Token | naughty |
Token ID | 45128 |
Feature activation | -1.59e-02 |
Token | cat |
Token ID | 3797 |
Feature activation | +7.91e-02 |
Token | had |
Token ID | 550 |
Feature activation | -9.13e-03 |
Token | out |
Token ID | 503 |
Feature activation | +2.78e-01 |
Token | sm |
Token ID | 5796 |
Feature activation | -2.10e-02 |
Token | arted |
Token ID | 19112 |
Feature activation | +5.75e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.01e-02 |
Token | bees |
Token ID | 17002 |
Feature activation | +4.22e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | learned |
Token ID | 4499 |
Feature activation | +9.55e-02 |
Token | a |
Token ID | 257 |
Feature activation | +7.59e-02 |
Token |
|
Token ID | 198 |
Feature activation | -1.01e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.02e-01 |
Token | The |
Token ID | 464 |
Feature activation | -2.05e-01 |
Token | fish |
Token ID | 5916 |
Feature activation | +1.20e-02 |
Token | d |
Token ID | 288 |
Feature activation | +5.73e-02 |
Token | arted |
Token ID | 19112 |
Feature activation | +5.75e-01 |
Token | away |
Token ID | 1497 |
Feature activation | +1.26e-01 |
Token | , |
Token ID | 11 |
Feature activation | -8.21e-02 |
Token | but |
Token ID | 475 |
Feature activation | +1.67e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.00e-02 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.52e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.01e-01 |
Token | M |
Token ID | 44 |
Feature activation | -9.87e-02 |
Token | add |
Token ID | 2860 |
Feature activation | +1.38e-01 |
Token | ie |
Token ID | 494 |
Feature activation | -1.40e-01 |
Token | per |
Token ID | 583 |
Feature activation | -4.96e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | up |
Token ID | 510 |
Feature activation | +2.71e-01 |
Token | when |
Token ID | 618 |
Feature activation | +1.68e-02 |
Token | she |
Token ID | 673 |
Feature activation | -1.04e-01 |
Token | heard |
Token ID | 2982 |
Feature activation | +9.76e-02 |
Token | this |
Token ID | 428 |
Feature activation | -2.44e-01 |
Token | a |
Token ID | 257 |
Feature activation | +7.53e-02 |
Token | lull |
Token ID | 42858 |
Feature activation | -4.34e-02 |
Token | aby |
Token ID | 3930 |
Feature activation | -1.10e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | stro |
Token ID | 9660 |
Feature activation | +3.38e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | their |
Token ID | 511 |
Feature activation | +3.42e-02 |
Token | hair |
Token ID | 4190 |
Feature activation | +2.72e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | Tom |
Token ID | 4186 |
Feature activation | +3.16e-03 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | car |
Token ID | 1097 |
Feature activation | +4.38e-02 |
Token | saw |
Token ID | 2497 |
Feature activation | +1.34e-01 |
Token | her |
Token ID | 607 |
Feature activation | +8.78e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | bra |
Token ID | 8290 |
Feature activation | +1.13e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.86e-01 |
Token | stopped |
Token ID | 5025 |
Feature activation | +2.09e-01 |
Token | just |
Token ID | 655 |
Feature activation | +3.73e-02 |
Token | in |
Token ID | 287 |
Feature activation | +1.52e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.05e-02 |
Token | trees |
Token ID | 7150 |
Feature activation | +1.90e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | Jack |
Token ID | 3619 |
Feature activation | +9.57e-02 |
Token | stro |
Token ID | 9660 |
Feature activation | +3.38e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.04e-02 |
Token | cat |
Token ID | 3797 |
Feature activation | +7.90e-02 |
Token | 's |
Token ID | 338 |
Feature activation | +6.98e-03 |
Token | back |
Token ID | 736 |
Feature activation | +3.14e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.02e-01 |
Token | The |
Token ID | 464 |
Feature activation | -2.05e-01 |
Token | plant |
Token ID | 4618 |
Feature activation | -8.25e-02 |
Token | instantly |
Token ID | 11101 |
Feature activation | +2.38e-01 |
Token | per |
Token ID | 583 |
Feature activation | -5.01e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | up |
Token ID | 510 |
Feature activation | +2.70e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | " |
Token ID | 366 |
Feature activation | +2.42e-01 |
Token | Thank |
Token ID | 10449 |
Feature activation | +6.03e-03 |
Token | you |
Token ID | 345 |
Feature activation | -6.25e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | kind |
Token ID | 1611 |
Feature activation | -3.07e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | She |
Token ID | 1375 |
Feature activation | +2.68e-01 |
Token | stro |
Token ID | 9660 |
Feature activation | +3.38e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.75e-01 |
Token | their |
Token ID | 511 |
Feature activation | +3.39e-02 |
Token | hair |
Token ID | 4190 |
Feature activation | +2.72e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | kissed |
