Token | bie |
Token ID | 12590 |
Feature activation | +9.00e-01 |
Pos logit contributions | |
ole | +0.022 |
roc | +0.020 |
Jak | +0.020 |
idea | +0.019 |
ny | +0.019 |
Neg logit contributions | |
doorstep | -0.023 |
salon | -0.022 |
Sweet | -0.021 |
arrest | -0.020 |
uddled | -0.020 |
Token | got |
Token ID | 1392 |
Feature activation | -2.57e-01 |
Pos logit contributions | |
sidelines | +0.028 |
salon | +0.028 |
doorstep | +0.028 |
Sweet | +0.027 |
arrest | +0.027 |
Neg logit contributions | |
ole | -0.020 |
ny | -0.018 |
roc | -0.016 |
Jak | -0.015 |
idea | -0.015 |
Token | up |
Token ID | 510 |
Feature activation | +1.65e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
zo | +0.006 |
suit | +0.006 |
Neg logit contributions | |
Sweet | -0.006 |
sidelines | -0.006 |
salon | -0.006 |
arrest | -0.006 |
freshly | -0.006 |
Token | on |
Token ID | 319 |
Feature activation | -2.45e-01 |
Pos logit contributions | |
salon | +0.042 |
arrest | +0.042 |
doorstep | +0.042 |
uddled | +0.041 |
Sweet | +0.041 |
Neg logit contributions | |
ole | -0.041 |
roc | -0.040 |
ny | -0.039 |
Jak | -0.035 |
ac | -0.035 |
Token | his |
Token ID | 465 |
Feature activation | -1.64e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
ac | +0.004 |
suit | +0.004 |
Neg logit contributions | |
arrest | -0.009 |
doorstep | -0.008 |
sidelines | -0.008 |
salon | -0.008 |
couch | -0.008 |
Token | hind |
Token ID | 16222 |
Feature activation | +5.04e+00 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
idea | +0.003 |
boy | +0.003 |
roc | +0.003 |
Neg logit contributions | |
Sweet | -0.005 |
sidelines | -0.005 |
salon | -0.005 |
uddled | -0.005 |
Acc | -0.005 |
Token | legs |
Token ID | 7405 |
Feature activation | +3.82e-01 |
Pos logit contributions | |
doorstep | +0.133 |
sidelines | +0.133 |
arrest | +0.131 |
salon | +0.131 |
Sweet | +0.128 |
Neg logit contributions | |
ole | -0.132 |
ny | -0.122 |
roc | -0.115 |
Jak | -0.108 |
idea | -0.106 |
Token | and |
Token ID | 290 |
Feature activation | -1.21e+00 |
Pos logit contributions | |
doorstep | +0.012 |
salon | +0.012 |
sidelines | +0.012 |
arrest | +0.011 |
Sweet | +0.011 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | asked |
Token ID | 1965 |
Feature activation | -5.19e-01 |
Pos logit contributions | |
ole | +0.034 |
ny | +0.030 |
roc | +0.030 |
idea | +0.027 |
Jak | +0.027 |
Neg logit contributions | |
doorstep | -0.031 |
salon | -0.030 |
sidelines | -0.030 |
arrest | -0.030 |
Sweet | -0.029 |
Token | the |
Token ID | 262 |
Feature activation | -7.24e-01 |
Pos logit contributions | |
ole | +0.013 |
roc | +0.012 |
ny | +0.012 |
suit | +0.011 |
Jak | +0.011 |
Neg logit contributions | |
Sweet | -0.013 |
Her | -0.013 |
Her | -0.013 |
couch | -0.012 |
doorstep | -0.012 |
Token | peanut |
Token ID | 26636 |
Feature activation | -1.92e+00 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.018 |
Jak | +0.016 |
roc | +0.016 |
boat | +0.015 |
Neg logit contributions | |
doorstep | -0.019 |
salon | -0.019 |
sidelines | -0.019 |
arrest | -0.019 |
Sweet | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | +2.42e-01 |
Pos logit contributions | |
doorstep | +0.022 |
salon | +0.022 |
sidelines | +0.021 |
arrest | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.017 |
roc | -0.015 |
ny | -0.015 |
Jak | -0.014 |
idea | -0.013 |
Token | It |
Token ID | 632 |
Feature activation | +4.90e-01 |
Pos logit contributions | |
sidelines | +0.005 |
arrest | +0.005 |
doorstep | +0.005 |
salon | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.007 |
ac | -0.006 |
idea | -0.006 |
Token | stood |
Token ID | 6204 |
Feature activation | -3.99e-01 |
Pos logit contributions | |
doorstep | +0.014 |
sidelines | +0.014 |
salon | +0.014 |
Sweet | +0.014 |
arrest | +0.014 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.010 |
Jak | -0.009 |
idea | -0.009 |
Token | on |
Token ID | 319 |
Feature activation | -2.54e-01 |
Pos logit contributions | |
ole | +0.011 |
roc | +0.011 |
ny | +0.011 |
guys | +0.011 |
idea | +0.010 |
Neg logit contributions | |
Acc | -0.008 |
balcony | -0.008 |
sidelines | -0.008 |
doorstep | -0.008 |
uddled | -0.008 |
Token | its |
Token ID | 663 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.005 |
roc | +0.005 |
ac | +0.005 |
suit | +0.004 |
Neg logit contributions | |
arrest | -0.008 |
sidelines | -0.008 |
doorstep | -0.007 |
couch | -0.007 |
freshly | -0.007 |
Token | hind |
Token ID | 16222 |
Feature activation | +5.01e+00 |
Pos logit contributions | |
salon | +0.004 |
Sweet | +0.004 |
uddled | +0.004 |
Acc | +0.004 |
sidelines | +0.004 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.003 |
idea | -0.003 |
roc | -0.003 |
boy | -0.002 |
Token | legs |
Token ID | 7405 |
Feature activation | -5.64e-02 |
Pos logit contributions | |
salon | +0.168 |
Sweet | +0.160 |
uddled | +0.159 |
sidelines | +0.159 |
doorstep | +0.158 |
Neg logit contributions | |
ole | -0.104 |
ny | -0.093 |
roc | -0.085 |
idea | -0.080 |
Jak | -0.079 |
Token | and |
Token ID | 290 |
Feature activation | -8.54e-01 |
Pos logit contributions | |
ole | +0.001 |
roc | +0.001 |
ny | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
doorstep | -0.002 |
salon | -0.002 |
sidelines | -0.002 |
arrest | -0.002 |
Sweet | -0.002 |
Token | started |
Token ID | 2067 |
Feature activation | +1.81e-01 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.019 |
roc | +0.019 |
idea | +0.017 |
Jak | +0.017 |
Neg logit contributions | |
sidelines | -0.024 |
doorstep | -0.023 |
salon | -0.023 |
arrest | -0.023 |
Sweet | -0.023 |
Token | to |
Token ID | 284 |
Feature activation | -3.77e-01 |
Pos logit contributions | |
doorstep | +0.006 |
salon | +0.006 |
sidelines | +0.006 |
Sweet | +0.006 |
uddled | +0.005 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token | stretch |
Token ID | 7539 |
Feature activation | -6.73e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.009 |
idea | +0.008 |
roc | +0.008 |
Jak | +0.008 |
Neg logit contributions | |
uddled | -0.010 |
salon | -0.010 |
Sweet | -0.010 |
doorstep | -0.010 |
sidelines | -0.010 |
Token | i |
Token ID | 72 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
sidelines | +0.149 |
salon | +0.147 |
Sweet | +0.146 |
doorstep | +0.146 |
arrest | +0.142 |
Neg logit contributions | |
ole | -0.102 |
ny | -0.093 |
roc | -0.085 |
Jak | -0.080 |
idea | -0.077 |
Token | . |
Token ID | 13 |
Feature activation | -5.42e-01 |
Pos logit contributions | |
arrest | +0.004 |
salon | +0.003 |
doorstep | +0.003 |
sidelines | +0.003 |
Sweet | +0.003 |
Neg logit contributions | |
ole | -0.003 |
roc | -0.003 |
ny | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token |
|
Token ID | 198 |
Feature activation | +2.91e-01 |
Pos logit contributions | |
ole | +0.016 |
roc | +0.014 |
ny | +0.013 |
ac | +0.013 |
idea | +0.013 |
Neg logit contributions | |
doorstep | -0.013 |
Sweet | -0.012 |
uddled | -0.012 |
sidelines | -0.012 |
salon | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | +3.73e-01 |
Pos logit contributions | |
sidelines | +0.008 |
salon | +0.008 |
doorstep | +0.008 |
Sweet | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | B |
Token ID | 33 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
Sweet | +0.010 |
sidelines | +0.010 |
doorstep | +0.010 |
uddled | +0.010 |
salon | +0.010 |
Neg logit contributions | |
ole | -0.009 |
roc | -0.008 |
ny | -0.008 |
Jak | -0.008 |
idea | -0.008 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.99e+00 |
Pos logit contributions | |
doorstep | +0.024 |
Sweet | +0.024 |
salon | +0.024 |
sidelines | +0.024 |
arrest | +0.023 |
Neg logit contributions | |
ole | -0.037 |
ny | -0.034 |
roc | -0.033 |
Jak | -0.032 |
idea | -0.032 |
Token | i |
Token ID | 72 |
Feature activation | -2.84e-01 |
Pos logit contributions | |
sidelines | +0.171 |
salon | +0.165 |
Sweet | +0.165 |
doorstep | +0.164 |
arrest | +0.163 |
Neg logit contributions | |
ole | -0.098 |
ny | -0.088 |
roc | -0.080 |
Jak | -0.072 |
idea | -0.071 |
Token | ran |
Token ID | 4966 |
Feature activation | +7.85e-01 |
Pos logit contributions | |
ole | +0.008 |
Jak | +0.007 |
idea | +0.007 |
roc | +0.007 |
ac | +0.006 |
Neg logit contributions | |
uddled | -0.007 |
arrest | -0.006 |
doorstep | -0.006 |
salon | -0.006 |
Sweet | -0.006 |
Token | as |
Token ID | 355 |
Feature activation | +4.50e-01 |
Pos logit contributions | |
sidelines | +0.021 |
arrest | +0.021 |
doorstep | +0.021 |
salon | +0.019 |
Sweet | +0.019 |
Neg logit contributions | |
ole | -0.020 |
roc | -0.018 |
ac | -0.017 |
Jak | -0.017 |
idea | -0.016 |
Token | fast |
Token ID | 3049 |
Feature activation | +1.66e-01 |
Pos logit contributions | |
arrest | +0.012 |
sidelines | +0.012 |
Sweet | +0.012 |
salon | +0.011 |
Acc | +0.011 |
Neg logit contributions | |
ole | -0.011 |
roc | -0.010 |
idea | -0.010 |
Jak | -0.009 |
ny | -0.009 |
Token | as |
Token ID | 355 |
Feature activation | -7.76e-01 |
Pos logit contributions | |
arrest | +0.004 |
sidelines | +0.004 |
salon | +0.004 |
doorstep | +0.004 |
Sweet | +0.004 |
Neg logit contributions | |
ole | -0.004 |
roc | -0.004 |
ny | -0.004 |
Jak | -0.004 |
idea | -0.004 |
Token | ens |
Token ID | 641 |
Feature activation | +3.28e-01 |
Pos logit contributions | |
Sweet | +0.032 |
sidelines | +0.031 |
uddled | +0.031 |
salon | +0.031 |
arrest | +0.030 |
Neg logit contributions | |
ole | -0.035 |
roc | -0.031 |
ny | -0.030 |
idea | -0.029 |
Jak | -0.029 |
Token | stood |
Token ID | 6204 |
Feature activation | -2.65e-01 |
Pos logit contributions | |
sidelines | +0.010 |
salon | +0.010 |
Sweet | +0.009 |
doorstep | +0.009 |
arrest | +0.009 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | up |
Token ID | 510 |
Feature activation | +1.11e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.007 |
doorstep | -0.007 |
salon | -0.007 |
Sweet | -0.007 |
balcony | -0.007 |
Token | on |
Token ID | 319 |
Feature activation | -4.52e-01 |
Pos logit contributions | |
sidelines | +0.032 |
salon | +0.032 |
doorstep | +0.031 |
arrest | +0.031 |
Sweet | +0.030 |
Neg logit contributions | |
ole | -0.026 |
roc | -0.024 |
ny | -0.024 |
Jak | -0.022 |
idea | -0.021 |
Token | his |
Token ID | 465 |
Feature activation | -3.62e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
roc | +0.009 |
ac | +0.008 |
beard | +0.008 |
Neg logit contributions | |
arrest | -0.014 |
sidelines | -0.014 |
doorstep | -0.014 |
salon | -0.013 |
Sweet | -0.013 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.96e+00 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
idea | +0.009 |
boy | +0.008 |
gear | +0.008 |
Neg logit contributions | |
Sweet | -0.010 |
uddled | -0.009 |
Acc | -0.009 |
salon | -0.009 |
sidelines | -0.008 |
Token | legs |
Token ID | 7405 |
Feature activation | -3.26e-01 |
Pos logit contributions | |
salon | +0.167 |
Sweet | +0.161 |
uddled | +0.159 |
sidelines | +0.158 |
Acc | +0.158 |
Neg logit contributions | |
ole | -0.101 |
ny | -0.092 |
roc | -0.082 |
idea | -0.078 |
Jak | -0.076 |
Token | and |
Token ID | 290 |
Feature activation | -9.54e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.010 |
salon | -0.010 |
doorstep | -0.010 |
Sweet | -0.010 |
arrest | -0.010 |
Token | brave |
Token ID | 14802 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
ole | +0.024 |
ny | +0.022 |
roc | +0.020 |
Jak | +0.019 |
idea | +0.019 |
Neg logit contributions | |
sidelines | -0.027 |
salon | -0.027 |
doorstep | -0.027 |
Sweet | -0.026 |
arrest | -0.026 |
Token | ly |
Token ID | 306 |
Feature activation | -1.12e+00 |
Pos logit contributions | |
arrest | +0.005 |
sidelines | +0.005 |
doorstep | +0.005 |
Sweet | +0.005 |
Her | +0.004 |
Neg logit contributions | |
ole | -0.007 |
roc | -0.006 |
ac | -0.006 |
Jak | -0.006 |
boat | -0.006 |
Token | faced |
Token ID | 7452 |
Feature activation | -1.14e+00 |
Pos logit contributions | |
ole | +0.031 |
roc | +0.029 |
ny | +0.027 |
idea | +0.025 |
Jak | +0.025 |
Neg logit contributions | |
arrest | -0.028 |
doorstep | -0.028 |
sidelines | -0.027 |
salon | -0.027 |
uddled | -0.027 |
Token | where |
Token ID | 810 |
Feature activation | +2.03e-01 |
Pos logit contributions | |
ole | +0.004 |
roc | +0.004 |
ny | +0.003 |
Jak | +0.003 |
beard | +0.003 |
Neg logit contributions | |
salon | -0.004 |
doorstep | -0.004 |
sidelines | -0.003 |
arrest | -0.003 |
uddled | -0.003 |
Token | he |
Token ID | 339 |
Feature activation | +6.91e-02 |
Pos logit contributions | |
arrest | +0.006 |
salon | +0.006 |
Sweet | +0.006 |
sidelines | +0.006 |
doorstep | +0.006 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.004 |
roc | -0.004 |
idea | -0.004 |
Jak | -0.004 |
Token | stood |
Token ID | 6204 |
Feature activation | -9.37e-01 |
Pos logit contributions | |
sidelines | +0.002 |
salon | +0.002 |
doorstep | +0.002 |
Sweet | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
idea | -0.001 |
Token | on |
Token ID | 319 |
Feature activation | -1.68e-01 |
Pos logit contributions | |
roc | +0.026 |
ole | +0.025 |
ny | +0.025 |
Jak | +0.022 |
guys | +0.021 |
Neg logit contributions | |
balcony | -0.019 |
freshly | -0.019 |
Acc | -0.019 |
salon | -0.019 |
She | -0.018 |
Token | his |
Token ID | 465 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.003 |
roc | +0.003 |
ac | +0.003 |
Jak | +0.003 |
Neg logit contributions | |
arrest | -0.005 |
sidelines | -0.005 |
doorstep | -0.005 |
couch | -0.005 |
salon | -0.005 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.92e+00 |
Pos logit contributions | |
Sweet | +0.001 |
salon | +0.001 |
sidelines | +0.001 |
uddled | +0.001 |
arrest | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
idea | -0.001 |
Jak | -0.001 |
Token | legs |
Token ID | 7405 |
Feature activation | +2.51e-01 |
Pos logit contributions | |
salon | +0.139 |
sidelines | +0.139 |
doorstep | +0.138 |
Sweet | +0.136 |
arrest | +0.133 |
Neg logit contributions | |
ole | -0.120 |
ny | -0.110 |
roc | -0.103 |
Jak | -0.098 |
idea | -0.096 |
Token | and |
Token ID | 290 |
Feature activation | -9.62e-01 |
Pos logit contributions | |
salon | +0.007 |
doorstep | +0.007 |
sidelines | +0.007 |
arrest | +0.006 |
balcony | +0.006 |
Neg logit contributions | |
ole | -0.006 |
roc | -0.006 |
ny | -0.005 |
Jak | -0.005 |
idea | -0.005 |
Token | spun |
Token ID | 26843 |
Feature activation | -4.28e-01 |
Pos logit contributions | |
ole | +0.025 |
ny | +0.022 |
roc | +0.021 |
idea | +0.019 |
Jak | +0.019 |
Neg logit contributions | |
sidelines | -0.027 |
salon | -0.027 |
doorstep | -0.027 |
arrest | -0.026 |
Sweet | -0.026 |
Token | around |
Token ID | 1088 |
Feature activation | +1.44e+00 |
Pos logit contributions | |
ole | +0.012 |
roc | +0.012 |
ac | +0.010 |
Jak | +0.010 |
boat | +0.010 |
Neg logit contributions | |
sidelines | -0.009 |
couch | -0.009 |
arrest | -0.008 |
swings | -0.008 |
chair | -0.008 |
Token | in |
Token ID | 287 |
Feature activation | -4.41e-01 |
Pos logit contributions | |
sidelines | +0.038 |
arrest | +0.038 |
doorstep | +0.036 |
couch | +0.035 |
balcony | +0.035 |
Neg logit contributions | |
ole | -0.038 |
roc | -0.035 |
ny | -0.034 |
Jak | -0.032 |
ac | -0.031 |
Token | from |
Token ID | 422 |
Feature activation | -9.31e-02 |
Pos logit contributions | |
ole | +0.000 |
roc | +0.000 |
ny | +0.000 |
Jak | +0.000 |
idea | +0.000 |
Neg logit contributions | |
arrest | -0.000 |
doorstep | -0.000 |
sidelines | -0.000 |
salon | -0.000 |
Sweet | -0.000 |
Token | the |
Token ID | 262 |
Feature activation | +1.30e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
ac | +0.002 |
idea | +0.002 |
Neg logit contributions | |
arrest | -0.003 |
sidelines | -0.003 |
freshly | -0.003 |
doorstep | -0.003 |
salon | -0.003 |
Token | trap |
Token ID | 12840 |
Feature activation | -5.02e-01 |
Pos logit contributions | |
sidelines | +0.004 |
doorstep | +0.004 |
arrest | +0.004 |
salon | +0.003 |
Sweet | +0.003 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.003 |
ac | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | -2.11e-01 |
Pos logit contributions | |
ole | +0.012 |
roc | +0.011 |
ny | +0.011 |
Jak | +0.010 |
ac | +0.010 |
Neg logit contributions | |
doorstep | -0.014 |
arrest | -0.014 |
sidelines | -0.014 |
salon | -0.013 |
uddled | -0.013 |
Token | B |
Token ID | 347 |
Feature activation | +7.88e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
salon | -0.006 |
sidelines | -0.006 |
doorstep | -0.006 |
Sweet | -0.006 |
arrest | -0.006 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.90e+00 |
Pos logit contributions | |
doorstep | +0.016 |
sidelines | +0.016 |
Sweet | +0.016 |
salon | +0.016 |
arrest | +0.015 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.024 |
roc | -0.023 |
Jak | -0.022 |
idea | -0.022 |
Token | i |
Token ID | 72 |
Feature activation | -8.85e-02 |
Pos logit contributions | |
sidelines | +0.167 |
salon | +0.164 |
doorstep | +0.163 |
Sweet | +0.161 |
arrest | +0.159 |
Neg logit contributions | |
ole | -0.096 |
ny | -0.086 |
roc | -0.079 |
Jak | -0.072 |
ac | -0.068 |
Token | was |
Token ID | 373 |
Feature activation | +5.17e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
doorstep | -0.003 |
arrest | -0.003 |
sidelines | -0.003 |
salon | -0.003 |
Sweet | -0.003 |
Token | very |
Token ID | 845 |
Feature activation | +2.29e+00 |
Pos logit contributions | |
arrest | +0.013 |
Sweet | +0.013 |
doorstep | +0.013 |
sidelines | +0.013 |
hugging | +0.013 |
Neg logit contributions | |
ole | -0.013 |
roc | -0.013 |
ny | -0.012 |
guys | -0.011 |
Jak | -0.011 |
Token | happy |
Token ID | 3772 |
Feature activation | +8.20e-01 |
Pos logit contributions | |
Her | +0.046 |
Sweet | +0.045 |
hugging | +0.044 |
sidelines | +0.044 |
arrest | +0.043 |
Neg logit contributions | |
ny | -0.067 |
roc | -0.066 |
ole | -0.066 |
guys | -0.066 |
Jak | -0.062 |
Token | and |
Token ID | 290 |
Feature activation | -1.36e-01 |
Pos logit contributions | |
doorstep | +0.024 |
sidelines | +0.024 |
salon | +0.024 |
arrest | +0.024 |
Sweet | +0.024 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.018 |
roc | -0.017 |
Jak | -0.015 |
idea | -0.015 |
Token | in |
Token ID | 287 |
Feature activation | +2.07e-02 |
Pos logit contributions | |
roc | +0.026 |
ole | +0.024 |
ny | +0.022 |
Jak | +0.021 |
idea | +0.021 |
Neg logit contributions | |
arrest | -0.024 |
sidelines | -0.024 |
uddled | -0.024 |
doorstep | -0.023 |
salon | -0.023 |
Token | a |
Token ID | 257 |
Feature activation | -5.90e-01 |
Pos logit contributions | |
arrest | +0.001 |
couch | +0.001 |
sidelines | +0.001 |
freshly | +0.001 |
doorstep | +0.001 |
Neg logit contributions | |
ole | -0.001 |
roc | -0.000 |
ny | -0.000 |
suit | -0.000 |
ac | -0.000 |
Token | trap |
Token ID | 12840 |
Feature activation | -1.17e+00 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.013 |
roc | +0.012 |
Jak | +0.011 |
idea | +0.011 |
Neg logit contributions | |
sidelines | -0.017 |
salon | -0.017 |
doorstep | -0.017 |
Sweet | -0.017 |
arrest | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | -4.10e-01 |
Pos logit contributions | |
ole | +0.026 |
ny | +0.023 |
roc | +0.022 |
Jak | +0.020 |
idea | +0.020 |
Neg logit contributions | |
sidelines | -0.036 |
salon | -0.036 |
doorstep | -0.036 |
Sweet | -0.035 |
arrest | -0.035 |
Token | B |
Token ID | 347 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
ole | +0.011 |
roc | +0.010 |
ny | +0.010 |
ac | +0.010 |
idea | +0.009 |
Neg logit contributions | |
sidelines | -0.010 |
doorstep | -0.010 |
salon | -0.010 |
Sweet | -0.010 |
uddled | -0.010 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.89e+00 |
Pos logit contributions | |
doorstep | +0.020 |
sidelines | +0.019 |
Sweet | +0.019 |
salon | +0.019 |
uddled | +0.019 |
Neg logit contributions | |
ole | -0.033 |
ny | -0.030 |
roc | -0.029 |
idea | -0.028 |
Jak | -0.028 |
Token | i |
Token ID | 72 |
Feature activation | -1.28e-01 |
Pos logit contributions | |
sidelines | +0.157 |
salon | +0.156 |
doorstep | +0.156 |
Sweet | +0.154 |
arrest | +0.152 |
Neg logit contributions | |
ole | -0.103 |
ny | -0.093 |
roc | -0.086 |
Jak | -0.080 |
idea | -0.077 |
Token | tried |
Token ID | 3088 |
Feature activation | +7.13e-01 |
Pos logit contributions | |
ole | +0.003 |
roc | +0.003 |
ny | +0.003 |
Jak | +0.003 |
idea | +0.002 |
Neg logit contributions | |
uddled | -0.004 |
doorstep | -0.003 |
arrest | -0.003 |
Sweet | -0.003 |
salon | -0.003 |
Token | to |
Token ID | 284 |
Feature activation | -1.24e-01 |
Pos logit contributions | |
salon | +0.020 |
sidelines | +0.020 |
doorstep | +0.020 |
Sweet | +0.020 |
arrest | +0.020 |
Neg logit contributions | |
ole | -0.018 |
ny | -0.016 |
roc | -0.015 |
Jak | -0.014 |
idea | -0.014 |
Token | help |
Token ID | 1037 |
Feature activation | +1.18e+00 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.003 |
idea | +0.003 |
Jak | +0.003 |
Neg logit contributions | |
salon | -0.003 |
sidelines | -0.003 |
doorstep | -0.003 |
Sweet | -0.003 |
uddled | -0.003 |
Token | Th |
Token ID | 536 |
Feature activation | -2.07e+00 |
Pos logit contributions | |
arrest | +0.031 |
sidelines | +0.031 |
Her | +0.031 |
doorstep | +0.029 |
Her | +0.029 |
Neg logit contributions | |
ole | -0.028 |
roc | -0.027 |
ny | -0.025 |
ac | -0.025 |
idea | -0.024 |
Token | want |
Token ID | 765 |
Feature activation | +1.85e-01 |
Pos logit contributions | |
doorstep | +0.025 |
arrest | +0.025 |
sidelines | +0.024 |
salon | +0.023 |
Her | +0.022 |
Neg logit contributions | |
ole | -0.036 |
roc | -0.034 |
Jak | -0.033 |
ac | -0.030 |
ny | -0.030 |
Token | to |
Token ID | 284 |
Feature activation | +4.06e-01 |
Pos logit contributions | |
arrest | +0.005 |
doorstep | +0.005 |
salon | +0.004 |
Sweet | +0.004 |
sidelines | +0.004 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.005 |
roc | -0.004 |
idea | -0.004 |
ac | -0.004 |
Token | help |
Token ID | 1037 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
doorstep | +0.009 |
arrest | +0.009 |
sidelines | +0.009 |
salon | +0.009 |
Sweet | +0.008 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.011 |
Jak | -0.011 |
ac | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | -2.92e-01 |
Pos logit contributions | |
arrest | +0.032 |
doorstep | +0.032 |
sidelines | +0.032 |
salon | +0.032 |
Her | +0.031 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.029 |
roc | -0.028 |
idea | -0.026 |
ac | -0.026 |
Token | B |
Token ID | 347 |
Feature activation | +7.53e-01 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.007 |
idea | +0.006 |
ac | +0.006 |
Neg logit contributions | |
doorstep | -0.008 |
sidelines | -0.008 |
Sweet | -0.008 |
uddled | -0.007 |
salon | -0.007 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.87e+00 |
Pos logit contributions | |
doorstep | +0.016 |
sidelines | +0.016 |
salon | +0.016 |
Sweet | +0.016 |
arrest | +0.015 |
Neg logit contributions | |
ole | -0.024 |
ny | -0.022 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.020 |
Token | i |
Token ID | 72 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
