7.69e-02 |
Token | famous |
Token ID | 5863 |
Feature activation | +4.09e-03 |
Token | . |
Token ID | 13 |
Feature activation | +5.60e-03 |
Token | is |
Token ID | 318 |
Feature activation | +1.96e-02 |
Token | worth |
Token ID | 2861 |
Feature activation | +9.66e-02 |
Token | it |
Token ID | 340 |
Feature activation | -9.97e-03 |
Token | in |
Token ID | 287 |
Feature activation | +6.16e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.21e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.68e-01 |
Token | . |
Token ID | 13 |
Feature activation | +3.81e-04 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -3.11e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.56e-02 |
Token | there |
Token ID | 612 |
Feature activation | -9.78e-02 |
Token | was |
Token ID | 373 |
Feature activation | -4.54e-03 |
Token | should |
Token ID | 815 |
Feature activation | -4.73e-02 |
Token | have |
Token ID | 423 |
Feature activation | +5.76e-02 |
Token | valued |
Token ID | 17560 |
Feature activation | +6.06e-02 |
Token | our |
Token ID | 674 |
Feature activation | +1.78e-02 |
Token | friendship |
Token ID | 14738 |
Feature activation | +4.10e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.68e-01 |
Token | than |
Token ID | 621 |
Feature activation | +4.26e-02 |
Token | our |
Token ID | 674 |
Feature activation | +1.50e-02 |
Token | snacks |
Token ID | 28487 |
Feature activation | +4.89e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +5.84e-02 |
Token |
|
Token ID | 198 |
Feature activation | +3.90e-02 |
Token | wished |
Token ID | 16555 |
Feature activation | -3.90e-02 |
Token | he |
Token ID | 339 |
Feature activation | +4.05e-03 |
Token | had |
Token ID | 550 |
Feature activation | +4.05e-02 |
Token | valued |
Token ID | 17560 |
Feature activation | +6.33e-02 |
Token | friendship |
Token ID | 14738 |
Feature activation | +2.91e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.67e-01 |
Token | than |
Token ID | 621 |
Feature activation | +4.31e-02 |
Token | tw |
Token ID | 665 |
Feature activation | +3.66e-02 |
Token | igs |
Token ID | 9235 |
Feature activation | +7.19e-02 |
Token | . |
Token ID | 13 |
Feature activation | +9.96e-03 |
Token | He |
Token ID | 679 |
Feature activation | -1.82e-02 |
Token | a |
Token ID | 257 |
Feature activation | +4.94e-02 |
Token | chance |
Token ID | 2863 |
Feature activation | +1.83e-02 |
Token | to |
Token ID | 284 |
Feature activation | -2.25e-02 |
Token | become |
Token ID | 1716 |
Feature activation | +7.95e-02 |
Token | something |
Token ID | 1223 |
Feature activation | +5.01e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.66e-01 |
Token | - |
Token ID | 532 |
Feature activation | +4.72e-03 |
Token | so |
Token ID | 523 |
Feature activation | +3.48e-02 |
Token | I |
Token ID | 314 |
Feature activation | +6.56e-02 |
Token | 'm |
Token ID | 1101 |
Feature activation | +3.89e-02 |
Token | cooking |
Token ID | 10801 |
Feature activation | +1.06e-02 |
Token | could |
Token ID | 714 |
Feature activation | +2.56e-02 |
Token | see |
Token ID | 766 |
Feature activation | +5.03e-03 |
Token | him |
Token ID | 683 |
Feature activation | -1.64e-02 |
Token | smile |
Token ID | 8212 |
Feature activation | +6.83e-02 |
Token | one |
Token ID | 530 |
Feature activation | +6.33e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.66e-01 |
Token | time |
Token ID | 640 |
Feature activation | +3.94e-02 |
Token | . |
Token ID | 13 |
Feature activation | +4.80e-03 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.88e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.42e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.72e-02 |
Token | he |
Token ID | 339 |
Feature activation | -1.62e-02 |
Token | wishes |
Token ID | 12802 |
Feature activation | -4.96e-02 |
Token | she |
Token ID | 673 |
Feature activation | +3.14e-02 |
Token | would |
Token ID | 561 |
Feature activation | +2.09e-02 |
Token | be |
Token ID | 307 |
Feature activation | +7.25e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.65e-01 |
Token | polite |
Token ID | 23507 |
Feature activation | +4.49e-02 |
Token | . |
Token ID | 13 |
Feature activation | +3.04e-03 |
Token | He |
Token ID | 679 |
Feature activation | -1.93e-02 |
Token | says |
Token ID | 1139 |
Feature activation | -6.87e-02 |
Token | , |
Token ID | 11 |
Feature activation | +4.60e-02 |
Token | find |
Token ID | 1064 |
Feature activation | +1.14e-02 |
Token | something |
Token ID | 1223 |
Feature activation | +4.21e-02 |
Token | special |
Token ID | 2041 |
Feature activation | +7.66e-03 |
Token | at |
Token ID | 379 |
Feature activation | -3.92e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.16e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.65e-01 |
Token | of |
Token ID | 286 |
Feature activation | +1.04e-01 |
Token | the |
Token ID | 262 |
Feature activation | +2.09e-02 |
Token | stream |
Token ID | 4269 |
Feature activation | +5.19e-04 |
Token | ". |
Token ID | 1911 |
Feature activation | +7.53e-03 |
Token |
|
Token ID | 198 |
