Token | we |
Token ID | 356 |
Feature activation | +1.26e-02 |
Pos logit contributions | |
Blue | +0.062 |
over | +0.061 |
back | +0.057 |
isk | +0.057 |
ety | +0.057 |
Neg logit contributions | |
inations | -0.047 |
rd | -0.046 |
arest | -0.046 |
iden | -0.046 |
very | -0.045 |
Token | stay |
Token ID | 2652 |
Feature activation | -5.60e-03 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.013 |
barking | +0.013 |
Spot | +0.013 |
back | +0.013 |
Neg logit contributions | |
inations | -0.028 |
arest | -0.027 |
rd | -0.027 |
iden | -0.027 |
nexpected | -0.027 |
Token | here |
Token ID | 994 |
Feature activation | +3.98e-04 |
Pos logit contributions | |
inations | +0.014 |
rd | +0.014 |
iden | +0.014 |
very | +0.014 |
nexpected | +0.014 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.005 |
Spot | -0.004 |
back | -0.004 |
ety | -0.004 |
Token | ?" |
Token ID | 1701 |
Feature activation | -1.56e-02 |
Pos logit contributions | |
Blue | +0.001 |
ety | +0.001 |
barking | +0.001 |
Spot | +0.001 |
over | +0.001 |
Neg logit contributions | |
inations | -0.001 |
very | -0.001 |
iden | -0.001 |
rd | -0.001 |
iest | -0.001 |
Token | The |
Token ID | 383 |
Feature activation | +6.53e-02 |
Pos logit contributions | |
inations | +0.039 |
rd | +0.039 |
iest | +0.038 |
arest | +0.038 |
iden | +0.038 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.016 |
Spot | -0.015 |
back | -0.014 |
from | -0.014 |
Token | other |
Token ID | 584 |
Feature activation | +8.86e-02 |
Pos logit contributions | |
ter | +0.080 |
ety | +0.079 |
ried | +0.078 |
over | +0.076 |
Blue | +0.076 |
Neg logit contributions | |
rd | -0.130 |
oc | -0.122 |
very | -0.121 |
iden | -0.120 |
inations | -0.119 |
Token | cat |
Token ID | 3797 |
Feature activation | -8.27e-03 |
Pos logit contributions | |
Blue | +0.091 |
asleep | +0.082 |
back | +0.079 |
pering | +0.077 |
blown | +0.076 |
Neg logit contributions | |
rd | -0.212 |
oc | -0.204 |
iest | -0.195 |
ises | -0.192 |
iden | -0.191 |
Token | said |
Token ID | 531 |
Feature activation | -1.04e-02 |
Pos logit contributions | |
rd | +0.013 |
inations | +0.013 |
iden | +0.013 |
nexpected | +0.013 |
very | +0.012 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.015 |
from | -0.014 |
Spot | -0.014 |
Token | , |
Token ID | 11 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
rd | +0.021 |
inations | +0.021 |
iden | +0.020 |
nexpected | +0.020 |
very | +0.020 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.014 |
from | -0.014 |
Spot | -0.014 |
Token | " |
Token ID | 366 |
Feature activation | +2.29e-02 |
Pos logit contributions | |
ety | +0.040 |
isk | +0.038 |
blown | +0.037 |
Blue | +0.036 |
burn | +0.034 |
Neg logit contributions | |
iden | -0.064 |
very | -0.064 |
nexpected | -0.060 |
rd | -0.059 |
mmm | -0.058 |
Token | Maybe |
Token ID | 13300 |
Feature activation | +1.66e-02 |
Pos logit contributions | |
Blue | +0.033 |
over | +0.030 |
back | +0.029 |
from | +0.028 |
Spot | +0.028 |
Neg logit contributions | |
inations | -0.049 |
rd | -0.049 |
iden | -0.048 |
nexpected | -0.047 |
very | -0.047 |
Token | again |
Token ID | 757 |
Feature activation | +1.87e-02 |
Pos logit contributions | |
rd | +0.032 |
iest | +0.032 |
inations | +0.032 |
very | +0.031 |
iden | +0.031 |
Neg logit contributions | |
Blue | -0.017 |
ety | -0.015 |
isk | -0.015 |
Spot | -0.014 |
over | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -1.28e-02 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.020 |
Spot | +0.019 |
back | +0.019 |
Sunny | +0.018 |
Neg logit contributions | |
rd | -0.043 |
inations | -0.043 |
iden | -0.042 |
very | -0.042 |
nexpected | -0.042 |
Token |
|
Token ID | 198 |
Feature activation | +6.02e-03 |
Pos logit contributions | |
rd | +0.028 |
iden | +0.028 |
very | +0.028 |
inations | +0.028 |
nexpected | +0.028 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.015 |
from | -0.014 |
barking | -0.014 |
Token |
|
Token ID | 198 |
Feature activation | +1.45e-02 |
Pos logit contributions | |
Blue | +0.006 |
over | +0.005 |
Spot | +0.005 |
from | +0.005 |
back | +0.005 |
Neg logit contributions | |
rd | -0.016 |
inations | -0.016 |
iden | -0.016 |
nexpected | -0.015 |
iest | -0.015 |
Token | The |
Token ID | 464 |
Feature activation | +5.48e-02 |
Pos logit contributions | |
Blue | +0.024 |
over | +0.022 |
Spot | +0.021 |
from | +0.021 |
back | +0.020 |
Neg logit contributions | |
inations | -0.028 |
rd | -0.028 |
iden | -0.028 |
nexpected | -0.027 |
iest | -0.027 |
Token | other |
Token ID | 584 |
Feature activation | +8.72e-02 |
Pos logit contributions | |
Blue | +0.079 |
over | +0.076 |
ety | +0.069 |
apart | +0.069 |
Spot | +0.068 |
Neg logit contributions | |
rd | -0.108 |
iest | -0.107 |
iden | -0.106 |
inations | -0.106 |
very | -0.106 |
Token | animals |
Token ID | 4695 |
Feature activation | -1.40e-03 |
Pos logit contributions | |
Blue | +0.091 |
over | +0.082 |
asleep | +0.073 |
back | +0.071 |
apart | +0.071 |
Neg logit contributions | |
rd | -0.212 |
iest | -0.212 |
oc | -0.207 |
nexpected | -0.203 |
iden | -0.201 |
Token | saw |
Token ID | 2497 |
Feature activation | -2.21e-02 |
Pos logit contributions | |
inations | +0.003 |
rd | +0.003 |
iden | +0.003 |
iest | +0.003 |
nexpected | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
Spot | -0.002 |
barking | -0.002 |
ety | -0.002 |
Token | how |
Token ID | 703 |
Feature activation | -3.67e-02 |
Pos logit contributions | |
nexpected | +0.048 |
iden | +0.048 |
iest | +0.048 |
inations | +0.047 |
rd | +0.047 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.024 |
ety | -0.023 |
back | -0.023 |
pering | -0.022 |
Token | the |
Token ID | 262 |
Feature activation | +4.35e-02 |
Pos logit contributions | |
rd | +0.081 |
inations | +0.081 |
nexpected | +0.080 |
iden | +0.080 |
arest | +0.079 |
Neg logit contributions | |
Blue | -0.051 |
over | -0.045 |
back | -0.044 |
Spot | -0.043 |
from | -0.043 |
Token | farmer |
Token ID | 18739 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
over | +0.051 |
Blue | +0.050 |
apart | +0.044 |
ety | +0.044 |
from | +0.043 |
Neg logit contributions | |
rd | -0.097 |
iest | -0.096 |
inations | -0.093 |
iden | -0.093 |
very | -0.092 |
Token | So |
Token ID | 1406 |
Feature activation | +3.50e-03 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.005 |
back | +0.005 |
from | +0.005 |
barking | +0.005 |
Neg logit contributions | |
rd | -0.008 |
iden | -0.008 |
inations | -0.008 |
nexpected | -0.007 |
very | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | +1.48e-02 |
Pos logit contributions | |
Blue | +0.004 |
from | +0.004 |
Spot | +0.004 |
over | +0.004 |
back | +0.004 |
Neg logit contributions | |
inations | -0.010 |
rd | -0.009 |
arest | -0.009 |
iden | -0.009 |
affles | -0.009 |
Token | he |
Token ID | 339 |
Feature activation | +1.45e-02 |
Pos logit contributions | |
Blue | +0.024 |
over | +0.022 |
back | +0.021 |
from | +0.021 |
Spot | +0.021 |
Neg logit contributions | |
rd | -0.029 |
inations | -0.029 |
iden | -0.029 |
nexpected | -0.028 |
arest | -0.028 |
Token | asked |
Token ID | 1965 |
Feature activation | +3.97e-03 |
Pos logit contributions | |
Blue | +0.022 |
over | +0.020 |
asleep | +0.020 |
Sunny | +0.020 |
pering | +0.019 |
Neg logit contributions | |
iden | -0.028 |
nexpected | -0.027 |
inations | -0.026 |
very | -0.026 |
rd | -0.025 |
Token | the |
Token ID | 262 |
Feature activation | +5.15e-02 |
Pos logit contributions | |
ety | +0.006 |
Blue | +0.006 |
blown | +0.006 |
pering | +0.005 |
barking | +0.005 |
Neg logit contributions | |
nexpected | -0.007 |
iest | -0.007 |
iden | -0.007 |
rd | -0.007 |
very | -0.007 |
Token | other |
Token ID | 584 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
Blue | +0.071 |
over | +0.067 |
from | +0.062 |
ety | +0.061 |
Spot | +0.061 |
Neg logit contributions | |
inations | -0.108 |
rd | -0.107 |
iden | -0.107 |
very | -0.105 |
nexpected | -0.105 |
Token | animals |
Token ID | 4695 |
Feature activation | +2.86e-02 |
Pos logit contributions | |
Blue | +0.083 |
over | +0.072 |
back | +0.064 |
Spot | +0.064 |
Sunny | +0.062 |
Neg logit contributions | |
rd | -0.215 |
inations | -0.212 |
iden | -0.211 |
iest | -0.211 |
nexpected | -0.207 |
Token | for |
Token ID | 329 |
Feature activation | +2.35e-03 |
Pos logit contributions | |
Blue | +0.038 |
blown | +0.033 |
ety | +0.032 |
Sunny | +0.031 |
Wh | +0.031 |
Neg logit contributions | |
inations | -0.059 |
iest | -0.058 |
iden | -0.057 |
rd | -0.055 |
ises | -0.054 |
Token | help |
Token ID | 1037 |
Feature activation | -1.84e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.004 |
ety | +0.004 |
back | +0.004 |
from | +0.004 |
Neg logit contributions | |
inations | -0.004 |
nexpected | -0.004 |
rd | -0.004 |
iest | -0.004 |
arest | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | -3.66e-03 |
Pos logit contributions | |
inations | +0.034 |
rd | +0.034 |
nexpected | +0.032 |
very | +0.032 |
iest | +0.032 |
Neg logit contributions | |
Blue | -0.028 |
ety | -0.025 |
Sunny | -0.025 |
blown | -0.024 |
asleep | -0.024 |
Token | The |
Token ID | 383 |
Feature activation | +5.11e-02 |
Pos logit contributions | |
iden | +0.008 |
rd | +0.008 |
very | +0.008 |
nexpected | +0.008 |
inations | +0.008 |
Neg logit contributions | |
Blue | -0.005 |
over | -0.004 |
back | -0.004 |
from | -0.004 |
barking | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +6.90e-03 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.005 |
back | +0.005 |
from | +0.005 |
barking | +0.004 |
Neg logit contributions | |
rd | -0.009 |
iden | -0.009 |
very | -0.009 |
nexpected | -0.008 |
inations | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | +1.30e-02 |
Pos logit contributions | |
Blue | +0.006 |
over | +0.005 |
Spot | +0.005 |
from | +0.004 |
back | +0.004 |
Neg logit contributions | |
inations | -0.019 |
nexpected | -0.019 |
iest | -0.019 |
rd | -0.019 |
ises | -0.019 |
Token | She |
Token ID | 3347 |
Feature activation | +8.84e-03 |
Pos logit contributions | |
Blue | +0.019 |
over | +0.018 |
Spot | +0.017 |
from | +0.016 |
back | +0.016 |
Neg logit contributions | |
inations | -0.027 |
iest | -0.027 |
rd | -0.027 |
iden | -0.026 |
nexpected | -0.026 |
Token | asked |
Token ID | 1965 |
Feature activation | -1.82e-03 |
Pos logit contributions | |
Blue | +0.014 |
over | +0.013 |
Sunny | +0.012 |
Spot | +0.012 |
back | +0.012 |
Neg logit contributions | |
nexpected | -0.017 |
rd | -0.017 |
iden | -0.016 |
inations | -0.016 |
very | -0.016 |
Token | the |
Token ID | 262 |
Feature activation | +5.03e-02 |
Pos logit contributions | |
nexpected | +0.004 |
rd | +0.003 |
iest | +0.003 |
very | +0.003 |
iden | +0.003 |
Neg logit contributions | |
ety | -0.002 |
blown | -0.002 |
pering | -0.002 |
Blue | -0.002 |
barking | -0.002 |
Token | other |
Token ID | 584 |
Feature activation | +8.49e-02 |
Pos logit contributions | |
Blue | +0.071 |
over | +0.064 |
ety | +0.059 |
apart | +0.057 |
from | +0.057 |
Neg logit contributions | |
iden | -0.105 |
ises | -0.102 |
very | -0.101 |
rd | -0.100 |
iest | -0.100 |
Token | sheep |
Token ID | 15900 |
Feature activation | +1.50e-02 |
Pos logit contributions | |
Blue | +0.108 |
over | +0.087 |
Sunny | +0.082 |
asleep | +0.080 |
blown | +0.080 |
Neg logit contributions | |
rd | -0.189 |
iest | -0.188 |
iden | -0.188 |
ises | -0.187 |
inations | -0.178 |
Token | for |
Token ID | 329 |
Feature activation | +6.62e-03 |
Pos logit contributions | |
blown | +0.020 |
Blue | +0.020 |
ety | +0.018 |
k | +0.017 |
Sunny | +0.017 |
Neg logit contributions | |
inations | -0.028 |
iden | -0.028 |
rd | -0.027 |
iest | -0.026 |
ises | -0.026 |
Token | help |
Token ID | 1037 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.011 |
ety | +0.010 |
back | +0.010 |
from | +0.010 |
Neg logit contributions | |
inations | -0.012 |
rd | -0.012 |
nexpected | -0.012 |
arest | -0.011 |
iest | -0.011 |
Token | , |
Token ID | 11 |
Feature activation | +5.87e-03 |
Pos logit contributions | |
rd | +0.040 |
inations | +0.040 |
very | +0.037 |
iest | +0.036 |
occasions | +0.036 |
Neg logit contributions | |
Blue | -0.037 |
ety | -0.035 |
blown | -0.033 |
Sunny | -0.032 |
Spot | -0.032 |
Token | but |
Token ID | 475 |
Feature activation | -9.54e-03 |
Pos logit contributions | |
Blue | +0.008 |
ety | +0.008 |
burn | +0.008 |
isk | +0.007 |
Sunny | +0.007 |
Neg logit contributions | |
inations | -0.011 |
iden | -0.011 |
very | -0.011 |
rd | -0.010 |
mmm | -0.010 |
Token | with |
Token ID | 351 |
Feature activation | +2.61e-03 |
Pos logit contributions | |
rd | +0.027 |
iden | +0.027 |
inations | +0.027 |
nexpected | +0.027 |
very | +0.026 |
Neg logit contributions | |
Blue | -0.022 |
over | -0.022 |
back | -0.021 |
from | -0.021 |
Spot | -0.020 |
Token | this |
Token ID | 428 |
Feature activation | -2.70e-02 |
Pos logit contributions | |
Blue | +0.003 |
over | +0.003 |
back | +0.003 |
from | +0.003 |
Spot | +0.003 |
Neg logit contributions | |
inations | -0.006 |
rd | -0.006 |
nexpected | -0.006 |
iden | -0.006 |
arest | -0.006 |
Token | stone |
Token ID | 7815 |
Feature activation | +2.87e-03 |
Pos logit contributions | |
rd | +0.052 |
inations | +0.051 |
iden | +0.050 |
nexpected | +0.050 |
very | +0.049 |
Neg logit contributions | |
Blue | -0.044 |
over | -0.041 |
back | -0.040 |
from | -0.039 |
Spot | -0.038 |
Token | !" |
Token ID | 2474 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
Blue | +0.006 |
over | +0.005 |
Spot | +0.005 |
barking | +0.005 |
ety | +0.005 |
Neg logit contributions | |
inations | -0.004 |
rd | -0.004 |
iest | -0.004 |
arest | -0.004 |
iden | -0.004 |
Token | The |
Token ID | 383 |
Feature activation | +5.88e-02 |
Pos logit contributions | |
inations | +0.029 |
rd | +0.029 |
iden | +0.029 |
nexpected | +0.028 |
arest | +0.028 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.014 |
back | -0.013 |
Spot | -0.013 |
from | -0.013 |
Token | other |
Token ID | 584 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
Blue | +0.088 |
Sunny | +0.079 |
over | +0.078 |
asleep | +0.074 |
from | +0.074 |
Neg logit contributions | |
iden | -0.122 |
rd | -0.118 |
inations | -0.116 |
very | -0.115 |
iest | -0.115 |
Token | kid |
Token ID | 5141 |
Feature activation | +5.08e-04 |
Pos logit contributions | |
Blue | +0.102 |
over | +0.091 |
back | +0.084 |