sidelines | +0.161 |
salon | +0.159 |
doorstep | +0.159 |
Sweet | +0.157 |
arrest | +0.155 |
Neg logit contributions | |
ole | -0.100 |
ny | -0.089 |
roc | -0.082 |
Jak | -0.076 |
idea | -0.073 |
Token | felt |
Token ID | 2936 |
Feature activation | +1.38e+00 |
Pos logit contributions | |
ole | +0.003 |
Jak | +0.002 |
roc | +0.002 |
ny | +0.002 |
idea | +0.002 |
Neg logit contributions | |
uddled | -0.003 |
doorstep | -0.003 |
arrest | -0.003 |
salon | -0.003 |
Sweet | -0.003 |
Token | sad |
Token ID | 6507 |
Feature activation | +1.04e+00 |
Pos logit contributions | |
sidelines | +0.040 |
arrest | +0.038 |
Sweet | +0.038 |
doorstep | +0.038 |
Acc | +0.037 |
Neg logit contributions | |
ole | -0.033 |
ny | -0.029 |
roc | -0.029 |
Jak | -0.028 |
ac | -0.026 |
Token | and |
Token ID | 290 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
sidelines | +0.034 |
salon | +0.033 |
doorstep | +0.033 |
Sweet | +0.033 |
arrest | +0.033 |
Neg logit contributions | |
ole | -0.021 |
ny | -0.019 |
roc | -0.017 |
idea | -0.016 |
Jak | -0.016 |
Token | scared |
Token ID | 12008 |
Feature activation | +1.48e+00 |
Pos logit contributions | |
sidelines | +0.039 |
doorstep | +0.037 |
arrest | +0.037 |
salon | +0.036 |
uddled | +0.036 |
Neg logit contributions | |
ole | -0.041 |
ny | -0.035 |
roc | -0.035 |
ac | -0.032 |
Jak | -0.032 |
Token | B |
Token ID | 347 |
Feature activation | +8.74e-01 |
Pos logit contributions | |
doorstep | +0.007 |
arrest | +0.007 |
sidelines | +0.007 |
couch | +0.006 |
Sweet | +0.006 |
Neg logit contributions | |
ole | -0.009 |
roc | -0.008 |
idea | -0.008 |
ac | -0.008 |
ny | -0.008 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.74e+00 |
Pos logit contributions | |
salon | +0.022 |
sidelines | +0.022 |
doorstep | +0.022 |
arrest | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.024 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.019 |
Token | i |
Token ID | 72 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
sidelines | +0.157 |
Sweet | +0.156 |
salon | +0.154 |
doorstep | +0.153 |
arrest | +0.149 |
Neg logit contributions | |
ole | -0.097 |
ny | -0.088 |
roc | -0.078 |
Jak | -0.074 |
idea | -0.072 |
Token | . |
Token ID | 13 |
Feature activation | -3.28e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
roc | +0.010 |
Jak | +0.009 |
boy | +0.009 |
Neg logit contributions | |
sidelines | -0.021 |
salon | -0.020 |
doorstep | -0.020 |
Sweet | -0.019 |
arrest | -0.019 |
Token | B |
Token ID | 347 |
Feature activation | +9.79e-01 |
Pos logit contributions | |
ole | +0.009 |
roc | +0.008 |
ny | +0.008 |
ac | +0.007 |
idea | +0.007 |
Neg logit contributions | |
Sweet | -0.008 |
freshly | -0.008 |
doorstep | -0.008 |
sidelines | -0.008 |
Her | -0.008 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.86e+00 |
Pos logit contributions | |
salon | +0.022 |
sidelines | +0.022 |
doorstep | +0.022 |
arrest | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.028 |
roc | -0.026 |
Jak | -0.025 |
idea | -0.025 |
Token | i |
Token ID | 72 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
sidelines | +0.178 |
Sweet | +0.172 |
salon | +0.172 |
doorstep | +0.166 |
arrest | +0.165 |
Neg logit contributions | |
ole | -0.088 |
ny | -0.079 |
roc | -0.068 |
ac | -0.061 |
idea | -0.061 |
Token | was |
Token ID | 373 |
Feature activation | +9.13e-02 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
doorstep | -0.004 |
salon | -0.004 |
Sweet | -0.004 |
arrest | -0.004 |
sidelines | -0.004 |
Token | gifted |
Token ID | 22527 |
Feature activation | -5.29e-01 |
Pos logit contributions | |
freshly | +0.002 |
Her | +0.002 |
arrest | +0.002 |
swinging | +0.002 |
balcony | +0.002 |
Neg logit contributions | |
ole | -0.002 |
roc | -0.002 |
ny | -0.002 |
beard | -0.002 |
boat | -0.002 |
Token | with |
Token ID | 351 |
Feature activation | +4.03e-01 |
Pos logit contributions | |
ole | +0.015 |
roc | +0.013 |
ny | +0.013 |
Jak | +0.013 |
idea | +0.012 |
Neg logit contributions | |
uddled | -0.013 |
salon | -0.012 |
arrest | -0.012 |
Sweet | -0.012 |
doorstep | -0.012 |
Token | the |
Token ID | 262 |
Feature activation | +4.68e-01 |
Pos logit contributions | |
arrest | +0.011 |
Sweet | +0.011 |
sidelines | +0.011 |
freshly | +0.011 |
Her | +0.010 |
Neg logit contributions | |
ole | -0.010 |
ny | -0.009 |
roc | -0.009 |
idea | -0.008 |
suit | -0.008 |
Token | It |
Token ID | 632 |
Feature activation | +3.36e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.007 |
roc | +0.006 |
ac | +0.006 |
boat | +0.006 |
Neg logit contributions | |
Her | -0.006 |
salon | -0.006 |
sidelines | -0.006 |
arrest | -0.005 |
Sweet | -0.005 |
Token | stood |
Token ID | 6204 |
Feature activation | -3.56e-01 |
Pos logit contributions | |
doorstep | +0.009 |
salon | +0.009 |
sidelines | +0.009 |
arrest | +0.009 |
Sweet | +0.009 |
Neg logit contributions | |
ole | -0.008 |
roc | -0.007 |
ny | -0.007 |
Jak | -0.007 |
idea | -0.006 |
Token | up |
Token ID | 510 |
Feature activation | +9.01e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.009 |
roc | +0.009 |
idea | +0.008 |
Jak | +0.008 |
Neg logit contributions | |
salon | -0.009 |
balcony | -0.009 |
doorstep | -0.009 |
sidelines | -0.009 |
uddled | -0.009 |
Token | on |
Token ID | 319 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
salon | +0.023 |
arrest | +0.022 |
sidelines | +0.022 |
doorstep | +0.022 |
balcony | +0.022 |
Neg logit contributions | |
ole | -0.023 |
roc | -0.022 |
ny | -0.022 |
Jak | -0.019 |
guys | -0.019 |
Token | its |
Token ID | 663 |
Feature activation | -3.92e-01 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.013 |
roc | +0.012 |
boat | +0.011 |
ac | +0.011 |
Neg logit contributions | |
arrest | -0.018 |
sidelines | -0.017 |
doorstep | -0.017 |
salon | -0.017 |
couch | -0.017 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.85e+00 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
idea | +0.008 |
roc | +0.008 |
boy | +0.008 |
Neg logit contributions | |
Sweet | -0.011 |
salon | -0.011 |
uddled | -0.011 |
Acc | -0.011 |
sidelines | -0.011 |
Token | legs |
Token ID | 7405 |
Feature activation | -4.95e-03 |
Pos logit contributions | |
salon | +0.145 |
sidelines | +0.143 |
doorstep | +0.142 |
Sweet | +0.141 |
uddled | +0.138 |
Neg logit contributions | |
ole | -0.114 |
ny | -0.103 |
roc | -0.097 |
Jak | -0.091 |
idea | -0.090 |
Token | and |
Token ID | 290 |
Feature activation | -8.43e-01 |
Pos logit contributions | |
ole | +0.000 |
roc | +0.000 |
ny | +0.000 |
Jak | +0.000 |
idea | +0.000 |
Neg logit contributions | |
salon | -0.000 |
doorstep | -0.000 |
sidelines | -0.000 |
arrest | -0.000 |
Sweet | -0.000 |
Token | showed |
Token ID | 3751 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.020 |
roc | +0.019 |
Jak | +0.017 |
idea | +0.017 |
Neg logit contributions | |
sidelines | -0.023 |
salon | -0.023 |
doorstep | -0.022 |
arrest | -0.022 |
Sweet | -0.022 |
Token | its |
Token ID | 663 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
Sweet | +0.003 |
salon | +0.002 |
arrest | +0.002 |
sidelines | +0.002 |
doorstep | +0.002 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
roc | -0.004 |
Jak | -0.004 |
idea | -0.003 |
Token | teeth |
Token ID | 9941 |
Feature activation | -1.68e+00 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
doorstep | -0.003 |
sidelines | -0.003 |
Sweet | -0.003 |
arrest | -0.003 |
salon | -0.003 |
Token | B |
Token ID | 347 |
Feature activation | +7.05e-01 |
Pos logit contributions | |
doorstep | +0.001 |
sidelines | +0.001 |
arrest | +0.001 |
salon | +0.001 |
Sweet | +0.001 |
Neg logit contributions | |
ole | -0.001 |
roc | -0.001 |
ny | -0.001 |
idea | -0.001 |
ac | -0.001 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.79e+00 |
Pos logit contributions | |
sidelines | +0.019 |
salon | +0.019 |
doorstep | +0.018 |
arrest | +0.018 |
Sweet | +0.018 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.019 |
roc | -0.016 |
Jak | -0.016 |
boy | -0.015 |
Token | i |
Token ID | 72 |
Feature activation | -9.32e-01 |
Pos logit contributions | |
Sweet | +0.165 |
sidelines | +0.164 |
salon | +0.159 |
doorstep | +0.159 |
arrest | +0.153 |
Neg logit contributions | |
ole | -0.095 |
ny | -0.085 |
roc | -0.076 |
Jak | -0.072 |
idea | -0.068 |
Token | . |
Token ID | 13 |
Feature activation | -2.80e-01 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.017 |
roc | +0.015 |
Jak | +0.014 |
boy | +0.014 |
Neg logit contributions | |
sidelines | -0.033 |
salon | -0.032 |
doorstep | -0.031 |
Sweet | -0.030 |
arrest | -0.030 |
Token | B |
Token ID | 347 |
Feature activation | +9.86e-01 |
Pos logit contributions | |
ole | +0.007 |
roc | +0.006 |
ny | +0.006 |
ac | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.007 |
Sweet | -0.007 |
doorstep | -0.007 |
salon | -0.007 |
uddled | -0.007 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.84e+00 |
Pos logit contributions | |
sidelines | +0.023 |
salon | +0.023 |
doorstep | +0.022 |
arrest | +0.022 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.028 |
roc | -0.026 |
Jak | -0.025 |
idea | -0.025 |
Token | i |
Token ID | 72 |
Feature activation | -2.83e-01 |
Pos logit contributions | |
sidelines | +0.177 |
Sweet | +0.174 |
salon | +0.171 |
doorstep | +0.169 |
arrest | +0.165 |
Neg logit contributions | |
ole | -0.087 |
ny | -0.077 |
roc | -0.068 |
Jak | -0.062 |
idea | -0.060 |
Token | loved |
Token ID | 6151 |
Feature activation | +1.64e+00 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.007 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
doorstep | -0.007 |
salon | -0.007 |
sidelines | -0.007 |
Sweet | -0.007 |
arrest | -0.007 |
Token | to |
Token ID | 284 |
Feature activation | -4.71e-01 |
Pos logit contributions | |
sidelines | +0.047 |
salon | +0.047 |
doorstep | +0.047 |
Sweet | +0.046 |
arrest | +0.046 |
Neg logit contributions | |
ole | -0.039 |
ny | -0.036 |
roc | -0.033 |
Jak | -0.031 |
idea | -0.031 |
Token | play |
Token ID | 711 |
Feature activation | +7.06e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.012 |
idea | +0.011 |
roc | +0.010 |
Jak | +0.010 |
Neg logit contributions | |
Sweet | -0.013 |
uddled | -0.013 |
salon | -0.013 |
doorstep | -0.013 |
sidelines | -0.012 |
Token | and |
Token ID | 290 |
Feature activation | -1.80e-01 |
Pos logit contributions | |
sidelines | +0.021 |
arrest | +0.021 |
doorstep | +0.021 |
salon | +0.021 |
uddled | +0.021 |
Neg logit contributions | |
ole | -0.016 |
ny | -0.014 |
roc | -0.014 |
Jak | -0.012 |
idea | -0.012 |
Token | j |
Token ID | 474 |
Feature activation | +1.45e+00 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.003 |
Jak | +0.002 |
wood | +0.002 |
Neg logit contributions | |
salon | -0.004 |
sidelines | -0.004 |
Sweet | -0.004 |
doorstep | -0.004 |
arrest | -0.004 |
Token | olly |
Token ID | 5098 |
Feature activation | -2.31e-02 |
Pos logit contributions | |
sidelines | +0.027 |
salon | +0.026 |
doorstep | +0.025 |
arrest | +0.024 |
Sweet | +0.024 |
Neg logit contributions | |
ole | -0.049 |
ny | -0.047 |
roc | -0.045 |
Jak | -0.044 |
idea | -0.043 |
Token | deer |
Token ID | 20096 |
Feature activation | +2.05e+00 |
Pos logit contributions | |
ole | +0.001 |
roc | +0.001 |
ny | +0.000 |
Jak | +0.000 |
wood | +0.000 |
Neg logit contributions | |
salon | -0.001 |
doorstep | -0.001 |
sidelines | -0.001 |
arrest | -0.001 |
Sweet | -0.001 |
Token | named |
Token ID | 3706 |
Feature activation | +4.68e-02 |
Pos logit contributions | |
sidelines | +0.072 |
Sweet | +0.070 |
doorstep | +0.068 |
salon | +0.067 |
uddled | +0.065 |
Neg logit contributions | |
ole | -0.041 |
ny | -0.040 |
boy | -0.034 |
boat | -0.030 |
idea | -0.030 |
Token | B |
Token ID | 347 |
Feature activation | +7.05e-01 |
Pos logit contributions | |
doorstep | +0.001 |
sidelines | +0.001 |
arrest | +0.001 |
salon | +0.001 |
Sweet | +0.001 |
Neg logit contributions | |
ole | -0.001 |
roc | -0.001 |
ny | -0.001 |
idea | -0.001 |
ac | -0.001 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.79e+00 |
Pos logit contributions | |
sidelines | +0.019 |
salon | +0.019 |
doorstep | +0.018 |
arrest | +0.018 |
Sweet | +0.018 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.019 |
roc | -0.016 |
Jak | -0.016 |
boy | -0.015 |
Token | i |
Token ID | 72 |
Feature activation | -9.32e-01 |
Pos logit contributions | |
Sweet | +0.165 |
sidelines | +0.164 |
salon | +0.159 |
doorstep | +0.159 |
arrest | +0.153 |
Neg logit contributions | |
ole | -0.095 |
ny | -0.085 |
roc | -0.076 |
Jak | -0.072 |
idea | -0.068 |
Token | . |
Token ID | 13 |
Feature activation | -2.80e-01 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.017 |
roc | +0.015 |
Jak | +0.014 |
boy | +0.014 |
Neg logit contributions | |
sidelines | -0.033 |
salon | -0.032 |
doorstep | -0.031 |
Sweet | -0.030 |
arrest | -0.030 |
Token | B |
Token ID | 347 |
Feature activation | +9.86e-01 |
Pos logit contributions | |
ole | +0.007 |
roc | +0.006 |
ny | +0.006 |
ac | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.007 |
Sweet | -0.007 |
doorstep | -0.007 |
salon | -0.007 |
uddled | -0.007 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.84e+00 |
Pos logit contributions | |
sidelines | +0.023 |
salon | +0.023 |
doorstep | +0.022 |
arrest | +0.022 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.028 |
roc | -0.026 |
Jak | -0.025 |
idea | -0.025 |
Token | i |
Token ID | 72 |
Feature activation | -2.83e-01 |
Pos logit contributions | |
sidelines | +0.177 |
Sweet | +0.174 |
salon | +0.171 |
doorstep | +0.169 |
arrest | +0.165 |
Neg logit contributions | |
ole | -0.087 |
ny | -0.077 |
roc | -0.068 |
Jak | -0.062 |
idea | -0.060 |
Token | Then |
Token ID | 3244 |
Feature activation | +5.70e-02 |
Pos logit contributions | |
ole | +0.000 |
ny | +0.000 |
Jak | +0.000 |
roc | +0.000 |
boy | +0.000 |
Neg logit contributions | |
doorstep | -0.001 |
sidelines | -0.001 |
salon | -0.001 |
Sweet | -0.001 |
arrest | -0.001 |
Token | he |
Token ID | 339 |
Feature activation | -1.01e-02 |
Pos logit contributions | |
salon | +0.002 |
arrest | +0.001 |
sidelines | +0.001 |
Sweet | +0.001 |
doorstep | +0.001 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.001 |
roc | -0.001 |
ac | -0.001 |
idea | -0.001 |
Token | stands |
Token ID | 6296 |
Feature activation | -8.15e-01 |
Pos logit contributions | |
ny | +0.000 |
ole | +0.000 |
idea | +0.000 |
boy | +0.000 |
roc | +0.000 |
Neg logit contributions | |
sidelines | -0.000 |
doorstep | -0.000 |
Sweet | -0.000 |
Acc | -0.000 |
Her | -0.000 |
Token | on |
Token ID | 319 |
Feature activation | -5.36e-01 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.017 |
roc | +0.017 |
idea | +0.015 |
Jak | +0.014 |
Neg logit contributions | |
salon | -0.024 |
sidelines | -0.023 |
uddled | -0.023 |
Sweet | -0.023 |
doorstep | -0.023 |
Token | his |
Token ID | 465 |
Feature activation | -5.24e-01 |
Pos logit contributions | |
ole | +0.013 |
ny | +0.012 |
roc | +0.011 |
idea | +0.010 |
ac | +0.010 |
Neg logit contributions | |
arrest | -0.016 |
salon | -0.016 |
sidelines | -0.016 |
couch | -0.015 |
doorstep | -0.015 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.79e+00 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.013 |
idea | +0.012 |
boy | +0.012 |
roc | +0.011 |
Neg logit contributions | |
Sweet | -0.014 |
sidelines | -0.014 |
uddled | -0.013 |
salon | -0.013 |
Acc | -0.013 |
Token | legs |
Token ID | 7405 |
Feature activation | -2.34e-01 |
Pos logit contributions | |
salon | +0.144 |
sidelines | +0.144 |
Sweet | +0.143 |
doorstep | +0.142 |
uddled | +0.138 |
Neg logit contributions | |
ole | -0.110 |
ny | -0.100 |
roc | -0.092 |
idea | -0.089 |
Jak | -0.089 |
Token | and |
Token ID | 290 |
Feature activation | -8.44e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
salon | -0.007 |
doorstep | -0.007 |
sidelines | -0.007 |
arrest | -0.007 |
Sweet | -0.007 |
Token | puts |
Token ID | 7584 |
Feature activation | -6.51e-01 |
Pos logit contributions | |
ny | +0.019 |
ole | +0.018 |
boy | +0.017 |
idea | +0.017 |
Jak | +0.017 |
Neg logit contributions | |
doorstep | -0.025 |
sidelines | -0.024 |
Sweet | -0.024 |
uddled | -0.023 |
salon | -0.023 |
Token | his |
Token ID | 465 |
Feature activation | +3.16e-01 |
Pos logit contributions | |
roc | +0.019 |
ole | +0.019 |
ny | +0.017 |
zo | +0.016 |
idea | +0.016 |
Neg logit contributions | |
arrest | -0.014 |
freshly | -0.014 |
sidelines | -0.014 |
couch | -0.013 |
salon | -0.012 |
Token | front |
Token ID | 2166 |
Feature activation | +1.19e+00 |
Pos logit contributions | |
sidelines | +0.009 |
salon | +0.008 |
doorstep | +0.008 |
Sweet | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.008 |
roc | -0.007 |
idea | -0.007 |
Jak | -0.007 |
Token | He |
Token ID | 679 |
Feature activation | -4.14e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.002 |
roc | +0.002 |
idea | +0.002 |
Jak | +0.002 |
Neg logit contributions | |
doorstep | -0.004 |
sidelines | -0.004 |
salon | -0.004 |
Sweet | -0.004 |
arrest | -0.004 |
Token | stood |
Token ID | 6204 |
Feature activation | -3.93e-01 |
Pos logit contributions | |
ole | +0.011 |
roc | +0.010 |
ny | +0.009 |
Jak | +0.009 |
idea | +0.009 |
Neg logit contributions | |
doorstep | -0.011 |
salon | -0.011 |
Sweet | -0.011 |
uddled | -0.011 |
arrest | -0.010 |
Token | up |
Token ID | 510 |
Feature activation | +1.11e+00 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.008 |
roc | +0.007 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.012 |
salon | -0.012 |
doorstep | -0.012 |
balcony | -0.012 |
Sweet | -0.012 |
Token | on |
Token ID | 319 |
Feature activation | -5.07e-01 |
Pos logit contributions | |
salon | +0.032 |
doorstep | +0.032 |
sidelines | +0.032 |
arrest | +0.031 |
Sweet | +0.030 |
Neg logit contributions | |
ole | -0.026 |
roc | -0.024 |
ny | -0.024 |
Jak | -0.021 |
idea | -0.021 |
Token | his |
Token ID | 465 |
Feature activation | -3.73e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
roc | +0.010 |
idea | +0.008 |
ac | +0.008 |
Neg logit contributions | |
arrest | -0.016 |
doorstep | -0.016 |
salon | -0.016 |
sidelines | -0.016 |
couch | -0.015 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.76e+00 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
idea | +0.008 |
boy | +0.007 |
Jak | +0.007 |
Neg logit contributions | |
Sweet | -0.011 |
salon | -0.011 |
uddled | -0.010 |
Acc | -0.010 |
sidelines | -0.010 |
Token | legs |
Token ID | 7405 |
Feature activation | +1.60e-01 |
Pos logit contributions | |
salon | +0.154 |
Sweet | +0.151 |
sidelines | +0.149 |
Acc | +0.147 |
uddled | +0.147 |
Neg logit contributions | |
ole | -0.099 |
ny | -0.088 |
idea | -0.080 |
roc | -0.080 |
Jak | -0.078 |
Token | and |
Token ID | 290 |
Feature activation | -1.28e+00 |
Pos logit contributions | |
doorstep | +0.005 |
salon | +0.005 |
sidelines | +0.005 |
arrest | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token | waved |
Token ID | 26834 |
Feature activation | +5.20e-01 |
Pos logit contributions | |
ole | +0.034 |
ny | +0.031 |
roc | +0.030 |
Jak | +0.027 |
idea | +0.027 |
Neg logit contributions | |
doorstep | -0.035 |
sidelines | -0.034 |
salon | -0.034 |
Sweet | -0.033 |
arrest | -0.033 |
Token | his |
Token ID | 465 |
Feature activation | -1.58e-01 |
Pos logit contributions | |
sidelines | +0.013 |
doorstep | +0.012 |
Her | +0.012 |
balcony | +0.012 |
Sweet | +0.012 |
Neg logit contributions | |
ole | -0.014 |
roc | -0.014 |
ny | -0.013 |
Jak | -0.011 |
beard | -0.011 |
Token | front |
Token ID | 2166 |
Feature activation | +7.51e-01 |
Pos logit contributions | |
ole | +0.004 |
roc | +0.004 |
ny | +0.004 |
Jak | +0.004 |
idea | +0.003 |
Neg logit contributions | |
doorstep | -0.004 |
sidelines | -0.004 |
balcony | -0.004 |
arrest | -0.004 |
Sweet | -0.004 |
Token | was |
Token ID | 373 |
Feature activation | -6.29e-01 |
Pos logit contributions | |
ole | +0.033 |
ny | +0.030 |
roc | +0.029 |
Jak | +0.027 |
idea | +0.025 |
Neg logit contributions | |
doorstep | -0.032 |
salon | -0.032 |
Sweet | -0.031 |
uddled | -0.031 |
arrest | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -3.91e-02 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.008 |
roc | +0.007 |
Jak | +0.007 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.024 |
salon | -0.023 |
doorstep | -0.023 |
Sweet | -0.023 |
arrest | -0.023 |
Token | deer |
Token ID | 20096 |
Feature activation | +2.24e+00 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
ac | +0.001 |
Neg logit contributions | |
salon | -0.001 |
sidelines | -0.001 |
doorstep | -0.001 |
Sweet | -0.001 |
arrest | -0.001 |
Token | named |
Token ID | 3706 |
Feature activation | +2.99e-01 |
Pos logit contributions | |
sidelines | +0.078 |
Sweet | +0.074 |
doorstep | +0.073 |
salon | +0.072 |
uddled | +0.069 |
Neg logit contributions | |
ole | -0.047 |
ny | -0.046 |
boy | -0.039 |
boat | -0.035 |
roc | -0.035 |
Token | B |
Token ID | 347 |
Feature activation | +8.74e-01 |
Pos logit contributions | |
doorstep | +0.007 |
arrest | +0.007 |
sidelines | +0.007 |
couch | +0.006 |
Sweet | +0.006 |
Neg logit contributions | |
ole | -0.009 |
roc | -0.008 |
idea | -0.008 |
ac | -0.008 |
ny | -0.008 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.74e+00 |
Pos logit contributions | |
salon | +0.022 |
sidelines | +0.022 |
doorstep | +0.022 |
arrest | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.024 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.019 |
Token | i |
Token ID | 72 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
sidelines | +0.157 |
Sweet | +0.156 |
salon | +0.154 |
doorstep | +0.153 |
arrest | +0.149 |
Neg logit contributions | |
ole | -0.097 |
ny | -0.088 |
roc | -0.078 |
Jak | -0.074 |
idea | -0.072 |
Token | . |
Token ID | 13 |
Feature activation | -3.28e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
roc | +0.010 |
Jak | +0.009 |
boy | +0.009 |
Neg logit contributions | |
sidelines | -0.021 |
salon | -0.020 |
doorstep | -0.020 |
Sweet | -0.019 |
arrest | -0.019 |
Token | B |
Token ID | 347 |
Feature activation | +9.79e-01 |
Pos logit contributions | |
ole | +0.009 |
roc | +0.008 |
ny | +0.008 |
ac | +0.007 |
idea | +0.007 |
Neg logit contributions | |
Sweet | -0.008 |
freshly | -0.008 |
doorstep | -0.008 |
sidelines | -0.008 |
Her | -0.008 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.86e+00 |
Pos logit contributions | |
salon | +0.022 |
sidelines | +0.022 |
doorstep | +0.022 |
arrest | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.028 |
roc | -0.026 |
Jak | -0.025 |
idea | -0.025 |
Token | i |
Token ID | 72 |
Feature activation | -1.83e-01 |
Pos logit contributions | |
sidelines | +0.178 |
Sweet | +0.172 |
salon | +0.172 |
doorstep | +0.166 |
arrest | +0.165 |
Neg logit contributions | |
ole | -0.088 |
ny | -0.079 |
roc | -0.068 |
ac | -0.061 |
idea | -0.061 |
Token | unk |
Token ID | 2954 |
Feature activation | +7.99e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