Feature activation | +1.93e-02 |
Token | but |
Token ID | 475 |
Feature activation | -7.40e-02 |
Token | she |
Token ID | 673 |
Feature activation | +1.37e-02 |
Token | liked |
Token ID | 8288 |
Feature activation | +3.63e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.00e-02 |
Token | monkeys |
Token ID | 26697 |
Feature activation | +5.14e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.65e-01 |
Token | . |
Token ID | 13 |
Feature activation | +1.15e-02 |
Token | She |
Token ID | 1375 |
Feature activation | +1.58e-02 |
Token | and |
Token ID | 290 |
Feature activation | -1.66e-02 |
Token | Sam |
Token ID | 3409 |
Feature activation | -3.27e-02 |
Token | gave |
Token ID | 2921 |
Feature activation | +2.82e-02 |
Token | bell |
Token ID | 8966 |
Feature activation | -5.31e-02 |
Token | together |
Token ID | 1978 |
Feature activation | +1.63e-02 |
Token | to |
Token ID | 284 |
Feature activation | -4.00e-02 |
Token | celebrate |
Token ID | 10648 |
Feature activation | +4.90e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.82e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.65e-01 |
Token | of |
Token ID | 286 |
Feature activation | +1.00e-01 |
Token | the |
Token ID | 262 |
Feature activation | +2.02e-02 |
Token | war |
Token ID | 1175 |
Feature activation | -7.19e-03 |
Token | . |
Token ID | 13 |
Feature activation | -6.12e-04 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.17e-02 |
Token | wished |
Token ID | 16555 |
Feature activation | -4.05e-02 |
Token | he |
Token ID | 339 |
Feature activation | +2.74e-03 |
Token | had |
Token ID | 550 |
Feature activation | +3.61e-02 |
Token | loved |
Token ID | 6151 |
Feature activation | +7.00e-02 |
Token | them |
Token ID | 606 |
Feature activation | +4.20e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.64e-01 |
Token | . |
Token ID | 13 |
Feature activation | +1.23e-02 |
Token | They |
Token ID | 1119 |
Feature activation | +3.52e-02 |
Token | wished |
Token ID | 16555 |
Feature activation | -3.84e-02 |
Token | he |
Token ID | 339 |
Feature activation | +3.39e-03 |
Token | had |
Token ID | 550 |
Feature activation | +3.73e-02 |
Token | was |
Token ID | 373 |
Feature activation | +1.85e-02 |
Token | worth |
Token ID | 2861 |
Feature activation | +9.48e-02 |
Token | it |
Token ID | 340 |
Feature activation | -1.49e-02 |
Token | in |
Token ID | 287 |
Feature activation | +5.89e-02 |
Token | the |
Token ID | 262 |
Feature activation | +8.11e-03 |
Token | end |
Token ID | 886 |
Feature activation | +1.64e-01 |
Token | . |
Token ID | 13 |
Feature activation | +5.46e-04 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.86e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.29e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.63e-02 |
Token | a |
Token ID | 257 |
Feature activation | +4.42e-03 |
Token | said |
Token ID | 531 |
Feature activation | -7.70e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.86e-02 |
Token | I |
Token ID | 40 |
Feature activation | +6.38e-02 |
Token | love |
Token ID | 1842 |
Feature activation | +6.84e-02 |
Token | carrots |
Token ID | 34397 |
Feature activation | +3.43e-02 |
Token | more |
Token ID | 517 |
Feature activation | +1.64e-01 |
Token | than |
Token ID | 621 |
Feature activation | +4.32e-02 |
Token | anything |
Token ID | 1997 |
Feature activation | +6.60e-02 |
Token | !" |
Token ID | 2474 |
Feature activation | +3.26e-02 |
Token | and |
Token ID | 290 |
Feature activation | -4.08e-02 |
Token | he |
Token ID | 339 |
Feature activation | -2.00e-02 |
Token | your |
Token ID | 534 |
Feature activation | -1.55e-02 |
Token | best |
Token ID | 1266 |
Feature activation | +7.56e-02 |
Token | to |
Token ID | 284 |
Feature activation | -6.33e-03 |
Token | find |
Token ID | 1064 |
Feature activation | +1.80e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.22e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.63e-01 |
Token | of |
Token ID | 286 |
Feature activation | +1.00e-01 |
Token | the |
Token ID | 262 |
Feature activation | +2.01e-02 |
Token | string |
Token ID | 4731 |
Feature activation | -1.84e-02 |
Token | - |
Token ID | 532 |
Feature activation | +9.21e-03 |
Token | it |
Token ID | 340 |
Feature activation | -5.35e-02 |
Token | to |
Token ID | 284 |
Feature activation | -5.02e-02 |
Token | make |
Token ID | 787 |
Feature activation | +4.12e-02 |
Token | it |
Token ID | 340 |
Feature activation | -3.29e-02 |
Token | sound |
Token ID | 2128 |
Feature activation | -1.73e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.57e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.63e-01 |
Token | . |
Token ID | 13 |
Feature activation | -9.21e-03 |
Token |
|
Token ID | 198 |
Feature activation | +1.63e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.15e-02 |
Token | On |
Token ID | 2202 |
Feature activation | -7.74e-02 |
Token | her |
Token ID | 607 |
Feature activation | -2.67e-02 |
Token | turn |
Token ID | 1210 |
Feature activation | +3.79e-02 |
Token | out |
Token ID | 503 |
Feature activation | +5.10e-02 |