Token ID | 27946 |
Feature activation | +1.17e-01 |
Token | their |
Token ID | 511 |
Feature activation | +3.39e-02 |
Token | Timothy |
Token ID | 22283 |
Feature activation | +7.27e-02 |
Token | smiled |
Token ID | 13541 |
Feature activation | +1.91e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | gently |
Token ID | 15165 |
Feature activation | +1.66e-01 |
Token | stro |
Token ID | 9660 |
Feature activation | +3.38e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.00e-02 |
Token | cat |
Token ID | 3797 |
Feature activation | +7.91e-02 |
Token | 's |
Token ID | 338 |
Feature activation | +7.49e-03 |
Token | head |
Token ID | 1182 |
Feature activation | +2.41e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | The |
Token ID | 383 |
Feature activation | +2.32e-01 |
Token | pigeon |
Token ID | 46387 |
Feature activation | -1.88e-02 |
Token | squ |
Token ID | 2809 |
Feature activation | +8.65e-02 |
Token | aw |
Token ID | 707 |
Feature activation | +1.03e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | bit |
Token ID | 1643 |
Feature activation | +5.99e-02 |
Token | her |
Token ID | 607 |
Feature activation | +8.84e-02 |
Token | finger |
Token ID | 7660 |
Feature activation | +1.12e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | ed |
Token ID | 276 |
Feature activation | +5.29e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | crashed |
Token ID | 14997 |
Feature activation | +8.83e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | hon |
Token ID | 3032 |
Feature activation | +6.06e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | But |
Token ID | 887 |
Feature activation | +5.92e-02 |
Token | they |
Token ID | 484 |
Feature activation | -1.59e-01 |
Token | did |
Token ID | 750 |
Feature activation | -6.21e-02 |
Token | not |
Token ID | 407 |
Feature activation | -1.82e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | his |
Token ID | 465 |
Feature activation | +1.71e-01 |
Token | room |
Token ID | 2119 |
Feature activation | +1.17e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | sul |
Token ID | 33154 |
Feature activation | +2.39e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.01e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.01e-01 |
Token | Then |
Token ID | 6423 |
Feature activation | -1.50e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.04e-02 |
Token | looked |
Token ID | 3114 |
Feature activation | +1.93e-01 |
Token | at |
Token ID | 379 |
Feature activation | +2.46e-01 |
Token | Sara |
Token ID | 24799 |
Feature activation | +1.04e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | smir |
Token ID | 42821 |
Feature activation | +2.17e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | He |
Token ID | 679 |
Feature activation | +3.49e-01 |
Token | said |
Token ID | 531 |
Feature activation | +2.02e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.24e-02 |
Token | " |
Token ID | 366 |
Feature activation | +2.41e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.09e-02 |
Token | room |
Token ID | 2119 |
Feature activation | +1.17e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | He |
Token ID | 679 |
Feature activation | +3.50e-01 |
Token | hon |
Token ID | 3032 |
Feature activation | +6.08e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.02e-02 |
Token | horn |
Token ID | 12718 |
Feature activation | +1.83e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | made |
Token ID | 925 |
Feature activation | +1.96e-02 |
Token | v |
Token ID | 410 |
Feature activation | +5.54e-02 |
Token | driver |
Token ID | 4639 |
Feature activation | +1.65e-01 |
Token | 's |
Token ID | 338 |
Feature activation | +6.81e-03 |
Token | seat |
Token ID | 5852 |
Feature activation | +1.40e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | hon |
Token ID | 3032 |
Feature activation | +6.07e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.07e-02 |
Token | horn |
Token ID | 12718 |
Feature activation | +1.83e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.33e-01 |
Token | Then |
Token ID | 3244 |
Feature activation | +3.08e-01 |
Token | he |
Token ID | 339 |
Feature activation | -4.93e-02 |
Token | all |
Token ID | 477 |
Feature activation | +1.53e-01 |
Token | qu |
Token ID | 627 |
Feature activation | +6.71e-02 |
Token | acked |
Token ID | 6021 |
Feature activation | +3.36e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | hon |
Token ID | 3032 |
Feature activation | +6.04e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | with |
Token ID | 351 |
Feature activation | +5.36e-02 |
Token | joy |
Token ID | 8716 |
Feature activation | +6.80e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | The |
Token ID | 383 |
Feature activation | +2.32e-01 |
Token | bunny |
Token ID | 44915 |
Feature activation | +1.39e-01 |
Token | a |
Token ID | 257 |
Feature activation | +7.52e-02 |
Token | fly |
Token ID | 6129 |