roc | +0.008 |
Jak | +0.007 |
ac | +0.007 |
Neg logit contributions | |
sidelines | -0.008 |
doorstep | -0.008 |
Sweet | -0.008 |
salon | -0.008 |
arrest | -0.007 |
Token | stood |
Token ID | 6204 |
Feature activation | -7.99e-01 |
Pos logit contributions | |
doorstep | +0.021 |
salon | +0.021 |
uddled | +0.021 |
Sweet | +0.020 |
arrest | +0.020 |
Neg logit contributions | |
ole | -0.020 |
ny | -0.018 |
roc | -0.018 |
idea | -0.017 |
Jak | -0.017 |
Token | up |
Token ID | 510 |
Feature activation | +8.94e-01 |
Pos logit contributions | |
ole | +0.021 |
roc | +0.020 |
ny | +0.020 |
guys | +0.018 |
idea | +0.018 |
Neg logit contributions | |
doorstep | -0.019 |
sidelines | -0.019 |
Acc | -0.019 |
salon | -0.019 |
balcony | -0.019 |
Token | on |
Token ID | 319 |
Feature activation | -8.81e-01 |
Pos logit contributions | |
doorstep | +0.025 |
sidelines | +0.025 |
salon | +0.025 |
Sweet | +0.024 |
arrest | +0.024 |
Neg logit contributions | |
ole | -0.021 |
roc | -0.020 |
ny | -0.020 |
Jak | -0.017 |
idea | -0.017 |
Token | its |
Token ID | 663 |
Feature activation | -4.41e-01 |
Pos logit contributions | |
ole | +0.021 |
ny | +0.019 |
ac | +0.018 |
roc | +0.018 |
suit | +0.017 |
Neg logit contributions | |
arrest | -0.027 |
couch | -0.025 |
sidelines | -0.025 |
doorstep | -0.025 |
salon | -0.023 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.71e+00 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
idea | +0.009 |
roc | +0.008 |
boy | +0.008 |
Neg logit contributions | |
salon | -0.013 |
Sweet | -0.013 |
uddled | -0.013 |
arrest | -0.012 |
Acc | -0.012 |
Token | legs |
Token ID | 7405 |
Feature activation | -6.80e-02 |
Pos logit contributions | |
salon | +0.139 |
sidelines | +0.137 |
doorstep | +0.136 |
Sweet | +0.135 |
uddled | +0.131 |
Neg logit contributions | |
ole | -0.113 |
ny | -0.102 |
roc | -0.095 |
Jak | -0.091 |
idea | -0.089 |
Token | and |
Token ID | 290 |
Feature activation | -1.13e+00 |
Pos logit contributions | |
ole | +0.002 |
roc | +0.001 |
ny | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
doorstep | -0.002 |
salon | -0.002 |
sidelines | -0.002 |
arrest | -0.002 |
Sweet | -0.002 |
Token | spoke |
Token ID | 5158 |
Feature activation | -1.54e+00 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.025 |
roc | +0.024 |
idea | +0.022 |
Jak | +0.022 |
Neg logit contributions | |
sidelines | -0.032 |
salon | -0.032 |
doorstep | -0.032 |
Sweet | -0.031 |
arrest | -0.031 |
Token | to |
Token ID | 284 |
Feature activation | -8.54e-01 |
Pos logit contributions | |
ole | +0.042 |
roc | +0.040 |
ny | +0.040 |
Jak | +0.037 |
idea | +0.035 |
Neg logit contributions | |
doorstep | -0.036 |
uddled | -0.035 |
Her | -0.035 |
Sweet | -0.034 |
balcony | -0.034 |
Token | Tim |
Token ID | 5045 |
Feature activation | -9.56e-01 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.015 |
roc | +0.014 |
Jak | +0.013 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.028 |
doorstep | -0.028 |
salon | -0.028 |
arrest | -0.027 |
Sweet | -0.027 |
Token | was |
Token ID | 373 |
Feature activation | -1.57e-01 |
Pos logit contributions | |
ole | +0.013 |
ny | +0.012 |
roc | +0.011 |
idea | +0.010 |
Jak | +0.010 |
Neg logit contributions | |
arrest | -0.018 |
doorstep | -0.018 |
sidelines | -0.018 |
salon | -0.018 |
Sweet | -0.018 |
Token | trying |
Token ID | 2111 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
ole | +0.004 |
roc | +0.004 |
ny | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
sidelines | -0.004 |
arrest | -0.004 |
salon | -0.004 |
freshly | -0.003 |
balcony | -0.003 |
Token | to |
Token ID | 284 |
Feature activation | +3.92e-02 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.027 |
roc | +0.026 |
Jak | +0.024 |
idea | +0.024 |
Neg logit contributions | |
sidelines | -0.031 |
doorstep | -0.031 |
salon | -0.031 |
Sweet | -0.031 |
arrest | -0.031 |
Token | shoot |
Token ID | 2686 |
Feature activation | -1.28e+00 |
Pos logit contributions | |
sidelines | +0.001 |
arrest | +0.001 |
doorstep | +0.001 |
salon | +0.001 |
Sweet | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
idea | -0.001 |
Token | B |
Token ID | 347 |
Feature activation | +1.36e+00 |
Pos logit contributions | |
bench | +0.034 |
ole | +0.034 |
idea | +0.033 |
Deliver | +0.032 |
iden | +0.032 |
Neg logit contributions | |
arrest | -0.026 |
Leave | -0.021 |
Sweet | -0.021 |
Her | -0.021 |
Her | -0.020 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.70e+00 |
Pos logit contributions | |
sidelines | +0.033 |
salon | +0.033 |
doorstep | +0.033 |
Sweet | +0.032 |
arrest | +0.032 |
Neg logit contributions | |
ole | -0.039 |
ny | -0.037 |
roc | -0.034 |
Jak | -0.033 |
idea | -0.032 |
Token | i |
Token ID | 72 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
sidelines | +0.149 |
salon | +0.147 |
Sweet | +0.146 |
doorstep | +0.146 |
arrest | +0.142 |
Neg logit contributions | |
ole | -0.102 |
ny | -0.093 |
roc | -0.085 |
Jak | -0.080 |
idea | -0.077 |
Token | . |
Token ID | 13 |
Feature activation | -5.42e-01 |
Pos logit contributions | |
arrest | +0.004 |
salon | +0.003 |
doorstep | +0.003 |
sidelines | +0.003 |
Sweet | +0.003 |
Neg logit contributions | |
ole | -0.003 |
roc | -0.003 |
ny | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token |
|
Token ID | 198 |
Feature activation | +2.91e-01 |
Pos logit contributions | |
ole | +0.016 |
roc | +0.014 |
ny | +0.013 |
ac | +0.013 |
idea | +0.013 |
Neg logit contributions | |
doorstep | -0.013 |
Sweet | -0.012 |
uddled | -0.012 |
sidelines | -0.012 |
salon | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | +3.73e-01 |
Pos logit contributions | |
sidelines | +0.008 |
salon | +0.008 |
doorstep | +0.008 |
Sweet | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | B |
Token ID | 33 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
Sweet | +0.010 |
sidelines | +0.010 |
doorstep | +0.010 |
uddled | +0.010 |
salon | +0.010 |
Neg logit contributions | |
ole | -0.009 |
roc | -0.008 |
ny | -0.008 |
Jak | -0.008 |
idea | -0.008 |
Token | He |
Token ID | 679 |
Feature activation | -4.79e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.007 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
Sweet | -0.008 |
arrest | -0.008 |
Token | stood |
Token ID | 6204 |
Feature activation | -3.82e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.010 |
roc | +0.009 |
Jak | +0.008 |
idea | +0.008 |
Neg logit contributions | |
sidelines | -0.015 |
doorstep | -0.015 |
salon | -0.015 |
Sweet | -0.015 |
arrest | -0.014 |
Token | up |
Token ID | 510 |
Feature activation | +9.98e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
roc | +0.009 |
idea | +0.008 |
guys | +0.008 |
Neg logit contributions | |
salon | -0.010 |
sidelines | -0.010 |
balcony | -0.010 |
Sweet | -0.010 |
doorstep | -0.010 |
Token | on |
Token ID | 319 |
Feature activation | -5.52e-01 |
Pos logit contributions | |
salon | +0.028 |
sidelines | +0.028 |
doorstep | +0.027 |
arrest | +0.027 |
Sweet | +0.027 |
Neg logit contributions | |
ole | -0.024 |
ny | -0.022 |
roc | -0.022 |
Jak | -0.019 |
idea | -0.019 |
Token | his |
Token ID | 465 |
Feature activation | -6.77e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
roc | +0.011 |
idea | +0.010 |
ac | +0.010 |
Neg logit contributions | |
arrest | -0.017 |
salon | -0.017 |
sidelines | -0.016 |
doorstep | -0.016 |
couch | -0.016 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.69e+00 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.016 |
idea | +0.014 |
boy | +0.014 |
roc | +0.013 |
Neg logit contributions | |
Sweet | -0.019 |
salon | -0.019 |
uddled | -0.019 |
Acc | -0.018 |
sidelines | -0.018 |
Token | legs |
Token ID | 7405 |
Feature activation | -1.14e-01 |
Pos logit contributions | |
salon | +0.147 |
sidelines | +0.143 |
doorstep | +0.143 |
Sweet | +0.142 |
uddled | +0.140 |
Neg logit contributions | |
ole | -0.104 |
ny | -0.095 |
roc | -0.087 |
Jak | -0.082 |
idea | -0.082 |
Token | and |
Token ID | 290 |
Feature activation | -9.63e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
sidelines | -0.004 |
doorstep | -0.004 |
salon | -0.004 |
Sweet | -0.004 |
arrest | -0.004 |
Token | cl |
Token ID | 537 |
Feature activation | -5.97e-02 |
Pos logit contributions | |
ole | +0.023 |
ny | +0.021 |
roc | +0.019 |
Jak | +0.019 |
idea | +0.018 |
Neg logit contributions | |
sidelines | -0.028 |
doorstep | -0.028 |
salon | -0.028 |
Sweet | -0.027 |
arrest | -0.027 |
Token | apped |
Token ID | 6320 |
Feature activation | -9.27e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.001 |
roc | +0.001 |
idea | +0.001 |
Jak | +0.001 |
Neg logit contributions | |
sidelines | -0.002 |
doorstep | -0.002 |
salon | -0.002 |
Sweet | -0.002 |
uddled | -0.002 |
Token | his |
Token ID | 465 |
Feature activation | -4.50e-01 |
Pos logit contributions | |
ole | +0.019 |
roc | +0.018 |
ny | +0.017 |
Jak | +0.015 |
wood | +0.014 |
Neg logit contributions | |
Her | -0.028 |
sidelines | -0.027 |
Sweet | -0.027 |
balcony | -0.026 |
arrest | -0.026 |
Token | He |
Token ID | 679 |
Feature activation | -3.22e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
sidelines | -0.002 |
salon | -0.002 |
doorstep | -0.002 |
Sweet | -0.002 |
arrest | -0.002 |
Token | stood |
Token ID | 6204 |
Feature activation | -4.21e-01 |
Pos logit contributions | |
ole | +0.009 |
roc | +0.008 |
ny | +0.007 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
salon | -0.009 |
doorstep | -0.008 |
uddled | -0.008 |
arrest | -0.008 |
sidelines | -0.008 |
Token | up |
Token ID | 510 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
ny | +0.012 |
ole | +0.011 |
roc | +0.011 |
idea | +0.010 |
guys | +0.010 |
Neg logit contributions | |
sidelines | -0.010 |
balcony | -0.010 |
uddled | -0.010 |
salon | -0.010 |
Her | -0.009 |
Token | on |
Token ID | 319 |
Feature activation | -4.70e-01 |
Pos logit contributions | |
salon | +0.029 |
sidelines | +0.028 |
arrest | +0.027 |
doorstep | +0.027 |
balcony | +0.027 |
Neg logit contributions | |
ole | -0.027 |
roc | -0.027 |
ny | -0.026 |
ac | -0.023 |
idea | -0.023 |
Token | his |
Token ID | 465 |
Feature activation | -3.97e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
roc | +0.010 |
ac | +0.010 |
idea | +0.009 |
Neg logit contributions | |
arrest | -0.015 |
sidelines | -0.014 |
couch | -0.014 |
doorstep | -0.014 |
salon | -0.013 |
Token | hind |
Token ID | 16222 |
Feature activation | +4.68e+00 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
idea | +0.009 |
boy | +0.009 |
roc | +0.008 |
Neg logit contributions | |
Sweet | -0.011 |
uddled | -0.010 |
Acc | -0.010 |
salon | -0.010 |
sidelines | -0.010 |
Token | legs |
Token ID | 7405 |
Feature activation | -3.00e-01 |
Pos logit contributions | |
salon | +0.145 |
Sweet | +0.141 |
sidelines | +0.140 |
doorstep | +0.140 |
uddled | +0.136 |
Neg logit contributions | |
ole | -0.106 |
ny | -0.095 |
roc | -0.089 |
Jak | -0.085 |
idea | -0.084 |
Token | and |
Token ID | 290 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
ole | +0.007 |
roc | +0.007 |
ny | +0.006 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
salon | -0.009 |
doorstep | -0.009 |
sidelines | -0.008 |
uddled | -0.008 |
arrest | -0.008 |
Token | held |
Token ID | 2714 |
Feature activation | +2.10e-01 |
Pos logit contributions | |
ole | +0.031 |
roc | +0.026 |
ny | +0.026 |
idea | +0.024 |
ac | +0.024 |
Neg logit contributions | |
salon | -0.029 |
doorstep | -0.028 |
sidelines | -0.028 |
arrest | -0.027 |
balcony | -0.027 |
Token | the |
Token ID | 262 |
Feature activation | +6.67e-02 |
Pos logit contributions | |
Her | +0.005 |
sidelines | +0.005 |
arrest | +0.005 |
freshly | +0.005 |
Her | +0.005 |
Neg logit contributions | |
roc | -0.005 |
ole | -0.005 |
ny | -0.004 |
beard | -0.004 |
zo | -0.004 |
Token | bone |
Token ID | 9970 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
salon | +0.002 |
doorstep | +0.002 |
sidelines | +0.002 |
Sweet | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.002 |
roc | -0.001 |
idea | -0.001 |
Jak | -0.001 |
Token |
|
Token ID | 198 |
Feature activation | +2.30e-01 |
Pos logit contributions | |
sidelines | +0.039 |
salon | +0.039 |
arrest | +0.038 |
doorstep | +0.038 |
balcony | +0.036 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.022 |
Jak | -0.019 |
roc | -0.019 |
idea | -0.018 |
Token |
|
Token ID | 198 |
Feature activation | +2.65e-01 |
Pos logit contributions | |
salon | +0.007 |
sidelines | +0.007 |
doorstep | +0.007 |
arrest | +0.007 |
Sweet | +0.007 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.005 |
roc | -0.004 |
Jak | -0.004 |
idea | -0.004 |
Token | Finally |
Token ID | 11158 |
Feature activation | +1.66e-01 |
Pos logit contributions | |
salon | +0.006 |
doorstep | +0.006 |
sidelines | +0.006 |
arrest | +0.006 |
Sweet | +0.006 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | -7.79e-01 |
Pos logit contributions | |
arrest | +0.005 |
uddled | +0.005 |
doorstep | +0.005 |
� | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
roc | -0.004 |
ole | -0.004 |
ny | -0.004 |
ac | -0.003 |
idea | -0.003 |
Token | B |
Token ID | 347 |
Feature activation | +5.21e-01 |
Pos logit contributions | |
ole | +0.020 |
roc | +0.020 |
ny | +0.020 |
ac | +0.018 |
idea | +0.017 |
Neg logit contributions | |
arrest | -0.019 |
Sweet | -0.019 |
sidelines | -0.019 |
doorstep | -0.019 |
� | -0.019 |
Token | amb |
Token ID | 4131 |
Feature activation | +4.67e+00 |
Pos logit contributions | |
doorstep | +0.011 |
Sweet | +0.010 |
sidelines | +0.010 |
salon | +0.010 |
arrest | +0.010 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.015 |
roc | -0.015 |
idea | -0.014 |
Jak | -0.014 |
Token | i |
Token ID | 72 |
Feature activation | -1.14e-01 |
Pos logit contributions | |
sidelines | +0.154 |
salon | +0.152 |
doorstep | +0.151 |
Sweet | +0.150 |
arrest | +0.147 |
Neg logit contributions | |
ole | -0.096 |
ny | -0.086 |
roc | -0.079 |
Jak | -0.073 |
idea | -0.070 |
Token | found |
Token ID | 1043 |
Feature activation | +4.39e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
doorstep | -0.003 |
arrest | -0.003 |
salon | -0.003 |
uddled | -0.003 |
Sweet | -0.003 |
Token | a |
Token ID | 257 |
Feature activation | -4.30e-01 |
Pos logit contributions | |
arrest | +0.013 |
Sweet | +0.013 |
uddled | +0.013 |
sidelines | +0.013 |
doorstep | +0.012 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.009 |
roc | -0.009 |
idea | -0.008 |
ac | -0.007 |
Token | kind |
Token ID | 1611 |
Feature activation | -3.05e+00 |
Pos logit contributions | |
ole | +0.010 |
roc | +0.009 |
ny | +0.009 |
Jak | +0.008 |
ac | +0.008 |
Neg logit contributions | |
doorstep | -0.013 |
sidelines | -0.012 |
salon | -0.012 |
balcony | -0.012 |
arrest | -0.012 |
Token | hunter |
Token ID | 19177 |
Feature activation | -5.00e-01 |
Pos logit contributions | |
ole | +0.075 |
roc | +0.073 |
Jak | +0.069 |
ny | +0.067 |
ac | +0.065 |
Neg logit contributions | |
doorstep | -0.079 |
sidelines | -0.077 |
salon | -0.076 |
arrest | -0.075 |
couch | -0.073 |
Token | They |
Token ID | 1119 |
Feature activation | +1.83e+00 |
Pos logit contributions | |
doorstep | +0.007 |
sidelines | +0.007 |
salon | +0.007 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.007 |
Jak | -0.006 |
idea | -0.006 |
Token | asked |
Token ID | 1965 |
Feature activation | -3.47e-01 |
Pos logit contributions | |
sidelines | +0.058 |
salon | +0.058 |
doorstep | +0.058 |
Sweet | +0.057 |
arrest | +0.057 |
Neg logit contributions | |
ole | -0.038 |
ny | -0.035 |
roc | -0.032 |
Jak | -0.029 |
idea | -0.029 |
Token | the |
Token ID | 262 |
Feature activation | -6.37e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.009 |
suit | +0.008 |
roc | +0.008 |
boat | +0.008 |
Neg logit contributions | |
Sweet | -0.009 |
doorstep | -0.009 |
salon | -0.008 |
Her | -0.008 |
uddled | -0.008 |
Token | festival |
Token ID | 10876 |
Feature activation | -7.36e-01 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.013 |
idea | +0.012 |
Neg logit contributions | |
sidelines | -0.018 |
doorstep | -0.018 |
salon | -0.018 |
Sweet | -0.018 |
arrest | -0.017 |
Token | organ |
Token ID | 1618 |
Feature activation | -3.03e+00 |
Pos logit contributions | |
ole | +0.019 |
roc | +0.018 |
ny | +0.018 |
Jak | +0.017 |
ac | +0.015 |
Neg logit contributions | |
salon | -0.019 |
doorstep | -0.019 |
Sweet | -0.018 |
balcony | -0.018 |
uddled | -0.018 |
Token | iser |
Token ID | 5847 |
Feature activation | -5.76e+00 |
Pos logit contributions | |
ole | +0.082 |
ny | +0.077 |
roc | +0.074 |
Jak | +0.070 |
idea | +0.067 |
Neg logit contributions | |
salon | -0.075 |
doorstep | -0.075 |
sidelines | -0.074 |
Sweet | -0.074 |
uddled | -0.073 |
Token | why |
Token ID | 1521 |
Feature activation | +1.77e-01 |
Pos logit contributions | |
ole | +0.108 |
Jak | +0.097 |
roc | +0.095 |
idea | +0.089 |
afloat | +0.084 |
Neg logit contributions | |
Sweet | -0.174 |
salon | -0.174 |
uddled | -0.169 |
Acc | -0.168 |
chair | -0.160 |
Token | it |
Token ID | 340 |
Feature activation | +6.47e-02 |
Pos logit contributions | |
salon | +0.005 |
doorstep | +0.005 |
sidelines | +0.005 |
arrest | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
roc | -0.004 |
Jak | -0.003 |
idea | -0.003 |
Token | was |
Token ID | 373 |
Feature activation | -4.44e-01 |
Pos logit contributions | |
sidelines | +0.002 |
salon | +0.002 |
doorstep | +0.002 |
Sweet | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
idea | -0.001 |
Jak | -0.001 |
Token | drying |
Token ID | 29621 |
Feature activation | -3.66e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
roc | +0.009 |
Jak | +0.008 |
idea | +0.008 |
Neg logit contributions | |
doorstep | -0.013 |
salon | -0.013 |
sidelines | -0.013 |
Sweet | -0.013 |
arrest | -0.013 |
Token | up |
Token ID | 510 |
Feature activation | +5.97e-01 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.008 |
roc | +0.008 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
doorstep | -0.010 |
salon | -0.010 |
Sweet | -0.010 |
uddled | -0.010 |
sidelines | -0.010 |
Token | man |
Token ID | 582 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.005 |
ac | +0.005 |
Neg logit contributions | |
sidelines | -0.009 |
salon | -0.009 |
doorstep | -0.009 |
Sweet | -0.009 |
arrest | -0.008 |
Token | stepped |
Token ID | 10764 |
Feature activation | -1.92e-01 |
Pos logit contributions | |
ole | +0.031 |
roc | +0.028 |
ny | +0.027 |
Jak | +0.026 |
idea | +0.025 |
Neg logit contributions | |
doorstep | -0.030 |
salon | -0.030 |
uddled | -0.030 |
Sweet | -0.029 |
arrest | -0.028 |
Token | closer |
Token ID | 5699 |
Feature activation | +6.82e-01 |
Pos logit contributions | |
ole | +0.004 |
roc | +0.004 |
ny | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
sidelines | -0.006 |
doorstep | -0.006 |
salon | -0.006 |
uddled | -0.006 |
Sweet | -0.005 |
Token | and |
Token ID | 290 |
Feature activation | -1.66e+00 |
Pos logit contributions | |
doorstep | +0.018 |
salon | +0.018 |
sidelines | +0.017 |
Sweet | +0.017 |
uddled | +0.017 |
Neg logit contributions | |
ole | -0.018 |
roc | -0.017 |
ny | -0.016 |
Jak | -0.015 |
ac | -0.014 |
Token | said |
Token ID | 531 |
Feature activation | -3.37e-01 |
Pos logit contributions | |
ole | +0.042 |
roc | +0.037 |
ny | +0.036 |
idea | +0.035 |
Jak | +0.033 |
Neg logit contributions | |
salon | -0.046 |
sidelines | -0.046 |
Sweet | -0.046 |
doorstep | -0.044 |
uddled | -0.044 |
Token | gru |
Token ID | 22848 |
Feature activation | -5.14e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.005 |
Neg logit contributions | |
salon | -0.010 |
uddled | -0.010 |
Sweet | -0.010 |
sidelines | -0.010 |
Her | -0.010 |
Token | ff |
Token ID | 487 |
Feature activation | -1.17e+00 |
Pos logit contributions | |
ole | +0.096 |
Jak | +0.088 |
idea | +0.084 |
boat | +0.083 |
afloat | +0.080 |
Neg logit contributions | |
Sweet | -0.147 |
doorstep | -0.141 |
Acc | -0.139 |
atters | -0.138 |
der | -0.138 |
Token | ly |
Token ID | 306 |
Feature activation | -1.32e+00 |
Pos logit contributions | |
roc | +0.027 |
Jak | +0.027 |
ole | +0.027 |
ny | +0.026 |
idea | +0.025 |
Neg logit contributions | |
doorstep | -0.027 |
uddled | -0.027 |
der | -0.027 |
salon | -0.027 |
balcony | -0.025 |
Token | , |
Token ID | 11 |
Feature activation | -1.18e+00 |
Pos logit contributions | |
ole | +0.028 |
roc | +0.027 |
ny | +0.027 |
Jak | +0.025 |
boat | +0.024 |
Neg logit contributions | |
uddled | -0.035 |
salon | -0.035 |
doorstep | -0.035 |
Her | -0.034 |
arrest | -0.034 |
Token | " |
Token ID | 366 |
Feature activation | +3.72e-01 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.028 |
roc | +0.026 |
Jak | +0.025 |
idea | +0.024 |
Neg logit contributions | |
sidelines | -0.032 |
salon | -0.031 |
doorstep | -0.031 |
Sweet | -0.031 |
arrest | -0.030 |
Token | You |
Token ID | 1639 |
Feature activation | -7.80e-01 |
Pos logit contributions | |
salon | +0.008 |
sidelines | +0.008 |
Sweet | +0.008 |
doorstep | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.010 |
idea | -0.010 |
Jak | -0.010 |
Token | ... |
Token ID | 986 |
Feature activation | -1.03e+00 |
Pos logit contributions | |
sidelines | +0.051 |
doorstep | +0.051 |
salon | +0.050 |
Sweet | +0.049 |
arrest | +0.049 |
Neg logit contributions | |
ole | -0.049 |
ny | -0.045 |
roc | -0.042 |
Jak | -0.040 |
idea | -0.039 |
Token | happy |
Token ID | 34191 |
Feature activation | -8.57e-01 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.026 |
roc | +0.025 |
Jak | +0.024 |
idea | +0.023 |
Neg logit contributions | |
sidelines | -0.026 |
doorstep | -0.026 |
salon | -0.026 |
Sweet | -0.025 |
arrest | -0.025 |
Token | . |
Token ID | 13 |
Feature activation | -7.28e-02 |
Pos logit contributions | |
ole | +0.024 |
ny | +0.022 |
roc | +0.021 |
Jak | +0.020 |
idea | +0.020 |
Neg logit contributions | |
sidelines | -0.021 |
salon | -0.021 |
doorstep | -0.021 |
Sweet | -0.020 |
arrest | -0.020 |
Token | I |
Token ID | 314 |
Feature activation | +2.20e+00 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
Jak | +0.002 |
roc | +0.002 |
boy | +0.002 |
Neg logit contributions | |
sidelines | -0.002 |
doorstep | -0.002 |
uddled | -0.002 |
salon | -0.002 |
arrest | -0.002 |
Token | ... |
Token ID | 986 |
Feature activation | -8.33e-01 |
Pos logit contributions | |
salon | +0.057 |
doorstep | +0.056 |
sidelines | +0.055 |
arrest | +0.055 |
Sweet | +0.055 |
Neg logit contributions | |
ole | -0.059 |
ny | -0.053 |
roc | -0.051 |
Jak | -0.049 |
idea | -0.046 |
Token | sent |
Token ID | 34086 |
Feature activation | -5.11e+00 |
Pos logit contributions | |
ole | +0.023 |
ny | +0.022 |
roc | +0.020 |
Jak | +0.019 |
idea | +0.019 |
Neg logit contributions | |
sidelines | -0.021 |
doorstep | -0.021 |
salon | -0.020 |
Sweet | -0.020 |
arrest | -0.020 |
Token | ... |
Token ID | 986 |
Feature activation | -9.91e-01 |
Pos logit contributions | |
ole | +0.117 |
ny | +0.107 |
roc | +0.100 |
idea | +0.093 |