Token | okay |
Token ID | 8788 |
Feature activation | +1.38e-02 |
Token | in |
Token ID | 287 |
Feature activation | +5.68e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.42e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.63e-01 |
Token | . |
Token ID | 13 |
Feature activation | -4.88e-03 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.80e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.30e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.60e-02 |
Token | a |
Token ID | 257 |
Feature activation | +5.81e-03 |
Token | would |
Token ID | 561 |
Feature activation | +2.26e-02 |
Token | be |
Token ID | 307 |
Feature activation | +7.21e-02 |
Token | okay |
Token ID | 8788 |
Feature activation | +1.75e-02 |
Token | in |
Token ID | 287 |
Feature activation | +6.03e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.66e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.63e-01 |
Token | . |
Token ID | 13 |
Feature activation | -6.95e-03 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.38e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -4.88e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.61e-02 |
Token | a |
Token ID | 257 |
Feature activation | +6.72e-03 |
Token | see |
Token ID | 766 |
Feature activation | +6.97e-04 |
Token | what |
Token ID | 644 |
Feature activation | +4.61e-02 |
Token | was |
Token ID | 373 |
Feature activation | +4.30e-02 |
Token | at |
Token ID | 379 |
Feature activation | -4.37e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.72e-02 |
Token | end |
Token ID | 886 |
Feature activation | +1.63e-01 |
Token | of |
Token ID | 286 |
Feature activation | +1.05e-01 |
Token | it |
Token ID | 340 |
Feature activation | -9.48e-03 |
Token | . |
Token ID | 13 |
Feature activation | -1.24e-02 |
Token | She |
Token ID | 1375 |
Feature activation | +2.67e-03 |
Token | was |
Token ID | 373 |
Feature activation | +3.01e-03 |
Token | a |
Token ID | 257 |
Feature activation | +3.48e-02 |
Token | place |
Token ID | 1295 |
Feature activation | +1.09e-02 |
Token | called |
Token ID | 1444 |
Feature activation | +3.57e-02 |
Token | jail |
Token ID | 7356 |
Feature activation | -2.60e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.31e-02 |
Token | Jail |
Token ID | 25715 |
Feature activation | -1.53e-01 |
Token | was |
Token ID | 373 |
Feature activation | -9.01e-04 |
Token | a |
Token ID | 257 |
Feature activation | +3.41e-02 |
Token | big |
Token ID | 1263 |
Feature activation | +1.08e-02 |
Token | house |
Token ID | 2156 |
Feature activation | -1.08e-02 |
Token | with |
Token ID | 351 |
Feature activation | +3.41e-02 |
Token | wall |
Token ID | 3355 |
Feature activation | -6.17e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.51e-02 |
Token | It |
Token ID | 632 |
Feature activation | -6.92e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.73e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.12e-02 |
Token | DO |
Token ID | 18227 |
Feature activation | -1.53e-01 |
Token | NOT |
Token ID | 5626 |
Feature activation | -6.60e-02 |
Token | T |
Token ID | 309 |
Feature activation | -5.41e-02 |
Token | OU |
Token ID | 2606 |
Feature activation | -8.41e-02 |
Token | CH |
Token ID | 3398 |
Feature activation | -5.45e-02 |
Token | ". |
Token ID | 1911 |
Feature activation | -6.57e-03 |
Token | red |
Token ID | 2266 |
Feature activation | -6.37e-02 |
Token | button |
Token ID | 4936 |
Feature activation | -9.30e-02 |
Token | that |
Token ID | 326 |
Feature activation | -2.33e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.17e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.27e-02 |
Token | DO |
Token ID | 18227 |
Feature activation | -1.52e-01 |
Token | NOT |
Token ID | 5626 |
Feature activation | -6.52e-02 |
Token | PRESS |
Token ID | 44310 |
Feature activation | -9.87e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +3.23e-02 |
Token | She |
Token ID | 1375 |
Feature activation | +5.07e-03 |
Token | couldn |
Token ID | 3521 |
Feature activation | -4.77e-02 |
Token | . |
Token ID | 13 |
Feature activation | -2.75e-02 |
Token |
|
Token ID | 198 |
Feature activation | +5.10e-03 |
Token |
|
Token ID | 198 |
Feature activation | +1.86e-03 |
Token | S |
Token ID | 50 |
Feature activation | -4.39e-02 |
Token | eb |
Token ID | 1765 |
Feature activation | -9.40e-02 |
Token | ast |
Token ID | 459 |
Feature activation | -1.51e-01 |
Token | ian |
Token ID | 666 |
Feature activation | -5.22e-02 |
Token | disagreed |
Token ID | 27884 |
Feature activation | -7.96e-02 |
Token | with |
Token ID | 351 |
Feature activation | +9.76e-03 |
Token | his |
Token ID | 465 |
Feature activation | -4.74e-02 |
Token | friend |
Token ID | 1545 |
Feature activation | +1.90e-02 |
Token | deep |
Token ID | 2769 |
Feature activation | -2.18e-02 |
Token | breath |
Token ID | 8033 |
Feature activation | +8.96e-03 |
Token | and |
Token ID | 290 |
Feature activation | -1.80e-02 |
Token | shouted |
Token ID | 17293 |
Feature activation | -6.68e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.26e-02 |