Feature activation | -1.53e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | He |
Token ID | 679 |
Feature activation | +3.49e-01 |
Token | d |
Token ID | 288 |
Feature activation | +5.73e-02 |
Token | arted |
Token ID | 19112 |
Feature activation | +5.74e-01 |
Token | over |
Token ID | 625 |
Feature activation | +3.14e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | look |
Token ID | 804 |
Feature activation | -4.34e-02 |
Token | more |
Token ID | 517 |
Feature activation | -3.09e-02 |
Token | closely |
Token ID | 7173 |
Feature activation | +1.71e-01 |
Token | F |
Token ID | 376 |
Feature activation | +4.47e-02 |
Token | ido |
Token ID | 17305 |
Feature activation | +5.53e-02 |
Token | 's |
Token ID | 338 |
Feature activation | +6.78e-03 |
Token | ears |
Token ID | 11368 |
Feature activation | +3.61e-01 |
Token | per |
Token ID | 583 |
Feature activation | -5.04e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | up |
Token ID | 510 |
Feature activation | +2.70e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | he |
Token ID | 339 |
Feature activation | -4.86e-02 |
Token | looked |
Token ID | 3114 |
Feature activation | +1.93e-01 |
Token | up |
Token ID | 510 |
Feature activation | +2.70e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | his |
Token ID | 465 |
Feature activation | +1.70e-01 |
Token | truck |
Token ID | 7779 |
Feature activation | +9.62e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.26e-02 |
Token | hon |
Token ID | 3032 |
Feature activation | +6.03e-02 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.07e-02 |
Token | horn |
Token ID | 12718 |
Feature activation | +1.83e-01 |
Token | as |
Token ID | 355 |
Feature activation | +2.23e-01 |
Token | loud |
Token ID | 7812 |
Feature activation | +8.47e-02 |
Token | as |
Token ID | 355 |
Feature activation | +2.22e-01 |
Token | stayed |
Token ID | 9658 |
Feature activation | +3.66e-02 |
Token | still |
Token ID | 991 |
Feature activation | +5.20e-03 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | Bob |
Token ID | 5811 |
Feature activation | +8.60e-02 |
Token | stro |
Token ID | 9660 |
Feature activation | +3.37e-01 |
Token | ked |
Token ID | 9091 |
Feature activation | +5.74e-01 |
Token | its |
Token ID | 663 |
Feature activation | +1.18e-01 |
Token | soft |
Token ID | 2705 |
Feature activation | +1.62e-01 |
Token | fur |
Token ID | 9230 |
Feature activation | +3.38e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | smiled |
Token ID | 13541 |
Feature activation | +1.91e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.76e-01 |
Token | imagining |
Token ID | 34140 |
Feature activation | +1.43e-01 |
Token | things |
Token ID | 1243 |
Feature activation | -1.67e-01 |
Token | can |
Token ID | 460 |
Feature activation | -2.60e-01 |
Token | help |
Token ID | 1037 |
Feature activation | -3.44e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | do |
Token ID | 466 |
Feature activation | -3.74e-01 |
Token | what |
Token ID | 644 |
Feature activation | -1.42e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.48e-01 |
Token | need |
Token ID | 761 |
Feature activation | -2.21e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.94e-01 |
Token | ok |
Token ID | 12876 |
Feature activation | -1.15e-01 |
Token | - |
Token ID | 532 |
Feature activation | -1.30e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.81e-01 |
Token | 'll |
Token ID | 1183 |
Feature activation | -1.19e-01 |
Token | let |
Token ID | 1309 |
Feature activation | -2.10e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | win |
Token ID | 1592 |
Feature activation | -2.69e-01 |
Token | next |
Token ID | 1306 |
Feature activation | -1.74e-01 |
Token | time |
Token ID | 640 |
Feature activation | -2.94e-01 |
Token | !" |
Token ID | 2474 |
Feature activation | -4.04e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.21e-01 |
Token | . |
Token ID | 13 |
Feature activation | -2.26e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.28e-01 |
Token | can |
Token ID | 460 |
Feature activation | -2.68e-01 |
Token | 't |
Token ID | 470 |
Feature activation | -4.43e-01 |
Token | hurt |
Token ID | 5938 |
Feature activation | -1.33e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.58e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.60e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -1.55e-02 |
Token | just |
Token ID | 655 |
Feature activation | +1.59e-02 |
Token | a |
Token ID | 257 |
Feature activation | +5.56e-02 |
Token | " |
Token ID | 366 |
Feature activation | +1.48e-01 |
Token | I |
Token ID | 40 |
Feature activation | -2.76e-01 |
Token | told |
Token ID | 1297 |
Feature activation | -2.69e-02 |
Token | you |
Token ID | 345 |
Feature activation | -6.61e-01 |
Token | that |
Token ID | 326 |
Feature activation | -2.21e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | would |
Token ID | 561 |
Feature activation | -1.16e-01 |
Token | regret |
Token ID | 13721 |