Spot | +0.083 |
from | +0.082 |
Neg logit contributions | |
rd | -0.200 |
inations | -0.198 |
iden | -0.196 |
nexpected | -0.194 |
iest | -0.193 |
Token | disagreed |
Token ID | 27884 |
Feature activation | -1.02e-03 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
from | +0.001 |
Spot | +0.001 |
Neg logit contributions | |
inations | -0.001 |
rd | -0.001 |
arest | -0.001 |
iest | -0.001 |
iden | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | -3.66e-04 |
Pos logit contributions | |
rd | +0.002 |
very | +0.002 |
nexpected | +0.002 |
iden | +0.002 |
iest | +0.002 |
Neg logit contributions | |
Blue | -0.001 |
Sunny | -0.001 |
Bird | -0.001 |
ety | -0.001 |
Pink | -0.001 |
Token | said |
Token ID | 531 |
Feature activation | -8.94e-03 |
Pos logit contributions | |
iden | +0.001 |
nexpected | +0.001 |
very | +0.001 |
rd | +0.001 |
inations | +0.001 |
Neg logit contributions | |
Blue | -0.001 |
over | -0.001 |
back | -0.001 |
from | -0.001 |
ety | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | +3.68e-02 |
Pos logit contributions | |
inations | +0.022 |
rd | +0.021 |
arest | +0.021 |
very | +0.020 |
iden | +0.020 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.010 |
back | -0.010 |
Spot | -0.010 |
from | -0.010 |
Token | . |
Token ID | 13 |
Feature activation | +3.93e-03 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
ety | +0.001 |
Sunny | +0.000 |
blown | +0.000 |
Neg logit contributions | |
rd | -0.001 |
oc | -0.001 |
iden | -0.001 |
inations | -0.001 |
iest | -0.001 |
Token | |
Token ID | 220 |
Feature activation | -4.12e-02 |
Pos logit contributions | |
over | +0.005 |
Blue | +0.005 |
back | +0.005 |
ter | +0.005 |
blown | +0.005 |
Neg logit contributions | |
iden | -0.008 |
very | -0.008 |
rd | -0.008 |
nexpected | -0.008 |
oc | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | +1.31e-02 |
Pos logit contributions | |
rd | +0.081 |
inations | +0.080 |
iden | +0.079 |
arest | +0.078 |
very | +0.077 |
Neg logit contributions | |
Blue | -0.060 |
over | -0.054 |
back | -0.053 |
ter | -0.053 |
barking | -0.052 |
Token |
|
Token ID | 198 |
Feature activation | +1.90e-02 |
Pos logit contributions | |
Blue | +0.014 |
over | +0.013 |
Spot | +0.012 |
back | +0.011 |
from | +0.011 |
Neg logit contributions | |
inations | -0.033 |
rd | -0.033 |
nexpected | -0.032 |
iest | -0.032 |
iden | -0.032 |
Token | The |
Token ID | 464 |
Feature activation | +6.05e-02 |
Pos logit contributions | |
Blue | +0.031 |
over | +0.029 |
back | +0.027 |
from | +0.027 |
Spot | +0.027 |
Neg logit contributions | |
rd | -0.037 |
inations | -0.037 |
iden | -0.036 |
nexpected | -0.036 |
very | -0.035 |
Token | young |
Token ID | 1862 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
ety | +0.089 |
over | +0.085 |
ried | +0.082 |
apart | +0.080 |
pering | +0.077 |
Neg logit contributions | |
iden | -0.108 |
oc | -0.108 |
rd | -0.107 |
arest | -0.103 |
iest | -0.103 |
Token | driver |
Token ID | 4639 |
Feature activation | -1.54e-02 |
Pos logit contributions | |
over | +0.139 |
pering | +0.132 |
ety | +0.132 |
isk | +0.129 |
apart | +0.126 |
Neg logit contributions | |
oc | -0.146 |
iden | -0.141 |
iest | -0.139 |
opter | -0.136 |
rd | -0.134 |
Token | smiled |
Token ID | 13541 |
Feature activation | -1.64e-03 |
Pos logit contributions | |
rd | +0.029 |
inations | +0.029 |
iden | +0.029 |
nexpected | +0.029 |
very | +0.028 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.023 |
back | -0.022 |
Spot | -0.022 |
from | -0.022 |
Token | and |
Token ID | 290 |
Feature activation | -1.46e-02 |
Pos logit contributions | |
very | +0.004 |
iden | +0.003 |
rd | +0.003 |
iest | +0.003 |
ement | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
isk | -0.002 |
blown | -0.002 |
Wh | -0.002 |
ety | -0.002 |
Token | picked |
Token ID | 6497 |
Feature activation | +1.68e-04 |
Pos logit contributions | |
iden | +0.024 |
very | +0.023 |
nexpected | +0.023 |
inations | +0.023 |
rd | +0.022 |
Neg logit contributions | |
Blue | -0.026 |
over | -0.025 |
back | -0.024 |
ety | -0.024 |
from | -0.023 |
Token | up |
Token ID | 510 |
Feature activation | +6.04e-03 |
Pos logit contributions | |
Blue | +0.000 |
over | +0.000 |
ety | +0.000 |
barking | +0.000 |
isk | +0.000 |
Neg logit contributions | |
nexpected | -0.000 |
iden | -0.000 |
rd | -0.000 |
very | -0.000 |
iest | -0.000 |
Token | than |
Token ID | 621 |
Feature activation | +3.47e-02 |
Pos logit contributions | |
iest | +0.013 |
inations | +0.011 |
rd | +0.011 |
very | +0.011 |
iden | +0.011 |
Neg logit contributions | |
Blue | -0.009 |
isk | -0.008 |
over | -0.008 |
Spot | -0.008 |
Sunny | -0.008 |
Token | the |
Token ID | 262 |
Feature activation | +4.81e-02 |
Pos logit contributions | |
Blue | +0.044 |
over | +0.041 |
barking | +0.040 |
back | +0.039 |
isk | +0.038 |
Neg logit contributions | |
iest | -0.074 |
inations | -0.073 |
iden | -0.071 |
very | -0.070 |
rd | -0.070 |
Token | others |
Token ID | 1854 |
Feature activation | +7.95e-03 |
Pos logit contributions | |
Blue | +0.063 |
over | +0.058 |
isk | +0.058 |
Spot | +0.053 |
Wh | +0.053 |
Neg logit contributions | |
iest | -0.098 |
rd | -0.095 |
arest | -0.094 |
inations | -0.093 |
iden | -0.091 |
Token | . |
Token ID | 13 |
Feature activation | -3.15e-04 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.009 |
isk | +0.009 |
back | +0.009 |
ety | +0.009 |
Neg logit contributions | |
inations | -0.016 |
iest | -0.016 |
rd | -0.016 |
iden | -0.016 |
very | -0.016 |
Token | The |
Token ID | 383 |
Feature activation | +4.87e-02 |
Pos logit contributions | |
rd | +0.001 |
iden | +0.001 |
inations | +0.001 |
very | +0.001 |
nexpected | +0.001 |
Neg logit contributions | |
Blue | -0.000 |
over | -0.000 |
back | -0.000 |
from | -0.000 |
barking | -0.000 |
Token | other |
Token ID | 584 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
Blue | +0.058 |
over | +0.052 |
isk | +0.048 |
apart | +0.048 |
from | +0.046 |
Neg logit contributions | |
iest | -0.108 |
rd | -0.107 |
inations | -0.105 |
arest | -0.103 |
iden | -0.102 |
Token | ducks |
Token ID | 39694 |
Feature activation | +1.98e-02 |
Pos logit contributions | |
Blue | +0.089 |
over | +0.072 |
asleep | +0.069 |
Sunny | +0.069 |
apart | +0.068 |
Neg logit contributions | |
rd | -0.211 |
iest | -0.206 |
oc | -0.196 |
nexpected | -0.194 |
inations | -0.193 |
Token | didn |
Token ID | 1422 |
Feature activation | -7.29e-03 |
Pos logit contributions | |
Blue | +0.034 |
Sunny | +0.030 |
Spot | +0.030 |
over | +0.030 |
back | +0.030 |
Neg logit contributions | |
inations | -0.034 |
iest | -0.033 |
iden | -0.033 |
very | -0.032 |
rd | -0.032 |
Token | 't |
Token ID | 470 |
Feature activation | +2.26e-02 |
Pos logit contributions | |
arest | +0.012 |
nexpected | +0.011 |
iest | +0.011 |
mmm | +0.011 |
inations | +0.011 |
Neg logit contributions | |
ety | -0.013 |
Blue | -0.013 |
ter | -0.012 |
isk | -0.012 |
asleep | -0.012 |
Token | mind |
Token ID | 2000 |
Feature activation | -4.14e-03 |
Pos logit contributions | |
Blue | +0.029 |
over | +0.026 |
Spot | +0.025 |
back | +0.025 |
Sunny | +0.025 |
Neg logit contributions | |
inations | -0.051 |
iest | -0.049 |
rd | -0.048 |
iden | -0.048 |
very | -0.047 |
Token | , |
Token ID | 11 |
Feature activation | +8.93e-03 |
Pos logit contributions | |
iden | +0.009 |
very | +0.009 |
iest | +0.009 |
inations | +0.009 |
rd | +0.009 |
Neg logit contributions | |
Blue | -0.005 |
back | -0.004 |
over | -0.004 |
isk | -0.004 |
Sunny | -0.004 |
Token |
|
Token ID | 198 |
Feature activation | +4.98e-03 |
Pos logit contributions | |
rd | +0.089 |
inations | +0.088 |
arest | +0.086 |
iden | +0.086 |
nexpected | +0.085 |
Neg logit contributions | |
Blue | -0.066 |
over | -0.060 |
back | -0.059 |
barking | -0.058 |
ter | -0.057 |
Token |
|
Token ID | 198 |
Feature activation | +1.11e-02 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.005 |
Spot | +0.004 |
back | +0.004 |
from | +0.004 |
Neg logit contributions | |
inations | -0.012 |
rd | -0.012 |
iden | -0.012 |
nexpected | -0.012 |
iest | -0.012 |
Token | Meanwhile |
Token ID | 10294 |
Feature activation | +1.68e-03 |
Pos logit contributions | |
Blue | +0.022 |
over | +0.021 |
back | +0.020 |
from | +0.019 |
Spot | +0.019 |
Neg logit contributions | |
rd | -0.017 |
inations | -0.017 |
nexpected | -0.017 |
iden | -0.017 |
arest | -0.017 |
Token | , |
Token ID | 11 |
Feature activation | -8.51e-03 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.002 |
back | +0.002 |
Spot | +0.002 |
from | +0.002 |
Neg logit contributions | |
inations | -0.004 |
rd | -0.004 |
arest | -0.004 |
iden | -0.004 |
nexpected | -0.004 |
Token | the |
Token ID | 262 |
Feature activation | +5.35e-02 |
Pos logit contributions | |
nexpected | +0.021 |
very | +0.021 |
iden | +0.021 |
iest | +0.021 |
arest | +0.021 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.007 |
back | -0.006 |
ety | -0.006 |
barking | -0.006 |
Token | other |
Token ID | 584 |
Feature activation | +8.45e-02 |
Pos logit contributions | |
over | +0.070 |
ety | +0.070 |
Blue | +0.069 |
from | +0.060 |
isk | +0.059 |
Neg logit contributions | |
rd | -0.109 |
iest | -0.108 |
inations | -0.108 |
iden | -0.107 |
arest | -0.106 |
Token | animals |
Token ID | 4695 |
Feature activation | +4.03e-03 |
Pos logit contributions | |
Blue | +0.095 |
over | +0.092 |
Sunny | +0.079 |
undone | +0.079 |
ety | +0.076 |
Neg logit contributions | |
oc | -0.193 |
iest | -0.190 |
rd | -0.185 |
ises | -0.177 |
nexpected | -0.172 |
Token | on |
Token ID | 319 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
Blue | +0.007 |
over | +0.006 |
Sunny | +0.006 |
ety | +0.006 |
Spot | +0.006 |
Neg logit contributions | |
inations | -0.007 |
iden | -0.007 |
rd | -0.007 |
nexpected | -0.007 |
iest | -0.006 |
Token | the |
Token ID | 262 |
Feature activation | -1.18e-03 |
Pos logit contributions | |
ety | +0.013 |
over | +0.012 |
back | +0.012 |
Blue | +0.011 |
ter | +0.011 |
Neg logit contributions | |
nexpected | -0.026 |
iden | -0.026 |
very | -0.025 |
rd | -0.025 |
iest | -0.025 |
Token | farm |
Token ID | 5318 |
Feature activation | -1.20e-02 |
Pos logit contributions | |
rd | +0.003 |
inations | +0.003 |
iden | +0.002 |
nexpected | +0.002 |
very | +0.002 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.001 |
ety | -0.001 |
Spot | -0.001 |
from | -0.001 |
Token | noticed |
Token ID | 6810 |
Feature activation | -2.58e-02 |
Pos logit contributions | |
rd | +0.021 |
iden | +0.020 |
nexpected | +0.020 |
inations | +0.020 |
very | +0.019 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.019 |
ety | -0.018 |
Sunny | -0.018 |
back | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -2.22e-04 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.011 |
back | +0.011 |
ety | +0.011 |
Spot | +0.011 |
Neg logit contributions | |
rd | -0.021 |
inations | -0.021 |
iest | -0.020 |
iden | -0.020 |
arest | -0.019 |
Token | |
Token ID | 220 |
Feature activation | -3.40e-02 |
Pos logit contributions | |
rd | +0.000 |
iden | +0.000 |
very | +0.000 |
nexpected | +0.000 |
inations | +0.000 |
Neg logit contributions | |
Blue | -0.000 |
over | -0.000 |
blown | -0.000 |
from | -0.000 |
back | -0.000 |
Token |
|
Token ID | 198 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
rd | +0.069 |
inations | +0.068 |
iden | +0.067 |
arest | +0.066 |
nexpected | +0.066 |
Neg logit contributions | |
Blue | -0.049 |
over | -0.044 |
back | -0.043 |
from | -0.042 |
barking | -0.042 |
Token |
|
Token ID | 198 |
Feature activation | +1.67e-02 |
Pos logit contributions | |
Blue | +0.012 |
over | +0.011 |
Spot | +0.010 |
back | +0.009 |
from | +0.009 |
Neg logit contributions | |
inations | -0.030 |
rd | -0.030 |
nexpected | -0.030 |
iest | -0.029 |
iden | -0.029 |
Token | The |
Token ID | 464 |
Feature activation | +6.12e-02 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.024 |
back | +0.023 |
from | +0.022 |
Spot | +0.022 |
Neg logit contributions | |
rd | -0.034 |
inations | -0.034 |
iden | -0.033 |
nexpected | -0.033 |
very | -0.033 |
Token | other |
Token ID | 584 |
Feature activation | +8.45e-02 |
Pos logit contributions | |
Blue | +0.084 |
isk | +0.077 |
over | +0.075 |
blown | +0.075 |
ety | +0.075 |
Neg logit contributions | |
rd | -0.121 |
iest | -0.118 |
very | -0.116 |
iden | -0.114 |
nexpected | -0.113 |
Token | frogs |
Token ID | 37475 |
Feature activation | +2.15e-02 |
Pos logit contributions | |
Blue | +0.101 |
over | +0.084 |
blown | +0.083 |
apart | +0.080 |
asleep | +0.080 |
Neg logit contributions | |
rd | -0.195 |
iest | -0.191 |
nexpected | -0.184 |
iden | -0.179 |
very | -0.178 |
Token | could |
Token ID | 714 |
Feature activation | +6.06e-03 |
Pos logit contributions | |
Blue | +0.036 |
over | +0.032 |
back | +0.032 |
Spot | +0.032 |
Sunny | +0.031 |
Neg logit contributions | |
inations | -0.037 |
rd | -0.037 |
nexpected | -0.037 |
iden | -0.037 |
very | -0.036 |
Token | jump |
Token ID | 4391 |
Feature activation | +3.98e-02 |
Pos logit contributions | |
Blue | +0.008 |
ety | +0.007 |
ter | +0.007 |
over | +0.007 |
apart | +0.007 |
Neg logit contributions | |
inations | -0.012 |
rd | -0.012 |
arest | -0.012 |
nexpected | -0.011 |
ises | -0.011 |
Token | high |
Token ID | 1029 |
Feature activation | +7.24e-03 |
Pos logit contributions | |
Blue | +0.045 |
ety | +0.045 |
vet | +0.043 |
Johnson | +0.042 |
isk | +0.042 |
Neg logit contributions | |
iest | -0.081 |
very | -0.077 |
rd | -0.075 |
nexpected | -0.075 |
inations | -0.075 |
Token | , |
Token ID | 11 |
Feature activation | -4.89e-03 |
Pos logit contributions | |
Blue | +0.008 |
isk | +0.007 |
vet | +0.007 |
ety | +0.007 |
Wh | +0.007 |
Neg logit contributions | |
very | -0.016 |
iest | -0.016 |
nexpected | -0.016 |
ises | -0.016 |
rd | -0.015 |
Token | dirt |
Token ID | 13647 |
Feature activation | +3.48e-02 |
Pos logit contributions | |
rd | +0.012 |
inations | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
arest | +0.012 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.007 |
from | -0.007 |
Spot | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | -1.23e-02 |
Pos logit contributions | |
Blue | +0.043 |
ety | +0.041 |
isk | +0.040 |
barking | +0.037 |
Squirrel | +0.036 |
Neg logit contributions | |
iest | -0.076 |
inations | -0.069 |
rd | -0.068 |
very | -0.067 |
ises | -0.067 |