Token | ST |
Token ID | 2257 |
Feature activation | -1.51e-01 |
Token | OP |
Token ID | 3185 |
Feature activation | -4.54e-02 |
Token | !" |
Token ID | 2474 |
Feature activation | +3.08e-02 |
Token | |
Token ID | 220 |
Feature activation | +9.98e-03 |
Token |
|
Token ID | 198 |
Feature activation | +1.19e-02 |
Token |
|
Token ID | 198 |
Feature activation | +8.01e-03 |
Token | visited |
Token ID | 8672 |
Feature activation | -1.93e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.15e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.16e-02 |
Token |
|
Token ID | 198 |
Feature activation | +7.45e-03 |
Token | Lyn |
Token ID | 37207 |
Feature activation | -9.69e-02 |
Token | n |
Token ID | 77 |
Feature activation | -1.51e-01 |
Token | loved |
Token ID | 6151 |
Feature activation | +5.04e-02 |
Token | to |
Token ID | 284 |
Feature activation | -3.57e-02 |
Token | eat |
Token ID | 4483 |
Feature activation | +1.66e-02 |
Token | the |
Token ID | 262 |
Feature activation | +4.38e-03 |
Token | food |
Token ID | 2057 |
Feature activation | +3.31e-02 |
Token | A |
Token ID | 317 |
Feature activation | -1.00e-02 |
Token | voice |
Token ID | 3809 |
Feature activation | -4.29e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.78e-02 |
Token | , |
Token ID | 11 |
Feature activation | +3.10e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.18e-02 |
Token | Warning |
Token ID | 20361 |
Feature activation | -1.49e-01 |
Token | , |
Token ID | 11 |
Feature activation | +4.15e-02 |
Token | warning |
Token ID | 6509 |
Feature activation | -9.68e-02 |
Token | , |
Token ID | 11 |
Feature activation | +4.17e-02 |
Token | bomb |
Token ID | 5194 |
Feature activation | -1.22e-01 |
Token | activated |
Token ID | 13906 |
Feature activation | +2.05e-03 |
Token |
|
Token ID | 198 |
Feature activation | +2.87e-03 |
Token | One |
Token ID | 3198 |
Feature activation | -1.28e-02 |
Token | day |
Token ID | 1110 |
Feature activation | -5.02e-02 |
Token | Alan |
Token ID | 12246 |
Feature activation | -3.19e-02 |
Token | and |
Token ID | 290 |
Feature activation | -5.03e-02 |
Token | Aub |
Token ID | 19723 |
Feature activation | -1.47e-01 |
Token | rey |
Token ID | 4364 |
Feature activation | -5.40e-02 |
Token | decided |
Token ID | 3066 |
Feature activation | -6.83e-02 |
Token | to |
Token ID | 284 |
Feature activation | -6.19e-02 |
Token | turn |
Token ID | 1210 |
Feature activation | +3.27e-03 |
Token | the |
Token ID | 262 |
Feature activation | -8.53e-03 |
Token | sign |
Token ID | 1051 |
Feature activation | -4.69e-02 |
Token | that |
Token ID | 326 |
Feature activation | -2.25e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.22e-02 |
Token | : |
Token ID | 25 |
Feature activation | +1.42e-02 |
Token | " |
Token ID | 366 |
Feature activation | +4.81e-02 |
Token | Warning |
Token ID | 20361 |
Feature activation | -1.46e-01 |
Token | ! |
Token ID | 0 |
Feature activation | -4.48e-02 |
Token | There |
Token ID | 1318 |
Feature activation | -7.97e-02 |
Token | is |
Token ID | 318 |
Feature activation | -5.81e-03 |
Token | bad |
Token ID | 2089 |
Feature activation | -2.02e-02 |
Token | news |
Token ID | 1705 |
Feature activation | -2.99e-02 |
Token | play |
Token ID | 711 |
Feature activation | +3.80e-02 |
Token | safely |
Token ID | 11512 |
Feature activation | +6.05e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +4.47e-02 |
Token |
|
Token ID | 198 |
Feature activation | +2.54e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.95e-02 |
Token | De |
Token ID | 5005 |
Feature activation | -1.45e-01 |
Token | ena |
Token ID | 8107 |
Feature activation | -2.40e-02 |
Token | grabbed |
Token ID | 13176 |
Feature activation | -2.45e-02 |
Token | the |
Token ID | 262 |
Feature activation | +9.20e-04 |
Token | matches |
Token ID | 7466 |
Feature activation | +5.48e-02 |
Token | and |
Token ID | 290 |
Feature activation | -2.86e-02 |
Token | wanted |
Token ID | 2227 |
Feature activation | -5.24e-02 |
Token | to |
Token ID | 284 |
Feature activation | -4.59e-02 |
Token | see |
Token ID | 766 |
Feature activation | -2.88e-03 |
Token | it |
Token ID | 340 |
Feature activation | -4.36e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.77e-02 |
Token | Tour |
Token ID | 9852 |
Feature activation | -1.45e-01 |
Token | ists |
Token ID | 1023 |
Feature activation | +2.13e-02 |
Token | from |
Token ID | 422 |
Feature activation | -3.04e-02 |
Token | all |
Token ID | 477 |
Feature activation | +4.79e-02 |
Token | over |
Token ID | 625 |
Feature activation | +4.24e-02 |
Token | the |
Token ID | 262 |
Feature activation | -2.23e-03 |
Token | . |
Token ID | 13 |
Feature activation | +1.39e-03 |
Token | It |
Token ID | 632 |
Feature activation | -6.41e-02 |
Token | says |
Token ID | 1139 |
Feature activation | -7.43e-02 |
Token | : |
Token ID | 25 |
Feature activation | +2.32e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.74e-02 |
Token | Warning |
Token ID | 20361 |
Feature activation | -1.44e-01 |
Token | : |