Feature activation | -1.18e-01 |
Token | it |
Token ID | 340 |
Feature activation | -2.40e-01 |
Token | ." |
Token ID | 526 |
Feature activation | -3.21e-01 |
Token | Jack |
Token ID | 3619 |
Feature activation | +7.58e-02 |
Token | Let |
Token ID | 3914 |
Feature activation | -1.13e-02 |
Token | 's |
Token ID | 338 |
Feature activation | -4.97e-02 |
Token | see |
Token ID | 766 |
Feature activation | -1.64e-01 |
Token | how |
Token ID | 703 |
Feature activation | +2.65e-02 |
Token | brave |
Token ID | 14802 |
Feature activation | -2.04e-02 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | can |
Token ID | 460 |
Feature activation | -2.51e-01 |
Token | be |
Token ID | 307 |
Feature activation | -2.48e-01 |
Token | !" |
Token ID | 2474 |
Feature activation | -4.05e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.23e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.21e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.76e-01 |
Token | Ben |
Token ID | 3932 |
Feature activation | -5.07e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.75e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.73e-01 |
Token | forgive |
Token ID | 20927 |
Feature activation | -3.01e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.58e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.63e-01 |
Token | love |
Token ID | 1842 |
Feature activation | -2.03e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.47e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.52e-01 |
Token | lunch |
Token ID | 9965 |
Feature activation | -2.03e-01 |
Token | ever |
Token ID | 1683 |
Feature activation | -3.03e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.75e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.73e-01 |
Token | love |
Token ID | 1842 |
Feature activation | -2.13e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | ." |
Token ID | 526 |
Feature activation | -3.27e-01 |
Token | Her |
Token ID | 2332 |
Feature activation | +2.06e-01 |
Token | mom |
Token ID | 1995 |
Feature activation | -5.42e-02 |
Token | said |
Token ID | 531 |
Feature activation | -8.52e-04 |
Token | , |
Token ID | 11 |
Feature activation | -1.02e-01 |
Token | She |
Token ID | 1375 |
Feature activation | +1.74e-01 |
Token | says |
Token ID | 1139 |
Feature activation | -1.04e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.25e-01 |
Token | " |
Token ID | 366 |
Feature activation | +2.06e-01 |
Token | Thank |
Token ID | 10449 |
Feature activation | -2.58e-02 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.09e-01 |
Token | Ben |
Token ID | 3932 |
Feature activation | -1.88e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.55e-01 |
Token | You |
Token ID | 921 |
Feature activation | -1.58e-01 |
Token | are |
Token ID | 389 |
Feature activation | -3.58e-01 |
Token | Mary |
Token ID | 24119 |
Feature activation | -2.08e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.39e-01 |
Token | why |
Token ID | 1521 |
Feature activation | -1.14e-01 |
Token | don |
Token ID | 836 |
Feature activation | -3.88e-01 |
Token | 't |
Token ID | 470 |
Feature activation | -4.40e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | go |
Token ID | 467 |
Feature activation | -2.00e-01 |
Token | outside |
Token ID | 2354 |
Feature activation | +1.96e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.40e-01 |
Token | play |
Token ID | 711 |
Feature activation | -1.53e-01 |
Token | ?" |
Token ID | 1701 |
Feature activation | -4.59e-01 |
Token | said |
Token ID | 531 |
Feature activation | -7.37e-02 |
Token | , |
Token ID | 11 |
Feature activation | -1.39e-01 |
Token | " |
Token ID | 366 |
Feature activation | +1.99e-01 |
Token | No |
Token ID | 2949 |
Feature activation | -2.52e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.14e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | can |
Token ID | 460 |
Feature activation | -2.51e-01 |
Token | 't |
Token ID | 470 |
Feature activation | -4.33e-01 |
Token | go |
Token ID | 467 |
Feature activation | -1.97e-01 |
Token | zoom |
Token ID | 19792 |
Feature activation | +2.92e-02 |
Token | in |
Token ID | 287 |
Feature activation | +1.32e-01 |
Token | " |
Token ID | 366 |
Feature activation | +1.48e-01 |
Token | It |
Token ID | 1026 |
Feature activation | -2.43e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -3.58e-02 |
Token | important |
Token ID | 1593 |
Feature activation | -1.00e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.14e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | have |
Token ID | 423 |
Feature activation | -3.36e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.89e-01 |
Token | !" |
Token ID | 2474 |
Feature activation | -4.05e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.23e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.21e-01 |
Token | have |
Token ID | 423 |
Feature activation | -4.03e-01 |
Token | big |
Token ID | 1263 |
Feature activation | -3.76e-02 |