Jak | +0.093 |
Neg logit contributions | |
salon | -0.150 |
sidelines | -0.149 |
doorstep | -0.149 |
arrest | -0.148 |
Sweet | -0.147 |
Token | you |
Token ID | 5832 |
Feature activation | -1.29e+00 |
Pos logit contributions | |
ole | +0.027 |
ny | +0.025 |
roc | +0.023 |
Jak | +0.023 |
idea | +0.022 |
Neg logit contributions | |
doorstep | -0.025 |
sidelines | -0.025 |
salon | -0.025 |
Sweet | -0.024 |
arrest | -0.024 |
Token | ... |
Token ID | 986 |
Feature activation | -8.06e-01 |
Pos logit contributions | |
ole | +0.029 |
ny | +0.027 |
roc | +0.023 |
idea | +0.023 |
ac | +0.022 |
Neg logit contributions | |
sidelines | -0.041 |
Sweet | -0.039 |
salon | -0.039 |
doorstep | -0.039 |
Acc | -0.038 |
Token | a |
Token ID | 64 |
Feature activation | -1.27e+00 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.021 |
roc | +0.020 |
Jak | +0.019 |
idea | +0.018 |
Neg logit contributions | |
sidelines | -0.020 |
doorstep | -0.020 |
salon | -0.020 |
Sweet | -0.019 |
arrest | -0.019 |
Token | ... |
Token ID | 986 |
Feature activation | -7.99e-01 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.027 |
roc | +0.025 |
Jak | +0.023 |
idea | +0.023 |
Neg logit contributions | |
sidelines | -0.038 |
doorstep | -0.038 |
Sweet | -0.037 |
salon | -0.037 |
uddled | -0.036 |
Token | he |
Token ID | 339 |
Feature activation | +4.18e-01 |
Pos logit contributions | |
ole | +0.021 |
roc | +0.019 |
ny | +0.019 |
Jak | +0.017 |
idea | +0.017 |
Neg logit contributions | |
salon | -0.023 |
doorstep | -0.022 |
sidelines | -0.022 |
Sweet | -0.022 |
arrest | -0.021 |
Token | had |
Token ID | 550 |
Feature activation | -1.13e+00 |
Pos logit contributions | |
sidelines | +0.013 |
salon | +0.013 |
doorstep | +0.013 |
Sweet | +0.013 |
arrest | +0.013 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | seen |
Token ID | 1775 |
Feature activation | +6.26e-02 |
Pos logit contributions | |
ole | +0.025 |
ny | +0.022 |
Jak | +0.020 |
roc | +0.019 |
idea | +0.019 |
Neg logit contributions | |
sidelines | -0.036 |
Sweet | -0.035 |
doorstep | -0.034 |
salon | -0.034 |
uddled | -0.033 |
Token | a |
Token ID | 257 |
Feature activation | -2.09e-01 |
Pos logit contributions | |
sidelines | +0.002 |
salon | +0.002 |
doorstep | +0.002 |
Sweet | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
idea | -0.001 |
Token | year |
Token ID | 614 |
Feature activation | -1.86e+00 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
idea | +0.004 |
boy | +0.004 |
roc | +0.004 |
Neg logit contributions | |
uddled | -0.006 |
Sweet | -0.006 |
arrest | -0.006 |
doorstep | -0.006 |
sidelines | -0.006 |
Token | ago |
Token ID | 2084 |
Feature activation | -5.08e+00 |
Pos logit contributions | |
ole | +0.044 |
roc | +0.044 |
ny | +0.042 |
Jak | +0.039 |
idea | +0.037 |
Neg logit contributions | |
doorstep | -0.051 |
salon | -0.050 |
sidelines | -0.049 |
balcony | -0.047 |
arrest | -0.046 |
Token | . |
Token ID | 13 |
Feature activation | -6.29e-02 |
Pos logit contributions | |
ole | +0.110 |
ny | +0.100 |
roc | +0.093 |
Jak | +0.087 |
idea | +0.084 |
Neg logit contributions | |
sidelines | -0.157 |
salon | -0.156 |
doorstep | -0.156 |
Sweet | -0.153 |
arrest | -0.152 |
Token | He |
Token ID | 679 |
Feature activation | -4.84e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.001 |
roc | +0.001 |
idea | +0.001 |
Jak | +0.001 |
Neg logit contributions | |
doorstep | -0.002 |
sidelines | -0.002 |
Sweet | -0.002 |
salon | -0.002 |
arrest | -0.002 |
Token | went |
Token ID | 1816 |
Feature activation | -4.30e-02 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
roc | +0.009 |
Jak | +0.009 |
idea | +0.008 |
Neg logit contributions | |
doorstep | -0.015 |
salon | -0.015 |
sidelines | -0.015 |
Sweet | -0.014 |
arrest | -0.014 |
Token | home |
Token ID | 1363 |
Feature activation | +2.09e+00 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
sidelines | -0.001 |
doorstep | -0.001 |
Sweet | -0.001 |
salon | -0.001 |
arrest | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | +4.36e-01 |
Pos logit contributions | |
sidelines | +0.069 |
doorstep | +0.068 |
Sweet | +0.068 |
salon | +0.068 |
arrest | +0.067 |
Neg logit contributions | |
ole | -0.042 |
ny | -0.037 |
roc | -0.033 |
Jak | -0.032 |
idea | -0.031 |
Token | He |
Token ID | 679 |
Feature activation | -2.84e-01 |
Pos logit contributions | |
sidelines | +0.037 |
uddled | +0.035 |
doorstep | +0.035 |
salon | +0.034 |
arrest | +0.034 |
Neg logit contributions | |
ole | -0.041 |
ny | -0.041 |
roc | -0.038 |
boy | -0.038 |
Jak | -0.037 |
Token | was |
Token ID | 373 |
Feature activation | +4.07e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
idea | +0.005 |
Jak | +0.005 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
Sweet | -0.008 |
arrest | -0.008 |
Token | here |
Token ID | 994 |
Feature activation | -1.49e+00 |
Pos logit contributions | |
sidelines | +0.012 |
salon | +0.012 |
doorstep | +0.012 |
arrest | +0.012 |
Sweet | +0.012 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.008 |
roc | -0.008 |
Jak | -0.007 |
idea | -0.007 |
Token | a |
Token ID | 257 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.025 |
idea | +0.021 |
roc | +0.020 |
boy | +0.020 |
Neg logit contributions | |
sidelines | -0.052 |
salon | -0.051 |
arrest | -0.051 |
doorstep | -0.050 |
Sweet | -0.050 |
Token | minute |
Token ID | 5664 |
Feature activation | -1.28e+00 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.003 |
Jak | +0.003 |
idea | +0.003 |
Neg logit contributions | |
sidelines | -0.005 |
salon | -0.005 |
doorstep | -0.005 |
Sweet | -0.005 |
arrest | -0.005 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.93e+00 |
Pos logit contributions | |
ole | +0.024 |
ny | +0.022 |
idea | +0.018 |
boy | +0.017 |
roc | +0.017 |
Neg logit contributions | |
sidelines | -0.045 |
Sweet | -0.044 |
doorstep | -0.043 |
uddled | -0.043 |
arrest | -0.043 |
Token | . |
Token ID | 13 |
Feature activation | +5.17e-01 |
Pos logit contributions | |
ole | +0.095 |
ny | +0.085 |
roc | +0.076 |
Jak | +0.071 |
idea | +0.071 |
Neg logit contributions | |
sidelines | -0.166 |
doorstep | -0.164 |
Sweet | -0.163 |
salon | -0.163 |
arrest | -0.159 |
Token | He |
Token ID | 679 |
Feature activation | -7.84e-01 |
Pos logit contributions | |
sidelines | +0.014 |
salon | +0.014 |
doorstep | +0.014 |
uddled | +0.014 |
arrest | +0.014 |
Neg logit contributions | |
ole | -0.013 |
ny | -0.012 |
roc | -0.011 |
Jak | -0.011 |
boy | -0.011 |
Token | was |
Token ID | 373 |
Feature activation | +2.46e-01 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
idea | +0.012 |
Jak | +0.012 |
Neg logit contributions | |
sidelines | -0.026 |
salon | -0.025 |
doorstep | -0.025 |
Sweet | -0.025 |
arrest | -0.025 |
Token | chasing |
Token ID | 20023 |
Feature activation | +7.23e-01 |
Pos logit contributions | |
sidelines | +0.008 |
doorstep | +0.008 |
salon | +0.008 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.005 |
roc | -0.005 |
Jak | -0.004 |
idea | -0.004 |
Token | the |
Token ID | 262 |
Feature activation | +5.43e-02 |
Pos logit contributions | |
arrest | +0.018 |
sidelines | +0.018 |
Sweet | +0.017 |
doorstep | +0.017 |
couch | +0.017 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.016 |
idea | -0.016 |
Jak | -0.016 |
roc | -0.015 |
Token | does |
Token ID | 857 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.010 |
roc | +0.009 |
idea | +0.008 |
boy | +0.008 |
Neg logit contributions | |
sidelines | -0.013 |
doorstep | -0.012 |
Sweet | -0.012 |
arrest | -0.012 |
uddled | -0.012 |
Token | not |
Token ID | 407 |
Feature activation | -5.26e-01 |
Pos logit contributions | |
ole | +0.033 |
ny | +0.030 |
roc | +0.028 |
idea | +0.028 |
Jak | +0.027 |
Neg logit contributions | |
doorstep | -0.033 |
Sweet | -0.032 |
sidelines | -0.032 |
uddled | -0.032 |
arrest | -0.031 |
Token | like |
Token ID | 588 |
Feature activation | +7.67e-01 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.013 |
roc | +0.012 |
Jak | +0.011 |
idea | +0.011 |
Neg logit contributions | |
salon | -0.014 |
sidelines | -0.014 |
doorstep | -0.014 |
arrest | -0.013 |
Sweet | -0.013 |
Token | exams |
Token ID | 26420 |
Feature activation | +8.67e-01 |
Pos logit contributions | |
doorstep | +0.024 |
sidelines | +0.024 |
salon | +0.024 |
Sweet | +0.024 |
arrest | +0.024 |
Neg logit contributions | |
ole | -0.016 |
ny | -0.014 |
roc | -0.013 |
Jak | -0.013 |
idea | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | +4.98e-02 |
Pos logit contributions | |
salon | +0.024 |
sidelines | +0.022 |
arrest | +0.021 |
doorstep | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ole | -0.023 |
roc | -0.021 |
ny | -0.020 |
Jak | -0.020 |
ac | -0.017 |
Token | Ex |
Token ID | 1475 |
Feature activation | -4.92e+00 |
Pos logit contributions | |
salon | +0.002 |
Sweet | +0.001 |
sidelines | +0.001 |
doorstep | +0.001 |
arrest | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
idea | -0.001 |
Jak | -0.001 |
Token | ams |
Token ID | 4105 |
Feature activation | -1.82e+00 |
Pos logit contributions | |
ole | +0.130 |
afloat | +0.125 |
idea | +0.123 |
boy | +0.116 |
suit | +0.111 |
Neg logit contributions | |
salon | -0.108 |
Sweet | -0.105 |
arrest | -0.104 |
atters | -0.102 |
anna | -0.099 |
Token | are |
Token ID | 389 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
ole | +0.051 |
ny | +0.047 |
roc | +0.044 |
Jak | +0.042 |
idea | +0.040 |
Neg logit contributions | |
salon | -0.046 |
sidelines | -0.045 |
doorstep | -0.044 |
arrest | -0.043 |
Sweet | -0.042 |
Token | hard |
Token ID | 1327 |
Feature activation | +8.58e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
boat | +0.001 |
Neg logit contributions | |
salon | -0.001 |
sidelines | -0.001 |
arrest | -0.001 |
freshly | -0.001 |
Sweet | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | +7.32e-02 |
Pos logit contributions | |
salon | +0.024 |
arrest | +0.023 |
doorstep | +0.023 |
sidelines | +0.022 |
Sweet | +0.022 |
Neg logit contributions | |
ole | -0.022 |
roc | -0.020 |
ny | -0.019 |
Jak | -0.018 |
boat | -0.017 |
Token | scary |
Token ID | 14343 |
Feature activation | -7.53e-01 |
Pos logit contributions | |
salon | +0.002 |
sidelines | +0.002 |
freshly | +0.002 |
Sweet | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.002 |
roc | -0.002 |
ny | -0.001 |
Jak | -0.001 |
ull | -0.001 |
Token | the |
Token ID | 262 |
Feature activation | -1.65e-02 |
Pos logit contributions | |
sidelines | +0.010 |
doorstep | +0.010 |
salon | +0.010 |
Sweet | +0.010 |
arrest | +0.010 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | cars |
Token ID | 5006 |
Feature activation | -3.46e-01 |
Pos logit contributions | |
ole | +0.000 |
ny | +0.000 |
roc | +0.000 |
Jak | +0.000 |
idea | +0.000 |
Neg logit contributions | |
sidelines | -0.000 |
salon | -0.000 |
doorstep | -0.000 |
Sweet | -0.000 |
arrest | -0.000 |
Token | stopped |
Token ID | 5025 |
Feature activation | -1.09e-02 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.007 |
Jak | +0.007 |
idea | +0.006 |
Neg logit contributions | |
salon | -0.010 |
sidelines | -0.010 |
doorstep | -0.010 |
arrest | -0.010 |
Sweet | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | -2.58e-01 |
Pos logit contributions | |
ole | +0.000 |
ny | +0.000 |
roc | +0.000 |
Jak | +0.000 |
idea | +0.000 |
Neg logit contributions | |
sidelines | -0.000 |
salon | -0.000 |
doorstep | -0.000 |
Sweet | -0.000 |
arrest | -0.000 |
Token | the |
Token ID | 262 |
Feature activation | +2.98e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.008 |
doorstep | -0.008 |
salon | -0.008 |
Sweet | -0.008 |
arrest | -0.007 |
Token | pedestrian |
Token ID | 22382 |
Feature activation | -4.83e+00 |
Pos logit contributions | |
salon | +0.008 |
sidelines | +0.008 |
doorstep | +0.008 |
Sweet | +0.008 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | light |
Token ID | 1657 |
Feature activation | -9.19e-01 |
Pos logit contributions | |
ole | +0.108 |
roc | +0.103 |
Jak | +0.101 |
ny | +0.096 |
boy | +0.086 |
Neg logit contributions | |
sidelines | -0.135 |
doorstep | -0.135 |
arrest | -0.128 |
freshly | -0.127 |
balcony | -0.127 |
Token | turned |
Token ID | 2900 |
Feature activation | -1.17e+00 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.020 |
roc | +0.019 |
Jak | +0.018 |
idea | +0.017 |
Neg logit contributions | |
doorstep | -0.027 |
salon | -0.027 |
sidelines | -0.026 |
arrest | -0.026 |
Sweet | -0.026 |
Token | green |
Token ID | 4077 |
Feature activation | +6.18e-01 |
Pos logit contributions | |
ole | +0.028 |
roc | +0.027 |
ny | +0.026 |
Jak | +0.024 |
idea | +0.023 |
Neg logit contributions | |
sidelines | -0.032 |
salon | -0.031 |
doorstep | -0.031 |
uddled | -0.031 |
arrest | -0.031 |
Token | . |
Token ID | 13 |
Feature activation | -2.54e-01 |
Pos logit contributions | |
sidelines | +0.018 |
doorstep | +0.018 |
salon | +0.018 |
arrest | +0.018 |
Sweet | +0.018 |
Neg logit contributions | |
ole | -0.015 |
ny | -0.013 |
roc | -0.012 |
Jak | -0.012 |
idea | -0.011 |
Token | Then |
Token ID | 3244 |
Feature activation | -6.28e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
Jak | +0.005 |
boy | +0.005 |
roc | +0.005 |
Neg logit contributions | |
arrest | -0.007 |
doorstep | -0.007 |
sidelines | -0.007 |
salon | -0.007 |
balcony | -0.007 |
Token | our |
Token ID | 674 |
Feature activation | +1.18e+00 |
Pos logit contributions | |
arrest | +0.014 |
sidelines | +0.014 |
salon | +0.013 |
freshly | +0.013 |
Sweet | +0.012 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.009 |
suit | -0.009 |
Jak | -0.008 |
Token | country |
Token ID | 1499 |
Feature activation | -4.35e-01 |
Pos logit contributions | |
sidelines | +0.032 |
doorstep | +0.031 |
salon | +0.031 |
arrest | +0.031 |
balcony | +0.029 |
Neg logit contributions | |
ole | -0.031 |
ny | -0.029 |
roc | -0.026 |
Jak | -0.026 |
idea | -0.025 |
Token | a |
Token ID | 257 |
Feature activation | -3.52e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
roc | +0.007 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
sidelines | -0.014 |
salon | -0.014 |
doorstep | -0.014 |
Sweet | -0.014 |
arrest | -0.013 |
Token | long |
Token ID | 890 |
Feature activation | -1.81e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.007 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.011 |
doorstep | -0.011 |
salon | -0.011 |
arrest | -0.011 |
Sweet | -0.011 |
Token | time |
Token ID | 640 |
Feature activation | -3.06e+00 |
Pos logit contributions | |
ole | +0.036 |
roc | +0.032 |
ny | +0.031 |
Jak | +0.029 |
idea | +0.026 |
Neg logit contributions | |
doorstep | -0.058 |
arrest | -0.058 |
sidelines | -0.058 |
salon | -0.058 |
balcony | -0.055 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.79e+00 |
Pos logit contributions | |
ole | +0.070 |
ny | +0.063 |
roc | +0.059 |
Jak | +0.055 |
idea | +0.054 |
Neg logit contributions | |
sidelines | -0.092 |
salon | -0.091 |
doorstep | -0.091 |
Sweet | -0.089 |
arrest | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | +2.28e-01 |
Pos logit contributions | |
ole | +0.092 |
ny | +0.082 |
roc | +0.074 |
Jak | +0.070 |
idea | +0.068 |
Neg logit contributions | |
sidelines | -0.161 |
doorstep | -0.159 |
salon | -0.159 |
Sweet | -0.158 |
arrest | -0.155 |
Token | He |
Token ID | 679 |
Feature activation | -6.05e-01 |
Pos logit contributions | |
sidelines | +0.006 |
salon | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
roc | -0.005 |
idea | -0.005 |
Jak | -0.005 |
Token | helped |
Token ID | 4193 |
Feature activation | +7.22e-01 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.013 |
roc | +0.012 |
Jak | +0.011 |
idea | +0.011 |
Neg logit contributions | |
salon | -0.018 |
doorstep | -0.018 |
sidelines | -0.017 |
arrest | -0.017 |
Sweet | -0.017 |
Token | many |
Token ID | 867 |
Feature activation | +5.90e-02 |
Pos logit contributions | |
arrest | +0.021 |
doorstep | +0.021 |
sidelines | +0.021 |
salon | +0.020 |
couch | +0.019 |
Neg logit contributions | |
ole | -0.016 |
ny | -0.015 |
suit | -0.014 |
roc | -0.013 |
Jak | -0.013 |
Token | people |
Token ID | 661 |
Feature activation | -7.72e-01 |
Pos logit contributions | |
arrest | +0.002 |
sidelines | +0.002 |
doorstep | +0.002 |
salon | +0.002 |
balcony | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
suit | -0.001 |
Token | born |
Token ID | 4642 |
Feature activation | -2.23e+00 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.003 |
idea | +0.003 |
Jak | +0.003 |
Neg logit contributions | |
salon | -0.005 |
doorstep | -0.005 |
sidelines | -0.005 |
Sweet | -0.005 |
arrest | -0.005 |
Token | only |
Token ID | 691 |
Feature activation | -2.31e+00 |
Pos logit contributions | |
ole | +0.045 |
ny | +0.041 |
roc | +0.040 |
Jak | +0.037 |
idea | +0.033 |
Neg logit contributions | |
sidelines | -0.070 |
doorstep | -0.070 |
salon | -0.070 |
arrest | -0.069 |
Sweet | -0.067 |
Token | a |
Token ID | 257 |
Feature activation | -9.40e-01 |
Pos logit contributions | |
ole | +0.054 |
ny | +0.048 |
roc | +0.047 |
Jak | +0.044 |
idea | +0.040 |
Neg logit contributions | |
salon | -0.067 |
sidelines | -0.067 |
doorstep | -0.067 |
arrest | -0.066 |
Sweet | -0.065 |
Token | few |
Token ID | 1178 |
Feature activation | -9.03e-01 |
Pos logit contributions | |
ole | +0.021 |
ny | +0.018 |
roc | +0.018 |
Jak | +0.017 |
idea | +0.016 |
Neg logit contributions | |
sidelines | -0.029 |
salon | -0.029 |
doorstep | -0.029 |
Sweet | -0.028 |
arrest | -0.028 |
Token | days |
Token ID | 1528 |
Feature activation | -7.98e-01 |
Pos logit contributions | |
ole | +0.021 |
ny | +0.019 |
roc | +0.017 |
Jak | +0.016 |
idea | +0.016 |
Neg logit contributions | |
sidelines | -0.027 |
salon | -0.027 |
doorstep | -0.027 |
Sweet | -0.026 |
arrest | -0.026 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.79e+00 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.012 |
Jak | +0.011 |
idea | +0.011 |
Neg logit contributions | |
sidelines | -0.027 |
doorstep | -0.027 |
salon | -0.027 |
Sweet | -0.026 |
arrest | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | +7.17e-01 |
Pos logit contributions | |
ole | +0.082 |
ny | +0.072 |
roc | +0.063 |
Jak | +0.059 |
idea | +0.058 |
Neg logit contributions | |
sidelines | -0.171 |
doorstep | -0.169 |
salon | -0.169 |
Sweet | -0.168 |
arrest | -0.165 |
Token | The |
Token ID | 383 |
Feature activation | +4.54e-01 |
Pos logit contributions | |
sidelines | +0.021 |
salon | +0.021 |
doorstep | +0.021 |
arrest | +0.020 |
Sweet | +0.020 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.016 |
roc | -0.014 |
Jak | -0.014 |
idea | -0.013 |
Token | deer |
Token ID | 20096 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
sidelines | +0.013 |
salon | +0.013 |
doorstep | +0.013 |
Sweet | +0.013 |
arrest | +0.013 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.009 |
Jak | -0.009 |
idea | -0.008 |
Token | felt |
Token ID | 2936 |
Feature activation | +1.05e+00 |
Pos logit contributions | |
sidelines | +0.046 |
salon | +0.046 |
doorstep | +0.045 |
Sweet | +0.045 |
arrest | +0.044 |
Neg logit contributions | |
ole | -0.031 |
ny | -0.029 |
roc | -0.026 |
Jak | -0.024 |
idea | -0.024 |
Token | sorry |
Token ID | 7926 |
Feature activation | +1.45e+00 |
Pos logit contributions | |
sidelines | +0.035 |
doorstep | +0.034 |
salon | +0.033 |
Sweet | +0.033 |
arrest | +0.033 |
Neg logit contributions | |
ole | -0.021 |
ny | -0.018 |
roc | -0.017 |
Jak | -0.016 |
idea | -0.015 |
Token | big |
Token ID | 1263 |
Feature activation | +1.41e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
doorstep | -0.008 |
salon | -0.008 |
sidelines | -0.008 |
Sweet | -0.008 |
arrest | -0.008 |
Token | farm |
Token ID | 5318 |
Feature activation | -1.07e-01 |
Pos logit contributions | |
salon | +0.004 |
doorstep | +0.004 |
sidelines | +0.004 |
arrest | +0.004 |
balcony | +0.004 |
Neg logit contributions | |
ole | -0.004 |
roc | -0.003 |
ny | -0.003 |
Jak | -0.003 |
ac | -0.003 |
Token | a |
Token ID | 257 |
Feature activation | -3.02e-01 |
Pos logit contributions | |
ole | +0.003 |
roc | +0.003 |
ny | +0.002 |
Jak | +0.002 |
ac | +0.002 |
Neg logit contributions | |
salon | -0.003 |
doorstep | -0.003 |
arrest | -0.003 |
sidelines | -0.003 |
balcony | -0.003 |
Token | long |
Token ID | 890 |
Feature activation | -1.52e+00 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.010 |
doorstep | -0.010 |
salon | -0.010 |
Sweet | -0.009 |
arrest | -0.009 |
Token | time |
Token ID | 640 |
Feature activation | -2.98e+00 |
Pos logit contributions | |
ole | +0.028 |
roc | +0.024 |
ny | +0.024 |
Jak | +0.022 |
ac | +0.019 |
Neg logit contributions | |
doorstep | -0.052 |
salon | -0.051 |
arrest | -0.051 |
sidelines | -0.051 |
balcony | -0.049 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.76e+00 |
Pos logit contributions | |
ole | +0.071 |
ny | +0.065 |
roc | +0.060 |
Jak | +0.058 |
idea | +0.057 |
Neg logit contributions | |
sidelines | -0.086 |
doorstep | -0.085 |
salon | -0.085 |
Sweet | -0.084 |
arrest | -0.082 |
Token | . |
Token ID | 13 |
Feature activation | -5.00e-01 |
Pos logit contributions | |
ole | +0.098 |
ny | +0.088 |
roc | +0.080 |
Jak | +0.075 |
idea | +0.074 |
Neg logit contributions | |
sidelines | -0.153 |
salon | -0.152 |
doorstep | -0.152 |
Sweet | -0.150 |
arrest | -0.148 |
Token | Lily |
Token ID | 20037 |
Feature activation | -7.60e-02 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
Jak | +0.008 |
roc | +0.008 |
boy | +0.008 |
Neg logit contributions | |
doorstep | -0.017 |
sidelines | -0.016 |
arrest | -0.016 |
salon | -0.016 |
balcony | -0.016 |
Token | listened |
Token ID | 16399 |
Feature activation | -8.48e-02 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
salon | -0.002 |
doorstep | -0.002 |
sidelines | -0.002 |
Sweet | -0.002 |
arrest | -0.002 |
Token | as |
Token ID | 355 |
Feature activation | +3.17e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
sidelines | -0.003 |
salon | -0.003 |
doorstep | -0.003 |
Sweet | -0.002 |
arrest | -0.002 |
Token | the |
Token ID | 262 |
Feature activation | -2.22e-01 |
Pos logit contributions | |
sidelines | +0.007 |
doorstep | +0.007 |
salon | +0.007 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.009 |
roc | -0.008 |
Jak | -0.008 |
idea | -0.008 |
Token | lived |
Token ID | 5615 |
Feature activation | -7.48e-01 |
Pos logit contributions | |
sidelines | +0.022 |
salon | +0.021 |
doorstep | +0.021 |
Sweet | +0.021 |
arrest | +0.020 |
Neg logit contributions | |
ny | -0.014 |
ole | -0.014 |
roc | -0.012 |
boy | -0.011 |
idea | -0.011 |
Token | here |
Token ID | 994 |
Feature activation | -2.06e+00 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.014 |
roc | +0.014 |
Jak | +0.013 |
idea | +0.012 |
Neg logit contributions | |
doorstep | -0.023 |
salon | -0.023 |
sidelines | -0.023 |
Sweet | -0.023 |
arrest | -0.022 |
Token | a |
Token ID | 257 |
Feature activation | -4.12e-01 |
Pos logit contributions | |
ole | +0.041 |