Token | hay |
Token ID | 27678 |
Feature activation | +1.64e-02 |
Pos logit contributions | |
rd | +0.028 |
inations | +0.028 |
iden | +0.028 |
arest | +0.028 |
nexpected | +0.027 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.014 |
back | -0.013 |
from | -0.013 |
Spot | -0.013 |
Token | . |
Token ID | 13 |
Feature activation | +6.80e-03 |
Pos logit contributions | |
Blue | +0.024 |
ety | +0.022 |
over | +0.021 |
Sunny | +0.020 |
isk | +0.020 |
Neg logit contributions | |
iest | -0.032 |
rd | -0.031 |
inations | -0.031 |
nexpected | -0.029 |
iden | -0.029 |
Token | The |
Token ID | 383 |
Feature activation | +4.21e-02 |
Pos logit contributions | |
Blue | +0.009 |
over | +0.009 |
back | +0.009 |
from | +0.008 |
blown | +0.008 |
Neg logit contributions | |
rd | -0.014 |
iden | -0.014 |
very | -0.014 |
inations | -0.014 |
nexpected | -0.014 |
Token | other |
Token ID | 584 |
Feature activation | +8.44e-02 |
Pos logit contributions | |
Blue | +0.055 |
over | +0.050 |
back | +0.046 |
from | +0.045 |
Spot | +0.045 |
Neg logit contributions | |
rd | -0.095 |
inations | -0.094 |
iden | -0.094 |
nexpected | -0.092 |
iest | -0.092 |
Token | animals |
Token ID | 4695 |
Feature activation | +6.23e-03 |
Pos logit contributions | |
Blue | +0.080 |
over | +0.067 |
back | +0.060 |
from | +0.058 |
Spot | +0.057 |
Neg logit contributions | |
rd | -0.222 |
iden | -0.218 |
inations | -0.218 |
iest | -0.217 |
nexpected | -0.215 |
Token | did |
Token ID | 750 |
Feature activation | -3.23e-03 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.008 |
back | +0.007 |
Spot | +0.007 |
from | +0.007 |
Neg logit contributions | |
inations | -0.013 |
rd | -0.013 |
iden | -0.013 |
nexpected | -0.013 |
iest | -0.013 |
Token | not |
Token ID | 407 |
Feature activation | +3.44e-02 |
Pos logit contributions | |
rd | +0.010 |
inations | +0.010 |
iden | +0.010 |
nexpected | +0.010 |
very | +0.009 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.001 |
back | -0.001 |
from | -0.001 |
Spot | -0.001 |
Token | like |
Token ID | 588 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
Blue | +0.047 |
over | +0.041 |
Spot | +0.040 |
Sunny | +0.040 |
back | +0.039 |
Neg logit contributions | |
inations | -0.076 |
rd | -0.073 |
iest | -0.072 |
iden | -0.071 |
very | -0.071 |
Token | to |
Token ID | 284 |
Feature activation | -1.01e-02 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.022 |
ety | +0.022 |
ter | +0.021 |
back | +0.021 |
Neg logit contributions | |
inations | -0.050 |
iden | -0.049 |
nexpected | -0.049 |
rd | -0.048 |
very | -0.047 |
Token | scared |
Token ID | 12008 |
Feature activation | -1.35e-03 |
Pos logit contributions | |
rd | +0.047 |
iden | +0.047 |
inations | +0.047 |
nexpected | +0.047 |
very | +0.046 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.018 |
back | -0.017 |
from | -0.017 |
Spot | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | -2.41e-03 |
Pos logit contributions | |
rd | +0.003 |
iest | +0.003 |
inations | +0.003 |
iden | +0.003 |
very | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
ety | -0.002 |
over | -0.002 |
blown | -0.002 |
Sunny | -0.002 |
Token |
|
Token ID | 198 |
Feature activation | +9.78e-03 |
Pos logit contributions | |
rd | +0.005 |
iden | +0.005 |
nexpected | +0.005 |
inations | +0.005 |
very | +0.005 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.003 |
back | -0.003 |
from | -0.003 |
ety | -0.003 |
Token |
|
Token ID | 198 |
Feature activation | +1.40e-02 |
Pos logit contributions | |
Blue | +0.009 |
over | +0.009 |
Spot | +0.008 |
from | +0.007 |
back | +0.007 |
Neg logit contributions | |
inations | -0.026 |
rd | -0.025 |
iest | -0.025 |
nexpected | -0.025 |
iden | -0.025 |
Token | The |
Token ID | 464 |
Feature activation | +5.58e-02 |
Pos logit contributions | |
Blue | +0.024 |
over | +0.023 |
Spot | +0.021 |
back | +0.021 |
from | +0.021 |
Neg logit contributions | |
inations | -0.026 |
iest | -0.026 |
rd | -0.026 |
iden | -0.025 |
nexpected | -0.025 |
Token | other |
Token ID | 584 |
Feature activation | +8.41e-02 |
Pos logit contributions | |
Blue | +0.067 |
over | +0.060 |
from | +0.054 |
back | +0.054 |
Spot | +0.054 |
Neg logit contributions | |
rd | -0.131 |
inations | -0.131 |
iden | -0.129 |
nexpected | -0.127 |
iest | -0.127 |
Token | animals |
Token ID | 4695 |
Feature activation | +7.79e-04 |
Pos logit contributions | |
Blue | +0.070 |
over | +0.059 |
back | +0.053 |
from | +0.051 |
Spot | +0.051 |
Neg logit contributions | |
rd | -0.229 |
inations | -0.227 |
iden | -0.225 |
nexpected | -0.224 |
iest | -0.222 |
Token | got |
Token ID | 1392 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
Sunny | +0.001 |
Neg logit contributions | |
inations | -0.001 |
rd | -0.001 |
iden | -0.001 |
nexpected | -0.001 |
iest | -0.001 |
Token | together |
Token ID | 1978 |
Feature activation | +2.27e-02 |
Pos logit contributions | |
Blue | +0.029 |
isk | +0.026 |
ety | +0.026 |
Sunny | +0.025 |
ter | +0.024 |
Neg logit contributions | |
iden | -0.045 |
inations | -0.045 |
nexpected | -0.044 |
rd | -0.044 |
ises | -0.043 |
Token | and |
Token ID | 290 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
Blue | +0.026 |
ety | +0.023 |
ter | +0.020 |
Sunny | +0.020 |
Wh | +0.020 |
Neg logit contributions | |
inations | -0.051 |
rd | -0.051 |
iden | -0.051 |
very | -0.049 |
nexpected | -0.049 |
Token | decided |
Token ID | 3066 |
Feature activation | +2.70e-02 |
Pos logit contributions | |
iden | +0.041 |
nexpected | +0.041 |
inations | +0.040 |
rd | +0.040 |
very | +0.039 |
Neg logit contributions | |
Blue | -0.031 |
over | -0.028 |
back | -0.027 |
ety | -0.027 |
Sunny | -0.026 |
Token | faster |
Token ID | 5443 |
Feature activation | +4.62e-02 |
Pos logit contributions | |
Blue | +0.037 |
over | +0.033 |
back | +0.031 |
Spot | +0.031 |
from | +0.030 |
Neg logit contributions | |
inations | -0.064 |
rd | -0.062 |
iden | -0.062 |
very | -0.061 |
nexpected | -0.061 |
Token | than |
Token ID | 621 |
Feature activation | +4.06e-02 |
Pos logit contributions | |
Blue | +0.068 |
over | +0.062 |
back | +0.059 |
Spot | +0.058 |
from | +0.057 |
Neg logit contributions | |
rd | -0.096 |
inations | -0.095 |
iden | -0.094 |
nexpected | -0.093 |
iest | -0.092 |
Token | you |
Token ID | 345 |
Feature activation | -7.06e-03 |
Pos logit contributions | |
ety | +0.079 |
Blue | +0.076 |
isk | +0.075 |
burn | +0.075 |
over | +0.075 |
Neg logit contributions | |
inations | -0.052 |
nexpected | -0.047 |
iest | -0.047 |
very | -0.047 |
mmm | -0.045 |
Token | !" |
Token ID | 2474 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.011 |
arest | +0.011 |
iest | +0.011 |
iden | +0.011 |
Neg logit contributions | |
Blue | -0.013 |
over | -0.012 |
Sunny | -0.011 |
ety | -0.011 |
Spot | -0.011 |
Token | So |
Token ID | 1406 |
Feature activation | +1.22e-03 |
Pos logit contributions | |
rd | +0.038 |
iden | +0.037 |
very | +0.037 |
inations | +0.037 |
nexpected | +0.037 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.020 |
back | -0.019 |
from | -0.018 |
ter | -0.018 |
Token | the |
Token ID | 262 |
Feature activation | +8.40e-02 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.001 |
from | +0.001 |
Spot | +0.001 |
back | +0.001 |
Neg logit contributions | |
inations | -0.003 |
rd | -0.003 |
arest | -0.003 |
iden | -0.003 |
nexpected | -0.003 |
Token | two |
Token ID | 734 |
Feature activation | +6.00e-02 |
Pos logit contributions | |
over | +0.111 |
apart | +0.103 |
Blue | +0.099 |
asleep | +0.097 |
ried | +0.097 |
Neg logit contributions | |
iden | -0.175 |
rd | -0.174 |
arest | -0.167 |
very | -0.167 |
opus | -0.165 |
Token | of |
Token ID | 286 |
Feature activation | +1.57e-02 |
Pos logit contributions | |
over | +0.106 |
Blue | +0.103 |
asleep | +0.096 |
Sunny | +0.095 |
isk | +0.093 |
Neg logit contributions | |
iest | -0.099 |
ises | -0.095 |
inations | -0.094 |
iden | -0.094 |
rd | -0.092 |
Token | them |
Token ID | 606 |
Feature activation | +3.11e-02 |
Pos logit contributions | |
Blue | +0.018 |
ety | +0.016 |
over | +0.016 |
back | +0.016 |
barking | +0.016 |
Neg logit contributions | |
nexpected | -0.036 |
inations | -0.034 |
iest | -0.034 |
iden | -0.034 |
rd | -0.034 |
Token | decided |
Token ID | 3066 |
Feature activation | +2.59e-02 |
Pos logit contributions | |
Blue | +0.047 |
Sunny | +0.042 |
over | +0.041 |
barking | +0.040 |
Spot | +0.040 |
Neg logit contributions | |
inations | -0.061 |
iden | -0.061 |
rd | -0.061 |
nexpected | -0.060 |
iest | -0.058 |
Token | to |
Token ID | 284 |
Feature activation | +2.04e-02 |
Pos logit contributions | |
Blue | +0.032 |
isk | +0.032 |
ety | +0.031 |
Sunny | +0.030 |
ter | +0.028 |
Neg logit contributions | |
iden | -0.053 |
nexpected | -0.051 |
rd | -0.051 |
inations | -0.050 |
very | -0.049 |
Token | fun |
Token ID | 1257 |
Feature activation | -4.75e-04 |
Pos logit contributions | |
rd | +0.005 |
iden | +0.005 |
very | +0.005 |
inations | +0.005 |
arest | +0.005 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.004 |
back | -0.004 |
Spot | -0.004 |
barking | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +3.26e-03 |
Pos logit contributions | |
iden | +0.001 |
arest | +0.001 |
inations | +0.001 |
very | +0.001 |
nexpected | +0.001 |
Neg logit contributions | |
over | -0.000 |
Blue | -0.000 |
from | -0.000 |
near | -0.000 |
barking | -0.000 |
Token |
|
Token ID | 198 |
Feature activation | +1.25e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.004 |
back | +0.004 |
Spot | +0.004 |
from | +0.004 |
Neg logit contributions | |
inations | -0.007 |
rd | -0.007 |
iden | -0.007 |
nexpected | -0.007 |
arest | -0.007 |
Token |
|
Token ID | 198 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.010 |
Spot | +0.009 |
from | +0.008 |
back | +0.008 |
Neg logit contributions | |
inations | -0.034 |
rd | -0.034 |
nexpected | -0.034 |
iest | -0.033 |
arest | -0.033 |
Token | The |
Token ID | 464 |
Feature activation | +5.44e-02 |
Pos logit contributions | |
over | +0.031 |
Blue | +0.030 |
also | +0.027 |
back | +0.027 |
Spot | +0.027 |
Neg logit contributions | |
iest | -0.032 |
inations | -0.032 |
arest | -0.031 |
rd | -0.031 |
ocol | -0.030 |
Token | other |
Token ID | 584 |
Feature activation | +8.40e-02 |
Pos logit contributions | |
Blue | +0.084 |
Sunny | +0.078 |
over | +0.077 |
from | +0.072 |
asleep | +0.072 |
Neg logit contributions | |
iden | -0.104 |
rd | -0.102 |
inations | -0.100 |
iest | -0.098 |
very | -0.097 |
Token | friend |
Token ID | 1545 |
Feature activation | -1.53e-03 |
Pos logit contributions | |
Blue | +0.137 |
Sunny | +0.125 |
over | +0.119 |
back | +0.113 |
asleep | +0.112 |
Neg logit contributions | |
rd | -0.164 |
iest | -0.159 |
iden | -0.159 |
nexpected | -0.154 |
inations | -0.152 |
Token | , |
Token ID | 11 |
Feature activation | -5.28e-03 |
Pos logit contributions | |
inations | +0.003 |
rd | +0.003 |
arest | +0.003 |
iden | +0.003 |
nexpected | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
Spot | -0.002 |
from | -0.002 |
Token | Jill |
Token ID | 22772 |
Feature activation | -1.32e-02 |
Pos logit contributions | |
iden | +0.011 |
rd | +0.011 |
inations | +0.011 |
nexpected | +0.011 |
very | +0.011 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.007 |
back | -0.006 |
from | -0.006 |
Spot | -0.006 |
Token | , |
Token ID | 11 |
Feature activation | -1.74e-03 |
Pos logit contributions | |
rd | +0.028 |
inations | +0.028 |
iden | +0.028 |
nexpected | +0.028 |
very | +0.027 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.017 |
back | -0.016 |
Spot | -0.016 |
from | -0.016 |
Token | was |
Token ID | 373 |
Feature activation | +1.48e-02 |
Pos logit contributions | |
iden | +0.003 |
very | +0.003 |
nexpected | +0.003 |
rd | +0.003 |
inations | +0.003 |
Neg logit contributions | |
Blue | -0.003 |
back | -0.003 |
Sunny | -0.003 |
pering | -0.003 |
over | -0.003 |
Token | selling |
Token ID | 6301 |
Feature activation | +3.04e-03 |
Pos logit contributions | |
rd | +0.021 |
inations | +0.020 |
iden | +0.019 |
very | +0.019 |
arest | +0.019 |
Neg logit contributions | |
Blue | -0.010 |
over | -0.008 |
ety | -0.008 |
Sunny | -0.007 |
Spot | -0.007 |
Token | ice |
Token ID | 4771 |
Feature activation | +1.18e-02 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.004 |
ety | +0.004 |
isk | +0.004 |
barking | +0.004 |
Neg logit contributions | |
iest | -0.006 |
rd | -0.006 |
nexpected | -0.005 |
iden | -0.005 |
inations | -0.005 |
Token | cream |
Token ID | 8566 |
Feature activation | -1.55e-03 |
Pos logit contributions | |
Blue | +0.024 |
barking | +0.021 |
over | +0.021 |
Sunny | +0.021 |
Spot | +0.021 |
Neg logit contributions | |
iest | -0.018 |
rd | -0.018 |
nexpected | -0.017 |
inations | -0.017 |
iden | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
iest | +0.003 |
iden | +0.003 |
rd | +0.003 |
oc | +0.003 |
very | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
ety | -0.002 |
Bird | -0.002 |
Tweet | -0.002 |
isk | -0.002 |
Token | They |
Token ID | 1119 |
Feature activation | +2.55e-02 |
Pos logit contributions | |
Blue | +0.022 |
over | +0.022 |
ter | +0.021 |
back | +0.021 |
ety | +0.020 |
Neg logit contributions | |
iden | -0.036 |
rd | -0.035 |
very | -0.035 |
inations | -0.034 |
nexpected | -0.034 |
Token | ran |
Token ID | 4966 |
Feature activation | +8.40e-02 |
Pos logit contributions | |
Blue | +0.046 |
over | +0.042 |
back | +0.042 |
pering | +0.041 |
Sunny | +0.041 |
Neg logit contributions | |
iden | -0.041 |
nexpected | -0.040 |
inations | -0.039 |
very | -0.038 |
Jamie | -0.038 |
Token | to |
Token ID | 284 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
Blue | +0.106 |
ter | +0.101 |
isk | +0.094 |
ety | +0.093 |
Sunny | +0.092 |
Neg logit contributions | |
iden | -0.161 |
rd | -0.159 |
very | -0.156 |
nexpected | -0.152 |
oc | -0.147 |
Token | him |
Token ID | 683 |
Feature activation | +2.96e-02 |
Pos logit contributions | |
ety | +0.022 |
Blue | +0.022 |
blown | +0.021 |
ter | +0.020 |
over | +0.020 |
Neg logit contributions | |
iden | -0.032 |
nexpected | -0.032 |
inations | -0.032 |
goodness | -0.031 |
rd | -0.031 |
Token | and |
Token ID | 290 |
Feature activation | +7.86e-03 |
Pos logit contributions | |
blown | +0.046 |
Blue | +0.044 |
Dove | +0.043 |
Woo | +0.041 |
ety | +0.041 |
Neg logit contributions | |