Token ID | 25 |
Feature activation | +3.71e-02 |
Token | Do |
Token ID | 2141 |
Feature activation | -4.22e-02 |
Token | not |
Token ID | 407 |
Feature activation | -1.73e-02 |
Token | touch |
Token ID | 3638 |
Feature activation | -3.91e-02 |
Token | . |
Token ID | 13 |
Feature activation | +5.73e-03 |
Token | Aub |
Token ID | 19723 |
Feature activation | -1.34e-01 |
Token | rey |
Token ID | 4364 |
Feature activation | -4.15e-02 |
Token | . |
Token ID | 13 |
Feature activation | -2.50e-02 |
Token | Alan |
Token ID | 12246 |
Feature activation | -2.88e-02 |
Token | and |
Token ID | 290 |
Feature activation | -4.11e-02 |
Token | Aub |
Token ID | 19723 |
Feature activation | -1.43e-01 |
Token | rey |
Token ID | 4364 |
Feature activation | -5.04e-02 |
Token | would |
Token ID | 561 |
Feature activation | -3.31e-03 |
Token | play |
Token ID | 711 |
Feature activation | +2.52e-02 |
Token | together |
Token ID | 1978 |
Feature activation | +1.48e-02 |
Token | every |
Token ID | 790 |
Feature activation | +3.66e-02 |
Token | chilly |
Token ID | 49018 |
Feature activation | -1.81e-02 |
Token | ground |
Token ID | 2323 |
Feature activation | -6.01e-02 |
Token | gave |
Token ID | 2921 |
Feature activation | +2.15e-02 |
Token | him |
Token ID | 683 |
Feature activation | -1.01e-02 |
Token | Goose |
Token ID | 46317 |
Feature activation | -1.01e-02 |
Token | bum |
Token ID | 4435 |
Feature activation | -1.43e-01 |
Token | ps |
Token ID | 862 |
Feature activation | -1.22e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.99e-02 |
Token | He |
Token ID | 679 |
Feature activation | -3.63e-02 |
Token | was |
Token ID | 373 |
Feature activation | -3.99e-03 |
Token | so |
Token ID | 523 |
Feature activation | +3.59e-02 |
Token | voice |
Token ID | 3809 |
Feature activation | -2.87e-02 |
Token | that |
Token ID | 326 |
Feature activation | -2.86e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.23e-02 |
Token | : |
Token ID | 25 |
Feature activation | +1.72e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.04e-02 |
Token | Warning |
Token ID | 20361 |
Feature activation | -1.43e-01 |
Token | ! |
Token ID | 0 |
Feature activation | -3.30e-02 |
Token | It |
Token ID | 632 |
Feature activation | -6.34e-02 |
Token | is |
Token ID | 318 |
Feature activation | +3.37e-03 |
Token | not |
Token ID | 407 |
Feature activation | -2.20e-02 |
Token | safe |
Token ID | 3338 |
Feature activation | +4.50e-02 |
Token | sc |
Token ID | 629 |
Feature activation | -3.21e-02 |
Token | ooter |
Token ID | 25141 |
Feature activation | -1.04e-01 |
Token | . |
Token ID | 13 |
Feature activation | -2.36e-02 |
Token | She |
Token ID | 1375 |
Feature activation | -5.47e-03 |
Token | tr |
Token ID | 491 |
Feature activation | -4.43e-02 |
Token | und |
Token ID | 917 |
Feature activation | -1.40e-01 |
Token | led |
Token ID | 992 |
Feature activation | -3.62e-03 |
Token | along |
Token ID | 1863 |
Feature activation | -2.67e-02 |
Token | the |
Token ID | 262 |
Feature activation | -9.78e-03 |
Token | sidewalk |
Token ID | 24058 |
Feature activation | -3.21e-02 |
Token | all |
Token ID | 477 |
Feature activation | +5.67e-02 |
Token | Tim |
Token ID | 5045 |
Feature activation | -1.58e-02 |
Token | my |
Token ID | 1820 |
Feature activation | +2.26e-02 |
Token | quickly |
Token ID | 2952 |
Feature activation | +1.33e-02 |
Token | yelled |
Token ID | 22760 |
Feature activation | -6.15e-02 |
Token | " |
Token ID | 366 |
Feature activation | +5.21e-02 |
Token | ST |
Token ID | 2257 |
Feature activation | -1.40e-01 |
Token | OP |
Token ID | 3185 |
Feature activation | -4.89e-02 |
Token | !" |
Token ID | 2474 |
Feature activation | +2.75e-02 |
Token | and |
Token ID | 290 |
Feature activation | -5.07e-02 |
Token | ran |
Token ID | 4966 |
Feature activation | -2.79e-02 |
Token | to |
Token ID | 284 |
Feature activation | -5.19e-02 |
Token | her |
Token ID | 607 |
Feature activation | -2.20e-02 |
Token | mom |
Token ID | 1995 |
Feature activation | +5.95e-02 |
Token | explained |
Token ID | 4893 |
Feature activation | -7.89e-02 |
Token | , |
Token ID | 11 |
Feature activation | +3.37e-02 |
Token | " |
Token ID | 366 |
Feature activation | +4.50e-02 |
Token | Weight |
Token ID | 25844 |
Feature activation | -1.40e-01 |
Token | is |
Token ID | 318 |
Feature activation | +5.41e-03 |
Token | how |
Token ID | 703 |
Feature activation | +8.99e-02 |
Token | heavy |
Token ID | 4334 |
Feature activation | -3.24e-02 |
Token | something |
Token ID | 1223 |
Feature activation | +3.36e-02 |
Token | is |
Token ID | 318 |
Feature activation | +2.40e-02 |
Token | smile |
Token ID | 8212 |
Feature activation | +6.22e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.02e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.71e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.48e-02 |
Token | S |
Token ID | 50 |
Feature activation | -3.88e-02 |
Token | aman |
Token ID | 10546 |
Feature activation | -1.40e-01 |
Token | tha |
Token ID | 12898 |
Feature activation | -2.84e-02 |