ny | +0.037 |
idea | +0.030 |
roc | +0.030 |
Jak | +0.029 |
Neg logit contributions | |
sidelines | -0.070 |
doorstep | -0.068 |
salon | -0.068 |
arrest | -0.068 |
Sweet | -0.067 |
Token | long |
Token ID | 890 |
Feature activation | -1.84e+00 |
Pos logit contributions | |
ole | +0.009 |
roc | +0.008 |
ny | +0.008 |
Jak | +0.008 |
ac | +0.007 |
Neg logit contributions | |
sidelines | -0.013 |
doorstep | -0.012 |
salon | -0.012 |
balcony | -0.012 |
Sweet | -0.012 |
Token | time |
Token ID | 640 |
Feature activation | -3.09e+00 |
Pos logit contributions | |
ole | +0.036 |
roc | +0.035 |
Jak | +0.032 |
ny | +0.029 |
ac | +0.027 |
Neg logit contributions | |
arrest | -0.059 |
doorstep | -0.058 |
sidelines | -0.058 |
salon | -0.057 |
couch | -0.055 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.75e+00 |
Pos logit contributions | |
ole | +0.063 |
ny | +0.057 |
idea | +0.049 |
roc | +0.048 |
Jak | +0.048 |
Neg logit contributions | |
sidelines | -0.100 |
salon | -0.100 |
doorstep | -0.099 |
Sweet | -0.099 |
uddled | -0.097 |
Token | ." |
Token ID | 526 |
Feature activation | +2.03e-01 |
Pos logit contributions | |
ole | +0.090 |
ny | +0.079 |
roc | +0.071 |
Jak | +0.067 |
idea | +0.066 |
Neg logit contributions | |
sidelines | -0.162 |
Sweet | -0.160 |
salon | -0.159 |
doorstep | -0.159 |
arrest | -0.155 |
Token |
|
Token ID | 198 |
Feature activation | +2.67e-01 |
Pos logit contributions | |
sidelines | +0.006 |
doorstep | +0.006 |
salon | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.005 |
roc | -0.004 |
Jak | -0.004 |
idea | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +3.34e-01 |
Pos logit contributions | |
sidelines | +0.007 |
salon | +0.007 |
doorstep | +0.007 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.006 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.005 |
Token | " |
Token ID | 1 |
Feature activation | +3.24e-01 |
Pos logit contributions | |
doorstep | +0.009 |
sidelines | +0.009 |
salon | +0.009 |
uddled | +0.009 |
arrest | +0.009 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.008 |
roc | -0.008 |
Jak | -0.007 |
idea | -0.007 |
Token | Can |
Token ID | 6090 |
Feature activation | +9.75e-01 |
Pos logit contributions | |
doorstep | +0.007 |
sidelines | +0.007 |
salon | +0.007 |
arrest | +0.007 |
Sweet | +0.007 |
Neg logit contributions | |
ole | -0.010 |
ny | -0.009 |
roc | -0.009 |
Jak | -0.008 |
idea | -0.008 |
Token | story |
Token ID | 1621 |
Feature activation | -3.07e+00 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.003 |
roc | +0.003 |
Jak | +0.003 |
idea | +0.003 |
Neg logit contributions | |
doorstep | -0.005 |
salon | -0.005 |
sidelines | -0.005 |
Sweet | -0.005 |
arrest | -0.005 |
Token | . |
Token ID | 13 |
Feature activation | +7.79e-01 |
Pos logit contributions | |
ole | +0.086 |
ny | +0.079 |
roc | +0.074 |
Jak | +0.071 |
idea | +0.070 |
Neg logit contributions | |
sidelines | -0.077 |
doorstep | -0.076 |
salon | -0.076 |
Sweet | -0.074 |
arrest | -0.074 |
Token | A |
Token ID | 317 |
Feature activation | -1.24e+00 |
Pos logit contributions | |
sidelines | +0.017 |
salon | +0.017 |
doorstep | +0.017 |
Sweet | +0.017 |
arrest | +0.016 |
Neg logit contributions | |
ole | -0.024 |
ny | -0.022 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.020 |
Token | long |
Token ID | 890 |
Feature activation | -1.54e+00 |
Pos logit contributions | |
ole | +0.029 |
ny | +0.027 |
roc | +0.026 |
Jak | +0.024 |
ac | +0.023 |
Neg logit contributions | |
salon | -0.037 |
doorstep | -0.035 |
sidelines | -0.035 |
Sweet | -0.035 |
balcony | -0.034 |
Token | time |
Token ID | 640 |
Feature activation | -2.85e+00 |
Pos logit contributions | |
ole | +0.037 |
roc | +0.033 |
ny | +0.033 |
Jak | +0.031 |
wood | +0.029 |
Neg logit contributions | |
salon | -0.044 |
arrest | -0.043 |
doorstep | -0.043 |
sidelines | -0.043 |
balcony | -0.041 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.71e+00 |
Pos logit contributions | |
ole | +0.052 |
ny | +0.047 |
idea | +0.042 |
roc | +0.040 |
Jak | +0.039 |
Neg logit contributions | |
sidelines | -0.099 |
doorstep | -0.098 |
salon | -0.098 |
Sweet | -0.097 |
uddled | -0.095 |
Token | , |
Token ID | 11 |
Feature activation | +3.52e-01 |
Pos logit contributions | |
ole | +0.073 |
ny | +0.063 |
roc | +0.055 |
Jak | +0.051 |
idea | +0.050 |
Neg logit contributions | |
sidelines | -0.177 |
Sweet | -0.174 |
doorstep | -0.174 |
salon | -0.174 |
arrest | -0.169 |
Token | there |
Token ID | 612 |
Feature activation | -1.53e+00 |
Pos logit contributions | |
sidelines | +0.012 |
salon | +0.012 |
doorstep | +0.012 |
Sweet | +0.012 |
arrest | +0.012 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.006 |
roc | -0.005 |
Jak | -0.005 |
idea | -0.005 |
Token | was |
Token ID | 373 |
Feature activation | -1.26e+00 |
Pos logit contributions | |
ole | +0.041 |
ny | +0.038 |
roc | +0.036 |
Jak | +0.034 |
idea | +0.033 |
Neg logit contributions | |
salon | -0.039 |
doorstep | -0.039 |
sidelines | -0.039 |
Sweet | -0.038 |
arrest | -0.038 |
Token | an |
Token ID | 281 |
Feature activation | +5.03e-02 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.016 |
roc | +0.014 |
Jak | +0.013 |
idea | +0.012 |
Neg logit contributions | |
sidelines | -0.048 |
doorstep | -0.048 |
salon | -0.048 |
Sweet | -0.047 |
arrest | -0.047 |
Token | adventurous |
Token ID | 39770 |
Feature activation | -4.56e-01 |
Pos logit contributions | |
doorstep | +0.001 |
sidelines | +0.001 |
salon | +0.001 |
arrest | +0.001 |
Sweet | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
suit | -0.001 |
Token | is |
Token ID | 318 |
Feature activation | -1.26e+00 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.028 |
roc | +0.027 |
Jak | +0.026 |
idea | +0.025 |
Neg logit contributions | |
salon | -0.025 |
doorstep | -0.024 |
arrest | -0.023 |
sidelines | -0.023 |
uddled | -0.023 |
Token | from |
Token ID | 422 |
Feature activation | -2.70e-01 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.028 |
roc | +0.027 |
Jak | +0.024 |
idea | +0.023 |
Neg logit contributions | |
doorstep | -0.036 |
sidelines | -0.035 |
salon | -0.035 |
arrest | -0.035 |
Sweet | -0.034 |
Token | a |
Token ID | 257 |
Feature activation | -2.61e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
Sweet | -0.008 |
arrest | -0.008 |
Token | long |
Token ID | 890 |
Feature activation | -1.83e+00 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.008 |
doorstep | -0.008 |
salon | -0.008 |
arrest | -0.008 |
balcony | -0.007 |
Token | time |
Token ID | 640 |
Feature activation | -2.70e+00 |
Pos logit contributions | |
roc | +0.037 |
ole | +0.036 |
Jak | +0.034 |
ny | +0.031 |
ac | +0.029 |
Neg logit contributions | |
arrest | -0.060 |
doorstep | -0.057 |
sidelines | -0.055 |
salon | -0.054 |
balcony | -0.051 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.67e+00 |
Pos logit contributions | |
ole | +0.060 |
ny | +0.055 |
roc | +0.048 |
idea | +0.047 |
Jak | +0.047 |
Neg logit contributions | |
sidelines | -0.082 |
Sweet | -0.082 |
salon | -0.081 |
doorstep | -0.080 |
uddled | -0.079 |
Token | , |
Token ID | 11 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
ole | +0.087 |
ny | +0.076 |
roc | +0.068 |
Jak | +0.064 |
idea | +0.064 |
Neg logit contributions | |
sidelines | -0.161 |
Sweet | -0.159 |
salon | -0.158 |
doorstep | -0.158 |
arrest | -0.154 |
Token | from |
Token ID | 422 |
Feature activation | -6.97e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
sidelines | -0.003 |
doorstep | -0.003 |
salon | -0.003 |
Sweet | -0.003 |
arrest | -0.003 |
Token | when |
Token ID | 618 |
Feature activation | -7.78e-01 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.012 |
idea | +0.012 |
Neg logit contributions | |
sidelines | -0.021 |
doorstep | -0.021 |
salon | -0.021 |
Sweet | -0.021 |
arrest | -0.020 |
Token | dinosaurs |
Token ID | 27199 |
Feature activation | -6.69e-01 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.020 |
Jak | +0.019 |
roc | +0.018 |
idea | +0.018 |
Neg logit contributions | |
sidelines | -0.019 |
doorstep | -0.019 |
Sweet | -0.019 |
salon | -0.018 |
uddled | -0.018 |
Token | lived |
Token ID | 5615 |
Feature activation | -1.44e+00 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.012 |
idea | +0.012 |
Neg logit contributions | |
sidelines | -0.020 |
doorstep | -0.020 |
Sweet | -0.020 |
salon | -0.020 |
arrest | -0.019 |
Token | had |
Token ID | 550 |
Feature activation | -1.26e+00 |
Pos logit contributions | |
ole | +0.029 |
ny | +0.026 |
roc | +0.024 |
Jak | +0.023 |
idea | +0.022 |
Neg logit contributions | |
salon | -0.041 |
doorstep | -0.041 |
sidelines | -0.040 |
Sweet | -0.040 |
arrest | -0.039 |
Token | bought |
Token ID | 5839 |
Feature activation | -8.48e-01 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.026 |
roc | +0.023 |
Jak | +0.023 |
idea | +0.022 |
Neg logit contributions | |
sidelines | -0.039 |
Sweet | -0.038 |
salon | -0.037 |
doorstep | -0.037 |
Acc | -0.036 |
Token | a |
Token ID | 257 |
Feature activation | -7.79e-01 |
Pos logit contributions | |
ole | +0.021 |
roc | +0.021 |
ny | +0.020 |
Jak | +0.018 |
zo | +0.017 |
Neg logit contributions | |
doorstep | -0.021 |
salon | -0.020 |
arrest | -0.020 |
freshly | -0.020 |
sidelines | -0.020 |
Token | long |
Token ID | 890 |
Feature activation | -2.06e+00 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.016 |
roc | +0.014 |
idea | +0.014 |
boy | +0.014 |
Neg logit contributions | |
sidelines | -0.023 |
Sweet | -0.023 |
arrest | -0.023 |
uddled | -0.023 |
salon | -0.023 |
Token | time |
Token ID | 640 |
Feature activation | -2.60e+00 |
Pos logit contributions | |
ole | +0.048 |
ny | +0.044 |
roc | +0.042 |
Jak | +0.039 |
idea | +0.036 |
Neg logit contributions | |
doorstep | -0.061 |
salon | -0.060 |
sidelines | -0.060 |
arrest | -0.059 |
Sweet | -0.058 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.67e+00 |
Pos logit contributions | |
ole | +0.056 |
ny | +0.050 |
roc | +0.045 |
Jak | +0.043 |
idea | +0.043 |
Neg logit contributions | |
sidelines | -0.082 |
salon | -0.081 |
doorstep | -0.080 |
Sweet | -0.080 |
arrest | -0.078 |
Token | . |
Token ID | 13 |
Feature activation | -4.77e-01 |
Pos logit contributions | |
ole | +0.099 |
ny | +0.089 |
roc | +0.081 |
Jak | +0.076 |
idea | +0.075 |
Neg logit contributions | |
sidelines | -0.148 |
salon | -0.146 |
doorstep | -0.146 |
Sweet | -0.145 |
arrest | -0.143 |
Token | They |
Token ID | 1119 |
Feature activation | +9.94e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
roc | +0.011 |
Jak | +0.010 |
idea | +0.010 |
Neg logit contributions | |
sidelines | -0.013 |
salon | -0.013 |
doorstep | -0.013 |
Sweet | -0.013 |
arrest | -0.012 |
Token | did |
Token ID | 750 |
Feature activation | -6.70e-01 |
Pos logit contributions | |
salon | +0.030 |
sidelines | +0.030 |
doorstep | +0.030 |
Sweet | +0.029 |
arrest | +0.029 |
Neg logit contributions | |
ole | -0.023 |
ny | -0.020 |
roc | -0.019 |
Jak | -0.018 |
idea | -0.017 |
Token | not |
Token ID | 407 |
Feature activation | +2.99e-01 |
Pos logit contributions | |
ole | +0.016 |
idea | +0.014 |
ny | +0.013 |
roc | +0.013 |
Jak | +0.013 |
Neg logit contributions | |
sidelines | -0.020 |
Sweet | -0.020 |
doorstep | -0.019 |
salon | -0.019 |
uddled | -0.019 |
Token | know |
Token ID | 760 |
Feature activation | +5.52e-01 |
Pos logit contributions | |
arrest | +0.007 |
salon | +0.007 |
sidelines | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
Neg logit contributions | |
ole | -0.009 |
roc | -0.008 |
ny | -0.008 |
Jak | -0.008 |
beard | -0.007 |
Token | it |
Token ID | 340 |
Feature activation | +3.01e-01 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.016 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.025 |
doorstep | -0.024 |
salon | -0.024 |
arrest | -0.024 |
Sweet | -0.024 |
Token | here |
Token ID | 994 |
Feature activation | -2.06e+00 |
Pos logit contributions | |
sidelines | +0.011 |
Sweet | +0.010 |
doorstep | +0.010 |
salon | +0.010 |
uddled | +0.010 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
ac | -0.004 |
idea | -0.004 |
Jak | -0.004 |
Token | a |
Token ID | 257 |
Feature activation | -9.07e-01 |
Pos logit contributions | |
ole | +0.051 |
ny | +0.047 |
roc | +0.042 |
idea | +0.041 |
Jak | +0.040 |
Neg logit contributions | |
sidelines | -0.059 |
salon | -0.057 |
doorstep | -0.057 |
Sweet | -0.057 |
arrest | -0.056 |
Token | long |
Token ID | 890 |
Feature activation | -2.20e+00 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.020 |
roc | +0.020 |
Jak | +0.018 |
idea | +0.017 |
Neg logit contributions | |
salon | -0.026 |
doorstep | -0.026 |
sidelines | -0.026 |
arrest | -0.024 |
balcony | -0.024 |
Token | time |
Token ID | 640 |
Feature activation | -2.84e+00 |
Pos logit contributions | |
ole | +0.045 |
roc | +0.044 |
Jak | +0.038 |
ny | +0.036 |
wood | +0.035 |
Neg logit contributions | |
arrest | -0.071 |
salon | -0.069 |
doorstep | -0.068 |
sidelines | -0.066 |
couch | -0.064 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.67e+00 |
Pos logit contributions | |
ole | +0.078 |
ny | +0.072 |
roc | +0.067 |
Jak | +0.064 |
idea | +0.063 |
Neg logit contributions | |
sidelines | -0.072 |
salon | -0.071 |
doorstep | -0.071 |
Sweet | -0.070 |
arrest | -0.069 |
Token | ." |
Token ID | 526 |
Feature activation | -6.70e-01 |
Pos logit contributions | |
ole | +0.116 |
ny | +0.107 |
roc | +0.102 |
Jak | +0.095 |
idea | +0.092 |
Neg logit contributions | |
salon | -0.128 |
doorstep | -0.128 |
sidelines | -0.127 |
arrest | -0.125 |
Sweet | -0.123 |
Token |
|
Token ID | 198 |
Feature activation | -2.57e-01 |
Pos logit contributions | |
ole | +0.013 |
ny | +0.013 |
roc | +0.012 |
Jak | +0.012 |
boy | +0.011 |
Neg logit contributions | |
salon | -0.022 |
doorstep | -0.022 |
sidelines | -0.022 |
arrest | -0.021 |
uddled | -0.021 |
Token |
|
Token ID | 198 |
Feature activation | -9.30e-02 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.004 |
boy | +0.004 |
Neg logit contributions | |
salon | -0.009 |
sidelines | -0.009 |
doorstep | -0.009 |
arrest | -0.009 |
uddled | -0.008 |
Token | They |
Token ID | 2990 |
Feature activation | +4.81e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
doorstep | -0.003 |
salon | -0.003 |
sidelines | -0.003 |
arrest | -0.003 |
Sweet | -0.003 |
Token | decided |
Token ID | 3066 |
Feature activation | +1.16e-01 |
Pos logit contributions | |
doorstep | +0.013 |
salon | +0.013 |
arrest | +0.012 |
sidelines | +0.012 |
balcony | +0.012 |
Neg logit contributions | |
ole | -0.013 |
ny | -0.012 |
roc | -0.012 |
idea | -0.011 |
Jak | -0.010 |
Token | if |
Token ID | 611 |
Feature activation | -7.82e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.010 |
roc | +0.009 |
Jak | +0.008 |
idea | +0.008 |
Neg logit contributions | |
salon | -0.014 |
sidelines | -0.014 |
Sweet | -0.014 |
doorstep | -0.014 |
arrest | -0.014 |
Token | to |
Token ID | 284 |
Feature activation | -7.62e-01 |
Pos logit contributions | |
ole | +0.020 |
ny | +0.018 |
roc | +0.018 |
idea | +0.015 |
Jak | +0.015 |
Neg logit contributions | |
salon | -0.022 |
arrest | -0.021 |
sidelines | -0.021 |
doorstep | -0.021 |
Sweet | -0.021 |
Token | say |
Token ID | 910 |
Feature activation | -1.70e+00 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.018 |
idea | +0.016 |
roc | +0.015 |
Jak | +0.015 |
Neg logit contributions | |
Sweet | -0.021 |
uddled | -0.021 |
sidelines | -0.020 |
doorstep | -0.020 |
Acc | -0.020 |
Token | , |
Token ID | 11 |
Feature activation | -7.41e-01 |
Pos logit contributions | |
ole | +0.042 |
ny | +0.039 |
roc | +0.037 |
Jak | +0.033 |
idea | +0.031 |
Neg logit contributions | |
salon | -0.048 |
sidelines | -0.046 |
Sweet | -0.046 |
doorstep | -0.045 |
arrest | -0.045 |
Token | " |
Token ID | 366 |
Feature activation | +7.24e-01 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.016 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.014 |
Neg logit contributions | |
doorstep | -0.021 |
sidelines | -0.021 |
salon | -0.021 |
arrest | -0.021 |
Sweet | -0.021 |
Token | sh |
Token ID | 1477 |
Feature activation | -4.66e+00 |
Pos logit contributions | |
sidelines | +0.017 |
doorstep | +0.017 |
salon | +0.016 |
arrest | +0.016 |
Sweet | +0.016 |
Neg logit contributions | |
ole | -0.022 |
ny | -0.020 |
roc | -0.019 |
Jak | -0.018 |
idea | -0.018 |
Token | hhhh |
Token ID | 36607 |
Feature activation | -1.87e+00 |
Pos logit contributions | |
ole | +0.128 |
ny | +0.120 |
roc | +0.113 |
Jak | +0.110 |
idea | +0.106 |
Neg logit contributions | |
salon | -0.114 |
sidelines | -0.114 |
doorstep | -0.113 |
Sweet | -0.112 |
arrest | -0.108 |
Token | ". |
Token ID | 1911 |
Feature activation | +2.27e-01 |
Pos logit contributions | |
ole | +0.051 |
ny | +0.048 |
roc | +0.047 |
Jak | +0.046 |
idea | +0.045 |
Neg logit contributions | |
salon | -0.045 |
sidelines | -0.045 |
Sweet | -0.043 |
arrest | -0.043 |
doorstep | -0.042 |
Token |
|
Token ID | 198 |
Feature activation | +2.62e-01 |
Pos logit contributions | |
sidelines | +0.007 |
doorstep | +0.007 |
salon | +0.007 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.005 |
Jak | -0.004 |
roc | -0.004 |
idea | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +9.31e-02 |
Pos logit contributions | |
sidelines | +0.008 |
doorstep | +0.008 |
salon | +0.008 |
arrest | +0.008 |
uddled | +0.007 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.006 |
roc | -0.005 |
Jak | -0.005 |
idea | -0.005 |
Token | Mom |
Token ID | 29252 |
Feature activation | -4.21e-01 |
Pos logit contributions | |
doorstep | +0.002 |
salon | +0.002 |
sidelines | +0.002 |
arrest | +0.002 |
uddled | +0.002 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.002 |
roc | -0.002 |
Jak | -0.002 |
ac | -0.002 |
Token | me |
Token ID | 502 |
Feature activation | +6.15e-01 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.015 |
roc | +0.015 |
idea | +0.014 |
Jak | +0.013 |
Neg logit contributions | |
arrest | -0.015 |
sidelines | -0.015 |
freshly | -0.014 |
salon | -0.014 |
Sweet | -0.014 |
Token | here |
Token ID | 994 |
Feature activation | -1.36e+00 |
Pos logit contributions | |
salon | +0.020 |
doorstep | +0.020 |
sidelines | +0.020 |
Sweet | +0.020 |
arrest | +0.020 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.010 |
roc | -0.010 |
Jak | -0.009 |
idea | -0.009 |
Token | a |
Token ID | 257 |
Feature activation | -5.85e-01 |
Pos logit contributions | |
ole | +0.030 |
ny | +0.027 |
roc | +0.025 |
Jak | +0.023 |
idea | +0.023 |
Neg logit contributions | |
sidelines | -0.042 |
salon | -0.042 |
doorstep | -0.042 |
Sweet | -0.041 |
arrest | -0.041 |
Token | long |
Token ID | 890 |
Feature activation | -2.19e+00 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.012 |
idea | +0.012 |
Neg logit contributions | |
salon | -0.016 |
sidelines | -0.016 |
doorstep | -0.016 |
Sweet | -0.015 |
arrest | -0.015 |
Token | time |
Token ID | 640 |
Feature activation | -2.62e+00 |
Pos logit contributions | |
ole | +0.056 |
roc | +0.053 |
ny | +0.050 |
Jak | +0.049 |
idea | +0.044 |
Neg logit contributions | |
salon | -0.058 |
arrest | -0.058 |
doorstep | -0.056 |
sidelines | -0.055 |
balcony | -0.053 |
Token | ago |
Token ID | 2084 |
Feature activation | -4.66e+00 |
Pos logit contributions | |
ole | +0.063 |
ny | +0.057 |
roc | +0.052 |
Jak | +0.050 |
idea | +0.050 |
Neg logit contributions | |
sidelines | -0.076 |
salon | -0.075 |
doorstep | -0.075 |
Sweet | -0.075 |
arrest | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | +4.61e-01 |
Pos logit contributions | |
ole | +0.097 |
ny | +0.088 |
roc | +0.082 |
Jak | +0.076 |
idea | +0.072 |
Neg logit contributions | |
salon | -0.147 |
doorstep | -0.147 |
sidelines | -0.146 |
arrest | -0.144 |
Sweet | -0.142 |
Token | They |
Token ID | 1119 |
Feature activation | +7.33e-01 |
Pos logit contributions | |
sidelines | +0.012 |
salon | +0.011 |
doorstep | +0.011 |
arrest | +0.011 |
uddled | +0.011 |
Neg logit contributions | |
ole | -0.013 |
ny | -0.012 |
roc | -0.011 |
Jak | -0.011 |
boy | -0.011 |
Token | forgot |
Token ID | 16453 |
Feature activation | +1.33e+00 |
Pos logit contributions | |
sidelines | +0.021 |
Sweet | +0.021 |
salon | +0.021 |
doorstep | +0.021 |
uddled | +0.020 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.016 |
roc | -0.015 |
idea | -0.014 |
Jak | -0.014 |
Token | about |
Token ID | 546 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
sidelines | +0.051 |
� | +0.050 |
salon | +0.048 |
doorstep | +0.048 |
Sweet | +0.047 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.017 |
boy | -0.016 |
idea | -0.015 |
roc | -0.015 |
Token | me |
Token ID | 502 |
Feature activation | +6.75e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
doorstep | -0.005 |
sidelines | -0.005 |
salon | -0.005 |
Sweet | -0.005 |
uddled | -0.005 |
Token | you |
Token ID | 345 |
Feature activation | -6.53e-01 |
Pos logit contributions | |
ole | +0.029 |
ny | +0.027 |
roc | +0.026 |
Jak | +0.025 |
idea | +0.025 |
Neg logit contributions | |
sidelines | -0.018 |
doorstep | -0.017 |
salon | -0.017 |
Sweet | -0.017 |
arrest | -0.017 |
Token | want |
Token ID | 765 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.015 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.013 |
Neg logit contributions | |
doorstep | -0.018 |
sidelines | -0.018 |
salon | -0.017 |
arrest | -0.017 |
Sweet | -0.017 |
Token | to |
Token ID | 284 |
Feature activation | +9.70e-02 |
Pos logit contributions | |
ole | +0.024 |
ny | +0.022 |
roc | +0.019 |
Jak | +0.019 |
idea | +0.018 |
Neg logit contributions | |
sidelines | -0.037 |
Acc | -0.036 |
Sweet | -0.036 |
salon | -0.036 |
doorstep | -0.035 |
Token | go |
Token ID | 467 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
Sweet | +0.003 |
sidelines | +0.003 |
Acc | +0.003 |
uddled | +0.003 |
salon | +0.003 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.002 |
idea | -0.002 |
roc | -0.002 |
Jak | -0.002 |
Token | , |
Token ID | 11 |
Feature activation | -2.76e-01 |
Pos logit contributions | |
ole | +0.031 |
ny | +0.029 |
roc | +0.027 |
Jak | +0.026 |
idea | +0.025 |
Neg logit contributions | |
sidelines | -0.027 |
salon | -0.027 |
doorstep | -0.027 |
Sweet | -0.026 |
arrest | -0.026 |
Token | Blank |
Token ID | 31990 |
Feature activation | -4.65e+00 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.007 |
Jak | +0.006 |
beard | +0.006 |
Neg logit contributions | |
salon | -0.007 |
doorstep | -0.007 |
sidelines | -0.007 |
Sweet | -0.007 |
couch | -0.006 |
Token | et |
Token ID | 316 |
Feature activation | -4.01e-01 |
Pos logit contributions | |
ole | +0.122 |
Jak | +0.109 |
roc | +0.105 |
ny | +0.102 |
boy | +0.100 |
Neg logit contributions | |
Acc | -0.112 |
doorstep | -0.110 |
salon | -0.105 |
Sweet | -0.104 |
balcony | -0.100 |
Token | ?" |
Token ID | 1701 |
Feature activation | +1.64e-02 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.010 |
roc | +0.010 |
Jak | +0.010 |
idea | +0.009 |