oc | -0.056 |
ry | -0.047 |
rd | -0.046 |
on | -0.043 |
!' | -0.042 |
Token | asked |
Token ID | 1965 |
Feature activation | +1.38e-02 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.012 |
back | +0.011 |
ety | +0.011 |
ter | +0.011 |
Neg logit contributions | |
iden | -0.014 |
inations | -0.014 |
nexpected | -0.014 |
rd | -0.014 |
very | -0.014 |
Token | for |
Token ID | 329 |
Feature activation | +1.12e-02 |
Pos logit contributions | |
Blue | +0.020 |
over | +0.018 |
blown | +0.018 |
ety | +0.017 |
barking | +0.017 |
Neg logit contributions | |
rd | -0.028 |
iden | -0.027 |
inations | -0.027 |
nexpected | -0.027 |
iest | -0.027 |
Token | felt |
Token ID | 2936 |
Feature activation | +8.83e-04 |
Pos logit contributions | |
iden | +0.015 |
rd | +0.015 |
inations | +0.015 |
nexpected | +0.015 |
very | +0.014 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.014 |
back | -0.014 |
ety | -0.014 |
from | -0.013 |
Token | happy |
Token ID | 3772 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
from | +0.001 |
Neg logit contributions | |
inations | -0.002 |
rd | -0.002 |
iden | -0.002 |
nexpected | -0.002 |
arest | -0.002 |
Token | and |
Token ID | 290 |
Feature activation | -1.67e-02 |
Pos logit contributions | |
inations | +0.040 |
rd | +0.040 |
iden | +0.040 |
very | +0.039 |
nexpected | +0.039 |
Neg logit contributions | |
Blue | -0.015 |
ety | -0.012 |
Sunny | -0.012 |
isk | -0.012 |
over | -0.011 |
Token | welcomed |
Token ID | 15846 |
Feature activation | +2.93e-03 |
Pos logit contributions | |
inations | +0.033 |
rd | +0.033 |
iden | +0.033 |
very | +0.032 |
nexpected | +0.032 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.023 |
ety | -0.022 |
Sunny | -0.021 |
back | -0.021 |
Token | the |
Token ID | 262 |
Feature activation | +4.36e-02 |
Pos logit contributions | |
Blue | +0.003 |
blown | +0.003 |
ety | +0.003 |
also | +0.003 |
ter | +0.003 |
Neg logit contributions | |
iden | -0.007 |
very | -0.006 |
rd | -0.006 |
nexpected | -0.006 |
inations | -0.006 |
Token | tow |
Token ID | 6094 |
Feature activation | +8.40e-02 |
Pos logit contributions | |
ety | +0.067 |
Blue | +0.063 |
Sunny | +0.059 |
ter | +0.059 |
isk | +0.057 |
Neg logit contributions | |
iest | -0.081 |
iden | -0.080 |
rd | -0.079 |
inations | -0.079 |
arest | -0.079 |
Token | truck |
Token ID | 7779 |
Feature activation | +2.18e-02 |
Pos logit contributions | |
Blue | +0.116 |
ety | +0.113 |
Sunny | +0.112 |
pering | +0.111 |
barking | +0.104 |
Neg logit contributions | |
iest | -0.169 |
oc | -0.163 |
iden | -0.155 |
rd | -0.151 |
ement | -0.145 |
Token | 's |
Token ID | 338 |
Feature activation | +1.47e-03 |
Pos logit contributions | |
Blue | +0.031 |
Sunny | +0.030 |
ety | +0.028 |
Dove | +0.027 |
Woo | +0.026 |
Neg logit contributions | |
rd | -0.042 |
inations | -0.040 |
oc | -0.039 |
iest | -0.038 |
arest | -0.037 |
Token | help |
Token ID | 1037 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
Blue | +0.002 |
Sunny | +0.002 |
ety | +0.001 |
over | +0.001 |
ter | +0.001 |
Neg logit contributions | |
rd | -0.003 |
inations | -0.003 |
oc | -0.003 |
arest | -0.003 |
iest | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | -7.25e-03 |
Pos logit contributions | |
rd | +0.027 |
inations | +0.027 |
iest | +0.026 |
gh | +0.026 |
oc | +0.026 |
Neg logit contributions | |
ety | -0.023 |
Blue | -0.022 |
Sunny | -0.021 |
ter | -0.020 |
blown | -0.019 |
Token |
|
Token ID | 198 |
Feature activation | +9.29e-03 |
Pos logit contributions | |
rd | +0.016 |
inations | +0.016 |
iden | +0.016 |
very | +0.016 |
nexpected | +0.015 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.009 |
back | -0.008 |
from | -0.008 |
ter | -0.008 |
Token | is |
Token ID | 318 |
Feature activation | +1.27e-02 |
Pos logit contributions | |
Blue | +0.059 |
ety | +0.055 |
Sunny | +0.053 |
ter | +0.053 |
isk | +0.052 |
Neg logit contributions | |
inations | -0.060 |
nexpected | -0.059 |
rd | -0.059 |
arest | -0.058 |
iden | -0.058 |
Token | fast |
Token ID | 3049 |
Feature activation | +3.95e-02 |
Pos logit contributions | |
Blue | +0.017 |
isk | +0.017 |
ety | +0.017 |
Sunny | +0.016 |
ter | +0.015 |
Neg logit contributions | |
inations | -0.024 |
iden | -0.023 |
arest | -0.023 |
inkle | -0.022 |
iest | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +4.06e-03 |
Pos logit contributions | |
Blue | +0.056 |
isk | +0.053 |
ety | +0.053 |
Sunny | +0.053 |
Bird | +0.049 |
Neg logit contributions | |
rd | -0.076 |
inations | -0.072 |
iest | -0.072 |
oc | -0.072 |
iden | -0.069 |
Token | The |
Token ID | 383 |
Feature activation | +2.61e-02 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.005 |
ety | +0.005 |
back | +0.005 |
ter | +0.005 |
Neg logit contributions | |
rd | -0.008 |
iden | -0.008 |
very | -0.008 |
inations | -0.008 |
arest | -0.008 |
Token | car |
Token ID | 1097 |
Feature activation | +3.51e-02 |
Pos logit contributions | |
Blue | +0.036 |
over | +0.033 |
Sunny | +0.032 |
ety | +0.030 |
from | +0.029 |
Neg logit contributions | |
iden | -0.055 |
inations | -0.053 |
rd | -0.053 |
iest | -0.053 |
arest | -0.052 |
Token | runs |
Token ID | 4539 |
Feature activation | +8.38e-02 |
Pos logit contributions | |
Blue | +0.058 |
ety | +0.054 |
ter | +0.052 |
Sunny | +0.052 |
isk | +0.051 |
Neg logit contributions | |
inations | -0.062 |
nexpected | -0.060 |
rd | -0.060 |
arest | -0.059 |
iden | -0.058 |
Token | over |
Token ID | 625 |
Feature activation | +2.72e-02 |
Pos logit contributions | |
Sunny | +0.103 |
Blue | +0.102 |
ety | +0.097 |
Bird | +0.096 |
isk | +0.095 |
Neg logit contributions | |
ar | -0.160 |
iden | -0.156 |
rd | -0.155 |
oc | -0.154 |
very | -0.153 |
Token | the |
Token ID | 262 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
Blue | +0.030 |
ety | +0.025 |
isk | +0.025 |
ter | +0.024 |
over | +0.024 |
Neg logit contributions | |
rd | -0.063 |
nexpected | -0.062 |
iden | -0.062 |
very | -0.061 |
inations | -0.061 |
Token | rag |
Token ID | 34232 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
Blue | +0.026 |
Sunny | +0.024 |
isk | +0.023 |
ety | +0.021 |
over | +0.020 |
Neg logit contributions | |
iden | -0.046 |
rd | -0.045 |
oc | -0.044 |
goodness | -0.043 |
very | -0.043 |
Token | . |
Token ID | 13 |
Feature activation | +9.44e-03 |
Pos logit contributions | |
ety | +0.022 |
Blue | +0.022 |
isk | +0.021 |
Sunny | +0.021 |
blown | +0.021 |
Neg logit contributions | |
rd | -0.021 |
oc | -0.020 |
iest | -0.019 |
opter | -0.018 |
robat | -0.018 |
Token | The |
Token ID | 383 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
ety | +0.012 |
Blue | +0.012 |
over | +0.012 |
ter | +0.012 |
back | +0.012 |
Neg logit contributions | |
iden | -0.019 |
rd | -0.019 |
very | -0.018 |
nexpected | -0.018 |
inations | -0.018 |
Token | cat |
Token ID | 3797 |
Feature activation | +1.60e-02 |
Pos logit contributions | |
over | +0.034 |
Blue | +0.031 |
from | +0.030 |
back | +0.028 |
apart | +0.028 |
Neg logit contributions | |
iden | -0.066 |
very | -0.064 |
inations | -0.062 |
rd | -0.062 |
arest | -0.061 |
Token | . |
Token ID | 13 |
Feature activation | +2.47e-02 |
Pos logit contributions | |
Blue | +0.021 |
back | +0.019 |
over | +0.019 |
blown | +0.019 |
Sunny | +0.018 |
Neg logit contributions | |
rd | -0.033 |
oc | -0.032 |
inations | -0.032 |
iden | -0.031 |
very | -0.031 |
Token | Every |
Token ID | 3887 |
Feature activation | +2.88e-02 |
Pos logit contributions | |
ter | +0.034 |
blown | +0.033 |
lying | +0.032 |
faster | +0.032 |
back | +0.031 |
Neg logit contributions | |
very | -0.047 |
iden | -0.047 |
rd | -0.045 |
oc | -0.044 |
nexpected | -0.044 |
Token | day |
Token ID | 1110 |
Feature activation | -1.64e-02 |
Pos logit contributions | |
Blue | +0.060 |
over | +0.056 |
apart | +0.055 |
ter | +0.055 |
ety | +0.054 |
Neg logit contributions | |
rd | -0.041 |
iden | -0.039 |
iest | -0.038 |
very | -0.038 |
nexpected | -0.037 |
Token | , |
Token ID | 11 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
inations | +0.036 |
rd | +0.036 |
iden | +0.035 |
nexpected | +0.035 |
arest | +0.035 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.022 |
Spot | -0.021 |
back | -0.020 |
from | -0.020 |
Token | the |
Token ID | 262 |
Feature activation | +8.38e-02 |
Pos logit contributions | |
isk | +0.014 |
over | +0.013 |
pering | +0.013 |
back | +0.013 |
ety | +0.013 |
Neg logit contributions | |
iden | -0.019 |
very | -0.019 |
nexpected | -0.019 |
inations | -0.018 |
iest | -0.018 |
Token | two |
Token ID | 734 |
Feature activation | +6.31e-02 |
Pos logit contributions | |
ried | +0.140 |
pering | +0.125 |
over | +0.125 |
isk | +0.120 |
apart | +0.120 |
Neg logit contributions | |
iden | -0.131 |
very | -0.130 |
sudden | -0.130 |
ay | -0.128 |
rd | -0.127 |
Token | friends |
Token ID | 2460 |
Feature activation | +2.21e-02 |
Pos logit contributions | |
over | +0.106 |
pering | +0.100 |
asleep | +0.097 |
Blue | +0.096 |
barking | +0.096 |
Neg logit contributions | |
iest | -0.106 |
ises | -0.102 |
ons | -0.095 |
inations | -0.094 |
iden | -0.093 |
Token | would |
Token ID | 561 |
Feature activation | +3.23e-02 |
Pos logit contributions | |
Blue | +0.039 |
over | +0.035 |
back | +0.035 |
Spot | +0.034 |
barking | +0.034 |
Neg logit contributions | |
inations | -0.038 |
rd | -0.038 |
iden | -0.037 |
nexpected | -0.037 |
iest | -0.037 |
Token | run |
Token ID | 1057 |
Feature activation | +6.68e-02 |
Pos logit contributions | |
ter | +0.044 |
Blue | +0.044 |
ety | +0.042 |
ters | +0.042 |
blown | +0.041 |
Neg logit contributions | |
inations | -0.060 |
rd | -0.058 |
iden | -0.056 |
ises | -0.056 |
nexpected | -0.056 |
Token | around |
Token ID | 1088 |
Feature activation | +2.90e-02 |
Pos logit contributions | |
Blue | +0.075 |
ter | +0.070 |
lying | +0.064 |
blown | +0.063 |
Sunny | +0.062 |
Neg logit contributions | |
rd | -0.149 |
iden | -0.147 |
very | -0.147 |
inations | -0.146 |
nexpected | -0.139 |
Token | sorry |
Token ID | 7926 |
Feature activation | -5.79e-03 |
Pos logit contributions | |
inations | +0.042 |
rd | +0.041 |
iden | +0.040 |
nexpected | +0.039 |
mmm | +0.039 |
Neg logit contributions | |
ety | -0.042 |
isk | -0.041 |
Blue | -0.041 |
over | -0.040 |
ried | -0.039 |
Token | , |
Token ID | 11 |
Feature activation | +3.70e-02 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
arest | +0.012 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
from | -0.007 |
Spot | -0.007 |
Token | Anna |
Token ID | 11735 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
Blue | +0.043 |
over | +0.039 |
ety | +0.037 |
back | +0.036 |
isk | +0.035 |
Neg logit contributions | |
inations | -0.084 |
rd | -0.084 |
iden | -0.083 |
very | -0.082 |
nexpected | -0.082 |
Token | , |
Token ID | 11 |
Feature activation | +3.46e-02 |
Pos logit contributions | |
Blue | +0.019 |
over | +0.017 |
back | +0.016 |
from | +0.016 |
Spot | +0.016 |
Neg logit contributions | |
rd | -0.033 |
inations | -0.033 |
iden | -0.033 |
nexpected | -0.032 |
very | -0.032 |
Token | the |
Token ID | 262 |
Feature activation | +5.95e-02 |
Pos logit contributions | |
Blue | +0.053 |
ety | +0.050 |
over | +0.050 |
isk | +0.048 |
blown | +0.048 |
Neg logit contributions | |
inations | -0.063 |
rd | -0.062 |
iden | -0.062 |
nexpected | -0.061 |
iest | -0.061 |
Token | old |
Token ID | 1468 |
Feature activation | +8.37e-02 |
Pos logit contributions | |
over | +0.073 |
Blue | +0.072 |
from | +0.066 |
apart | +0.065 |
blown | +0.064 |
Neg logit contributions | |
rd | -0.129 |
iden | -0.123 |
inations | -0.121 |
iest | -0.120 |
nexpected | -0.119 |
Token | monkey |
Token ID | 21657 |
Feature activation | +2.08e-02 |
Pos logit contributions | |
Blue | +0.102 |
over | +0.097 |
blown | +0.090 |
from | +0.088 |
apart | +0.087 |
Neg logit contributions | |
rd | -0.187 |
iden | -0.180 |
iest | -0.178 |
very | -0.176 |
oc | -0.174 |
Token | was |
Token ID | 373 |
Feature activation | +5.37e-03 |
Pos logit contributions | |
Blue | +0.031 |
over | +0.029 |
back | +0.027 |
from | +0.026 |
Spot | +0.026 |
Neg logit contributions | |
rd | -0.042 |
inations | -0.042 |
nexpected | -0.041 |
iden | -0.041 |
iest | -0.040 |
Token | mean |
Token ID | 1612 |
Feature activation | +1.24e-02 |
Pos logit contributions | |
Blue | +0.006 |
isk | +0.006 |
ety | +0.006 |
over | +0.006 |
from | +0.006 |
Neg logit contributions | |
inations | -0.011 |
rd | -0.011 |
nexpected | -0.011 |
very | -0.010 |
iden | -0.010 |
Token | to |
Token ID | 284 |
Feature activation | -2.56e-02 |
Pos logit contributions | |
Blue | +0.018 |
isk | +0.016 |
over | +0.015 |
Sunny | +0.015 |
blown | +0.015 |
Neg logit contributions | |
rd | -0.026 |
inations | -0.025 |
iest | -0.025 |
iden | -0.024 |
very | -0.024 |
Token | you |
Token ID | 345 |
Feature activation | +2.14e-03 |
Pos logit contributions | |
rd | +0.051 |
inations | +0.051 |
nexpected | +0.051 |
iest | +0.050 |
arest | +0.050 |
Neg logit contributions | |
Blue | -0.034 |
over | -0.033 |
back | -0.031 |
from | -0.030 |
asleep | -0.029 |
Token | feeling |
Token ID | 4203 |
Feature activation | +4.02e-03 |
Pos logit contributions | |
Blue | +0.030 |
isk | +0.027 |
ety | +0.026 |
back | +0.025 |
over | +0.025 |
Neg logit contributions | |
very | -0.049 |
iden | -0.049 |
rd | -0.046 |
inations | -0.046 |
apy | -0.045 |
Token | gr |
Token ID | 1036 |
Feature activation | -4.09e-02 |
Pos logit contributions | |
Blue | +0.005 |
ety | +0.004 |
isk | +0.004 |
over | +0.004 |
Sunny | +0.004 |
Neg logit contributions | |
iden | -0.009 |
very | -0.009 |
inations | -0.009 |
iest | -0.009 |
rd | -0.008 |
Token | umpy |
Token ID | 32152 |
Feature activation | +1.63e-02 |
Pos logit contributions | |
rd | +0.073 |
inations | +0.072 |
very | +0.070 |
iden | +0.070 |
iest | +0.070 |
Neg logit contributions | |
Blue | -0.074 |
over | -0.067 |
back | -0.067 |
Spot | -0.066 |
Sunny | -0.064 |
Token | . |
Token ID | 13 |
Feature activation | +1.20e-02 |
Pos logit contributions | |
Blue | +0.019 |
ety | +0.017 |
Sunny | +0.016 |
blown | +0.016 |
isk | +0.016 |
Neg logit contributions | |
rd | -0.037 |
iden | -0.037 |
very | -0.036 |
iest | -0.036 |
inations | -0.035 |
Token | The |
Token ID | 383 |
Feature activation | +5.34e-02 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.014 |
back | +0.013 |
from | +0.013 |
Spot | +0.012 |