Token | laughed |
Token ID | 13818 |
Feature activation | -8.04e-03 |
Token | , |
Token ID | 11 |
Feature activation | +4.05e-02 |
Token | happy |
Token ID | 3772 |
Feature activation | -3.62e-03 |
Token | that |
Token ID | 326 |
Feature activation | -1.82e-02 |
Token | a |
Token ID | 257 |
Feature activation | -1.80e-03 |
Token | big |
Token ID | 1263 |
Feature activation | -2.01e-02 |
Token | light |
Token ID | 1657 |
Feature activation | -3.60e-02 |
Token | came |
Token ID | 1625 |
Feature activation | -1.53e-02 |
Token | from |
Token ID | 422 |
Feature activation | -3.78e-02 |
Token | overhead |
Token ID | 16965 |
Feature activation | -1.40e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.99e-02 |
Token | It |
Token ID | 632 |
Feature activation | -7.82e-02 |
Token | was |
Token ID | 373 |
Feature activation | -5.02e-03 |
Token | the |
Token ID | 262 |
Feature activation | -9.46e-03 |
Token | bright |
Token ID | 6016 |
Feature activation | +8.26e-03 |
Token | , |
Token ID | 11 |
Feature activation | +6.10e-02 |
Token | they |
Token ID | 484 |
Feature activation | +2.56e-02 |
Token | can |
Token ID | 460 |
Feature activation | +2.53e-02 |
Token | still |
Token ID | 991 |
Feature activation | +2.80e-02 |
Token | do |
Token ID | 466 |
Feature activation | -1.47e-03 |
Token | good |
Token ID | 922 |
Feature activation | +8.23e-02 |
Token | things |
Token ID | 1243 |
Feature activation | +3.24e-02 |
Token | if |
Token ID | 611 |
Feature activation | +5.33e-03 |
Token | they |
Token ID | 484 |
Feature activation | +4.54e-02 |
Token | try |
Token ID | 1949 |
Feature activation | -9.51e-03 |
Token | hard |
Token ID | 1327 |
Feature activation | +1.02e-02 |
Token | are |
Token ID | 389 |
Feature activation | +7.84e-02 |
Token | for |
Token ID | 329 |
Feature activation | +4.22e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +3.86e-02 |
Token | They |
Token ID | 1119 |
Feature activation | +2.16e-02 |
Token | both |
Token ID | 1111 |
Feature activation | +5.60e-02 |
Token | continued |
Token ID | 3767 |
Feature activation | +8.04e-02 |
Token | on |
Token ID | 319 |
Feature activation | -2.44e-02 |
Token | their |
Token ID | 511 |
Feature activation | -1.20e-02 |
Token | way |
Token ID | 835 |
Feature activation | -1.40e-02 |
Token | , |
Token ID | 11 |
Feature activation | +4.02e-02 |
Token | adm |
Token ID | 6178 |
Feature activation | -3.72e-02 |
Token |
|
Token ID | 198 |
Feature activation | +2.13e-02 |
Token | " |
Token ID | 1 |
Feature activation | +6.10e-02 |
Token | Thank |
Token ID | 10449 |
Feature activation | +3.20e-02 |
Token | you |
Token ID | 345 |
Feature activation | +3.13e-02 |
Token | so |
Token ID | 523 |
Feature activation | +6.69e-02 |
Token | much |
Token ID | 881 |
Feature activation | +1.16e-01 |
Token | for |
Token ID | 329 |
Feature activation | +4.26e-02 |
Token | helping |
Token ID | 5742 |
Feature activation | +1.48e-02 |
Token | me |
Token ID | 502 |
Feature activation | +2.18e-02 |
Token | ," |
Token ID | 553 |
Feature activation | +7.92e-02 |
Token | he |
Token ID | 339 |
Feature activation | +1.18e-03 |
Token | the |
Token ID | 262 |
Feature activation | -3.03e-02 |
Token | rabbit |
Token ID | 22746 |
Feature activation | -2.92e-02 |
Token | spent |
Token ID | 3377 |
Feature activation | +6.82e-02 |
Token | a |
Token ID | 257 |
Feature activation | +3.68e-02 |
Token | lot |
Token ID | 1256 |
Feature activation | +5.59e-02 |
Token | of |
Token ID | 286 |
Feature activation | +8.92e-02 |
Token | time |
Token ID | 640 |
Feature activation | -6.13e-03 |
Token | together |
Token ID | 1978 |
Feature activation | +1.21e-02 |
Token | playing |
Token ID | 2712 |
Feature activation | +3.11e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.05e-02 |
Token | The |
Token ID | 383 |
Feature activation | -5.31e-02 |
Token | to |
Token ID | 284 |
Feature activation | -6.17e-02 |
Token | the |
Token ID | 262 |
Feature activation | -1.63e-02 |
Token | safety |
Token ID | 3747 |
Feature activation | -1.62e-02 |
Token | of |
Token ID | 286 |
Feature activation | +7.26e-02 |
Token | their |
Token ID | 511 |
Feature activation | -2.76e-03 |
Token | cos |
Token ID | 8615 |
Feature activation | +8.01e-02 |
Token | y |
Token ID | 88 |
Feature activation | +2.89e-02 |
Token | house |
Token ID | 2156 |
Feature activation | -2.70e-03 |
Token | . |
Token ID | 13 |
Feature activation | -1.27e-02 |
Token | |
Token ID | 220 |
Feature activation | +1.09e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.82e-02 |
Token | , |
Token ID | 11 |
Feature activation | +5.66e-02 |
Token | bee |
Token ID | 20697 |
Feature activation | +1.09e-02 |
Token | ! |
Token ID | 0 |
Feature activation | -2.48e-02 |
Token | I |
Token ID | 314 |
Feature activation | +7.52e-02 |
Token | love |
Token ID | 1842 |
Feature activation | +6.78e-02 |
Token | honey |
Token ID | 12498 |
Feature activation | +2.13e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +3.46e-02 |
Token | The |
Token ID | 383 |
Feature activation | -5.15e-02 |
Token | bee |