Neg logit contributions | |
salon | -0.010 |
doorstep | -0.009 |
sidelines | -0.009 |
arrest | -0.009 |
Sweet | -0.009 |
Token | |
Token ID | 220 |
Feature activation | +5.53e-01 |
Pos logit contributions | |
sidelines | +0.001 |
uddled | +0.001 |
salon | +0.001 |
doorstep | +0.001 |
arrest | +0.001 |
Neg logit contributions | |
ny | -0.000 |
ole | -0.000 |
Jak | -0.000 |
boy | -0.000 |
roc | -0.000 |
Token |
|
Token ID | 198 |
Feature activation | -3.11e-01 |
Pos logit contributions | |
sidelines | +0.019 |
salon | +0.018 |
arrest | +0.018 |
doorstep | +0.018 |
Sweet | +0.017 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
Jak | -0.009 |
roc | -0.009 |
idea | -0.009 |
Token | The |
Token ID | 464 |
Feature activation | -3.30e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.007 |
roc | +0.006 |
Jak | +0.006 |
boy | +0.005 |
Neg logit contributions | |
sidelines | -0.010 |
salon | -0.010 |
doorstep | -0.009 |
arrest | -0.009 |
uddled | -0.009 |
Token | � |
Token ID | 95 |
Feature activation | -6.24e-01 |
Pos logit contributions | |
doorstep | +0.029 |
sidelines | +0.029 |
salon | +0.029 |
arrest | +0.028 |
Sweet | +0.027 |
Neg logit contributions | |
ole | -0.026 |
ny | -0.024 |
roc | -0.022 |
idea | -0.021 |
Jak | -0.021 |
Token | € |
Token ID | 26391 |
Feature activation | +5.43e-01 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.016 |
roc | +0.015 |
idea | +0.014 |
Jak | +0.014 |
Neg logit contributions | |
doorstep | -0.016 |
sidelines | -0.016 |
salon | -0.015 |
arrest | -0.015 |
Sweet | -0.015 |
Token | ˜ |
Token ID | 41185 |
Feature activation | -1.35e+00 |
Pos logit contributions | |
doorstep | +0.013 |
arrest | +0.013 |
salon | +0.013 |
sidelines | +0.013 |
Sweet | +0.013 |
Neg logit contributions | |
ole | -0.016 |
ny | -0.014 |
roc | -0.013 |
idea | -0.013 |
Jak | -0.013 |
Token | Come |
Token ID | 16773 |
Feature activation | -9.16e-01 |
Pos logit contributions | |
ole | +0.031 |
roc | +0.029 |
ny | +0.028 |
idea | +0.026 |
Jak | +0.026 |
Neg logit contributions | |
doorstep | -0.038 |
salon | -0.037 |
uddled | -0.036 |
arrest | -0.035 |
balcony | -0.035 |
Token | and |
Token ID | 290 |
Feature activation | -2.79e-01 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.016 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.030 |
doorstep | -0.030 |
salon | -0.030 |
Sweet | -0.030 |
arrest | -0.030 |
Token | Ex |
Token ID | 1475 |
Feature activation | -4.64e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.007 |
Jak | +0.006 |
roc | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.008 |
Sweet | -0.008 |
salon | -0.008 |
uddled | -0.008 |
doorstep | -0.007 |
Token | amine |
Token ID | 9862 |
Feature activation | -1.36e+00 |
Pos logit contributions | |
idea | +0.120 |
ole | +0.106 |
boy | +0.105 |
afloat | +0.101 |
roc | +0.090 |
Neg logit contributions | |
arrest | -0.111 |
Sweet | -0.111 |
Her | -0.107 |
uddled | -0.105 |
salon | -0.103 |
Token | ! |
Token ID | 0 |
Feature activation | +1.75e-01 |
Pos logit contributions | |
ole | +0.032 |
ny | +0.031 |
roc | +0.030 |
idea | +0.027 |
Jak | +0.026 |
Neg logit contributions | |
salon | -0.035 |
doorstep | -0.033 |
arrest | -0.033 |
couch | -0.032 |
balcony | -0.032 |
Token | â |
Token ID | 22940 |
Feature activation | -7.73e-01 |
Pos logit contributions | |
sidelines | +0.004 |
doorstep | +0.004 |
salon | +0.004 |
Sweet | +0.004 |
arrest | +0.004 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.004 |
roc | -0.004 |
idea | -0.004 |
Jak | -0.004 |
Token | € |
Token ID | 26391 |
Feature activation | -2.38e-01 |
Pos logit contributions | |
ole | +0.022 |
roc | +0.019 |
ny | +0.019 |
idea | +0.019 |
Jak | +0.018 |
Neg logit contributions | |
doorstep | -0.018 |
sidelines | -0.018 |
salon | -0.017 |
Sweet | -0.016 |
uddled | -0.016 |
Token | â„¢ |
Token ID | 8151 |
Feature activation | +1.04e+00 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.006 |
salon | -0.006 |
doorstep | -0.006 |
Sweet | -0.006 |
arrest | -0.006 |
Token | lips |
Token ID | 11914 |
Feature activation | -1.63e+00 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
idea | +0.008 |
roc | +0.008 |
Jak | +0.007 |
Neg logit contributions | |
Sweet | -0.009 |
sidelines | -0.009 |
salon | -0.009 |
arrest | -0.009 |
uddled | -0.009 |
Token | and |
Token ID | 290 |
Feature activation | -1.20e+00 |
Pos logit contributions | |
ole | +0.039 |
roc | +0.038 |
ny | +0.034 |
Jak | +0.033 |
idea | +0.031 |
Neg logit contributions | |
salon | -0.046 |
doorstep | -0.046 |
sidelines | -0.044 |
uddled | -0.043 |
Sweet | -0.042 |
Token | made |
Token ID | 925 |
Feature activation | -2.00e-01 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.025 |
roc | +0.023 |
Jak | +0.022 |
idea | +0.022 |
Neg logit contributions | |
sidelines | -0.036 |
salon | -0.035 |
doorstep | -0.035 |
Sweet | -0.035 |
arrest | -0.034 |
Token | the |
Token ID | 262 |
Feature activation | +9.99e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
suit | +0.004 |
ac | +0.004 |
Neg logit contributions | |
sidelines | -0.006 |
arrest | -0.005 |
Sweet | -0.005 |
doorstep | -0.005 |
salon | -0.005 |
Token | " |
Token ID | 366 |
Feature activation | +1.03e+00 |
Pos logit contributions | |
salon | +0.028 |
Sweet | +0.028 |
sidelines | +0.027 |
doorstep | +0.027 |
arrest | +0.027 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.023 |
roc | -0.021 |
idea | -0.021 |
boy | -0.020 |
Token | sh |
Token ID | 1477 |
Feature activation | -4.64e+00 |
Pos logit contributions | |
Sweet | +0.025 |
sidelines | +0.025 |
salon | +0.025 |
doorstep | +0.024 |
arrest | +0.024 |
Neg logit contributions | |
ole | -0.029 |
ny | -0.026 |
roc | -0.025 |
idea | -0.024 |
Jak | -0.024 |
Token | hh |
Token ID | 12337 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
ole | +0.119 |
ny | +0.112 |
roc | +0.107 |
Jak | +0.106 |
boy | +0.100 |
Neg logit contributions | |
Sweet | -0.115 |
salon | -0.115 |
sidelines | -0.115 |
doorstep | -0.113 |
uddled | -0.112 |
Token | " |
Token ID | 1 |
Feature activation | +1.98e-01 |
Pos logit contributions | |
ole | +0.032 |
ny | +0.029 |
roc | +0.028 |
Jak | +0.027 |
idea | +0.026 |
Neg logit contributions | |
sidelines | -0.032 |
salon | -0.032 |
doorstep | -0.032 |
arrest | -0.031 |
Sweet | -0.031 |
Token | sound |
Token ID | 2128 |
Feature activation | -2.38e+00 |
Pos logit contributions | |
Sweet | +0.005 |
uddled | +0.005 |
salon | +0.005 |
She | +0.005 |
Her | +0.005 |
Neg logit contributions | |
ole | -0.005 |
roc | -0.005 |
ny | -0.004 |
beard | -0.004 |
idea | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +3.39e-01 |
Pos logit contributions | |
ole | +0.053 |
roc | +0.048 |
ny | +0.047 |
Jak | +0.044 |
idea | +0.041 |
Neg logit contributions | |
sidelines | -0.069 |
salon | -0.069 |
doorstep | -0.068 |
Sweet | -0.068 |
uddled | -0.067 |
Token | All |
Token ID | 1439 |
Feature activation | -2.07e-01 |
Pos logit contributions | |
sidelines | +0.008 |
Sweet | +0.008 |
doorstep | +0.008 |
salon | +0.008 |
� | +0.008 |
Neg logit contributions | |
ole | -0.010 |
ny | -0.009 |
roc | -0.009 |
idea | -0.008 |
ac | -0.008 |
Token | named |
Token ID | 3706 |
Feature activation | +2.30e-01 |
Pos logit contributions | |
ole | +0.000 |
ny | +0.000 |
roc | +0.000 |
Jak | +0.000 |
boy | +0.000 |
Neg logit contributions | |
sidelines | -0.001 |
salon | -0.001 |
Sweet | -0.001 |
doorstep | -0.000 |
uddled | -0.000 |
Token | Lily |
Token ID | 20037 |
Feature activation | -9.58e-01 |
Pos logit contributions | |
uddled | +0.008 |
sidelines | +0.008 |
arrest | +0.008 |
doorstep | +0.008 |
salon | +0.008 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
Jak | -0.003 |
boy | -0.003 |
roc | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | +3.96e-01 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.017 |
roc | +0.013 |
boy | +0.013 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.035 |
salon | -0.033 |
arrest | -0.033 |
doorstep | -0.032 |
Sweet | -0.032 |
Token | She |
Token ID | 1375 |
Feature activation | +6.38e-01 |
Pos logit contributions | |
doorstep | +0.008 |
arrest | +0.008 |
sidelines | +0.008 |
uddled | +0.008 |
salon | +0.008 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.012 |
roc | -0.011 |
Jak | -0.011 |
boy | -0.011 |
Token | woke |
Token ID | 19092 |
Feature activation | +3.26e+00 |
Pos logit contributions | |
sidelines | +0.025 |
salon | +0.023 |
arrest | +0.022 |
doorstep | +0.022 |
Sweet | +0.022 |
Neg logit contributions | |
ny | -0.012 |
ole | -0.011 |
boy | -0.010 |
idea | -0.009 |
zo | -0.008 |
Token | up |
Token ID | 510 |
Feature activation | +2.64e+00 |
Pos logit contributions | |
sidelines | +0.122 |
arrest | +0.121 |
doorstep | +0.120 |
Sweet | +0.120 |
salon | +0.119 |
Neg logit contributions | |
ole | -0.051 |
roc | -0.039 |
ny | -0.038 |
idea | -0.036 |
boy | -0.034 |
Token | today |
Token ID | 1909 |
Feature activation | -1.19e+00 |
Pos logit contributions | |
sidelines | +0.088 |
doorstep | +0.085 |
salon | +0.084 |
Sweet | +0.083 |
arrest | +0.083 |
Neg logit contributions | |
ole | -0.054 |
ny | -0.047 |
idea | -0.043 |
boy | -0.041 |
roc | -0.041 |
Token | feeling |
Token ID | 4203 |
Feature activation | +1.13e+00 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.013 |
idea | +0.011 |
boy | +0.011 |
Jak | +0.010 |
Neg logit contributions | |
sidelines | -0.047 |
Sweet | -0.047 |
arrest | -0.046 |
doorstep | -0.045 |
salon | -0.045 |
Token | weak |
Token ID | 4939 |
Feature activation | -7.30e-01 |
Pos logit contributions | |
doorstep | +0.040 |
sidelines | +0.039 |
salon | +0.039 |
arrest | +0.039 |
Sweet | +0.038 |
Neg logit contributions | |
ole | -0.020 |
ny | -0.018 |
roc | -0.016 |
idea | -0.016 |
Jak | -0.015 |
Token | and |
Token ID | 290 |
Feature activation | +1.06e+00 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.015 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.013 |
Neg logit contributions | |
doorstep | -0.021 |
salon | -0.021 |
sidelines | -0.021 |
Sweet | -0.021 |
arrest | -0.021 |
Token | tired |
Token ID | 10032 |
Feature activation | +5.28e-01 |
Pos logit contributions | |
doorstep | +0.034 |
sidelines | +0.034 |
arrest | +0.034 |
salon | +0.034 |
Sweet | +0.034 |
Neg logit contributions | |
ole | -0.021 |
ny | -0.020 |
idea | -0.017 |
roc | -0.017 |
boy | -0.016 |
Token | had |
Token ID | 550 |
Feature activation | -5.99e-01 |
Pos logit contributions | |
sidelines | +0.024 |
salon | +0.022 |
arrest | +0.022 |
doorstep | +0.021 |
Sweet | +0.021 |
Neg logit contributions | |
ny | -0.011 |
ole | -0.010 |
boy | -0.009 |
idea | -0.008 |
roc | -0.008 |
Token | a |
Token ID | 257 |
Feature activation | +3.07e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.005 |
Jak | +0.005 |
idea | +0.005 |
Neg logit contributions | |
sidelines | -0.024 |
doorstep | -0.024 |
salon | -0.024 |
uddled | -0.024 |
Sweet | -0.023 |
Token | pair |
Token ID | 5166 |
Feature activation | +1.30e+00 |
Pos logit contributions | |
sidelines | +0.010 |
Sweet | +0.010 |
doorstep | +0.010 |
arrest | +0.010 |
salon | +0.010 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
roc | -0.005 |
idea | -0.005 |
boy | -0.005 |
Token | of |
Token ID | 286 |
Feature activation | +2.80e-01 |
Pos logit contributions | |
sidelines | +0.042 |
doorstep | +0.041 |
Sweet | +0.041 |
salon | +0.040 |
arrest | +0.040 |
Neg logit contributions | |
ole | -0.028 |
ny | -0.025 |
idea | -0.022 |
Jak | -0.021 |
roc | -0.021 |
Token | pink |
Token ID | 11398 |
Feature activation | +1.46e+00 |
Pos logit contributions | |
salon | +0.007 |
sidelines | +0.007 |
Sweet | +0.007 |
freshly | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.007 |
Jak | -0.007 |
beard | -0.006 |
Token | m |
Token ID | 285 |
Feature activation | +2.75e+00 |
Pos logit contributions | |
Acc | +0.047 |
Sweet | +0.047 |
sidelines | +0.046 |
doorstep | +0.045 |
arrest | +0.043 |
Neg logit contributions | |
ole | -0.031 |
ny | -0.028 |
boy | -0.027 |
idea | -0.026 |
guys | -0.024 |
Token | ittens |
Token ID | 34978 |
Feature activation | +8.96e-01 |
Pos logit contributions | |
Acc | +0.067 |
sidelines | +0.067 |
doorstep | +0.063 |
Sweet | +0.062 |
arrest | +0.061 |
Neg logit contributions | |
ole | -0.081 |
roc | -0.068 |
ny | -0.067 |
Jak | -0.066 |
ac | -0.063 |
Token | that |
Token ID | 326 |
Feature activation | -1.26e+00 |
Pos logit contributions | |
sidelines | +0.029 |
Sweet | +0.029 |
salon | +0.029 |
doorstep | +0.028 |
arrest | +0.028 |
Neg logit contributions | |
ole | -0.018 |
ny | -0.017 |
roc | -0.015 |
Jak | -0.014 |
idea | -0.014 |
Token | she |
Token ID | 673 |
Feature activation | +1.16e+00 |
Pos logit contributions | |
ole | +0.048 |
ny | +0.045 |
roc | +0.043 |
Jak | +0.042 |
idea | +0.042 |
Neg logit contributions | |
sidelines | -0.019 |
doorstep | -0.018 |
salon | -0.018 |
Sweet | -0.018 |
arrest | -0.018 |
Token | loved |
Token ID | 6151 |
Feature activation | +5.08e-01 |
Pos logit contributions | |
sidelines | +0.037 |
doorstep | +0.036 |
salon | +0.035 |
arrest | +0.035 |
Sweet | +0.035 |
Neg logit contributions | |
ny | -0.025 |
ole | -0.025 |
boy | -0.020 |
idea | -0.020 |
roc | -0.020 |
Token | very |
Token ID | 845 |
Feature activation | +1.11e+00 |
Pos logit contributions | |
sidelines | +0.015 |
salon | +0.015 |
doorstep | +0.015 |
Sweet | +0.015 |
arrest | +0.015 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.010 |
Jak | -0.009 |
idea | -0.009 |
Token | day |
Token ID | 1110 |
Feature activation | -6.96e-01 |
Pos logit contributions | |
ole | +0.015 |
roc | +0.013 |
Jak | +0.012 |
ny | +0.012 |
beard | +0.011 |
Neg logit contributions | |
Sweet | -0.022 |
doorstep | -0.022 |
salon | -0.022 |
sidelines | -0.021 |
balcony | -0.021 |
Token | , |
Token ID | 11 |
Feature activation | -1.19e+00 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
roc | +0.007 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.028 |
doorstep | -0.028 |
salon | -0.027 |
Sweet | -0.027 |
arrest | -0.027 |
Token | she |
Token ID | 673 |
Feature activation | +8.33e-01 |
Pos logit contributions | |
ole | +0.036 |
boy | +0.034 |
ny | +0.034 |
Jak | +0.033 |
idea | +0.033 |
Neg logit contributions | |
doorstep | -0.026 |
sidelines | -0.025 |
uddled | -0.024 |
arrest | -0.024 |
salon | -0.023 |
Token | was |
Token ID | 373 |
Feature activation | -3.45e-01 |
Pos logit contributions | |
sidelines | +0.026 |
salon | +0.026 |
doorstep | +0.026 |
Sweet | +0.026 |
arrest | +0.026 |
Neg logit contributions | |
ole | -0.018 |
ny | -0.016 |
roc | -0.015 |
Jak | -0.014 |
idea | -0.013 |
Token | playing |
Token ID | 2712 |
Feature activation | +1.70e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
idea | +0.009 |
roc | +0.008 |
Jak | +0.008 |
Neg logit contributions | |
Sweet | -0.008 |
doorstep | -0.008 |
sidelines | -0.008 |
arrest | -0.008 |
salon | -0.008 |
Token | outside |
Token ID | 2354 |
Feature activation | +2.93e+00 |
Pos logit contributions | |
sidelines | +0.005 |
salon | +0.005 |
doorstep | +0.005 |
arrest | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token | in |
Token ID | 287 |
Feature activation | +1.90e-03 |
Pos logit contributions | |
uddled | +0.123 |
doorstep | +0.120 |
Sweet | +0.119 |
sidelines | +0.119 |
salon | +0.114 |
Neg logit contributions | |
ny | -0.029 |
ole | -0.027 |
boy | -0.026 |
idea | -0.024 |
by | -0.023 |
Token | the |
Token ID | 262 |
Feature activation | -1.70e-01 |
Pos logit contributions | |
uddled | +0.000 |
doorstep | +0.000 |
sidelines | +0.000 |
salon | +0.000 |
Sweet | +0.000 |
Neg logit contributions | |
ole | -0.000 |
ny | -0.000 |
roc | -0.000 |
Jak | -0.000 |
boy | -0.000 |
Token | dirt |
Token ID | 13647 |
Feature activation | +2.53e-01 |
Pos logit contributions | |
ny | +0.004 |
ole | +0.004 |
roc | +0.004 |
boy | +0.004 |
Jak | +0.004 |
Neg logit contributions | |
salon | -0.004 |
uddled | -0.004 |
arrest | -0.004 |
doorstep | -0.004 |
sidelines | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | -2.50e-01 |
Pos logit contributions | |
Sweet | +0.009 |
sidelines | +0.009 |
doorstep | +0.009 |
arrest | +0.008 |
salon | +0.008 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.004 |
idea | -0.004 |
boy | -0.003 |
roc | -0.003 |
Token | It |
Token ID | 632 |
Feature activation | +2.55e-01 |
Pos logit contributions | |
ny | +0.006 |
ole | +0.006 |
boy | +0.006 |
Jak | +0.006 |
roc | +0.005 |
Neg logit contributions | |
arrest | -0.007 |
uddled | -0.007 |
salon | -0.007 |
sidelines | -0.007 |
doorstep | -0.006 |
Token | . |
Token ID | 13 |
Feature activation | -1.33e-01 |
Pos logit contributions | |
sidelines | +0.009 |
doorstep | +0.009 |
salon | +0.009 |
uddled | +0.009 |
Sweet | +0.008 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.006 |
boy | -0.006 |
Jak | -0.006 |
Token | She |
Token ID | 1375 |
Feature activation | +4.61e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.004 |
roc | +0.004 |
idea | +0.004 |
Jak | +0.004 |
Neg logit contributions | |
sidelines | -0.002 |
doorstep | -0.002 |
salon | -0.002 |
arrest | -0.002 |
uddled | -0.002 |
Token | was |
Token ID | 373 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
sidelines | +0.014 |
doorstep | +0.014 |
salon | +0.014 |
arrest | +0.013 |
Sweet | +0.013 |
Neg logit contributions | |
ny | -0.010 |
ole | -0.010 |
boy | -0.009 |
idea | -0.009 |
roc | -0.008 |
Token | sad |
Token ID | 6507 |
Feature activation | +8.54e-01 |
Pos logit contributions | |
salon | +0.003 |
sidelines | +0.003 |
Sweet | +0.003 |
doorstep | +0.003 |
arrest | +0.003 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.002 |
roc | -0.002 |
Jak | -0.002 |
beard | -0.002 |
Token | because |
Token ID | 780 |
Feature activation | +1.87e+00 |
Pos logit contributions | |
sidelines | +0.037 |
arrest | +0.036 |
Sweet | +0.035 |
salon | +0.034 |
doorstep | +0.034 |
Neg logit contributions | |
ole | -0.009 |
idea | -0.009 |
ny | -0.009 |
boy | -0.008 |
, | -0.007 |
Token | her |
Token ID | 607 |
Feature activation | +2.47e+00 |
Pos logit contributions | |
sidelines | +0.050 |
doorstep | +0.049 |
uddled | +0.048 |
arrest | +0.048 |
salon | +0.044 |
Neg logit contributions | |
ole | -0.045 |
boy | -0.042 |
there | -0.041 |
his | -0.041 |
Jak | -0.040 |
Token | house |
Token ID | 2156 |
Feature activation | +6.09e-01 |
Pos logit contributions | |
sidelines | +0.080 |
arrest | +0.077 |
Sweet | +0.074 |
salon | +0.074 |
Acc | +0.073 |
Neg logit contributions | |
idea | -0.050 |
ole | -0.048 |
boy | -0.047 |
ny | -0.045 |
thing | -0.040 |
Token | was |
Token ID | 373 |
Feature activation | +3.56e-01 |
Pos logit contributions | |
salon | +0.018 |
doorstep | +0.018 |
sidelines | +0.018 |
Sweet | +0.018 |
arrest | +0.018 |
Neg logit contributions | |
ole | -0.014 |
ny | -0.012 |
roc | -0.011 |
Jak | -0.011 |
idea | -0.010 |
Token | boring |
Token ID | 14262 |
Feature activation | -6.31e-01 |
Pos logit contributions | |
sidelines | +0.011 |
doorstep | +0.011 |
salon | +0.011 |
Sweet | +0.011 |
arrest | +0.011 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | and |
Token ID | 290 |
Feature activation | +2.45e-01 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.013 |
roc | +0.013 |
Jak | +0.012 |
idea | +0.011 |
Neg logit contributions | |
salon | -0.018 |
doorstep | -0.018 |
sidelines | -0.018 |
arrest | -0.018 |
Sweet | -0.018 |
Token | grey |
Token ID | 13791 |
Feature activation | +5.07e-02 |
Pos logit contributions | |
sidelines | +0.007 |
doorstep | +0.007 |
salon | +0.007 |
arrest | +0.007 |
Sweet | +0.007 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
roc | -0.005 |
Jak | -0.005 |
idea | -0.005 |
Token | was |
Token ID | 373 |
Feature activation | -1.24e-01 |
Pos logit contributions | |
Sweet | +0.003 |
doorstep | +0.003 |
Acc | +0.002 |
salon | +0.002 |
arrest | +0.002 |
Neg logit contributions | |
ole | -0.005 |
roc | -0.004 |
ny | -0.004 |
Jak | -0.004 |
ac | -0.004 |
Token | very |
Token ID | 845 |
Feature activation | +1.83e+00 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
salon | -0.004 |
sidelines | -0.004 |
doorstep | -0.004 |
Sweet | -0.004 |
arrest | -0.004 |
Token | excited |
Token ID | 6568 |
Feature activation | +1.16e+00 |
Pos logit contributions | |
doorstep | +0.063 |
sidelines | +0.063 |
salon | +0.062 |
arrest | +0.061 |
Sweet | +0.061 |
Neg logit contributions | |
ole | -0.034 |
ny | -0.031 |
roc | -0.028 |
idea | -0.026 |
Jak | -0.025 |
Token | to |
Token ID | 284 |
Feature activation | -4.69e-01 |
Pos logit contributions | |
sidelines | +0.040 |
Sweet | +0.040 |
salon | +0.039 |
arrest | +0.039 |
doorstep | +0.038 |
Neg logit contributions | |
ole | -0.022 |
ny | -0.020 |
idea | -0.017 |
boy | -0.016 |
roc | -0.016 |
Token | play |
Token ID | 711 |
Feature activation | +8.48e-01 |
Pos logit contributions | |
idea | +0.008 |
ole | +0.008 |
ny | +0.008 |
Jak | +0.007 |
guys | +0.007 |
Neg logit contributions | |
uddled | -0.017 |
Sweet | -0.016 |
Acc | -0.016 |
salon | -0.016 |
sidelines | -0.015 |
Token | outside |
Token ID | 2354 |
Feature activation | +2.86e+00 |
Pos logit contributions | |
sidelines | +0.025 |
salon | +0.024 |
doorstep | +0.024 |
Sweet | +0.024 |
arrest | +0.024 |
Neg logit contributions | |
ole | -0.020 |
ny | -0.018 |
roc | -0.017 |
idea | -0.016 |
Jak | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | +2.51e-01 |
Pos logit contributions | |
uddled | +0.119 |
sidelines | +0.119 |
doorstep | +0.118 |
Sweet | +0.117 |
salon | +0.115 |
Neg logit contributions | |
ole | -0.031 |
ny | -0.031 |
boy | -0.025 |
idea | -0.023 |
his | -0.021 |
Token | She |
Token ID | 1375 |
Feature activation | +6.27e-01 |
Pos logit contributions | |
arrest | +0.006 |
sidelines | +0.006 |
uddled | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
Neg logit contributions | |
ny | -0.007 |
ole | -0.007 |
boy | -0.006 |
idea | -0.006 |
Jak | -0.006 |
Token | asked |
Token ID | 1965 |
Feature activation | -2.87e-01 |
Pos logit contributions | |
sidelines | +0.023 |
doorstep | +0.022 |
salon | +0.022 |
Sweet | +0.021 |
arrest | +0.021 |
Neg logit contributions | |
ny | -0.012 |
ole | -0.010 |
boy | -0.009 |
idea | -0.009 |
Jak | -0.008 |
Token | her |
Token ID | 607 |
Feature activation | +1.72e-02 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.010 |
roc | +0.009 |
Jak | +0.009 |
idea | +0.008 |
Neg logit contributions | |
Sweet | -0.005 |
salon | -0.005 |
doorstep | -0.005 |
sidelines | -0.005 |
arrest | -0.005 |
Token | mom |
Token ID | 1995 |
Feature activation | -9.35e-01 |
Pos logit contributions | |
salon | +0.000 |
doorstep | +0.000 |
sidelines | +0.000 |
Sweet | +0.000 |
arrest | +0.000 |
Neg logit contributions | |
ole | -0.000 |
ny | -0.000 |
roc | -0.000 |
Jak | -0.000 |
idea | -0.000 |
Token | river |
Token ID | 7850 |
Feature activation | +4.56e-02 |
Pos logit contributions | |
ole | +0.025 |
ny | +0.023 |
roc | +0.021 |
Jak | +0.020 |
idea | +0.020 |
Neg logit contributions | |
sidelines | -0.029 |