Neg logit contributions | |
rd | -0.027 |
iden | -0.027 |
inations | -0.026 |
very | -0.026 |
nexpected | -0.026 |
Token | other |
Token ID | 584 |
Feature activation | +8.37e-02 |
Pos logit contributions | |
Blue | +0.068 |
over | +0.064 |
back | +0.061 |
Spot | +0.058 |
from | +0.058 |
Neg logit contributions | |
rd | -0.121 |
inations | -0.121 |
nexpected | -0.119 |
arest | -0.118 |
iden | -0.117 |
Token | brother |
Token ID | 3956 |
Feature activation | -6.92e-05 |
Pos logit contributions | |
over | +0.086 |
Blue | +0.082 |
back | +0.079 |
barking | +0.074 |
from | +0.073 |
Neg logit contributions | |
arest | -0.212 |
inations | -0.210 |
nexpected | -0.206 |
rd | -0.206 |
mmm | -0.201 |
Token | smiled |
Token ID | 13541 |
Feature activation | -2.23e-03 |
Pos logit contributions | |
rd | +0.000 |
inations | +0.000 |
iden | +0.000 |
nexpected | +0.000 |
very | +0.000 |
Neg logit contributions | |
Blue | -0.000 |
over | -0.000 |
back | -0.000 |
from | -0.000 |
Spot | -0.000 |
Token | , |
Token ID | 11 |
Feature activation | -3.29e-03 |
Pos logit contributions | |
very | +0.005 |
rd | +0.005 |
iden | +0.005 |
iest | +0.005 |
nexpected | +0.005 |
Neg logit contributions | |
Blue | -0.003 |
isk | -0.002 |
ety | -0.002 |
Sunny | -0.002 |
Wh | -0.002 |
Token | grabbed |
Token ID | 13176 |
Feature activation | +5.53e-03 |
Pos logit contributions | |
iden | +0.007 |
very | +0.007 |
inations | +0.006 |
rd | +0.006 |
nexpected | +0.006 |
Neg logit contributions | |
Blue | -0.004 |
isk | -0.004 |
pering | -0.004 |
ter | -0.004 |
back | -0.004 |
Token | his |
Token ID | 465 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
Blue | +0.005 |
over | +0.005 |
back | +0.004 |
Spot | +0.004 |
from | +0.004 |
Neg logit contributions | |
inations | -0.015 |
rd | -0.015 |
arest | -0.015 |
iden | -0.015 |
nexpected | -0.015 |
Token | to |
Token ID | 284 |
Feature activation | +8.05e-03 |
Pos logit contributions | |
Blue | +0.049 |
Spot | +0.045 |
Sunny | +0.045 |
isk | +0.045 |
Wh | +0.043 |
Neg logit contributions | |
ee | -0.043 |
very | -0.043 |
iest | -0.040 |
rd | -0.040 |
inations | -0.039 |
Token | the |
Token ID | 262 |
Feature activation | +3.45e-02 |
Pos logit contributions | |
over | +0.010 |
ety | +0.010 |
Blue | +0.010 |
pering | +0.010 |
back | +0.009 |
Neg logit contributions | |
inations | -0.015 |
iden | -0.015 |
nexpected | -0.015 |
very | -0.015 |
rd | -0.015 |
Token | farm |
Token ID | 5318 |
Feature activation | +5.44e-03 |
Pos logit contributions | |
Blue | +0.050 |
over | +0.047 |
ety | +0.046 |
Spot | +0.044 |
apart | +0.044 |
Neg logit contributions | |
very | -0.064 |
oc | -0.062 |
iden | -0.062 |
rd | -0.061 |
lifetime | -0.060 |
Token | . |
Token ID | 13 |
Feature activation | +1.80e-02 |
Pos logit contributions | |
Blue | +0.008 |
ety | +0.007 |
blown | +0.007 |
Wh | +0.007 |
Spot | +0.007 |
Neg logit contributions | |
rd | -0.010 |
very | -0.010 |
oc | -0.010 |
inations | -0.009 |
iden | -0.009 |
Token | The |
Token ID | 383 |
Feature activation | +5.59e-02 |
Pos logit contributions | |
back | +0.026 |
over | +0.025 |
blown | +0.025 |
ter | +0.025 |
Blue | +0.025 |
Neg logit contributions | |
iden | -0.035 |
very | -0.034 |
rd | -0.033 |
mmm | -0.032 |
sha | -0.031 |
Token | other |
Token ID | 584 |
Feature activation | +8.37e-02 |
Pos logit contributions | |
over | +0.071 |
Blue | +0.071 |
apart | +0.066 |
asleep | +0.061 |
ety | +0.060 |
Neg logit contributions | |
iden | -0.116 |
oc | -0.115 |
very | -0.113 |
rd | -0.113 |
iest | -0.112 |
Token | animals |
Token ID | 4695 |
Feature activation | +4.85e-03 |
Pos logit contributions | |
Blue | +0.093 |
over | +0.082 |
Sunny | +0.077 |
Spot | +0.074 |
asleep | +0.074 |
Neg logit contributions | |
iest | -0.201 |
rd | -0.201 |
oc | -0.195 |
iden | -0.194 |
nexpected | -0.192 |
Token | were |
Token ID | 547 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.007 |
back | +0.007 |
Spot | +0.007 |
Sunny | +0.007 |
Neg logit contributions | |
inations | -0.009 |
rd | -0.009 |
iest | -0.009 |
nexpected | -0.009 |
iden | -0.009 |
Token | curious |
Token ID | 11040 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
ety | +0.022 |
isk | +0.021 |
Blue | +0.019 |
burn | +0.018 |
undone | +0.018 |
Neg logit contributions | |
inations | -0.021 |
very | -0.020 |
iden | -0.020 |
rd | -0.020 |
inkle | -0.019 |
Token | about |
Token ID | 546 |
Feature activation | +1.11e-02 |
Pos logit contributions | |
rd | +0.032 |
inations | +0.032 |
iden | +0.031 |
very | +0.031 |
nexpected | +0.031 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.006 |
Spot | -0.006 |
ety | -0.005 |
Token | the |
Token ID | 262 |
Feature activation | +3.63e-02 |
Pos logit contributions | |
Blue | +0.012 |
ety | +0.012 |
over | +0.011 |
back | +0.011 |
blown | +0.011 |
Neg logit contributions | |
inations | -0.025 |
iden | -0.025 |
rd | -0.024 |
nexpected | -0.024 |
very | -0.024 |
Token | trick |
Token ID | 6908 |
Feature activation | -5.79e-02 |
Pos logit contributions | |
inations | +0.077 |
rd | +0.075 |
nexpected | +0.075 |
arest | +0.074 |
iden | +0.074 |
Neg logit contributions | |
Blue | -0.067 |
over | -0.063 |
back | -0.062 |
from | -0.060 |
Spot | -0.057 |
Token | . |
Token ID | 13 |
Feature activation | -2.44e-02 |
Pos logit contributions | |
inations | +0.129 |
rd | +0.128 |
nexpected | +0.126 |
iden | +0.126 |
arest | +0.124 |
Neg logit contributions | |
Blue | -0.077 |
over | -0.073 |
back | -0.069 |
from | -0.069 |
Spot | -0.066 |
Token | Everyone |
Token ID | 11075 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
rd | +0.051 |
inations | +0.051 |
iden | +0.050 |
nexpected | +0.050 |
very | +0.049 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.032 |
back | -0.031 |
from | -0.030 |
Spot | -0.030 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.34e-02 |
Pos logit contributions | |
rd | +0.030 |
inations | +0.030 |
mmm | +0.030 |
arest | +0.029 |
nexpected | +0.029 |
Neg logit contributions | |
over | -0.019 |
from | -0.017 |
back | -0.017 |
Blue | -0.017 |
also | -0.015 |
Token | in |
Token ID | 287 |
Feature activation | -7.20e-03 |
Pos logit contributions | |
inations | +0.073 |
rd | +0.073 |
iden | +0.072 |
very | +0.072 |
nexpected | +0.072 |
Neg logit contributions | |
Blue | -0.042 |
over | -0.038 |
ety | -0.036 |
back | -0.036 |
Spot | -0.036 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.24e-01 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
arest | +0.012 |
Neg logit contributions | |
Blue | -0.014 |
over | -0.013 |
back | -0.013 |
from | -0.013 |
Spot | -0.012 |
Token | ement |
Token ID | 972 |
Feature activation | -4.79e-02 |
Pos logit contributions | |
inations | +0.132 |
iden | +0.121 |
oc | +0.114 |
arest | +0.111 |
rd | +0.108 |
Neg logit contributions | |
back | -0.315 |
over | -0.307 |
from | -0.304 |
Blue | -0.301 |
up | -0.289 |
Token | and |
Token ID | 290 |
Feature activation | -1.41e-02 |
Pos logit contributions | |
rd | +0.120 |
inations | +0.120 |
iden | +0.117 |
nexpected | +0.117 |
very | +0.116 |
Neg logit contributions | |
Blue | -0.049 |
over | -0.043 |
back | -0.040 |
Spot | -0.039 |
from | -0.038 |
Token | applauded |
Token ID | 38911 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
rd | +0.030 |
inations | +0.030 |
iden | +0.029 |
arest | +0.029 |
nexpected | +0.029 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.019 |
back | -0.018 |
from | -0.018 |
Spot | -0.018 |
Token | the |
Token ID | 262 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
inations | +0.047 |
rd | +0.047 |
iden | +0.046 |
nexpected | +0.046 |
arest | +0.046 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.022 |
back | -0.021 |
from | -0.020 |
Spot | -0.020 |
Token | young |
Token ID | 1862 |
Feature activation | +5.82e-02 |
Pos logit contributions | |
Blue | +0.026 |
over | +0.024 |
back | +0.024 |
Spot | +0.023 |
from | +0.022 |
Neg logit contributions | |
rd | -0.035 |
inations | -0.035 |
nexpected | -0.034 |
iden | -0.033 |
arest | -0.033 |
Token | her |
Token ID | 607 |
Feature activation | -6.07e-03 |
Pos logit contributions | |
inations | +0.029 |
nexpected | +0.028 |
arest | +0.028 |
very | +0.028 |
rd | +0.028 |
Neg logit contributions | |
Blue | -0.011 |
ety | -0.010 |
over | -0.010 |
barking | -0.009 |
back | -0.009 |
Token | doll |
Token ID | 3654 |
Feature activation | +2.57e-03 |
Pos logit contributions | |
rd | +0.013 |
inations | +0.013 |
iden | +0.012 |
nexpected | +0.012 |
very | +0.012 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.008 |
back | -0.008 |
Spot | -0.008 |
from | -0.008 |
Token | and |
Token ID | 290 |
Feature activation | -1.44e-02 |
Pos logit contributions | |
Blue | +0.003 |
ety | +0.002 |
over | +0.002 |
blown | +0.002 |
Spot | +0.002 |
Neg logit contributions | |
rd | -0.006 |
inations | -0.006 |
iden | -0.006 |
very | -0.006 |
nexpected | -0.006 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.12e-02 |
Pos logit contributions | |
inations | +0.029 |
iden | +0.029 |
rd | +0.029 |
nexpected | +0.029 |
very | +0.028 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.019 |
back | -0.018 |
ety | -0.018 |
barking | -0.018 |
Token | in |
Token ID | 287 |
Feature activation | -9.72e-03 |
Pos logit contributions | |
inations | +0.069 |
rd | +0.068 |
nexpected | +0.068 |
very | +0.067 |
iest | +0.067 |
Neg logit contributions | |
Blue | -0.038 |
ety | -0.035 |
over | -0.033 |
barking | -0.032 |
Spot | -0.032 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.24e-01 |
Pos logit contributions | |
inations | +0.014 |
rd | +0.014 |
iden | +0.014 |
nexpected | +0.014 |
very | +0.014 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.020 |
back | -0.019 |
from | -0.019 |
Spot | -0.018 |
Token | ement |
Token ID | 972 |
Feature activation | -4.58e-02 |
Pos logit contributions | |
inations | +0.130 |
iden | +0.125 |
oc | +0.117 |
rd | +0.106 |
arest | +0.106 |
Neg logit contributions | |
back | -0.317 |
over | -0.307 |
from | -0.302 |
Blue | -0.300 |
up | -0.287 |
Token | as |
Token ID | 355 |
Feature activation | +1.82e-02 |
Pos logit contributions | |
rd | +0.111 |
inations | +0.111 |
iden | +0.108 |
nexpected | +0.108 |
very | +0.107 |
Neg logit contributions | |
Blue | -0.051 |
over | -0.046 |
back | -0.042 |
Spot | -0.041 |
from | -0.041 |
Token | it |
Token ID | 340 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.021 |
back | +0.020 |
from | +0.019 |
Spot | +0.019 |
Neg logit contributions | |
rd | -0.042 |
inations | -0.042 |
iden | -0.041 |
nexpected | -0.041 |
arest | -0.040 |
Token | slowly |
Token ID | 6364 |
Feature activation | +5.24e-03 |
Pos logit contributions | |
rd | +0.043 |
inations | +0.043 |
iden | +0.042 |
nexpected | +0.042 |
very | +0.042 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.023 |
back | -0.021 |
from | -0.021 |
Spot | -0.020 |
Token | flew |
Token ID | 13112 |
Feature activation | +4.23e-02 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.007 |
Spot | +0.007 |
ety | +0.007 |
Sunny | +0.006 |
Neg logit contributions | |
inations | -0.010 |
iest | -0.010 |
rd | -0.010 |
nexpected | -0.010 |
iden | -0.010 |
Token | just |
Token ID | 655 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
rd | +0.045 |
inations | +0.045 |
arest | +0.044 |
iden | +0.043 |
nexpected | +0.043 |
Neg logit contributions | |
Blue | -0.037 |
over | -0.035 |
back | -0.033 |
from | -0.033 |
Spot | -0.033 |
Token | stared |
Token ID | 18484 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
rd | +0.028 |
inations | +0.028 |
iden | +0.027 |
arest | +0.027 |
very | +0.027 |
Neg logit contributions | |
Blue | -0.026 |
over | -0.025 |
back | -0.024 |
from | -0.023 |
Spot | -0.023 |
Token | at |
Token ID | 379 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
rd | +0.059 |
inations | +0.059 |
iden | +0.058 |
very | +0.058 |
nexpected | +0.058 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.030 |
Spot | -0.029 |
Sunny | -0.028 |
back | -0.028 |
Token | him |
Token ID | 683 |
Feature activation | +7.23e-05 |
Pos logit contributions | |
rd | +0.058 |
inations | +0.058 |
nexpected | +0.058 |
iden | +0.057 |
iest | +0.057 |
Neg logit contributions | |
Blue | -0.026 |
over | -0.024 |
back | -0.022 |
from | -0.021 |
barking | -0.021 |
Token | in |
Token ID | 287 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
Blue | +0.000 |
over | +0.000 |
ety | +0.000 |
Spot | +0.000 |
back | +0.000 |
Neg logit contributions | |
rd | -0.000 |
inations | -0.000 |
iden | -0.000 |
very | -0.000 |
nexpected | -0.000 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.22e-01 |
Pos logit contributions | |
rd | +0.029 |
inations | +0.028 |
iden | +0.028 |
nexpected | +0.028 |
very | +0.027 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.033 |
back | -0.032 |
Spot | -0.032 |
from | -0.031 |
Token | ement |
Token ID | 972 |
Feature activation | -4.93e-02 |
Pos logit contributions | |
inations | +0.120 |
iden | +0.109 |
oc | +0.102 |
rd | +0.099 |
arest | +0.099 |
Neg logit contributions | |
back | -0.320 |
over | -0.315 |
from | -0.308 |
Blue | -0.304 |
up | -0.289 |
Token | . |
Token ID | 13 |
Feature activation | -2.08e-02 |
Pos logit contributions | |
inations | +0.120 |
rd | +0.120 |
iden | +0.118 |
nexpected | +0.117 |
arest | +0.116 |
Neg logit contributions | |
Blue | -0.054 |
over | -0.050 |
back | -0.047 |
from | -0.045 |
Spot | -0.044 |
Token | So |
Token ID | 1406 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
iden | +0.047 |
very | +0.047 |
rd | +0.047 |
nexpected | +0.046 |
inations | +0.045 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.023 |
back | -0.022 |
from | -0.021 |
blown | -0.020 |
Token | , |
Token ID | 11 |
Feature activation | -1.01e-03 |
Pos logit contributions | |
inations | +0.036 |
rd | +0.036 |
arest | +0.035 |
iden | +0.035 |
nexpected | +0.034 |
Neg logit contributions | |
Blue | -0.019 |
over | -0.017 |
from | -0.017 |
back | -0.017 |
Spot | -0.016 |
Token | he |
Token ID | 339 |
Feature activation | -2.64e-04 |
Pos logit contributions | |
rd | +0.002 |
inations | +0.002 |
iden | +0.002 |
nexpected | +0.002 |
arest | +0.002 |
Neg logit contributions | |
Blue | -0.001 |
over | -0.001 |
back | -0.001 |
from | -0.001 |
Spot | -0.001 |
Token | the |
Token ID | 262 |
Feature activation | -9.12e-03 |