Token ID | 20697 |
Feature activation | -2.05e-03 |
Token | replied |
Token ID | 8712 |
Feature activation | -8.50e-02 |
Token | , |
Token ID | 11 |
Feature activation | +3.14e-02 |
Token | good |
Token ID | 922 |
Feature activation | +8.06e-02 |
Token | girl |
Token ID | 2576 |
Feature activation | +7.75e-02 |
Token | . |
Token ID | 13 |
Feature activation | +1.02e-02 |
Token | But |
Token ID | 887 |
Feature activation | -7.61e-02 |
Token | you |
Token ID | 345 |
Feature activation | -4.05e-02 |
Token | have |
Token ID | 423 |
Feature activation | +2.94e-02 |
Token | to |
Token ID | 284 |
Feature activation | -5.24e-02 |
Token | be |
Token ID | 307 |
Feature activation | +4.90e-02 |
Token | careful |
Token ID | 8161 |
Feature activation | -5.02e-02 |
Token | . |
Token ID | 13 |
Feature activation | -3.32e-03 |
Token | You |
Token ID | 921 |
Feature activation | -4.53e-02 |
Token |
|
Token ID | 198 |
Feature activation | +2.80e-02 |
Token | " |
Token ID | 1 |
Feature activation | +6.16e-02 |
Token | You |
Token ID | 1639 |
Feature activation | -3.57e-02 |
Token | are |
Token ID | 389 |
Feature activation | +6.82e-02 |
Token | brave |
Token ID | 14802 |
Feature activation | +2.70e-02 |
Token | , |
Token ID | 11 |
Feature activation | +6.38e-02 |
Token | Tim |
Token ID | 5045 |
Feature activation | +1.81e-02 |
Token | and |
Token ID | 290 |
Feature activation | -1.59e-02 |
Token | Lily |
Token ID | 20037 |
Feature activation | +6.14e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.14e-03 |
Token | You |
Token ID | 921 |
Feature activation | -4.07e-02 |
Token | don |
Token ID | 836 |
Feature activation | -3.91e-03 |
Token | 't |
Token ID | 470 |
Feature activation | -1.18e-02 |
Token | want |
Token ID | 765 |
Feature activation | -1.51e-02 |
Token | to |
Token ID | 284 |
Feature activation | -3.96e-02 |
Token | ." |
Token ID | 526 |
Feature activation | +3.40e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.69e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.13e-02 |
Token | So |
Token ID | 2396 |
Feature activation | -3.25e-02 |
Token | , |
Token ID | 11 |
Feature activation | +2.78e-02 |
Token | the |
Token ID | 262 |
Feature activation | -3.50e-02 |
Token | mom |
Token ID | 1995 |
Feature activation | +5.64e-02 |
Token | and |
Token ID | 290 |
Feature activation | -3.09e-02 |
Token | make |
Token ID | 787 |
Feature activation | +4.20e-02 |
Token | something |
Token ID | 1223 |
Feature activation | +3.74e-02 |
Token | even |
Token ID | 772 |
Feature activation | +1.01e-01 |
Token | better |
Token ID | 1365 |
Feature activation | +1.26e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.99e-03 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.55e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.25e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.70e-02 |
Token | a |
Token ID | 257 |
Feature activation | +5.18e-03 |
Token | time |
Token ID | 640 |
Feature activation | -2.89e-02 |
Token | ' |
Token ID | 6 |
Feature activation | +3.40e-02 |
Token | bedroom |
Token ID | 14043 |
Feature activation | +5.48e-03 |
Token | . |
Token ID | 13 |
Feature activation | -1.31e-02 |
Token | He |
Token ID | 679 |
Feature activation | -2.92e-02 |
Token | was |
Token ID | 373 |
Feature activation | +1.33e-03 |
Token | curious |
Token ID | 11040 |
Feature activation | -4.35e-02 |
Token | about |
Token ID | 546 |
Feature activation | +2.61e-02 |
Token | what |
Token ID | 644 |
Feature activation | +5.07e-02 |
Token | was |
Token ID | 373 |
Feature activation | +3.98e-02 |
Token | inside |
Token ID | 2641 |
Feature activation | +3.34e-02 |
Token | the |
Token ID | 262 |
Feature activation | +1.02e-02 |
Token | sun |
Token ID | 4252 |
Feature activation | +6.67e-03 |
Token | went |
Token ID | 1816 |
Feature activation | +1.83e-02 |
Token | down |
Token ID | 866 |
Feature activation | -5.93e-04 |
Token | . |
Token ID | 13 |
Feature activation | -1.06e-02 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.96e-02 |
Token | Once |
Token ID | 7454 |
Feature activation | -5.64e-02 |
Token | upon |
Token ID | 2402 |
Feature activation | -5.85e-02 |
Token | a |
Token ID | 257 |
Feature activation | +3.58e-03 |
Token | time |
Token ID | 640 |
Feature activation | -3.01e-02 |
Token | , |
Token ID | 11 |
Feature activation | +2.77e-02 |
Token | there |
Token ID | 612 |
Feature activation | -9.50e-02 |
Token | they |
Token ID | 484 |
Feature activation | +2.02e-02 |
Token | each |
Token ID | 1123 |
Feature activation | +5.74e-02 |
Token | had |
Token ID | 550 |
Feature activation | +1.99e-02 |
Token | another |
Token ID | 1194 |
Feature activation | +3.77e-02 |
Token | big |
Token ID | 1263 |
Feature activation | +3.28e-03 |
Token | piece |
Token ID | 3704 |
Feature activation | -2.90e-02 |
Token | . |
Token ID | 13 |
Feature activation | -7.18e-03 |
Token |
|
Token ID | 198 |
Feature activation | +2.09e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.72e-02 |
Token | The |
Token ID | 464 |
Feature activation | -5.23e-02 |