Token | tasks |
Token ID | 8861 |
Feature activation | -2.09e-01 |
Token | ahead |
Token ID | 4058 |
Feature activation | +1.22e-02 |
Token | of |
Token ID | 286 |
Feature activation | -6.93e-02 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.09e-01 |
Token | it |
Token ID | 340 |
Feature activation | -2.43e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -1.55e-02 |
Token | okay |
Token ID | 8788 |
Feature activation | -1.01e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.94e-01 |
Token | You |
Token ID | 1639 |
Feature activation | -2.98e-01 |
Token | 'll |
Token ID | 1183 |
Feature activation | -1.40e-01 |
Token | never |
Token ID | 1239 |
Feature activation | -2.40e-01 |
Token | know |
Token ID | 760 |
Feature activation | -1.67e-01 |
Token | unless |
Token ID | 4556 |
Feature activation | -1.65e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | dare |
Token ID | 16498 |
Feature activation | -3.12e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.89e-01 |
Token | try |
Token ID | 1949 |
Feature activation | -2.11e-01 |
Token | ," |
Token ID | 553 |
Feature activation | -3.32e-01 |
Token | said |
Token ID | 531 |
Feature activation | +5.13e-04 |
Token | " |
Token ID | 1 |
Feature activation | -4.43e-01 |
Token | I |
Token ID | 40 |
Feature activation | -2.76e-01 |
Token | think |
Token ID | 892 |
Feature activation | -2.11e-01 |
Token | he |
Token ID | 339 |
Feature activation | -8.38e-02 |
Token | likes |
Token ID | 7832 |
Feature activation | +9.47e-03 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | ," |
Token ID | 553 |
Feature activation | -3.37e-01 |
Token | the |
Token ID | 262 |
Feature activation | -6.53e-02 |
Token | farmer |
Token ID | 18739 |
Feature activation | +7.33e-02 |
Token | says |
Token ID | 1139 |
Feature activation | -6.78e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.52e-01 |
Token | . |
Token ID | 13 |
Feature activation | -2.26e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.28e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -3.58e-02 |
Token | good |
Token ID | 922 |
Feature activation | -1.74e-01 |
Token | for |
Token ID | 329 |
Feature activation | +1.33e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | ." |
Token ID | 526 |
Feature activation | -3.27e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.26e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.24e-01 |
Token | L |
Token ID | 43 |
Feature activation | -2.16e-01 |
Token | ily |
Token ID | 813 |
Feature activation | -8.06e-02 |
Token | worry |
Token ID | 5490 |
Feature activation | -2.12e-01 |
Token | , |
Token ID | 11 |
Feature activation | -1.39e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.81e-01 |
Token | will |
Token ID | 481 |
Feature activation | -3.79e-01 |
Token | lead |
Token ID | 1085 |
Feature activation | -2.10e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.58e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.63e-01 |
Token | will |
Token ID | 481 |
Feature activation | -3.66e-01 |
Token | go |
Token ID | 467 |
Feature activation | -1.95e-01 |
Token | first |
Token ID | 717 |
Feature activation | -1.93e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -8.68e-02 |
Token | a |
Token ID | 257 |
Feature activation | +1.85e-02 |
Token | new |
Token ID | 649 |
Feature activation | -2.39e-02 |
Token | toy |
Token ID | 13373 |
Feature activation | +1.08e-01 |
Token | for |
Token ID | 329 |
Feature activation | +1.33e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.88e-01 |
Token | play |
Token ID | 711 |
Feature activation | -1.57e-01 |
Token | with |
Token ID | 351 |
Feature activation | +3.14e-02 |
Token | ," |
Token ID | 553 |
Feature activation | -3.32e-01 |
Token | mom |
Token ID | 1995 |
Feature activation | -5.14e-02 |
Token | tomorrow |
Token ID | 9439 |
Feature activation | -3.52e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.89e-01 |
Token | But |
Token ID | 887 |
Feature activation | +1.61e-02 |
Token | only |
Token ID | 691 |
Feature activation | -3.44e-02 |
Token | if |
Token ID | 611 |
Feature activation | -1.29e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | promise |
Token ID | 6991 |
Feature activation | -3.39e-01 |
Token | to |
Token ID | 284 |
Feature activation | +1.89e-01 |
Token | share |
Token ID | 2648 |
Feature activation | -2.71e-01 |
Token | it |
Token ID | 340 |
Feature activation | -2.39e-01 |
Token | nicely |
Token ID | 16576 |
Feature activation | -9.56e-02 |
Token | I |
Token ID | 314 |
Feature activation | -4.32e-01 |
Token | want |
Token ID | 765 |
Feature activation | -6.23e-02 |
Token | to |
Token ID | 284 |
Feature activation | +1.71e-01 |
Token | talk |
Token ID | 1561 |
Feature activation | -9.03e-02 |
Token | to |
Token ID | 284 |
Feature activation | +1.82e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.58e-01 |
Token | I |
Token ID | 314 |
Feature activation | -3.63e-01 |
Token | 'm |
Token ID | 1101 |