Pos logit contributions | |
ety | +0.012 |
Blue | +0.011 |
isk | +0.011 |
pering | +0.011 |
ters | +0.010 |
Neg logit contributions | |
nexpected | -0.024 |
very | -0.021 |
iest | -0.021 |
moments | -0.020 |
Jamie | -0.020 |
Token | bush |
Token ID | 24484 |
Feature activation | +6.46e-03 |
Pos logit contributions | |
rd | +0.020 |
inations | +0.020 |
iden | +0.020 |
nexpected | +0.020 |
arest | +0.019 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
Spot | -0.010 |
from | -0.010 |
Sunny | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | -1.08e-02 |
Pos logit contributions | |
Blue | +0.008 |
ety | +0.007 |
ter | +0.007 |
Sunny | +0.007 |
blown | +0.006 |
Neg logit contributions | |
rd | -0.014 |
inations | -0.013 |
oc | -0.013 |
very | -0.013 |
iest | -0.013 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.45e-02 |
Pos logit contributions | |
nexpected | +0.023 |
iden | +0.022 |
inations | +0.022 |
very | +0.022 |
iest | +0.022 |
Neg logit contributions | |
Blue | -0.013 |
ety | -0.012 |
barking | -0.012 |
pering | -0.012 |
over | -0.012 |
Token | in |
Token ID | 287 |
Feature activation | -6.01e-03 |
Pos logit contributions | |
nexpected | +0.066 |
inations | +0.065 |
rd | +0.064 |
very | +0.063 |
arest | +0.062 |
Neg logit contributions | |
ety | -0.050 |
Blue | -0.048 |
isk | -0.045 |
barking | -0.045 |
over | -0.044 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.21e-01 |
Pos logit contributions | |
rd | +0.013 |
inations | +0.013 |
nexpected | +0.013 |
iden | +0.012 |
iest | +0.012 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.007 |
ety | -0.007 |
Spot | -0.007 |
Token | ement |
Token ID | 972 |
Feature activation | -4.52e-02 |
Pos logit contributions | |
inations | +0.111 |
iden | +0.110 |
oc | +0.093 |
arest | +0.091 |
rd | +0.091 |
Neg logit contributions | |
back | -0.320 |
over | -0.313 |
Blue | -0.312 |
from | -0.307 |
Spot | -0.289 |
Token | as |
Token ID | 355 |
Feature activation | +2.05e-02 |
Pos logit contributions | |
rd | +0.104 |
inations | +0.103 |
nexpected | +0.101 |
very | +0.100 |
iest | +0.100 |
Neg logit contributions | |
Blue | -0.055 |
over | -0.048 |
back | -0.045 |
Spot | -0.045 |
ety | -0.044 |
Token | it |
Token ID | 340 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
Blue | +0.022 |
over | +0.021 |
ety | +0.020 |
back | +0.020 |
barking | +0.019 |
Neg logit contributions | |
inations | -0.048 |
nexpected | -0.047 |
iest | -0.047 |
rd | -0.046 |
iden | -0.046 |
Token | flew |
Token ID | 13112 |
Feature activation | +4.24e-02 |
Pos logit contributions | |
inations | +0.037 |
nexpected | +0.036 |
rd | +0.036 |
iest | +0.036 |
iden | +0.035 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.021 |
Spot | -0.020 |
back | -0.020 |
barking | -0.020 |
Token | away |
Token ID | 1497 |
Feature activation | +1.45e-02 |
Pos logit contributions | |
Blue | +0.055 |
isk | +0.050 |
barking | +0.050 |
ety | +0.049 |
Spot | +0.049 |
Neg logit contributions | |
iest | -0.083 |
very | -0.083 |
rd | -0.082 |
nexpected | -0.082 |
iden | -0.080 |
Token | gazed |
Token ID | 50255 |
Feature activation | -2.00e-02 |
Pos logit contributions | |
rd | +0.040 |
inations | +0.040 |
nexpected | +0.039 |
iden | +0.039 |
very | +0.039 |
Neg logit contributions | |
Blue | -0.030 |
over | -0.027 |
back | -0.026 |
Spot | -0.025 |
from | -0.025 |
Token | at |
Token ID | 379 |
Feature activation | -3.06e-02 |
Pos logit contributions | |
rd | +0.049 |
inations | +0.049 |
iden | +0.048 |
nexpected | +0.048 |
very | +0.048 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.019 |
back | -0.018 |
Spot | -0.017 |
from | -0.017 |
Token | Ad |
Token ID | 1215 |
Feature activation | -3.03e-02 |
Pos logit contributions | |
rd | +0.075 |
nexpected | +0.075 |
inations | +0.075 |
iden | +0.075 |
iest | +0.075 |
Neg logit contributions | |
Blue | -0.029 |
over | -0.026 |
barking | -0.024 |
back | -0.024 |
ety | -0.023 |
Token | eline |
Token ID | 4470 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
inations | +0.054 |
rd | +0.054 |
nexpected | +0.053 |
iden | +0.053 |
arest | +0.052 |
Neg logit contributions | |
Blue | -0.053 |
over | -0.049 |
back | -0.048 |
from | -0.047 |
Spot | -0.046 |
Token | in |
Token ID | 287 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
rd | +0.026 |
inations | +0.026 |
very | +0.025 |
nexpected | +0.025 |
iden | +0.024 |
Neg logit contributions | |
Blue | -0.015 |
Spot | -0.012 |
over | -0.012 |
ety | -0.012 |
Sunny | -0.012 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.20e-01 |
Pos logit contributions | |
inations | +0.031 |
rd | +0.031 |
iden | +0.030 |
arest | +0.030 |
nexpected | +0.030 |
Neg logit contributions | |
Blue | -0.034 |
over | -0.032 |
back | -0.031 |
from | -0.031 |
Spot | -0.031 |
Token | ement |
Token ID | 972 |
Feature activation | -5.41e-02 |
Pos logit contributions | |
inations | +0.117 |
iden | +0.109 |
oc | +0.101 |
rd | +0.096 |
arest | +0.096 |
Neg logit contributions | |
back | -0.314 |
over | -0.309 |
from | -0.303 |
Blue | -0.294 |
up | -0.283 |
Token | . |
Token ID | 13 |
Feature activation | -1.11e-02 |
Pos logit contributions | |
inations | +0.133 |
rd | +0.132 |
iden | +0.131 |
nexpected | +0.129 |
arest | +0.129 |
Neg logit contributions | |
Blue | -0.058 |
over | -0.054 |
back | -0.050 |
from | -0.049 |
Spot | -0.047 |
Token | As |
Token ID | 1081 |
Feature activation | +2.97e-02 |
Pos logit contributions | |
iden | +0.023 |
very | +0.023 |
rd | +0.022 |
nexpected | +0.022 |
arest | +0.021 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.013 |
Sunny | -0.013 |
Spot | -0.013 |
ety | -0.013 |
Token | she |
Token ID | 673 |
Feature activation | -1.97e-02 |
Pos logit contributions | |
Blue | +0.037 |
over | +0.033 |
back | +0.031 |
Spot | +0.030 |
from | +0.030 |
Neg logit contributions | |
rd | -0.068 |
inations | -0.068 |
iden | -0.067 |
nexpected | -0.067 |
very | -0.066 |
Token | sang |
Token ID | 25889 |
Feature activation | -1.95e-02 |
Pos logit contributions | |
rd | +0.041 |
inations | +0.041 |
iden | +0.041 |
iest | +0.040 |
arest | +0.040 |
Neg logit contributions | |
Blue | -0.028 |
over | -0.027 |
back | -0.026 |
from | -0.025 |
Spot | -0.024 |
Token | but |
Token ID | 475 |
Feature activation | -8.25e-03 |
Pos logit contributions | |
rd | +0.022 |
inations | +0.022 |
iden | +0.022 |
nexpected | +0.022 |
iest | +0.021 |
Neg logit contributions | |
Blue | -0.013 |
over | -0.011 |
back | -0.011 |
ety | -0.010 |
Spot | -0.010 |
Token | they |
Token ID | 484 |
Feature activation | +9.54e-03 |
Pos logit contributions | |
iden | +0.018 |
rd | +0.018 |
nexpected | +0.018 |
inations | +0.018 |
very | +0.017 |
Neg logit contributions | |
Blue | -0.010 |
ety | -0.009 |
isk | -0.009 |
over | -0.008 |
Sunny | -0.008 |
Token | ran |
Token ID | 4966 |
Feature activation | +5.61e-02 |
Pos logit contributions | |
Blue | +0.017 |
over | +0.015 |
back | +0.015 |
Spot | +0.015 |
from | +0.014 |
Neg logit contributions | |
inations | -0.017 |
rd | -0.017 |
iden | -0.017 |
nexpected | -0.017 |
arest | -0.016 |
Token | back |
Token ID | 736 |
Feature activation | +4.40e-02 |
Pos logit contributions | |
Blue | +0.072 |
Sunny | +0.069 |
ety | +0.067 |
isk | +0.065 |
Bird | +0.063 |
Neg logit contributions | |
rd | -0.108 |
iden | -0.107 |
nexpected | -0.103 |
inations | -0.103 |
very | -0.103 |
Token | in |
Token ID | 287 |
Feature activation | +6.64e-04 |
Pos logit contributions | |
Blue | +0.052 |
Sunny | +0.042 |
ter | +0.042 |
isk | +0.041 |
ety | +0.040 |
Neg logit contributions | |
rd | -0.098 |
iden | -0.096 |
nexpected | -0.095 |
inations | -0.095 |
very | -0.094 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.20e-01 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
from | +0.001 |
Spot | +0.001 |
Neg logit contributions | |
rd | -0.002 |
inations | -0.002 |
iden | -0.002 |
nexpected | -0.002 |
very | -0.002 |
Token | ement |
Token ID | 972 |
Feature activation | -4.21e-02 |
Pos logit contributions | |
inations | +0.111 |
iden | +0.106 |
oc | +0.100 |
arest | +0.094 |
rd | +0.091 |
Neg logit contributions | |
back | -0.313 |
over | -0.310 |
Blue | -0.302 |
from | -0.301 |
up | -0.286 |
Token | ! |
Token ID | 0 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
rd | +0.096 |
inations | +0.095 |
nexpected | +0.093 |
iden | +0.092 |
very | +0.092 |
Neg logit contributions | |
Blue | -0.053 |
over | -0.047 |
Spot | -0.044 |
back | -0.044 |
Sunny | -0.042 |
Token | The |
Token ID | 383 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
rd | +0.037 |
inations | +0.037 |
iden | +0.037 |
nexpected | +0.037 |
very | +0.036 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.014 |
back | -0.013 |
from | -0.013 |
Spot | -0.013 |
Token | post |
Token ID | 1281 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
Blue | +0.051 |
over | +0.049 |
Spot | +0.046 |
Sunny | +0.045 |
barking | +0.045 |
Neg logit contributions | |
arest | -0.057 |
inations | -0.056 |
nexpected | -0.056 |
iest | -0.056 |
rd | -0.056 |
Token | had |
Token ID | 550 |
Feature activation | +3.99e-03 |
Pos logit contributions | |
rd | +0.026 |
inations | +0.026 |
iden | +0.026 |
nexpected | +0.025 |
very | +0.025 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.020 |
back | -0.019 |
Spot | -0.019 |
from | -0.018 |
Token | towards |
Token ID | 3371 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
Blue | +0.065 |
ety | +0.058 |
isk | +0.058 |
Sunny | +0.053 |
burn | +0.052 |
Neg logit contributions | |
iden | -0.139 |
very | -0.138 |
rd | -0.136 |
inations | -0.132 |
nexpected | -0.131 |
Token | it |
Token ID | 340 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
rd | +0.032 |
nexpected | +0.031 |
inations | +0.031 |
iden | +0.031 |
iest | +0.031 |
Neg logit contributions | |
Blue | -0.013 |
ety | -0.011 |
over | -0.010 |
from | -0.010 |
Spot | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
Blue | +0.026 |
ety | +0.024 |
isk | +0.023 |
Wh | +0.023 |
burn | +0.022 |
Neg logit contributions | |
rd | -0.039 |
inations | -0.039 |
very | -0.038 |
iest | -0.037 |
nexpected | -0.036 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.09e-02 |
Pos logit contributions | |
rd | +0.052 |
inations | +0.052 |
iden | +0.051 |
nexpected | +0.051 |
very | +0.050 |
Neg logit contributions | |
Blue | -0.043 |
over | -0.040 |
back | -0.038 |
from | -0.037 |
Spot | -0.037 |
Token | in |
Token ID | 287 |
Feature activation | -7.84e-03 |
Pos logit contributions | |
inations | +0.070 |
rd | +0.070 |
nexpected | +0.070 |
iest | +0.069 |
very | +0.068 |
Neg logit contributions | |
Blue | -0.035 |
ety | -0.033 |
over | -0.030 |
isk | -0.029 |
Spot | -0.029 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
inations | +0.010 |
rd | +0.009 |
iden | +0.009 |
nexpected | +0.009 |
very | +0.009 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.018 |
back | -0.017 |
from | -0.017 |
Spot | -0.017 |
Token | ement |
Token ID | 972 |
Feature activation | -4.33e-02 |
Pos logit contributions | |
inations | +0.122 |
iden | +0.118 |
oc | +0.109 |
rd | +0.103 |
arest | +0.101 |
Neg logit contributions | |
back | -0.306 |
over | -0.300 |
from | -0.297 |
Blue | -0.293 |
up | -0.280 |
Token | as |
Token ID | 355 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
rd | +0.107 |
inations | +0.107 |
iden | +0.105 |
nexpected | +0.104 |
very | +0.103 |
Neg logit contributions | |
Blue | -0.046 |
over | -0.041 |
back | -0.038 |
Spot | -0.037 |
from | -0.036 |
Token | the |
Token ID | 262 |
Feature activation | +2.58e-02 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.014 |
back | +0.013 |
Spot | +0.012 |
from | +0.012 |
Neg logit contributions | |
inations | -0.032 |
rd | -0.032 |
iden | -0.032 |
nexpected | -0.032 |
iest | -0.031 |
Token | little |
Token ID | 1310 |
Feature activation | +3.75e-02 |
Pos logit contributions | |
Blue | +0.030 |
over | +0.027 |
from | +0.025 |
Spot | +0.025 |
back | +0.025 |
Neg logit contributions | |
inations | -0.060 |
rd | -0.060 |
iden | -0.060 |
iest | -0.059 |
very | -0.058 |
Token | frogs |
Token ID | 37475 |
Feature activation | +2.21e-02 |
Pos logit contributions | |
Blue | +0.053 |
over | +0.049 |
isk | +0.049 |
from | +0.045 |
barking | +0.045 |
Neg logit contributions | |
iden | -0.075 |
iest | -0.074 |
rd | -0.073 |
very | -0.073 |
inations | -0.072 |
Token | but |
Token ID | 475 |
Feature activation | -2.18e-02 |
Pos logit contributions | |
inations | +0.019 |
rd | +0.017 |
iden | +0.017 |
iest | +0.016 |
very | +0.016 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.009 |
blown | -0.008 |
ter | -0.008 |
back | -0.008 |
Token | they |
Token ID | 484 |
Feature activation | -3.64e-03 |
Pos logit contributions | |
inations | +0.046 |
rd | +0.045 |
iest | +0.045 |
nexpected | +0.045 |
iden | +0.045 |
Neg logit contributions | |
Blue | -0.026 |
isk | -0.025 |
over | -0.024 |
ety | -0.023 |
back | -0.023 |
Token | chose |
Token ID | 7690 |
Feature activation | +3.33e-03 |
Pos logit contributions | |
inations | +0.009 |
rd | +0.009 |
iden | +0.009 |
nexpected | +0.009 |
arest | +0.009 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.003 |
back | -0.003 |
Spot | -0.003 |
from | -0.003 |
Token | safer |
Token ID | 14178 |
Feature activation | +2.79e-02 |
Pos logit contributions | |
Blue | +0.003 |
pering | +0.003 |
barking | +0.003 |
isk | +0.003 |
ter | +0.003 |
Neg logit contributions | |
iest | -0.007 |
rd | -0.007 |
iden | -0.007 |
nexpected | -0.007 |
inations | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
Blue | +0.032 |
over | +0.029 |
back | +0.027 |
from | +0.026 |
Spot | +0.026 |
Neg logit contributions | |
inations | -0.067 |
rd | -0.067 |
iden | -0.066 |
nexpected | -0.066 |
arest | -0.065 |
Token | funn |
Token ID | 36090 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
inations | +0.029 |
rd | +0.029 |
iden | +0.028 |
nexpected | +0.028 |
very | +0.028 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.014 |
Spot | -0.013 |
back | -0.013 |
Sunny | -0.013 |
Token | ier |
Token ID | 959 |
Feature activation | +1.79e-02 |
Pos logit contributions | |
iden | +0.237 |
rd | +0.232 |
very | +0.225 |
inations | +0.225 |
nexpected | +0.216 |
Neg logit contributions | |
over | -0.183 |
from | -0.178 |
Blue | -0.176 |
back | -0.170 |
ter | -0.165 |
Token | ones |
Token ID | 3392 |