salon | -0.029 |
doorstep | -0.029 |
Sweet | -0.028 |
arrest | -0.028 |
Token | . |
Token ID | 13 |
Feature activation | -3.27e-01 |
Pos logit contributions | |
salon | +0.001 |
doorstep | +0.001 |
sidelines | +0.001 |
arrest | +0.001 |
balcony | +0.001 |
Neg logit contributions | |
ole | -0.001 |
roc | -0.001 |
Jak | -0.001 |
ny | -0.001 |
ac | -0.001 |
Token | The |
Token ID | 383 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
roc | +0.008 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
arrest | -0.008 |
uddled | -0.008 |
Token | bee |
Token ID | 20697 |
Feature activation | +1.20e+00 |
Pos logit contributions | |
sidelines | +0.004 |
salon | +0.004 |
doorstep | +0.004 |
Sweet | +0.004 |
arrest | +0.004 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
roc | -0.003 |
Jak | -0.003 |
ac | -0.003 |
Token | knew |
Token ID | 2993 |
Feature activation | +1.62e+00 |
Pos logit contributions | |
doorstep | +0.035 |
sidelines | +0.035 |
salon | +0.035 |
arrest | +0.034 |
Sweet | +0.034 |
Neg logit contributions | |
ole | -0.028 |
ny | -0.025 |
roc | -0.024 |
Jak | -0.023 |
idea | -0.022 |
Token | she |
Token ID | 673 |
Feature activation | +8.21e-01 |
Pos logit contributions | |
sidelines | +0.050 |
doorstep | +0.047 |
salon | +0.046 |
arrest | +0.046 |
Sweet | +0.045 |
Neg logit contributions | |
ole | -0.038 |
ny | -0.033 |
Jak | -0.031 |
roc | -0.031 |
idea | -0.030 |
Token | had |
Token ID | 550 |
Feature activation | -7.05e-01 |
Pos logit contributions | |
sidelines | +0.027 |
salon | +0.027 |
doorstep | +0.027 |
Sweet | +0.027 |
arrest | +0.026 |
Neg logit contributions | |
ole | -0.016 |
ny | -0.015 |
roc | -0.013 |
Jak | -0.012 |
idea | -0.012 |
Token | to |
Token ID | 284 |
Feature activation | -5.77e-01 |
Pos logit contributions | |
ole | +0.017 |
ny | +0.016 |
roc | +0.015 |
Jak | +0.014 |
idea | +0.014 |
Neg logit contributions | |
salon | -0.020 |
doorstep | -0.020 |
sidelines | -0.020 |
arrest | -0.019 |
Sweet | -0.019 |
Token | save |
Token ID | 3613 |
Feature activation | +3.14e-01 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.015 |
roc | +0.014 |
Jak | +0.013 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.014 |
salon | -0.014 |
doorstep | -0.014 |
Sweet | -0.014 |
arrest | -0.014 |
Token | it |
Token ID | 340 |
Feature activation | +5.01e-01 |
Pos logit contributions | |
doorstep | +0.009 |
sidelines | +0.009 |
salon | +0.009 |
arrest | +0.009 |
Sweet | +0.009 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
boat | -0.006 |
Token | . |
Token ID | 13 |
Feature activation | -1.24e-01 |
Pos logit contributions | |
doorstep | +0.015 |
sidelines | +0.015 |
salon | +0.015 |
arrest | +0.015 |
Sweet | +0.015 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.010 |
Jak | -0.009 |
idea | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | -5.29e-01 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.013 |
doorstep | -0.013 |
salon | -0.013 |
Sweet | -0.013 |
arrest | -0.013 |
Token | The |
Token ID | 383 |
Feature activation | +1.87e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
Jak | +0.010 |
roc | +0.010 |
boy | +0.009 |
Neg logit contributions | |
salon | -0.017 |
doorstep | -0.017 |
sidelines | -0.016 |
arrest | -0.016 |
Sweet | -0.016 |
Token | bird |
Token ID | 6512 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
sidelines | +0.006 |
salon | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
roc | -0.004 |
Jak | -0.003 |
idea | -0.003 |
Token | whist |
Token ID | 32262 |
Feature activation | -1.54e+00 |
Pos logit contributions | |
sidelines | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
salon | +0.006 |
Neg logit contributions | |
ny | -0.004 |
ole | -0.004 |
boy | -0.003 |
idea | -0.003 |
roc | -0.003 |
Token | les |
Token ID | 829 |
Feature activation | +3.60e-01 |
Pos logit contributions | |
ole | +0.051 |
Jak | +0.047 |
ny | +0.045 |
wood | +0.045 |
roc | +0.044 |
Neg logit contributions | |
uddled | -0.028 |
sidelines | -0.027 |
Sweet | -0.027 |
salon | -0.024 |
arrest | -0.024 |
Token | back |
Token ID | 736 |
Feature activation | +4.41e-01 |
Pos logit contributions | |
doorstep | +0.012 |
sidelines | +0.012 |
salon | +0.012 |
Sweet | +0.011 |
arrest | +0.011 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.007 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | and |
Token ID | 290 |
Feature activation | -7.45e-01 |
Pos logit contributions | |
sidelines | +0.015 |
doorstep | +0.015 |
salon | +0.015 |
Sweet | +0.014 |
arrest | +0.014 |
Neg logit contributions | |
ole | -0.009 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.006 |
idea | -0.006 |
Token | they |
Token ID | 484 |
Feature activation | +8.95e-01 |
Pos logit contributions | |
ole | +0.020 |
ny | +0.019 |
roc | +0.018 |
Jak | +0.017 |
idea | +0.017 |
Neg logit contributions | |
doorstep | -0.019 |
sidelines | -0.019 |
salon | -0.018 |
Sweet | -0.018 |
arrest | -0.018 |
Token | all |
Token ID | 477 |
Feature activation | +8.22e-01 |
Pos logit contributions | |
Sweet | +0.024 |
sidelines | +0.024 |
doorstep | +0.023 |
salon | +0.023 |
uddled | +0.023 |
Neg logit contributions | |
ole | -0.023 |
ny | -0.022 |
roc | -0.020 |
idea | -0.020 |
Jak | -0.019 |
Token | have |
Token ID | 423 |
Feature activation | -8.83e-01 |
Pos logit contributions | |
sidelines | +0.021 |
salon | +0.021 |
doorstep | +0.021 |
Sweet | +0.021 |
arrest | +0.021 |
Neg logit contributions | |
ole | -0.022 |
ny | -0.020 |
roc | -0.019 |
Jak | -0.018 |
idea | -0.017 |
Token | fun |
Token ID | 1257 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.020 |
roc | +0.018 |
boat | +0.018 |
Jak | +0.017 |
Neg logit contributions | |
arrest | -0.025 |
Sweet | -0.024 |
freshly | -0.023 |
doorstep | -0.023 |
sidelines | -0.023 |
Token | an |
Token ID | 272 |
Feature activation | +7.36e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
doorstep | -0.002 |
sidelines | -0.002 |
arrest | -0.002 |
uddled | -0.002 |
salon | -0.002 |
Token | . |
Token ID | 13 |
Feature activation | +2.64e-01 |
Pos logit contributions | |
sidelines | +0.022 |
salon | +0.022 |
doorstep | +0.022 |
Sweet | +0.022 |
arrest | +0.021 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.015 |
roc | -0.014 |
Jak | -0.013 |
idea | -0.013 |
Token | You |
Token ID | 921 |
Feature activation | -2.04e-01 |
Pos logit contributions | |
sidelines | +0.006 |
salon | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.007 |
roc | -0.007 |
Jak | -0.007 |
idea | -0.007 |
Token | have |
Token ID | 423 |
Feature activation | -1.21e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
salon | -0.006 |
doorstep | -0.006 |
arrest | -0.006 |
sidelines | -0.005 |
Sweet | -0.005 |
Token | me |
Token ID | 502 |
Feature activation | +5.57e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.002 |
roc | +0.002 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
sidelines | -0.004 |
doorstep | -0.004 |
salon | -0.004 |
Sweet | -0.004 |
arrest | -0.004 |
Token | and |
Token ID | 290 |
Feature activation | -4.86e-02 |
Pos logit contributions | |
sidelines | +0.017 |
doorstep | +0.017 |
salon | +0.017 |
Sweet | +0.017 |
arrest | +0.017 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.010 |
Jak | -0.009 |
idea | -0.009 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.21e+00 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
boy | +0.001 |
Neg logit contributions | |
doorstep | -0.001 |
sidelines | -0.001 |
Sweet | -0.001 |
uddled | -0.001 |
salon | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | +1.42e-01 |
Pos logit contributions | |
ole | +0.049 |
ny | +0.044 |
roc | +0.042 |
Jak | +0.040 |
idea | +0.038 |
Neg logit contributions | |
salon | -0.066 |
doorstep | -0.066 |
sidelines | -0.065 |
arrest | -0.065 |
Sweet | -0.064 |
Token | We |
Token ID | 775 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
Sweet | +0.003 |
sidelines | +0.003 |
doorstep | +0.003 |
salon | +0.003 |
arrest | +0.003 |
Neg logit contributions | |
ole | -0.004 |
ny | -0.004 |
roc | -0.004 |
idea | -0.004 |
Jak | -0.004 |
Token | love |
Token ID | 1842 |
Feature activation | +1.26e+00 |
Pos logit contributions | |
salon | +0.037 |
doorstep | +0.037 |
arrest | +0.037 |
sidelines | +0.037 |
Sweet | +0.036 |
Neg logit contributions | |
ole | -0.039 |
ny | -0.035 |
roc | -0.034 |
Jak | -0.033 |
idea | -0.031 |
Token | you |
Token ID | 345 |
Feature activation | -1.62e+00 |
Pos logit contributions | |
sidelines | +0.028 |
doorstep | +0.028 |
salon | +0.028 |
Sweet | +0.027 |
arrest | +0.027 |
Neg logit contributions | |
ole | -0.038 |
ny | -0.036 |
roc | -0.034 |
Jak | -0.032 |
idea | -0.032 |
Token | fun |
Token ID | 1257 |
Feature activation | -1.31e+00 |
Pos logit contributions | |
sidelines | +0.029 |
Sweet | +0.028 |
doorstep | +0.028 |
salon | +0.028 |
arrest | +0.028 |
Neg logit contributions | |
ole | -0.018 |
ny | -0.016 |
roc | -0.015 |
Jak | -0.014 |
boat | -0.012 |
Token | , |
Token ID | 11 |
Feature activation | -3.41e-01 |
Pos logit contributions | |
ole | +0.028 |
ny | +0.026 |
roc | +0.024 |
Jak | +0.022 |
idea | +0.022 |
Neg logit contributions | |
sidelines | -0.041 |
salon | -0.040 |
doorstep | -0.040 |
Sweet | -0.040 |
arrest | -0.039 |
Token | Mary |
Token ID | 5335 |
Feature activation | +9.46e-01 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
idea | +0.006 |
roc | +0.006 |
Jak | +0.006 |
Neg logit contributions | |
doorstep | -0.011 |
sidelines | -0.011 |
Sweet | -0.010 |
salon | -0.010 |
uddled | -0.010 |
Token | 's |
Token ID | 338 |
Feature activation | -1.46e-01 |
Pos logit contributions | |
sidelines | +0.031 |
doorstep | +0.030 |
salon | +0.030 |
Sweet | +0.030 |
arrest | +0.029 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.018 |
roc | -0.016 |
Jak | -0.015 |
idea | -0.015 |
Token | pretend |
Token ID | 16614 |
Feature activation | +2.43e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.003 |
ac | +0.003 |
Neg logit contributions | |
doorstep | -0.004 |
sidelines | -0.004 |
salon | -0.004 |
balcony | -0.004 |
arrest | -0.004 |
Token | pool |
Token ID | 5933 |
Feature activation | +8.49e-01 |
Pos logit contributions | |
sidelines | +0.007 |
salon | +0.007 |
doorstep | +0.007 |
Sweet | +0.007 |
arrest | +0.007 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
roc | -0.005 |
Jak | -0.004 |
idea | -0.004 |
Token | was |
Token ID | 373 |
Feature activation | -8.17e-01 |
Pos logit contributions | |
salon | +0.024 |
doorstep | +0.024 |
sidelines | +0.024 |
uddled | +0.023 |
arrest | +0.023 |
Neg logit contributions | |
ole | -0.020 |
roc | -0.018 |
ny | -0.018 |
Jak | -0.017 |
idea | -0.015 |
Token | the |
Token ID | 262 |
Feature activation | +1.47e+00 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.016 |
roc | +0.016 |
Jak | +0.014 |
idea | +0.013 |
Neg logit contributions | |
salon | -0.025 |
sidelines | -0.025 |
doorstep | -0.025 |
Sweet | -0.024 |
arrest | -0.024 |
Token | most |
Token ID | 749 |
Feature activation | -1.15e+00 |
Pos logit contributions | |
sidelines | +0.045 |
salon | +0.045 |
doorstep | +0.044 |
balcony | +0.043 |
Sweet | +0.042 |
Neg logit contributions | |
ole | -0.033 |
roc | -0.030 |
ny | -0.029 |
Jak | -0.027 |
ac | -0.024 |
Token | magical |
Token ID | 10883 |
Feature activation | -1.52e-01 |
Pos logit contributions | |
ole | +0.023 |
ny | +0.020 |
roc | +0.019 |
Jak | +0.018 |
idea | +0.016 |
Neg logit contributions | |
sidelines | -0.038 |
doorstep | -0.038 |
salon | -0.037 |
Sweet | -0.037 |
arrest | -0.037 |
Token | place |
Token ID | 1295 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.003 |
Jak | +0.002 |
idea | +0.002 |
Neg logit contributions | |
doorstep | -0.005 |
sidelines | -0.005 |
salon | -0.005 |
Sweet | -0.005 |
arrest | -0.005 |
Token | big |
Token ID | 1263 |
Feature activation | -5.80e-01 |
Pos logit contributions | |
ole | +0.009 |
ny | +0.008 |
roc | +0.008 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
sidelines | -0.010 |
doorstep | -0.010 |
salon | -0.010 |
Sweet | -0.010 |
arrest | -0.010 |
Token | park |
Token ID | 3952 |
Feature activation | +1.10e+00 |
Pos logit contributions | |
ole | +0.019 |
roc | +0.017 |
ny | +0.017 |
Jak | +0.016 |
ac | +0.016 |
Neg logit contributions | |
sidelines | -0.013 |
salon | -0.012 |
doorstep | -0.012 |
arrest | -0.012 |
balcony | -0.011 |
Token | and |
Token ID | 290 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
doorstep | +0.033 |
salon | +0.033 |
sidelines | +0.032 |
arrest | +0.032 |
balcony | +0.031 |
Neg logit contributions | |
ole | -0.026 |
ny | -0.023 |
roc | -0.023 |
Jak | -0.020 |
idea | -0.020 |
Token | the |
Token ID | 262 |
Feature activation | +4.30e-01 |
Pos logit contributions | |
ole | +0.026 |
ny | +0.023 |
roc | +0.023 |
idea | +0.021 |
Jak | +0.020 |
Neg logit contributions | |
sidelines | -0.032 |
doorstep | -0.032 |
salon | -0.032 |
Sweet | -0.031 |
arrest | -0.031 |
Token | dog |
Token ID | 3290 |
Feature activation | +3.68e-01 |
Pos logit contributions | |
sidelines | +0.012 |
doorstep | +0.012 |
salon | +0.012 |
arrest | +0.012 |
Sweet | +0.011 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.010 |
Jak | -0.009 |
ac | -0.009 |
Token | showed |
Token ID | 3751 |
Feature activation | +4.28e-01 |
Pos logit contributions | |
sidelines | +0.011 |
salon | +0.011 |
Sweet | +0.011 |
doorstep | +0.011 |
arrest | +0.011 |
Neg logit contributions | |
ole | -0.008 |
ny | -0.008 |
roc | -0.007 |
Jak | -0.006 |
idea | -0.006 |
Token | him |
Token ID | 683 |
Feature activation | -1.50e+00 |
Pos logit contributions | |
arrest | +0.011 |
Sweet | +0.010 |
doorstep | +0.010 |
sidelines | +0.010 |
Her | +0.010 |
Neg logit contributions | |
ole | -0.012 |
roc | -0.010 |
ny | -0.010 |
suit | -0.009 |
Jak | -0.009 |
Token | the |
Token ID | 262 |
Feature activation | +1.15e-02 |
Pos logit contributions | |
ole | +0.035 |
ny | +0.033 |
roc | +0.031 |
Jak | +0.029 |
idea | +0.028 |
Neg logit contributions | |
salon | -0.043 |
doorstep | -0.043 |
sidelines | -0.043 |
Sweet | -0.042 |
arrest | -0.042 |
Token | best |
Token ID | 1266 |
Feature activation | -7.21e-01 |
Pos logit contributions | |
doorstep | +0.000 |
sidelines | +0.000 |
arrest | +0.000 |
salon | +0.000 |
couch | +0.000 |
Neg logit contributions | |
ole | -0.000 |
roc | -0.000 |
ny | -0.000 |
Jak | -0.000 |
ac | -0.000 |
Token | place |
Token ID | 1295 |
Feature activation | +1.11e+00 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.012 |
ac | +0.011 |
Neg logit contributions | |
doorstep | -0.023 |
sidelines | -0.023 |
salon | -0.023 |
arrest | -0.023 |
Sweet | -0.022 |
Token | to |
Token ID | 284 |
Feature activation | -6.09e-01 |
Pos logit contributions | |
sidelines | +0.033 |
salon | +0.033 |
doorstep | +0.033 |
Sweet | +0.032 |
arrest | +0.032 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.023 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.020 |
Token | hear |
Token ID | 3285 |
Feature activation | +8.64e-01 |
Pos logit contributions | |
ole | +0.018 |
ny | +0.017 |
idea | +0.016 |
roc | +0.015 |
Jak | +0.015 |
Neg logit contributions | |
Sweet | -0.022 |
uddled | -0.022 |
sidelines | -0.022 |
salon | -0.021 |
arrest | -0.021 |
Token | some |
Token ID | 617 |
Feature activation | -8.57e-01 |
Pos logit contributions | |
sidelines | +0.029 |
doorstep | +0.029 |
salon | +0.028 |
uddled | +0.028 |
arrest | +0.028 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.014 |
roc | -0.014 |
Jak | -0.013 |
idea | -0.012 |
Token | of |
Token ID | 286 |
Feature activation | +6.83e-01 |
Pos logit contributions | |
ole | +0.021 |
ny | +0.018 |
idea | +0.018 |
roc | +0.017 |
boy | +0.016 |
Neg logit contributions | |
doorstep | -0.025 |
salon | -0.024 |
Sweet | -0.024 |
sidelines | -0.024 |
uddled | -0.024 |
Token | your |
Token ID | 534 |
Feature activation | +2.44e-01 |
Pos logit contributions | |
doorstep | +0.025 |
uddled | +0.025 |
Sweet | +0.025 |
salon | +0.024 |
Acc | +0.024 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.008 |
roc | -0.008 |
Jak | -0.008 |
idea | -0.008 |
Token | mighty |
Token ID | 18680 |
Feature activation | +7.35e-02 |
Pos logit contributions | |
salon | +0.007 |
sidelines | +0.007 |
doorstep | +0.006 |
Sweet | +0.006 |
uddled | +0.006 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.006 |
roc | -0.005 |
idea | -0.005 |
Jak | -0.005 |
Token | music |
Token ID | 2647 |
Feature activation | -1.04e+00 |
Pos logit contributions | |
sidelines | +0.002 |
doorstep | +0.002 |
salon | +0.002 |
arrest | +0.002 |
Sweet | +0.002 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.002 |
roc | -0.002 |
Jak | -0.002 |
idea | -0.001 |
Token | ." |
Token ID | 526 |
Feature activation | -2.96e-01 |
Pos logit contributions | |
ole | +0.023 |
ny | +0.021 |
idea | +0.018 |
roc | +0.018 |
boy | +0.017 |
Neg logit contributions | |
sidelines | -0.032 |
Sweet | -0.032 |
doorstep | -0.032 |
salon | -0.032 |
arrest | -0.031 |
Token | his |
Token ID | 465 |
Feature activation | -1.87e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
Jak | +0.005 |
boy | +0.005 |
roc | +0.005 |
Neg logit contributions | |
doorstep | -0.009 |
sidelines | -0.009 |
salon | -0.009 |
uddled | -0.009 |
arrest | -0.009 |
Token | mom |
Token ID | 1995 |
Feature activation | -2.07e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
sidelines | -0.005 |
salon | -0.005 |
doorstep | -0.005 |
Sweet | -0.005 |
arrest | -0.005 |
Token | says |
Token ID | 1139 |
Feature activation | -7.01e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
idea | +0.003 |
roc | +0.003 |
boy | +0.003 |
Neg logit contributions | |
sidelines | -0.007 |
doorstep | -0.007 |
arrest | -0.007 |
salon | -0.007 |
Sweet | -0.007 |
Token | . |
Token ID | 13 |
Feature activation | -1.10e+00 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.012 |
idea | +0.012 |
Neg logit contributions | |
sidelines | -0.022 |
salon | -0.021 |
doorstep | -0.021 |
Sweet | -0.021 |
arrest | -0.021 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +4.53e-01 |
Pos logit contributions | |
ole | +0.010 |
ny | +0.009 |
roc | +0.009 |
Jak | +0.008 |
boy | +0.008 |
Neg logit contributions | |
sidelines | -0.016 |
salon | -0.016 |
doorstep | -0.015 |
arrest | -0.015 |
balcony | -0.015 |
Token | Once |
Token ID | 7454 |
Feature activation | +5.85e-01 |
Pos logit contributions | |
doorstep | +0.010 |
Sweet | +0.010 |
arrest | +0.010 |
salon | +0.010 |
sidelines | +0.010 |
Neg logit contributions | |
ole | -0.014 |
roc | -0.012 |
ny | -0.012 |
Jak | -0.011 |
idea | -0.011 |
Token | upon |
Token ID | 2402 |
Feature activation | +1.45e-01 |
Pos logit contributions | |
sidelines | +0.017 |
doorstep | +0.017 |
salon | +0.017 |
arrest | +0.017 |
Sweet | +0.017 |
Neg logit contributions | |
ole | -0.014 |
ny | -0.013 |
roc | -0.012 |
Jak | -0.011 |
idea | -0.011 |
Token | a |
Token ID | 257 |
Feature activation | +2.52e-01 |
Pos logit contributions | |
sidelines | +0.006 |
salon | +0.006 |
doorstep | +0.006 |
arrest | +0.006 |
uddled | +0.006 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
idea | -0.001 |
boy | -0.001 |
Token | time |
Token ID | 640 |
Feature activation | -3.01e+00 |
Pos logit contributions | |
sidelines | +0.011 |
salon | +0.011 |
doorstep | +0.011 |
Sweet | +0.010 |
arrest | +0.010 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.002 |
roc | -0.002 |
idea | -0.002 |
Jak | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -1.01e+00 |
Pos logit contributions | |
ole | +0.052 |
ny | +0.048 |
roc | +0.044 |
Jak | +0.039 |
idea | +0.036 |
Neg logit contributions | |
salon | -0.103 |
Sweet | -0.102 |
arrest | -0.102 |
doorstep | -0.102 |
uddled | -0.101 |
Token | there |
Token ID | 612 |
Feature activation | -1.29e+00 |
Pos logit contributions | |
ole | +0.018 |
roc | +0.016 |
ny | +0.015 |
Jak | +0.015 |
wood | +0.013 |
Neg logit contributions | |
Sweet | -0.033 |
salon | -0.033 |
uddled | -0.032 |
Acc | -0.032 |
doorstep | -0.032 |
Token | was |
Token ID | 373 |
Feature activation | -9.01e-01 |
Pos logit contributions | |
ole | +0.033 |
ny | +0.029 |
roc | +0.028 |
Jak | +0.027 |
idea | +0.025 |
Neg logit contributions | |
doorstep | -0.035 |
salon | -0.035 |
Sweet | -0.034 |
arrest | -0.034 |
uddled | -0.034 |
Token | a |
Token ID | 257 |
Feature activation | -1.69e-01 |
Pos logit contributions | |
ole | +0.014 |
ny | +0.012 |
roc | +0.011 |
Jak | +0.010 |
idea | +0.009 |
Neg logit contributions | |
sidelines | -0.033 |
salon | -0.033 |
doorstep | -0.033 |
Sweet | -0.033 |
arrest | -0.033 |
Token | little |
Token ID | 1310 |
Feature activation | +8.21e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.003 |
roc | +0.003 |
Jak | +0.003 |
wood | +0.003 |
Neg logit contributions | |
salon | -0.005 |
sidelines | -0.005 |
Sweet | -0.005 |
doorstep | -0.005 |
arrest | -0.005 |
Token | boy |
Token ID | 2933 |
Feature activation | -6.80e-01 |
Pos logit contributions | |
salon | +0.025 |
sidelines | +0.024 |
Sweet | +0.024 |
couch | +0.023 |
porch | +0.023 |
Neg logit contributions | |
ole | -0.018 |
roc | -0.016 |
Jak | -0.016 |
ny | -0.015 |
ac | -0.014 |
Token | stripes |
Token ID | 28806 |
Feature activation | -1.96e+00 |
Pos logit contributions | |
sidelines | +0.016 |
doorstep | +0.016 |
salon | +0.016 |
arrest | +0.016 |
Sweet | +0.016 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.011 |
Jak | -0.010 |
idea | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | -5.05e-01 |
Pos logit contributions | |
ole | +0.043 |
ny | +0.039 |
roc | +0.035 |
Jak | +0.033 |
idea | +0.033 |
Neg logit contributions | |
sidelines | -0.061 |
salon | -0.060 |
doorstep | -0.060 |
Sweet | -0.059 |
arrest | -0.059 |
Token | Daisy |
Token ID | 40355 |
Feature activation | +6.14e-01 |
Pos logit contributions | |
ole | +0.012 |
ny | +0.011 |
roc | +0.010 |
idea | +0.010 |
Jak | +0.009 |
Neg logit contributions | |
salon | -0.016 |
doorstep | -0.015 |
sidelines | -0.015 |
arrest | -0.014 |
Sweet | -0.014 |
Token | wanted |
Token ID | 2227 |
Feature activation | +1.83e-01 |
Pos logit contributions | |
doorstep | +0.017 |
salon | +0.017 |
sidelines | +0.017 |
arrest | +0.017 |
uddled | +0.017 |
Neg logit contributions | |
ole | -0.015 |
ny | -0.013 |
roc | -0.013 |
Jak | -0.012 |
idea | -0.012 |
Token | to |
Token ID | 284 |
Feature activation | -1.44e-01 |
Pos logit contributions | |
salon | +0.005 |
sidelines | +0.005 |
doorstep | +0.005 |
Sweet | +0.005 |
arrest | +0.005 |
Neg logit contributions | |
ole | -0.005 |
ny | -0.004 |
roc | -0.004 |
Jak | -0.004 |
idea | -0.004 |
Token | go |
Token ID | 467 |
Feature activation | -6.85e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
idea | +0.003 |
Jak | +0.003 |
roc | +0.003 |
Neg logit contributions | |
uddled | -0.004 |
Sweet | -0.004 |
salon | -0.004 |
Acc | -0.004 |
sidelines | -0.004 |
Token | on |
Token ID | 319 |
Feature activation | -1.81e-01 |
Pos logit contributions | |
ole | +0.016 |
roc | +0.015 |