Feature activation | -2.52e-01 |
Token | sorry |
Token ID | 7926 |
Feature activation | -9.71e-03 |
Token | I |
Token ID | 314 |
Feature activation | -3.58e-01 |
Token | the |
Token ID | 262 |
Feature activation | -1.34e-01 |
Token | dog |
Token ID | 3290 |
Feature activation | +2.50e-02 |
Token | won |
Token ID | 1839 |
Feature activation | -9.12e-02 |
Token | 't |
Token ID | 470 |
Feature activation | -4.43e-01 |
Token | hurt |
Token ID | 5938 |
Feature activation | -1.33e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.54e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.58e-01 |
Token | He |
Token ID | 679 |
Feature activation | +3.24e-01 |
Token | 's |
Token ID | 338 |
Feature activation | -1.55e-02 |
Token | friendly |
Token ID | 8030 |
Feature activation | +1.40e-01 |
Token | ." |
Token ID | 526 |
Feature activation | -3.20e-01 |
Token | shoe |
Token ID | 17292 |
Feature activation | +6.55e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.39e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.79e-01 |
Token | was |
Token ID | 373 |
Feature activation | -6.31e-02 |
Token | blue |
Token ID | 4171 |
Feature activation | +2.05e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.56e-01 |
Token | shiny |
Token ID | 22441 |
Feature activation | +8.17e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.57e-01 |
Token | had |
Token ID | 550 |
Feature activation | -1.51e-02 |
Token | a |
Token ID | 257 |
Feature activation | +6.96e-02 |
Token | picture |
Token ID | 4286 |
Feature activation | -6.02e-02 |
Token | milk |
Token ID | 7545 |
Feature activation | +5.68e-02 |
Token | out |
Token ID | 503 |
Feature activation | +2.76e-01 |
Token | of |
Token ID | 286 |
Feature activation | -3.92e-02 |
Token | the |
Token ID | 262 |
Feature activation | -4.26e-02 |
Token | mug |
Token ID | 25152 |
Feature activation | +5.52e-02 |
Token | onto |
Token ID | 4291 |
Feature activation | +3.56e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.21e-02 |
Token | floor |
Token ID | 4314 |
Feature activation | +1.12e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.34e-01 |
Token | |
Token ID | 220 |
Feature activation | +3.51e-02 |
Token |
|
Token ID | 198 |
Feature activation | -1.03e-01 |
Token | stream |
Token ID | 4269 |
Feature activation | +3.54e-02 |
Token | ers |
Token ID | 364 |
Feature activation | +2.16e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.27e-01 |
Token | put |
Token ID | 1234 |
Feature activation | +1.66e-03 |
Token | spark |
Token ID | 9009 |
Feature activation | +2.17e-02 |
Token | ly |
Token ID | 306 |
Feature activation | +2.79e-01 |
Token | dots |
Token ID | 22969 |
Feature activation | +1.50e-01 |
Token | on |
Token ID | 319 |
Feature activation | +8.97e-02 |
Token | the |
Token ID | 262 |
Feature activation | -6.15e-02 |
Token | windows |
Token ID | 9168 |
Feature activation | +1.98e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.51e-01 |
Token | excited |
Token ID | 6568 |
Feature activation | +3.69e-01 |
Token | , |
Token ID | 11 |
Feature activation | -8.52e-02 |
Token | she |
Token ID | 673 |
Feature activation | -1.08e-01 |
Token | jumped |
Token ID | 11687 |
Feature activation | +1.27e-01 |
Token | up |
Token ID | 510 |
Feature activation | +2.68e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.60e-01 |
Token | down |
Token ID | 866 |
Feature activation | +4.43e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.35e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.83e-01 |
Token | was |
Token ID | 373 |
Feature activation | -5.88e-02 |
Token | her |
Token ID | 607 |
Feature activation | +8.59e-02 |
Token | rare |
Token ID | 4071 |
Feature activation | -3.44e-02 |
Token | cake |
Token ID | 12187 |
Feature activation | -1.66e-01 |
Token | with |
Token ID | 351 |
Feature activation | +5.39e-02 |
Token | her |
Token ID | 607 |
Feature activation | +8.78e-02 |
Token | friends |
Token ID | 2460 |
Feature activation | +3.43e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.62e-01 |
Token | she |
Token ID | 673 |
Feature activation | -1.05e-01 |
Token | couldn |
Token ID | 3521 |
Feature activation | +9.92e-02 |
Token | 't |
Token ID | 470 |
Feature activation | -4.08e-01 |
Token | wait |
Token ID | 4043 |
Feature activation | -3.45e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | wait |
Token ID | 4043 |
Feature activation | -3.46e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | play |
Token ID | 711 |
Feature activation | -1.32e-01 |
Token | with |
Token ID | 351 |
Feature activation | +5.41e-02 |
Token | her |
Token ID | 607 |
Feature activation | +8.82e-02 |
Token | t |
Token ID | 256 |
Feature activation | +2.07e-01 |
Token | eddy |
Token ID | 21874 |
Feature activation | +1.62e-01 |
Token | bear |
Token ID | 6842 |
Feature activation | +1.19e-01 |
Token | again |
Token ID | 757 |
Feature activation | +6.62e-02 |
Token | when |
Token ID | 618 |
Feature activation | +1.68e-02 |