Feature activation | -1.81e-02 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.022 |
back | +0.020 |
from | +0.020 |
Spot | +0.020 |
Neg logit contributions | |
inations | -0.040 |
rd | -0.040 |
iden | -0.039 |
nexpected | -0.039 |
arest | -0.039 |
Token | . |
Token ID | 13 |
Feature activation | -5.24e-03 |
Pos logit contributions | |
iest | +0.039 |
inations | +0.039 |
rd | +0.039 |
arest | +0.038 |
iden | +0.037 |
Neg logit contributions | |
Blue | -0.024 |
barking | -0.021 |
over | -0.020 |
ety | -0.020 |
Spot | -0.020 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.97e-02 |
Pos logit contributions | |
rd | +0.012 |
iden | +0.011 |
inations | +0.011 |
very | +0.011 |
nexpected | +0.011 |
Neg logit contributions | |
over | -0.006 |
Blue | -0.006 |
back | -0.006 |
ter | -0.005 |
from | -0.005 |
Token | Tom |
Token ID | 13787 |
Feature activation | -2.37e-02 |
Pos logit contributions | |
rd | +0.051 |
inations | +0.051 |
iden | +0.050 |
nexpected | +0.049 |
very | +0.049 |
Neg logit contributions | |
Blue | -0.053 |
over | -0.050 |
back | -0.048 |
from | -0.047 |
Spot | -0.047 |
Token | ice |
Token ID | 4771 |
Feature activation | +1.24e-02 |
Pos logit contributions | |
Blue | +0.026 |
over | +0.023 |
Sunny | +0.021 |
from | +0.020 |
Spot | +0.020 |
Neg logit contributions | |
rd | -0.049 |
iden | -0.049 |
inations | -0.048 |
iest | -0.048 |
arest | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
Blue | +0.021 |
over | +0.019 |
Spot | +0.018 |
Sunny | +0.018 |
back | +0.018 |
Neg logit contributions | |
rd | -0.023 |
iest | -0.023 |
inations | -0.023 |
iden | -0.022 |
very | -0.022 |
Token | They |
Token ID | 1119 |
Feature activation | +1.36e-02 |
Pos logit contributions | |
rd | +0.029 |
iden | +0.029 |
very | +0.029 |
inations | +0.029 |
nexpected | +0.029 |
Neg logit contributions | |
Blue | -0.014 |
over | -0.013 |
back | -0.012 |
ety | -0.012 |
from | -0.012 |
Token | stared |
Token ID | 18484 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.021 |
back | +0.020 |
Spot | +0.020 |
from | +0.020 |
Neg logit contributions | |
inations | -0.025 |
iden | -0.025 |
rd | -0.025 |
nexpected | -0.024 |
very | -0.024 |
Token | in |
Token ID | 287 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
rd | +0.048 |
inations | +0.047 |
iden | +0.047 |
very | +0.046 |
iest | +0.046 |
Neg logit contributions | |
Blue | -0.033 |
over | -0.028 |
Spot | -0.028 |
ety | -0.027 |
Sunny | -0.027 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
inations | +0.031 |
rd | +0.031 |
iden | +0.030 |
arest | +0.030 |
nexpected | +0.030 |
Neg logit contributions | |
Blue | -0.023 |
over | -0.022 |
back | -0.021 |
from | -0.020 |
Spot | -0.020 |
Token | ement |
Token ID | 972 |
Feature activation | -4.64e-02 |
Pos logit contributions | |
inations | +0.132 |
iden | +0.120 |
oc | +0.116 |
arest | +0.114 |
lifetime | +0.108 |
Neg logit contributions | |
back | -0.300 |
over | -0.294 |
from | -0.286 |
Blue | -0.276 |
and | -0.273 |
Token | . |
Token ID | 13 |
Feature activation | -1.95e-02 |
Pos logit contributions | |
rd | +0.118 |
inations | +0.118 |
iden | +0.116 |
nexpected | +0.115 |
arest | +0.114 |
Neg logit contributions | |
Blue | -0.046 |
over | -0.042 |
back | -0.039 |
from | -0.038 |
Spot | -0.037 |
Token | They |
Token ID | 1119 |
Feature activation | +1.40e-02 |
Pos logit contributions | |
rd | +0.042 |
inations | +0.041 |
iden | +0.041 |
nexpected | +0.041 |
very | +0.040 |
Neg logit contributions | |
Blue | -0.027 |
over | -0.024 |
back | -0.023 |
from | -0.023 |
Spot | -0.022 |
Token | wanted |
Token ID | 2227 |
Feature activation | +1.63e-02 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.021 |
back | +0.020 |
from | +0.020 |
Spot | +0.020 |
Neg logit contributions | |
rd | -0.027 |
inations | -0.027 |
iden | -0.026 |
nexpected | -0.026 |
arest | -0.026 |
Token | to |
Token ID | 284 |
Feature activation | +5.47e-03 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.013 |
back | +0.011 |
from | +0.011 |
Spot | +0.011 |
Neg logit contributions | |
inations | -0.042 |
rd | -0.042 |
iden | -0.042 |
nexpected | -0.041 |
very | -0.041 |
Token | He |
Token ID | 679 |
Feature activation | -6.52e-03 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.007 |
back | +0.007 |
ter | +0.007 |
from | +0.007 |
Neg logit contributions | |
rd | -0.013 |
iden | -0.013 |
very | -0.013 |
inations | -0.013 |
nexpected | -0.012 |
Token | would |
Token ID | 561 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
very | +0.012 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
back | -0.010 |
from | -0.009 |
Spot | -0.009 |
Token | always |
Token ID | 1464 |
Feature activation | -2.03e-03 |
Pos logit contributions | |
Blue | +0.017 |
over | +0.015 |
back | +0.015 |
ety | +0.015 |
Spot | +0.015 |
Neg logit contributions | |
inations | -0.030 |
rd | -0.029 |
iden | -0.029 |
nexpected | -0.029 |
very | -0.028 |
Token | stare |
Token ID | 24170 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
rd | +0.004 |
inations | +0.004 |
iden | +0.004 |
nexpected | +0.004 |
arest | +0.004 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.003 |
back | -0.002 |
from | -0.002 |
Spot | -0.002 |
Token | in |
Token ID | 287 |
Feature activation | -1.16e-02 |
Pos logit contributions | |
rd | +0.068 |
inations | +0.068 |
nexpected | +0.067 |
iden | +0.067 |
arest | +0.066 |
Neg logit contributions | |
Blue | -0.028 |
over | -0.027 |
back | -0.026 |
from | -0.025 |
Spot | -0.023 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
inations | +0.032 |
rd | +0.032 |
arest | +0.031 |
iden | +0.031 |
nexpected | +0.031 |
Neg logit contributions | |
Blue | -0.010 |
over | -0.009 |
back | -0.008 |
from | -0.008 |
Spot | -0.008 |
Token | ement |
Token ID | 972 |
Feature activation | -4.96e-02 |
Pos logit contributions | |
iden | +0.114 |
inations | +0.113 |
oc | +0.103 |
arest | +0.102 |
rd | +0.100 |
Neg logit contributions | |
back | -0.311 |
over | -0.302 |
from | -0.296 |
Blue | -0.291 |
up | -0.282 |
Token | at |
Token ID | 379 |
Feature activation | -3.28e-02 |
Pos logit contributions | |
rd | +0.128 |
inations | +0.127 |
iden | +0.127 |
nexpected | +0.125 |
arest | +0.125 |
Neg logit contributions | |
Blue | -0.047 |
over | -0.045 |
from | -0.042 |
back | -0.042 |
Spot | -0.037 |
Token | the |
Token ID | 262 |
Feature activation | -1.41e-02 |
Pos logit contributions | |
rd | +0.087 |
inations | +0.087 |
iden | +0.086 |
nexpected | +0.085 |
arest | +0.084 |
Neg logit contributions | |
Blue | -0.029 |
over | -0.026 |
back | -0.023 |
from | -0.022 |
Spot | -0.022 |
Token | l |
Token ID | 300 |
Feature activation | -2.25e-02 |
Pos logit contributions | |
inations | +0.031 |
rd | +0.030 |
nexpected | +0.030 |
iden | +0.030 |
arest | +0.029 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.018 |
back | -0.018 |
from | -0.017 |
Spot | -0.017 |
Token | ily |
Token ID | 813 |
Feature activation | +2.71e-03 |
Pos logit contributions | |
iden | +0.035 |
nexpected | +0.034 |
inations | +0.034 |
rd | +0.034 |
very | +0.033 |
Neg logit contributions | |
Blue | -0.041 |
over | -0.041 |
back | -0.040 |
from | -0.039 |
near | -0.038 |
Token | Lily |
Token ID | 20037 |
Feature activation | +1.03e-03 |
Pos logit contributions | |
rd | +0.025 |
inations | +0.025 |
iden | +0.025 |
nexpected | +0.025 |
very | +0.025 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.008 |
back | -0.008 |
from | -0.007 |
Spot | -0.007 |
Token | stopped |
Token ID | 5025 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.002 |
from | +0.002 |
back | +0.002 |
Spot | +0.002 |
Neg logit contributions | |
rd | -0.002 |
iden | -0.002 |
inations | -0.002 |
nexpected | -0.002 |
very | -0.002 |
Token | and |
Token ID | 290 |
Feature activation | -1.62e-02 |
Pos logit contributions | |
Blue | +0.019 |
Sunny | +0.016 |
Spot | +0.014 |
isk | +0.014 |
Red | +0.014 |
Neg logit contributions | |
nexpected | -0.031 |
very | -0.030 |
iden | -0.030 |
iley | -0.030 |
ocol | -0.030 |
Token | stared |
Token ID | 18484 |
Feature activation | -2.54e-02 |
Pos logit contributions | |
inations | +0.032 |
rd | +0.032 |
iest | +0.031 |
arest | +0.031 |
iden | +0.031 |
Neg logit contributions | |
Blue | -0.026 |
over | -0.024 |
back | -0.023 |
Spot | -0.023 |
from | -0.023 |
Token | in |
Token ID | 287 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
rd | +0.056 |
inations | +0.055 |
iden | +0.055 |
nexpected | +0.055 |
very | +0.054 |
Neg logit contributions | |
Blue | -0.034 |
over | -0.030 |
Spot | -0.029 |
back | -0.028 |
barking | -0.027 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
rd | +0.036 |
inations | +0.036 |
iden | +0.035 |
nexpected | +0.035 |
very | +0.035 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.022 |
back | -0.021 |
from | -0.020 |
Spot | -0.020 |
Token | ement |
Token ID | 972 |
Feature activation | -4.67e-02 |
Pos logit contributions | |
inations | +0.122 |
iden | +0.118 |
oc | +0.107 |
rd | +0.103 |
arest | +0.101 |
Neg logit contributions | |
back | -0.304 |
over | -0.295 |
from | -0.294 |
Blue | -0.293 |
too | -0.274 |
Token | . |
Token ID | 13 |
Feature activation | +4.72e-03 |
Pos logit contributions | |
inations | +0.115 |
rd | +0.115 |
iden | +0.114 |
nexpected | +0.112 |
arest | +0.111 |
Neg logit contributions | |
Blue | -0.050 |
over | -0.046 |
back | -0.043 |
from | -0.042 |
Spot | -0.040 |
Token | The |
Token ID | 383 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
Blue | +0.005 |
blown | +0.005 |
ter | +0.005 |
from | +0.005 |
ety | +0.005 |
Neg logit contributions | |
very | -0.010 |
iden | -0.010 |
rd | -0.010 |
nexpected | -0.010 |
mmm | -0.010 |
Token | bush |
Token ID | 24484 |
Feature activation | +9.92e-04 |
Pos logit contributions | |
Blue | +0.020 |
over | +0.017 |
from | +0.016 |
back | +0.016 |
Spot | +0.016 |
Neg logit contributions | |
rd | -0.037 |
iden | -0.036 |
nexpected | -0.036 |
iest | -0.035 |
very | -0.035 |
Token | was |
Token ID | 373 |
Feature activation | +8.51e-03 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.002 |
back | +0.001 |
from | +0.001 |
Spot | +0.001 |
Neg logit contributions | |
rd | -0.002 |
inations | -0.002 |
iden | -0.002 |
nexpected | -0.002 |
very | -0.002 |
Token | around |
Token ID | 1088 |
Feature activation | -4.57e-03 |
Pos logit contributions | |
over | +0.002 |
Blue | +0.002 |
back | +0.002 |
from | +0.002 |
Spot | +0.002 |
Neg logit contributions | |
rd | -0.003 |
inations | -0.003 |
nexpected | -0.003 |
arest | -0.003 |
iden | -0.003 |
Token | stopped |
Token ID | 5025 |
Feature activation | +2.08e-02 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
arest | +0.012 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.003 |
back | -0.003 |
from | -0.003 |
Spot | -0.003 |
Token | and |
Token ID | 290 |
Feature activation | -2.31e-02 |
Pos logit contributions | |
Blue | +0.024 |
Sunny | +0.020 |
Spot | +0.018 |
ety | +0.018 |
over | +0.018 |
Neg logit contributions | |
very | -0.048 |
inations | -0.048 |
iden | -0.048 |
rd | -0.047 |
nexpected | -0.047 |
Token | watched |
Token ID | 7342 |
Feature activation | -4.09e-02 |
Pos logit contributions | |
rd | +0.047 |
inations | +0.047 |
iden | +0.046 |
nexpected | +0.046 |
very | +0.045 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.032 |
back | -0.031 |
from | -0.030 |
Spot | -0.030 |
Token | in |
Token ID | 287 |
Feature activation | -5.27e-03 |
Pos logit contributions | |
rd | +0.102 |
inations | +0.102 |
iden | +0.100 |
nexpected | +0.100 |
very | +0.099 |
Neg logit contributions | |
Blue | -0.042 |
over | -0.037 |
back | -0.034 |
Spot | -0.033 |
from | -0.032 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
inations | +0.007 |
rd | +0.007 |
iden | +0.007 |
nexpected | +0.007 |
arest | +0.007 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.011 |
back | -0.011 |
from | -0.010 |
Spot | -0.010 |
Token | ement |
Token ID | 972 |
Feature activation | -4.57e-02 |
Pos logit contributions | |
inations | +0.129 |
iden | +0.122 |
oc | +0.112 |
arest | +0.105 |
rd | +0.104 |
Neg logit contributions | |
back | -0.303 |
over | -0.293 |
from | -0.289 |
Blue | -0.286 |
up | -0.273 |
Token | . |
Token ID | 13 |
Feature activation | -2.95e-02 |
Pos logit contributions | |
rd | +0.110 |
inations | +0.110 |
iden | +0.108 |
nexpected | +0.107 |
very | +0.106 |
Neg logit contributions | |
Blue | -0.051 |
over | -0.046 |
back | -0.043 |
Spot | -0.042 |
from | -0.042 |
Token | They |
Token ID | 1119 |
Feature activation | +6.35e-03 |
Pos logit contributions | |
rd | +0.058 |
iden | +0.057 |
very | +0.056 |
inations | +0.056 |
nexpected | +0.055 |
Neg logit contributions | |
Blue | -0.045 |
over | -0.042 |
back | -0.041 |
from | -0.040 |
barking | -0.039 |
Token | were |
Token ID | 547 |
Feature activation | -2.98e-03 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.008 |
back | +0.007 |
from | +0.007 |
Spot | +0.007 |
Neg logit contributions | |
rd | -0.014 |
inations | -0.014 |
nexpected | -0.014 |
iden | -0.014 |
arest | -0.014 |
Token | so |
Token ID | 523 |
Feature activation | -4.38e-02 |
Pos logit contributions | |
inations | +0.007 |
rd | +0.007 |
iden | +0.007 |
very | +0.007 |
nexpected | +0.007 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.003 |
ety | -0.003 |
Spot | -0.002 |
Sunny | -0.002 |
Token | . |
Token ID | 13 |
Feature activation | -1.48e-02 |
Pos logit contributions | |
rd | +0.019 |
inations | +0.018 |
iest | +0.018 |
arest | +0.018 |
nexpected | +0.018 |
Neg logit contributions | |
Blue | -0.010 |
Spot | -0.009 |
barking | -0.009 |
Sunny | -0.009 |
ety | -0.009 |
Token | Tim |
Token ID | 5045 |
Feature activation | -1.58e-02 |
Pos logit contributions | |
rd | +0.033 |
iden | +0.032 |
nexpected | +0.032 |
inations | +0.032 |
very | +0.031 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.017 |
back | -0.017 |
from | -0.016 |
blown | -0.016 |
Token | my |
Token ID | 1820 |
Feature activation | -4.01e-02 |
Pos logit contributions | |
inations | +0.044 |
rd | +0.044 |
iden | +0.043 |
nexpected | +0.043 |
arest | +0.043 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
back | -0.010 |
from | -0.009 |
Spot | -0.009 |
Token | watched |
Token ID | 7342 |
Feature activation | -1.99e-02 |
Pos logit contributions | |
arest | +0.086 |
inations | +0.086 |