ny | +0.014 |
idea | +0.013 |
Jak | +0.013 |
Neg logit contributions | |
sidelines | -0.020 |
doorstep | -0.019 |
salon | -0.019 |
arrest | -0.019 |
Sweet | -0.018 |
Token | - |
Token ID | 12 |
Feature activation | -1.11e+00 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.003 |
Jak | +0.003 |
idea | +0.003 |
Neg logit contributions | |
sidelines | -0.006 |
salon | -0.006 |
doorstep | -0.006 |
arrest | -0.006 |
Sweet | -0.006 |
Token | board |
Token ID | 3526 |
Feature activation | +9.29e-01 |
Pos logit contributions | |
ole | +0.035 |
ny | +0.032 |
roc | +0.031 |
Jak | +0.029 |
idea | +0.029 |
Neg logit contributions | |
sidelines | -0.024 |
salon | -0.024 |
doorstep | -0.024 |
Sweet | -0.023 |
arrest | -0.023 |
Token | , |
Token ID | 11 |
Feature activation | -4.26e-01 |
Pos logit contributions | |
sidelines | +0.032 |
doorstep | +0.032 |
Sweet | +0.032 |
salon | +0.031 |
arrest | +0.031 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.015 |
roc | -0.013 |
Jak | -0.013 |
idea | -0.013 |
Token | so |
Token ID | 523 |
Feature activation | +1.85e-01 |
Pos logit contributions | |
ole | +0.008 |
ny | +0.007 |
roc | +0.007 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.014 |
salon | -0.014 |
doorstep | -0.014 |
Sweet | -0.014 |
arrest | -0.014 |
Token | We |
Token ID | 775 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
salon | +0.009 |
sidelines | +0.009 |
doorstep | +0.009 |
uddled | +0.009 |
arrest | +0.009 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.010 |
roc | -0.010 |
Jak | -0.009 |
boy | -0.009 |
Token | can |
Token ID | 460 |
Feature activation | +1.09e+00 |
Pos logit contributions | |
doorstep | +0.031 |
arrest | +0.031 |
salon | +0.030 |
sidelines | +0.030 |
balcony | +0.028 |
Neg logit contributions | |
ole | -0.034 |
roc | -0.031 |
ny | -0.029 |
Jak | -0.029 |
idea | -0.026 |
Token | have |
Token ID | 423 |
Feature activation | -2.39e-01 |
Pos logit contributions | |
Sweet | +0.038 |
Acc | +0.037 |
uddled | +0.036 |
salon | +0.034 |
sidelines | +0.034 |
Neg logit contributions | |
ole | -0.021 |
ny | -0.020 |
idea | -0.018 |
guys | -0.017 |
boat | -0.017 |
Token | so |
Token ID | 523 |
Feature activation | +9.10e-01 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
arrest | -0.008 |
doorstep | -0.008 |
sidelines | -0.008 |
Sweet | -0.008 |
salon | -0.008 |
Token | much |
Token ID | 881 |
Feature activation | -1.51e+00 |
Pos logit contributions | |
doorstep | +0.029 |
sidelines | +0.029 |
salon | +0.029 |
Sweet | +0.028 |
arrest | +0.028 |
Neg logit contributions | |
ole | -0.019 |
ny | -0.017 |
roc | -0.016 |
idea | -0.015 |
Jak | -0.015 |
Token | fun |
Token ID | 1257 |
Feature activation | -1.49e+00 |
Pos logit contributions | |
ole | +0.022 |
ny | +0.019 |
idea | +0.016 |
roc | +0.016 |
Jak | +0.015 |
Neg logit contributions | |
salon | -0.058 |
doorstep | -0.057 |
sidelines | -0.057 |
Sweet | -0.056 |
uddled | -0.055 |
Token | here |
Token ID | 994 |
Feature activation | -1.56e+00 |
Pos logit contributions | |
ole | +0.041 |
ny | +0.038 |
roc | +0.035 |
idea | +0.034 |
Jak | +0.034 |
Neg logit contributions | |
sidelines | -0.038 |
salon | -0.038 |
doorstep | -0.038 |
Sweet | -0.037 |
arrest | -0.036 |
Token | !" |
Token ID | 2474 |
Feature activation | -7.54e-02 |
Pos logit contributions | |
ole | +0.037 |
ny | +0.033 |
idea | +0.029 |
roc | +0.028 |
boy | +0.027 |
Neg logit contributions | |
sidelines | -0.048 |
salon | -0.047 |
Sweet | -0.046 |
doorstep | -0.046 |
uddled | -0.045 |
Token | said |
Token ID | 531 |
Feature activation | -3.54e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
sidelines | -0.002 |
doorstep | -0.002 |
salon | -0.002 |
Sweet | -0.002 |
arrest | -0.002 |
Token | her |
Token ID | 607 |
Feature activation | +3.37e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.011 |
roc | +0.010 |
suit | +0.009 |
boat | +0.009 |
Neg logit contributions | |
Her | -0.007 |
Her | -0.007 |
doorstep | -0.007 |
Sweet | -0.007 |
balcony | -0.007 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.51e-01 |
Pos logit contributions | |
doorstep | +0.008 |
salon | +0.008 |
sidelines | +0.008 |
Sweet | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.010 |
ny | -0.009 |
roc | -0.009 |
Jak | -0.008 |
idea | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -1.89e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
Jak | +0.000 |
roc | +0.000 |
idea | +0.000 |
Neg logit contributions | |
doorstep | -0.001 |
salon | -0.001 |
sidelines | -0.001 |
arrest | -0.001 |
Sweet | -0.001 |
Token |
|
Token ID | 198 |
Feature activation | -6.42e-02 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.003 |
roc | +0.003 |
boy | +0.003 |
Jak | +0.003 |
Neg logit contributions | |
salon | -0.007 |
doorstep | -0.006 |
sidelines | -0.006 |
arrest | -0.006 |
balcony | -0.006 |
Token | L |
Token ID | 43 |
Feature activation | -9.86e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
boy | +0.001 |
Jak | +0.001 |
Neg logit contributions | |
doorstep | -0.002 |
sidelines | -0.002 |
salon | -0.002 |
arrest | -0.002 |
balcony | -0.002 |
Token | ily |
Token ID | 813 |
Feature activation | +5.02e-02 |
Pos logit contributions | |
ole | +0.021 |
ny | +0.020 |
idea | +0.017 |
roc | +0.017 |
ac | +0.016 |
Neg logit contributions | |
sidelines | -0.033 |
doorstep | -0.033 |
balcony | -0.032 |
salon | -0.032 |
Sweet | -0.030 |
Token | and |
Token ID | 290 |
Feature activation | -2.83e-01 |
Pos logit contributions | |
sidelines | +0.002 |
salon | +0.002 |
doorstep | +0.002 |
Sweet | +0.001 |
arrest | +0.001 |
Neg logit contributions | |
ole | -0.001 |
ny | -0.001 |
roc | -0.001 |
Jak | -0.001 |
idea | -0.001 |
Token | Tom |
Token ID | 4186 |
Feature activation | -1.58e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
salon | -0.008 |
sidelines | -0.008 |
doorstep | -0.008 |
arrest | -0.008 |
Sweet | -0.008 |
Token | met |
Token ID | 1138 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
ole | +0.039 |
ny | +0.035 |
roc | +0.033 |
Jak | +0.031 |
idea | +0.031 |
Neg logit contributions | |
sidelines | -0.045 |
salon | -0.045 |
doorstep | -0.045 |
Sweet | -0.044 |
arrest | -0.043 |
Token | at |
Token ID | 379 |
Feature activation | -3.28e-01 |
Pos logit contributions | |
ole | +0.003 |
ny | +0.003 |
roc | +0.002 |
Jak | +0.002 |
ac | +0.002 |
Neg logit contributions | |
sidelines | -0.004 |
salon | -0.004 |
doorstep | -0.003 |
Sweet | -0.003 |
uddled | -0.003 |
Token | the |
Token ID | 262 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.004 |
Jak | +0.004 |
idea | +0.004 |
Neg logit contributions | |
doorstep | -0.012 |
sidelines | -0.012 |
salon | -0.012 |
Sweet | -0.012 |
arrest | -0.012 |
Token | bottom |
Token ID | 4220 |
Feature activation | +1.06e+00 |
Pos logit contributions | |
ole | +0.000 |
boy | +0.000 |
idea | +0.000 |
ny | +0.000 |
roc | +0.000 |
Neg logit contributions | |
uddled | -0.000 |
Sweet | -0.000 |
Acc | -0.000 |
arrest | -0.000 |
doorstep | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | +1.77e-01 |
Pos logit contributions | |
sidelines | +0.031 |
doorstep | +0.031 |
salon | +0.031 |
Sweet | +0.030 |
arrest | +0.030 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.023 |
roc | -0.021 |
Jak | -0.020 |
idea | -0.020 |
Token | importance |
Token ID | 6817 |
Feature activation | -1.85e+00 |
Pos logit contributions | |
ole | +0.005 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.004 |
boat | +0.004 |
Neg logit contributions | |
salon | -0.005 |
sidelines | -0.005 |
arrest | -0.005 |
Sweet | -0.005 |
doorstep | -0.005 |
Token | of |
Token ID | 286 |
Feature activation | +1.02e+00 |
Pos logit contributions | |
ole | +0.049 |
ny | +0.045 |
roc | +0.044 |
Jak | +0.042 |
idea | +0.040 |
Neg logit contributions | |
salon | -0.047 |
sidelines | -0.047 |
doorstep | -0.047 |
Sweet | -0.046 |
arrest | -0.046 |
Token | standing |
Token ID | 5055 |
Feature activation | -2.21e+00 |
Pos logit contributions | |
sidelines | +0.024 |
salon | +0.024 |
doorstep | +0.023 |
arrest | +0.023 |
Sweet | +0.023 |
Neg logit contributions | |
ole | -0.030 |
ny | -0.028 |
roc | -0.026 |
Jak | -0.025 |
idea | -0.024 |
Token | up |
Token ID | 510 |
Feature activation | +1.15e+00 |
Pos logit contributions | |
ny | +0.049 |
ole | +0.047 |
roc | +0.044 |
Jak | +0.042 |
idea | +0.042 |
Neg logit contributions | |
sidelines | -0.060 |
arrest | -0.060 |
salon | -0.059 |
uddled | -0.058 |
doorstep | -0.056 |
Token | against |
Token ID | 1028 |
Feature activation | -1.94e-01 |
Pos logit contributions | |
sidelines | +0.034 |
salon | +0.034 |
arrest | +0.034 |
doorstep | +0.034 |
Sweet | +0.033 |
Neg logit contributions | |
ole | -0.025 |
ny | -0.023 |
roc | -0.022 |
Jak | -0.020 |
idea | -0.020 |
Token | rude |
Token ID | 22066 |
Feature activation | -3.31e+00 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.003 |
Jak | +0.003 |
idea | +0.003 |
Neg logit contributions | |
arrest | -0.006 |
sidelines | -0.006 |
salon | -0.006 |
doorstep | -0.006 |
Sweet | -0.006 |
Token | behavior |
Token ID | 4069 |
Feature activation | -8.22e-01 |
Pos logit contributions | |
ole | +0.083 |
ny | +0.073 |
roc | +0.071 |
Jak | +0.071 |
ac | +0.064 |
Neg logit contributions | |
arrest | -0.089 |
salon | -0.089 |
sidelines | -0.089 |
doorstep | -0.088 |
uddled | -0.081 |
Token | . |
Token ID | 13 |
Feature activation | -2.24e-01 |
Pos logit contributions | |
ole | +0.019 |
ny | +0.017 |
roc | +0.017 |
Jak | +0.015 |
idea | +0.015 |
Neg logit contributions | |
salon | -0.024 |
doorstep | -0.024 |
sidelines | -0.024 |
arrest | -0.023 |
Sweet | -0.023 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +3.83e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
Jak | +0.003 |
roc | +0.003 |
boy | +0.003 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
balcony | -0.008 |
arrest | -0.008 |
Token | Once |
Token ID | 7454 |
Feature activation | +5.73e-01 |
Pos logit contributions | |
doorstep | +0.008 |
Sweet | +0.008 |
arrest | +0.008 |
sidelines | +0.008 |
salon | +0.008 |
Neg logit contributions | |
ole | -0.012 |
roc | -0.011 |
ny | -0.010 |
Jak | -0.010 |
idea | -0.010 |
Token | upon |
Token ID | 2402 |
Feature activation | -9.95e-03 |
Pos logit contributions | |
sidelines | +0.016 |
doorstep | +0.016 |
salon | +0.016 |
Sweet | +0.016 |
arrest | +0.016 |
Neg logit contributions | |
ole | -0.014 |
ny | -0.013 |
roc | -0.012 |
Jak | -0.011 |
idea | -0.011 |
Token | I |
Token ID | 40 |
Feature activation | +2.23e+00 |
Pos logit contributions | |
doorstep | +0.015 |
sidelines | +0.015 |
salon | +0.015 |
arrest | +0.014 |
balcony | +0.014 |
Neg logit contributions | |
ny | -0.019 |
ole | -0.018 |
boy | -0.017 |
roc | -0.017 |
Jak | -0.016 |
Token | forgive |
Token ID | 20927 |
Feature activation | +8.68e-01 |
Pos logit contributions | |
doorstep | +0.051 |
Sweet | +0.051 |
salon | +0.051 |
arrest | +0.049 |
uddled | +0.048 |
Neg logit contributions | |
ole | -0.066 |
ny | -0.059 |
roc | -0.056 |
Jak | -0.056 |
idea | -0.051 |
Token | you |
Token ID | 345 |
Feature activation | -1.47e+00 |
Pos logit contributions | |
doorstep | +0.020 |
salon | +0.020 |
sidelines | +0.020 |
uddled | +0.019 |
arrest | +0.019 |
Neg logit contributions | |
ole | -0.026 |
ny | -0.025 |
roc | -0.024 |
Jak | -0.023 |
idea | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +3.75e-01 |
Pos logit contributions | |
ole | +0.031 |
ny | +0.029 |
roc | +0.026 |
idea | +0.024 |
Jak | +0.024 |
Neg logit contributions | |
sidelines | -0.047 |
salon | -0.047 |
doorstep | -0.047 |
Sweet | -0.046 |
arrest | -0.046 |
Token | But |
Token ID | 887 |
Feature activation | +4.09e-01 |
Pos logit contributions | |
salon | +0.009 |
doorstep | +0.009 |
sidelines | +0.009 |
uddled | +0.008 |
arrest | +0.008 |
Neg logit contributions | |
ole | -0.011 |
ny | -0.011 |
roc | -0.010 |
Jak | -0.010 |
boy | -0.010 |
Token | next |
Token ID | 1306 |
Feature activation | -3.43e+00 |
Pos logit contributions | |
salon | +0.009 |
sidelines | +0.009 |
doorstep | +0.009 |
Sweet | +0.009 |
arrest | +0.009 |
Neg logit contributions | |
ole | -0.013 |
ny | -0.012 |
roc | -0.011 |
Jak | -0.011 |
idea | -0.010 |
Token | time |
Token ID | 640 |
Feature activation | -1.65e+00 |
Pos logit contributions | |
ole | +0.061 |
ny | +0.055 |
roc | +0.054 |
Jak | +0.051 |
wood | +0.044 |
Neg logit contributions | |
salon | -0.118 |
doorstep | -0.117 |
sidelines | -0.114 |
Sweet | -0.112 |
arrest | -0.111 |
Token | , |
Token ID | 11 |
Feature activation | -2.44e-01 |
Pos logit contributions | |
ole | +0.024 |
ny | +0.021 |
roc | +0.019 |
Jak | +0.017 |
idea | +0.016 |
Neg logit contributions | |
sidelines | -0.063 |
salon | -0.063 |
doorstep | -0.062 |
Sweet | -0.062 |
arrest | -0.061 |
Token | don |
Token ID | 836 |
Feature activation | -1.05e+00 |
Pos logit contributions | |
ole | +0.007 |
ny | +0.006 |
roc | +0.006 |
Jak | +0.006 |
idea | +0.006 |
Neg logit contributions | |
sidelines | -0.006 |
doorstep | -0.006 |
salon | -0.006 |
Sweet | -0.006 |
arrest | -0.006 |
Token | 't |
Token ID | 470 |
Feature activation | -2.58e-02 |
Pos logit contributions | |
ole | +0.041 |
ny | +0.038 |
Jak | +0.036 |
idea | +0.036 |
roc | +0.036 |
Neg logit contributions | |
doorstep | -0.015 |
salon | -0.015 |
sidelines | -0.015 |
Sweet | -0.014 |
balcony | -0.014 |
Token | play |
Token ID | 711 |
Feature activation | +5.48e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
Jak | +0.001 |
idea | +0.001 |
Neg logit contributions | |
sidelines | -0.001 |
doorstep | -0.001 |
salon | -0.001 |
Sweet | -0.001 |
uddled | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | -2.70e-01 |
Pos logit contributions | |
sidelines | +0.004 |
salon | +0.004 |
doorstep | +0.004 |
Sweet | +0.004 |
arrest | +0.004 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.002 |
roc | -0.002 |
Jak | -0.002 |
idea | -0.002 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +3.10e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.006 |
Jak | +0.005 |
roc | +0.005 |
boy | +0.005 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
uddled | -0.008 |
doorstep | -0.008 |
arrest | -0.008 |
Token | Once |
Token ID | 7454 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
Sweet | +0.006 |
doorstep | +0.006 |
salon | +0.006 |
arrest | +0.006 |
� | +0.006 |
Neg logit contributions | |
ole | -0.010 |
roc | -0.010 |
ny | -0.009 |
idea | -0.009 |
Jak | -0.009 |
Token | upon |
Token ID | 2402 |
Feature activation | -1.99e-01 |
Pos logit contributions | |
sidelines | +0.016 |
salon | +0.016 |
doorstep | +0.016 |
Sweet | +0.015 |
arrest | +0.015 |
Neg logit contributions | |
ole | -0.015 |
ny | -0.014 |
roc | -0.013 |
Jak | -0.012 |
idea | -0.012 |
Token | a |
Token ID | 257 |
Feature activation | +1.52e-01 |
Pos logit contributions | |
ole | +0.002 |
ny | +0.002 |
roc | +0.002 |
idea | +0.002 |
Jak | +0.001 |
Neg logit contributions | |
sidelines | -0.008 |
salon | -0.008 |
doorstep | -0.008 |
arrest | -0.008 |
uddled | -0.008 |
Token | time |
Token ID | 640 |
Feature activation | -3.10e+00 |
Pos logit contributions | |
sidelines | +0.006 |
salon | +0.006 |
doorstep | +0.006 |
Sweet | +0.006 |
arrest | +0.006 |
Neg logit contributions | |
ole | -0.002 |
ny | -0.001 |
roc | -0.001 |
idea | -0.001 |
Jak | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -9.12e-01 |
Pos logit contributions | |
ole | +0.059 |
ny | +0.056 |
roc | +0.054 |
Jak | +0.046 |
idea | +0.044 |
Neg logit contributions | |
salon | -0.098 |
Sweet | -0.097 |
arrest | -0.097 |
doorstep | -0.096 |
uddled | -0.095 |
Token | there |
Token ID | 612 |
Feature activation | -1.40e+00 |
Pos logit contributions | |
ole | +0.016 |
roc | +0.014 |
ny | +0.013 |
Jak | +0.013 |
wood | +0.011 |
Neg logit contributions | |
Sweet | -0.031 |
salon | -0.030 |
doorstep | -0.030 |
Acc | -0.030 |
Her | -0.030 |
Token | was |
Token ID | 373 |
Feature activation | -6.85e-01 |
Pos logit contributions | |
ole | +0.040 |
ny | +0.035 |
roc | +0.034 |
Jak | +0.033 |
zo | +0.030 |
Neg logit contributions | |
doorstep | -0.034 |
salon | -0.034 |
Sweet | -0.033 |
uddled | -0.033 |
arrest | -0.032 |
Token | a |
Token ID | 257 |
Feature activation | -2.69e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.009 |
roc | +0.008 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
sidelines | -0.026 |
salon | -0.025 |
doorstep | -0.025 |
Sweet | -0.025 |
arrest | -0.025 |
Token | gray |
Token ID | 12768 |
Feature activation | +6.23e-01 |
Pos logit contributions | |
ole | +0.006 |
ny | +0.005 |
roc | +0.005 |
Jak | +0.005 |
ac | +0.004 |
Neg logit contributions | |
salon | -0.008 |
sidelines | -0.008 |
Sweet | -0.008 |
doorstep | -0.008 |
arrest | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | -7.18e-01 |
Pos logit contributions | |
salon | +0.044 |
sidelines | +0.044 |
doorstep | +0.044 |
arrest | +0.043 |
Sweet | +0.043 |
Neg logit contributions | |
ole | -0.034 |
ny | -0.031 |
roc | -0.029 |
Jak | -0.027 |
idea | -0.027 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +2.12e-01 |
Pos logit contributions | |
ole | +0.015 |
ny | +0.014 |
roc | +0.013 |
Jak | +0.013 |
boy | +0.012 |
Neg logit contributions | |
salon | -0.024 |
doorstep | -0.023 |
sidelines | -0.023 |
balcony | -0.022 |
arrest | -0.022 |
Token | Once |
Token ID | 7454 |
Feature activation | +4.49e-01 |
Pos logit contributions | |
Sweet | +0.004 |
doorstep | +0.004 |
salon | +0.004 |
arrest | +0.004 |
sidelines | +0.004 |
Neg logit contributions | |
ole | -0.007 |
roc | -0.006 |
ny | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | upon |
Token ID | 2402 |
Feature activation | -6.83e-02 |
Pos logit contributions | |
salon | +0.012 |
doorstep | +0.011 |
sidelines | +0.011 |
Sweet | +0.011 |
arrest | +0.011 |
Neg logit contributions | |
ole | -0.012 |
ny | -0.011 |
roc | -0.010 |
Jak | -0.010 |
idea | -0.009 |
Token | a |
Token ID | 257 |
Feature activation | +2.41e-01 |
Pos logit contributions | |
ole | +0.001 |
ny | +0.001 |
roc | +0.001 |
idea | +0.001 |
Jak | +0.001 |
Neg logit contributions | |
sidelines | -0.003 |
salon | -0.003 |
doorstep | -0.003 |
arrest | -0.003 |
uddled | -0.003 |
Token | time |
Token ID | 640 |
Feature activation | -3.11e+00 |
Pos logit contributions | |
sidelines | +0.010 |
salon | +0.010 |
doorstep | +0.010 |
Sweet | +0.010 |
arrest | +0.010 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.002 |
roc | -0.002 |
idea | -0.002 |
Jak | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -1.03e+00 |
Pos logit contributions | |
ole | +0.059 |
ny | +0.056 |
roc | +0.054 |
Jak | +0.048 |
wood | +0.045 |
Neg logit contributions | |
salon | -0.098 |
Sweet | -0.096 |
arrest | -0.096 |
doorstep | -0.094 |
uddled | -0.094 |
Token | there |
Token ID | 612 |
Feature activation | -1.30e+00 |
Pos logit contributions | |
ole | +0.018 |
roc | +0.017 |
ny | +0.015 |
Jak | +0.015 |
wood | +0.013 |
Neg logit contributions | |
Sweet | -0.034 |
salon | -0.033 |
Her | -0.033 |
Acc | -0.033 |
doorstep | -0.033 |
Token | was |
Token ID | 373 |
Feature activation | -6.75e-01 |
Pos logit contributions | |
ole | +0.036 |
ny | +0.032 |
roc | +0.031 |
Jak | +0.029 |
idea | +0.027 |
Neg logit contributions | |
doorstep | -0.033 |
salon | -0.033 |
Sweet | -0.032 |
uddled | -0.031 |
arrest | -0.031 |
Token | a |
Token ID | 257 |
Feature activation | -1.92e-01 |
Pos logit contributions | |
ole | +0.011 |
ny | +0.009 |
roc | +0.008 |
Jak | +0.007 |
idea | +0.007 |
Neg logit contributions | |
sidelines | -0.025 |
salon | -0.025 |
doorstep | -0.025 |
Sweet | -0.025 |
arrest | -0.024 |
Token | big |
Token ID | 1263 |
Feature activation | +6.04e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.003 |
wood | +0.003 |
Neg logit contributions | |
salon | -0.006 |
sidelines | -0.006 |
Sweet | -0.006 |
doorstep | -0.006 |
balcony | -0.006 |
Token | together |
Token ID | 1978 |
Feature activation | -1.65e-01 |
Pos logit contributions | |
sidelines | +0.016 |
doorstep | +0.016 |
salon | +0.016 |
arrest | +0.016 |
uddled | +0.016 |
Neg logit contributions | |
ole | -0.017 |
ny | -0.016 |
roc | -0.015 |
Jak | -0.014 |
wood | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -6.31e-01 |
Pos logit contributions | |
ole | +0.004 |
ny | +0.004 |
roc | +0.004 |
Jak | +0.003 |
idea | +0.003 |
Neg logit contributions | |
salon | -0.005 |
doorstep | -0.005 |
sidelines | -0.005 |
arrest | -0.004 |
Sweet | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +1.28e-01 |
Pos logit contributions | |
ole | +0.016 |
ny | +0.015 |
roc | +0.014 |
Jak | +0.013 |
idea | +0.013 |
Neg logit contributions | |
sidelines | -0.018 |
salon | -0.017 |
doorstep | -0.017 |
Sweet | -0.017 |
arrest | -0.017 |
Token |
|
Token ID | 198 |
Feature activation | +2.27e-01 |
Pos logit contributions | |
sidelines | +0.004 |
salon | +0.004 |
doorstep | +0.004 |
arrest | +0.003 |
Sweet | +0.003 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.003 |
idea | -0.003 |
Token | The |
Token ID | 464 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
doorstep | +0.005 |
sidelines | +0.005 |
salon | +0.005 |
arrest | +0.005 |
Sweet | +0.005 |
Neg logit contributions | |
ole | -0.007 |
ny | -0.006 |
roc | -0.006 |
Jak | -0.006 |
idea | -0.006 |
Token | next |
Token ID | 1306 |
Feature activation | -3.08e+00 |
Pos logit contributions | |
sidelines | +0.005 |
doorstep | +0.005 |
salon | +0.005 |
Sweet | +0.005 |
arrest | +0.005 |
Neg logit contributions | |
ole | -0.003 |
ny | -0.003 |
roc | -0.003 |
Jak | -0.002 |
idea | -0.002 |
Token | day |
Token ID | 1110 |
Feature activation | -3.67e-01 |
Pos logit contributions | |
ole | +0.061 |
roc | +0.054 |
ny | +0.051 |
Jak | +0.050 |
beard | +0.045 |
Neg logit contributions | |
salon | -0.101 |
Sweet | -0.098 |
sidelines | -0.097 |
doorstep | -0.097 |
arrest | -0.093 |
Token | , |
Token ID | 11 |
Feature activation | -1.51e+00 |
Pos logit contributions | |
ole | +0.009 |
roc | +0.008 |
ny | +0.008 |
idea | +0.007 |
Jak | +0.007 |
Neg logit contributions | |
arrest | -0.010 |
Sweet | -0.010 |
doorstep | -0.010 |
salon | -0.010 |
uddled | -0.010 |
Token | the |
Token ID | 262 |
Feature activation | +2.72e-01 |
Pos logit contributions | |
ole | +0.036 |
ny | +0.034 |
roc | +0.031 |
ac | +0.029 |
Jak | +0.027 |
Neg logit contributions | |
sidelines | -0.043 |
arrest | -0.043 |
salon | -0.043 |
Sweet | -0.042 |
doorstep | -0.041 |
Token | little |
Token ID | 1310 |
Feature activation | +7.04e-01 |
Pos logit contributions | |
sidelines | +0.009 |
doorstep | +0.009 |
salon | +0.009 |
Sweet | +0.009 |
arrest | +0.009 |
Neg logit contributions | |
ole | -0.006 |
ny | -0.005 |
roc | -0.005 |
Jak | -0.004 |
ac | -0.004 |
Token | boy |
Token ID | 2933 |
Feature activation | -7.16e-01 |
Pos logit contributions | |
salon | +0.020 |
porch | +0.020 |
couch | +0.019 |
swings | +0.019 |
swing | +0.019 |
Neg logit contributions | |
ole | -0.017 |
ac | -0.015 |
roc | -0.014 |
Jak | -0.013 |
ke | -0.012 |