Token | it |
Token ID | 340 |
Feature activation | -2.18e-01 |
Token | upon |
Token ID | 2402 |
Feature activation | +1.15e-01 |
Token | a |
Token ID | 257 |
Feature activation | +7.50e-02 |
Token | time |
Token ID | 640 |
Feature activation | -2.72e-01 |
Token | there |
Token ID | 612 |
Feature activation | -4.89e-02 |
Token | was |
Token ID | 373 |
Feature activation | -5.72e-02 |
Token | a |
Token ID | 257 |
Feature activation | +7.50e-02 |
Token | snowy |
Token ID | 46742 |
Feature activation | +9.27e-02 |
Token | day |
Token ID | 1110 |
Feature activation | -2.18e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.32e-01 |
Token | The |
Token ID | 383 |
Feature activation | +2.32e-01 |
Token | wind |
Token ID | 2344 |
Feature activation | +1.11e-01 |
Token | 't |
Token ID | 470 |
Feature activation | -4.13e-01 |
Token | listen |
Token ID | 6004 |
Feature activation | -2.63e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.09e-01 |
Token | his |
Token ID | 465 |
Feature activation | +1.67e-01 |
Token | mom |
Token ID | 1995 |
Feature activation | -3.65e-02 |
Token | my |
Token ID | 1820 |
Feature activation | +1.00e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.58e-01 |
Token | continued |
Token ID | 3767 |
Feature activation | +1.44e-01 |
Token | playing |
Token ID | 2712 |
Feature activation | +9.52e-02 |
Token | with |
Token ID | 351 |
Feature activation | +4.98e-02 |
Token | his |
Token ID | 465 |
Feature activation | +1.67e-01 |
Token | The |
Token ID | 464 |
Feature activation | -2.07e-01 |
Token | little |
Token ID | 1310 |
Feature activation | +8.33e-02 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.64e-01 |
Token | was |
Token ID | 373 |
Feature activation | -5.76e-02 |
Token | excited |
Token ID | 6568 |
Feature activation | +3.70e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.12e-01 |
Token | see |
Token ID | 766 |
Feature activation | -1.22e-01 |
Token | a |
Token ID | 257 |
Feature activation | +7.40e-02 |
Token | new |
Token ID | 649 |
Feature activation | +1.80e-02 |
Token | animal |
Token ID | 5044 |
Feature activation | +5.68e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.33e-01 |
Token | After |
Token ID | 2293 |
Feature activation | +2.77e-01 |
Token | a |
Token ID | 257 |
Feature activation | +7.23e-02 |
Token | few |
Token ID | 1178 |
Feature activation | +7.08e-02 |
Token | tries |
Token ID | 8404 |
Feature activation | +4.82e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.50e-02 |
Token | he |
Token ID | 339 |
Feature activation | -5.13e-02 |
Token | finally |
Token ID | 3443 |
Feature activation | +2.90e-02 |
Token | unlocked |
Token ID | 14838 |
Feature activation | +1.91e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.26e-02 |
Token | box |
Token ID | 3091 |
Feature activation | -2.64e-02 |
Token | ! |
Token ID | 0 |
Feature activation | -2.06e-01 |
Token | your |
Token ID | 534 |
Feature activation | -3.07e-01 |
Token | dreams |
Token ID | 10625 |
Feature activation | -3.68e-02 |
Token | , |
Token ID | 11 |
Feature activation | -9.05e-02 |
Token | and |
Token ID | 290 |
Feature activation | +2.54e-01 |
Token | maybe |
Token ID | 3863 |
Feature activation | -1.23e-01 |
Token | it |
Token ID | 340 |
Feature activation | -2.26e-01 |
Token | will |
Token ID | 481 |
Feature activation | -3.52e-01 |
Token | smile |
Token ID | 8212 |
Feature activation | +1.27e-01 |
Token | at |
Token ID | 379 |
Feature activation | +2.38e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.33e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.40e-01 |
Token | Lily |
Token ID | 20037 |
Feature activation | +4.05e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.37e-01 |
Token | We |
Token ID | 775 |
Feature activation | -1.71e-01 |
Token | can |
Token ID | 460 |
Feature activation | -2.30e-01 |
Token | fix |
Token ID | 4259 |
Feature activation | -8.83e-02 |
Token | your |
Token ID | 534 |
Feature activation | -3.03e-01 |
Token | toy |
Token ID | 13373 |
Feature activation | +1.38e-01 |
Token | car |
Token ID | 1097 |
Feature activation | +3.92e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.36e-01 |
Token | Let |
Token ID | 3914 |
Feature activation | +7.81e-02 |
Token | 's |
Token ID | 338 |
Feature activation | +2.94e-03 |
Token | remove |
Token ID | 4781 |
Feature activation | -1.06e-01 |
Token | his |
Token ID | 465 |
Feature activation | +1.60e-01 |
Token | sun |
Token ID | 4252 |
Feature activation | +4.63e-02 |
Token | fl |
Token ID | 2704 |
Feature activation | +1.23e-01 |
Token | owers |
Token ID | 3618 |
Feature activation | -5.01e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.43e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.12e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.12e-01 |
Token | The |
Token ID | 464 |
Feature activation | -2.15e-01 |
Token | moral |
Token ID | 6573 |
Feature activation | -2.20e-01 |
Token | of |