Token | girl |
Token ID | 2576 |
Feature activation | +4.04e-02 |
Token | shirt |
Token ID | 10147 |
Feature activation | -3.88e-02 |
Token | . |
Token ID | 13 |
Feature activation | -8.79e-03 |
Token | Her |
Token ID | 2332 |
Feature activation | -4.03e-02 |
Token | mom |
Token ID | 1995 |
Feature activation | +4.36e-02 |
Token | my |
Token ID | 1820 |
Feature activation | +2.85e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.10e-02 |
Token | it |
Token ID | 340 |
Feature activation | -6.01e-02 |
Token | was |
Token ID | 373 |
Feature activation | +1.89e-02 |
Token | okay |
Token ID | 8788 |
Feature activation | -2.96e-04 |
Token | and |
Token ID | 290 |
Feature activation | -2.57e-02 |
Token | they |
Token ID | 484 |
Feature activation | +3.18e-02 |
Token | not |
Token ID | 407 |
Feature activation | -1.40e-02 |
Token | been |
Token ID | 587 |
Feature activation | +2.27e-02 |
Token | so |
Token ID | 523 |
Feature activation | +5.39e-02 |
Token | small |
Token ID | 1402 |
Feature activation | +9.23e-03 |
Token | and |
Token ID | 290 |
Feature activation | -5.29e-03 |
Token | afraid |
Token ID | 7787 |
Feature activation | -4.91e-02 |
Token | . |
Token ID | 13 |
Feature activation | +5.40e-03 |
Token | He |
Token ID | 679 |
Feature activation | -2.36e-02 |
Token | wished |
Token ID | 16555 |
Feature activation | -4.33e-02 |
Token | he |
Token ID | 339 |
Feature activation | -5.13e-04 |
Token | had |
Token ID | 550 |
Feature activation | +3.43e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.65e-02 |
Token |
|
Token ID | 198 |
Feature activation | +1.24e-02 |
Token |
|
Token ID | 198 |
Feature activation | +9.14e-03 |
Token | " |
Token ID | 1 |
Feature activation | +5.40e-02 |
Token | Was |
Token ID | 16973 |
Feature activation | -5.00e-02 |
Token | n |
Token ID | 77 |
Feature activation | -1.11e-01 |
Token | 't |
Token ID | 470 |
Feature activation | -1.78e-02 |
Token | that |
Token ID | 326 |
Feature activation | -2.01e-02 |
Token | fun |
Token ID | 1257 |
Feature activation | +5.73e-02 |
Token | ?" |
Token ID | 1701 |
Feature activation | +4.21e-02 |
Token | the |
Token ID | 262 |
Feature activation | -1.83e-02 |
Token | face |
Token ID | 1986 |
Feature activation | +1.59e-02 |
Token | on |
Token ID | 319 |
Feature activation | -3.35e-02 |
Token | it |
Token ID | 340 |
Feature activation | -4.14e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.96e-02 |
Token | Spot |
Token ID | 15899 |
Feature activation | -1.07e-02 |
Token | thought |
Token ID | 1807 |
Feature activation | -1.14e-01 |
Token | it |
Token ID | 340 |
Feature activation | -6.30e-02 |
Token | was |
Token ID | 373 |
Feature activation | +1.66e-02 |
Token | interesting |
Token ID | 3499 |
Feature activation | -2.24e-02 |
Token | , |
Token ID | 11 |
Feature activation | +4.25e-02 |
Token | but |
Token ID | 475 |
Feature activation | -9.55e-02 |
Token | in |
Token ID | 287 |
Feature activation | +4.92e-02 |
Token | the |
Token ID | 262 |
Feature activation | +6.35e-03 |
Token | garden |
Token ID | 11376 |
Feature activation | +7.82e-04 |
Token | . |
Token ID | 13 |
Feature activation | +4.77e-03 |
Token | She |
Token ID | 1375 |
Feature activation | +8.46e-03 |
Token | thought |
Token ID | 1807 |
Feature activation | -1.06e-01 |
Token | it |
Token ID | 340 |
Feature activation | -5.27e-02 |
Token | was |
Token ID | 373 |
Feature activation | +2.68e-02 |
Token | so |
Token ID | 523 |
Feature activation | +5.27e-02 |
Token | lovely |
Token ID | 14081 |
Feature activation | +2.03e-02 |
Token | , |
Token ID | 11 |
Feature activation | +5.81e-02 |
Token | She |
Token ID | 1375 |
Feature activation | -4.62e-03 |
Token | had |
Token ID | 550 |
Feature activation | +5.94e-03 |
Token | a |
Token ID | 257 |
Feature activation | +2.20e-02 |
Token | birthday |
Token ID | 10955 |
Feature activation | +2.98e-03 |
Token | . |
Token ID | 13 |
Feature activation | -2.04e-02 |
Token | On |
Token ID | 1550 |
Feature activation | -1.13e-01 |
Token | her |
Token ID | 607 |
Feature activation | -2.87e-02 |
Token | birthday |
Token ID | 10955 |
Feature activation | -1.32e-02 |
Token | , |
Token ID | 11 |
Feature activation | +2.06e-02 |
Token | her |
Token ID | 607 |
Feature activation | -2.64e-02 |
Token | mom |
Token ID | 1995 |
Feature activation | +4.59e-02 |
Token | bug |
Token ID | 5434 |
Feature activation | -3.10e-02 |
Token | was |
Token ID | 373 |
Feature activation | -8.36e-03 |
Token | very |
Token ID | 845 |
Feature activation | +5.39e-02 |
Token | scared |
Token ID | 12008 |
Feature activation | -5.42e-02 |
Token | , |
Token ID | 11 |
Feature activation | +2.17e-02 |
Token | but |
Token ID | 475 |
Feature activation | -1.12e-01 |
Token | he |
Token ID | 339 |
Feature activation | -3.98e-02 |
Token | tried |
Token ID | 3088 |
Feature activation | -5.32e-02 |
Token | to |
Token ID | 284 |
Feature activation | -7.86e-02 |
Token | be |
Token ID | 307 |
Feature activation | +3.06e-02 |
Token | brave |
Token ID | 14802 |
Feature activation | +9.28e-03 |