Token ID | 286 |
Feature activation | -4.72e-02 |
Token | truck |
Token ID | 7779 |
Feature activation | +9.66e-02 |
Token | felt |
Token ID | 2936 |
Feature activation | +2.37e-01 |
Token | sad |
Token ID | 6507 |
Feature activation | +2.41e-01 |
Token | for |
Token ID | 329 |
Feature activation | +1.65e-01 |
Token | disturbing |
Token ID | 14851 |
Feature activation | -1.42e-02 |
Token | them |
Token ID | 606 |
Feature activation | -1.04e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.63e-01 |
Token | agreed |
Token ID | 4987 |
Feature activation | +1.96e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.14e-01 |
Token | be |
Token ID | 307 |
Feature activation | -2.23e-01 |
Token | quiet |
Token ID | 5897 |
Feature activation | +4.63e-02 |
Token | ," |
Token ID | 553 |
Feature activation | -3.13e-01 |
Token | Lily |
Token ID | 20037 |
Feature activation | +4.28e-02 |
Token | says |
Token ID | 1139 |
Feature activation | -4.98e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.35e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.04e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.04e-01 |
Token | Dad |
Token ID | 46270 |
Feature activation | -1.68e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.60e-01 |
Token | Lily |
Token ID | 20037 |
Feature activation | +4.33e-02 |
Token | go |
Token ID | 467 |
Feature activation | -1.76e-01 |
Token | to |
Token ID | 284 |
Feature activation | +2.12e-01 |
Token | and |
Token ID | 290 |
Feature activation | +2.59e-01 |
Token | said |
Token ID | 531 |
Feature activation | +1.71e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.55e-02 |
Token | " |
Token ID | 366 |
Feature activation | +2.39e-01 |
Token | Thank |
Token ID | 10449 |
Feature activation | +3.03e-03 |
Token | you |
Token ID | 345 |
Feature activation | -6.28e-01 |
Token | for |
Token ID | 329 |
Feature activation | +1.62e-01 |
Token | the |
Token ID | 262 |
Feature activation | -4.31e-02 |
Token | fruit |
Token ID | 8234 |
Feature activation | +1.96e-02 |
Token | . |
Token ID | 13 |
Feature activation | -1.35e-01 |
Token | It |
Token ID | 632 |
Feature activation | +1.83e-01 |
Token | let |
Token ID | 1309 |
Feature activation | -1.82e-01 |
Token | me |
Token ID | 502 |
Feature activation | -1.57e-01 |
Token | have |
Token ID | 423 |
Feature activation | -3.13e-01 |
Token | your |
Token ID | 534 |
Feature activation | -3.01e-01 |
Token | bicycle |
Token ID | 17026 |
Feature activation | +1.10e-01 |
Token | ?" |
Token ID | 1701 |
Feature activation | -4.43e-01 |
Token | Tom |
Token ID | 4186 |
Feature activation | +2.30e-04 |
Token | says |
Token ID | 1139 |
Feature activation | -4.99e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.55e-02 |
Token | " |
Token ID | 366 |
Feature activation | +2.38e-01 |
Token | Yes |
Token ID | 5297 |
Feature activation | -1.18e-01 |
Token | he |
Token ID | 339 |
Feature activation | -5.08e-02 |
Token | shouted |
Token ID | 17293 |
Feature activation | +9.01e-02 |
Token | " |
Token ID | 366 |
Feature activation | +2.39e-01 |
Token | I |
Token ID | 40 |
Feature activation | -2.21e-01 |
Token | hate |
Token ID | 5465 |
Feature activation | -1.61e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.28e-01 |
Token | !" |
Token ID | 2474 |
Feature activation | -3.85e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.04e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.03e-01 |
Token | But |
Token ID | 1537 |
Feature activation | -2.34e-01 |
Token | then |
Token ID | 788 |
Feature activation | +2.41e-02 |
Token | ? |
Token ID | 30 |
Feature activation | -4.49e-01 |
Token | What |
Token ID | 1867 |
Feature activation | +1.03e-01 |
Token | does |
Token ID | 857 |
Feature activation | -1.16e-01 |
Token | that |
Token ID | 326 |
Feature activation | -1.89e-01 |
Token | mean |
Token ID | 1612 |
Feature activation | -7.33e-02 |
Token | ?" |
Token ID | 1701 |
Feature activation | -4.40e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.02e-01 |
Token |
|
Token ID | 198 |
Feature activation | -1.02e-01 |
Token | Mom |
Token ID | 29252 |
Feature activation | -2.14e-01 |
Token | said |
Token ID | 531 |
Feature activation | +1.95e-02 |
Token | , |
Token ID | 11 |
Feature activation | -8.27e-02 |
Token | mouse |
Token ID | 10211 |
Feature activation | +1.34e-01 |
Token | . |
Token ID | 13 |
Feature activation | -1.43e-01 |
Token | Who |
Token ID | 5338 |
Feature activation | -3.68e-02 |
Token | are |
Token ID | 389 |
Feature activation | -3.49e-01 |
Token | you |
Token ID | 345 |
Feature activation | -6.36e-01 |
Token | ?" |
Token ID | 1701 |
Feature activation | -4.50e-01 |
Token | the |
Token ID | 262 |
Feature activation | -5.07e-02 |
Token | mouse |
Token ID | 10211 |
Feature activation | +1.36e-01 |
Token | said |
Token ID | 531 |
Feature activation | +1.01e-02 |
Token | , |
Token ID | 11 |
Feature activation | -9.21e-02 |
Token | as |
Token ID | 355 |
Feature activation | +2.13e-01 |