rd | +0.085 |
iest | +0.084 |
iden | +0.083 |
Neg logit contributions | |
Blue | -0.056 |
over | -0.055 |
back | -0.052 |
from | -0.050 |
Spot | -0.049 |
Token | in |
Token ID | 287 |
Feature activation | +9.25e-04 |
Pos logit contributions | |
inations | +0.046 |
nexpected | +0.046 |
very | +0.045 |
iest | +0.045 |
rd | +0.045 |
Neg logit contributions | |
Blue | -0.021 |
ety | -0.021 |
Sunny | -0.018 |
barking | -0.017 |
isk | -0.017 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.19e-01 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
from | +0.001 |
Neg logit contributions | |
rd | -0.002 |
inations | -0.002 |
iden | -0.002 |
nexpected | -0.002 |
very | -0.002 |
Token | ement |
Token ID | 972 |
Feature activation | -3.95e-02 |
Pos logit contributions | |
inations | +0.118 |
iden | +0.109 |
arest | +0.100 |
rd | +0.100 |
oc | +0.097 |
Neg logit contributions | |
back | -0.312 |
over | -0.306 |
Blue | -0.303 |
from | -0.302 |
Spot | -0.285 |
Token | as |
Token ID | 355 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
inations | +0.102 |
rd | +0.102 |
very | +0.100 |
nexpected | +0.100 |
iden | +0.099 |
Neg logit contributions | |
Blue | -0.036 |
over | -0.030 |
Spot | -0.027 |
back | -0.027 |
Sunny | -0.026 |
Token | the |
Token ID | 262 |
Feature activation | +2.56e-02 |
Pos logit contributions | |
Blue | +0.030 |
over | +0.029 |
back | +0.028 |
ety | +0.026 |
barking | +0.026 |
Neg logit contributions | |
inations | -0.047 |
iden | -0.046 |
nexpected | -0.046 |
rd | -0.046 |
iest | -0.046 |
Token | black |
Token ID | 2042 |
Feature activation | +8.05e-03 |
Pos logit contributions | |
back | +0.025 |
over | +0.023 |
barking | +0.023 |
Blue | +0.022 |
lying | +0.021 |
Neg logit contributions | |
inations | -0.068 |
arest | -0.067 |
opter | -0.067 |
mmm | -0.066 |
rd | -0.066 |
Token | board |
Token ID | 3526 |
Feature activation | -1.51e-02 |
Pos logit contributions | |
back | +0.017 |
over | +0.017 |
from | +0.016 |
Blue | +0.016 |
and | +0.016 |
Neg logit contributions | |
inations | -0.012 |
arest | -0.012 |
rd | -0.011 |
mmm | -0.011 |
prise | -0.010 |
Token | go |
Token ID | 467 |
Feature activation | +4.44e-02 |
Pos logit contributions | |
inations | +0.007 |
rd | +0.007 |
nexpected | +0.007 |
iden | +0.007 |
very | +0.007 |
Neg logit contributions | |
Blue | -0.004 |
ety | -0.004 |
over | -0.004 |
back | -0.004 |
barking | -0.003 |
Token | ! |
Token ID | 0 |
Feature activation | -2.36e-03 |
Pos logit contributions | |
Blue | +0.048 |
isk | +0.039 |
ety | +0.039 |
ter | +0.038 |
Sunny | +0.037 |
Neg logit contributions | |
inations | -0.105 |
rd | -0.103 |
iest | -0.102 |
iden | -0.102 |
very | -0.101 |
Token | Everyone |
Token ID | 11075 |
Feature activation | +2.46e-03 |
Pos logit contributions | |
rd | +0.005 |
iden | +0.005 |
inations | +0.005 |
very | +0.005 |
nexpected | +0.005 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.003 |
back | -0.003 |
from | -0.003 |
barking | -0.003 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.40e-02 |
Pos logit contributions | |
Blue | +0.003 |
over | +0.003 |
back | +0.003 |
from | +0.003 |
Spot | +0.003 |
Neg logit contributions | |
inations | -0.005 |
rd | -0.005 |
arest | -0.005 |
nexpected | -0.005 |
iden | -0.005 |
Token | in |
Token ID | 287 |
Feature activation | +6.26e-03 |
Pos logit contributions | |
inations | +0.084 |
rd | +0.084 |
iden | +0.083 |
nexpected | +0.082 |
very | +0.082 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.030 |
back | -0.028 |
Spot | -0.028 |
ety | -0.027 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
Blue | +0.007 |
over | +0.006 |
back | +0.006 |
from | +0.006 |
Spot | +0.006 |
Neg logit contributions | |
inations | -0.015 |
rd | -0.015 |
iden | -0.015 |
nexpected | -0.015 |
arest | -0.015 |
Token | ement |
Token ID | 972 |
Feature activation | -3.04e-02 |
Pos logit contributions | |
iden | +0.109 |
inations | +0.105 |
oc | +0.101 |
arest | +0.089 |
sha | +0.089 |
Neg logit contributions | |
back | -0.310 |
over | -0.309 |
from | -0.298 |
Blue | -0.293 |
up | -0.285 |
Token | as |
Token ID | 355 |
Feature activation | +1.81e-02 |
Pos logit contributions | |
rd | +0.074 |
inations | +0.074 |
iden | +0.073 |
nexpected | +0.072 |
very | +0.072 |
Neg logit contributions | |
Blue | -0.033 |
over | -0.029 |
back | -0.027 |
Spot | -0.027 |
from | -0.026 |
Token | it |
Token ID | 340 |
Feature activation | -5.23e-03 |
Pos logit contributions | |
Blue | +0.021 |
over | +0.019 |
from | +0.018 |
Spot | +0.017 |
back | +0.017 |
Neg logit contributions | |
rd | -0.045 |
inations | -0.044 |
nexpected | -0.043 |
iden | -0.043 |
arest | -0.043 |
Token | took |
Token ID | 1718 |
Feature activation | -2.41e-03 |
Pos logit contributions | |
rd | +0.010 |
inations | +0.010 |
iden | +0.010 |
nexpected | +0.010 |
very | +0.010 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
from | -0.007 |
Spot | -0.007 |
Token | off |
Token ID | 572 |
Feature activation | +3.02e-02 |
Pos logit contributions | |
inations | +0.007 |
iden | +0.007 |
rd | +0.007 |
very | +0.006 |
arest | +0.006 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
from | -0.002 |
Spot | -0.002 |
Token | calmed |
Token ID | 49566 |
Feature activation | -6.69e-03 |
Pos logit contributions | |
rd | +0.010 |
inations | +0.010 |
iden | +0.010 |
nexpected | +0.010 |
arest | +0.010 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.007 |
from | -0.007 |
Spot | -0.007 |
Token | down |
Token ID | 866 |
Feature activation | +2.73e-02 |
Pos logit contributions | |
iden | +0.018 |
very | +0.018 |
rd | +0.018 |
nexpected | +0.018 |
inations | +0.018 |
Neg logit contributions | |
Blue | -0.005 |
over | -0.004 |
from | -0.003 |
Sunny | -0.003 |
Spot | -0.003 |
Token | and |
Token ID | 290 |
Feature activation | -1.46e-02 |
Pos logit contributions | |
Blue | +0.025 |
ety | +0.020 |
over | +0.020 |
Spot | +0.019 |
Sunny | +0.019 |
Neg logit contributions | |
rd | -0.069 |
iden | -0.069 |
inations | -0.069 |
very | -0.068 |
nexpected | -0.068 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.52e-02 |
Pos logit contributions | |
inations | +0.031 |
rd | +0.031 |
iden | +0.031 |
nexpected | +0.030 |
very | +0.030 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.019 |
back | -0.018 |
Spot | -0.017 |
from | -0.017 |
Token | in |
Token ID | 287 |
Feature activation | -9.29e-04 |
Pos logit contributions | |
inations | +0.088 |
rd | +0.088 |
nexpected | +0.087 |
iden | +0.086 |
very | +0.086 |
Neg logit contributions | |
Blue | -0.034 |
over | -0.029 |
ety | -0.028 |
back | -0.027 |
Spot | -0.026 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
inations | +0.001 |
rd | +0.001 |
iden | +0.001 |
nexpected | +0.001 |
arest | +0.001 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
from | -0.002 |
Spot | -0.002 |
Token | ement |
Token ID | 972 |
Feature activation | -3.99e-02 |
Pos logit contributions | |
inations | +0.126 |
iden | +0.124 |
oc | +0.111 |
rd | +0.106 |
arest | +0.104 |
Neg logit contributions | |
back | -0.299 |
over | -0.293 |
from | -0.287 |
Blue | -0.284 |
up | -0.271 |
Token | as |
Token ID | 355 |
Feature activation | +2.01e-02 |
Pos logit contributions | |
rd | +0.097 |
inations | +0.097 |
iden | +0.095 |
nexpected | +0.095 |
very | +0.094 |
Neg logit contributions | |
Blue | -0.044 |
over | -0.039 |
back | -0.036 |
Spot | -0.035 |
from | -0.035 |
Token | the |
Token ID | 262 |
Feature activation | +1.25e-02 |
Pos logit contributions | |
Blue | +0.018 |
over | +0.016 |
back | +0.015 |
from | +0.014 |
Spot | +0.014 |
Neg logit contributions | |
rd | -0.054 |
inations | -0.053 |
iden | -0.053 |
nexpected | -0.052 |
arest | -0.052 |
Token | rainbow |
Token ID | 27223 |
Feature activation | -8.84e-03 |
Pos logit contributions | |
Blue | +0.018 |
back | +0.017 |
over | +0.017 |
from | +0.015 |
Spot | +0.015 |
Neg logit contributions | |
inations | -0.027 |
rd | -0.027 |
iden | -0.026 |
nexpected | -0.026 |
arest | -0.026 |
Token | melted |
Token ID | 24178 |
Feature activation | -2.94e-03 |
Pos logit contributions | |
inations | +0.017 |
rd | +0.017 |
iden | +0.017 |
nexpected | +0.017 |
iest | +0.017 |
Neg logit contributions | |
Blue | -0.014 |
over | -0.013 |
back | -0.012 |
Spot | -0.012 |
from | -0.012 |
Token | surprise |
Token ID | 5975 |
Feature activation | -4.78e-02 |
Pos logit contributions | |
rd | +0.009 |
inations | +0.009 |
iden | +0.009 |
nexpected | +0.009 |
very | +0.009 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.006 |
back | -0.005 |
Spot | -0.005 |
from | -0.005 |
Token | . |
Token ID | 13 |
Feature activation | -9.36e-03 |
Pos logit contributions | |
inations | +0.121 |
rd | +0.121 |
iden | +0.119 |
nexpected | +0.118 |
arest | +0.118 |
Neg logit contributions | |
Blue | -0.049 |
over | -0.045 |
back | -0.042 |
from | -0.041 |
Spot | -0.040 |
Token | She |
Token ID | 1375 |
Feature activation | -6.03e-03 |
Pos logit contributions | |
rd | +0.021 |
very | +0.021 |
iden | +0.021 |
nexpected | +0.021 |
inations | +0.020 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
back | -0.010 |
ety | -0.009 |
from | -0.009 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.89e-02 |
Pos logit contributions | |
rd | +0.013 |
inations | +0.012 |
arest | +0.012 |
iest | +0.012 |
iden | +0.012 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.009 |
back | -0.008 |
from | -0.008 |
Spot | -0.008 |
Token | in |
Token ID | 287 |
Feature activation | -3.54e-03 |
Pos logit contributions | |
nexpected | +0.061 |
inations | +0.060 |
very | +0.060 |
iest | +0.059 |
rd | +0.059 |
Neg logit contributions | |
Blue | -0.036 |
ety | -0.036 |
isk | -0.033 |
Spot | -0.031 |
over | -0.031 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
inations | +0.004 |
rd | +0.004 |
iden | +0.004 |
nexpected | +0.004 |
arest | +0.004 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.008 |
from | -0.008 |
Spot | -0.008 |
Token | ement |
Token ID | 972 |
Feature activation | -4.17e-02 |
Pos logit contributions | |
inations | +0.129 |
iden | +0.125 |
oc | +0.120 |
arest | +0.108 |
rd | +0.107 |
Neg logit contributions | |
back | -0.298 |
over | -0.292 |
from | -0.288 |
Blue | -0.283 |
up | -0.274 |
Token | as |
Token ID | 355 |
Feature activation | +1.86e-02 |
Pos logit contributions | |
rd | +0.103 |
inations | +0.103 |
nexpected | +0.101 |
very | +0.100 |
iden | +0.099 |
Neg logit contributions | |
Blue | -0.044 |
over | -0.037 |
Spot | -0.034 |
back | -0.034 |
Sunny | -0.033 |
Token | the |
Token ID | 262 |
Feature activation | +2.88e-02 |
Pos logit contributions | |
Blue | +0.021 |
over | +0.020 |
back | +0.019 |
ety | +0.018 |
barking | +0.018 |
Neg logit contributions | |
inations | -0.043 |
rd | -0.042 |
nexpected | -0.042 |
iden | -0.042 |
iest | -0.042 |
Token | yellow |
Token ID | 7872 |
Feature activation | +9.85e-03 |
Pos logit contributions | |
Blue | +0.038 |
over | +0.035 |
from | +0.032 |
Spot | +0.032 |
back | +0.031 |
Neg logit contributions | |
rd | -0.063 |
iden | -0.062 |
inations | -0.062 |
iest | -0.062 |
nexpected | -0.061 |
Token | knee |
Token ID | 10329 |
Feature activation | +4.11e-03 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.013 |
back | +0.012 |
Spot | +0.012 |
barking | +0.012 |
Neg logit contributions | |
iest | -0.020 |
rd | -0.020 |
iden | -0.020 |
nexpected | -0.020 |
inations | -0.019 |
Token | water |
Token ID | 1660 |
Feature activation | +9.72e-03 |
Pos logit contributions | |
Blue | +0.003 |
Spot | +0.003 |
over | +0.003 |
Sunny | +0.003 |
barking | +0.003 |
Neg logit contributions | |
very | -0.006 |
iden | -0.006 |
inations | -0.006 |
iest | -0.005 |
rd | -0.005 |
Token | . |
Token ID | 13 |
Feature activation | -1.05e-03 |
Pos logit contributions | |
Blue | +0.015 |
isk | +0.013 |
ety | +0.013 |
Wh | +0.013 |
Spot | +0.013 |
Neg logit contributions | |
iest | -0.019 |
rd | -0.018 |
very | -0.018 |
inations | -0.018 |
arest | -0.017 |
Token | She |
Token ID | 1375 |
Feature activation | -4.19e-03 |
Pos logit contributions | |
iden | +0.002 |
very | +0.002 |
nexpected | +0.002 |
opter | +0.002 |
oc | +0.002 |
Neg logit contributions | |
back | -0.001 |
Blue | -0.001 |
over | -0.001 |
gone | -0.001 |
faster | -0.001 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.55e-02 |
Pos logit contributions | |
rd | +0.008 |
inations | +0.008 |
iden | +0.008 |
nexpected | +0.008 |
arest | +0.008 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
back | -0.006 |
from | -0.006 |
Spot | -0.006 |
Token | in |
Token ID | 287 |
Feature activation | -2.54e-03 |
Pos logit contributions | |
nexpected | +0.051 |
very | +0.050 |
inations | +0.050 |
iest | +0.049 |
goodness | +0.048 |
Neg logit contributions | |
ety | -0.034 |
Blue | -0.033 |
isk | -0.033 |
veterinarian | -0.028 |
over | -0.028 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
rd | +0.003 |
inations | +0.003 |
nexpected | +0.003 |
iden | +0.003 |
very | +0.002 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.006 |
ety | -0.006 |
back | -0.006 |
Spot | -0.006 |
Token | ement |
Token ID | 972 |
Feature activation | -3.99e-02 |
Pos logit contributions | |
inations | +0.118 |
iden | +0.116 |
oc | +0.106 |
rd | +0.103 |
arest | +0.099 |
Neg logit contributions | |
back | -0.299 |
over | -0.298 |
from | -0.295 |
Blue | -0.294 |
up | -0.276 |
Token | as |
Token ID | 355 |
Feature activation | +2.05e-02 |
Pos logit contributions | |
rd | +0.101 |
inations | +0.101 |
nexpected | +0.100 |
very | +0.100 |
iden | +0.099 |
Neg logit contributions | |
Blue | -0.038 |
over | -0.032 |
Spot | -0.029 |
back | -0.029 |
Sunny | -0.028 |
Token | the |
Token ID | 262 |
Feature activation | +2.71e-02 |
Pos logit contributions | |
Blue | +0.026 |
over | +0.024 |
back | +0.023 |
barking | +0.022 |
ety | +0.022 |
Neg logit contributions | |
iden | -0.044 |
iest | -0.043 |
nexpected | -0.043 |
inations | -0.043 |
very | -0.043 |
Token | duck |
Token ID | 22045 |
Feature activation | +6.94e-03 |
Pos logit contributions | |
Blue | +0.033 |
over | +0.031 |
isk | +0.031 |
from | +0.028 |
ety | +0.027 |
Neg logit contributions | |
iest | -0.062 |
iden | -0.059 |
rd | -0.058 |
arest | -0.058 |
inations | -0.058 |
Token | effortlessly |
Token ID | 42241 |
Feature activation | -9.61e-03 |
Pos logit contributions | |
Blue | +0.010 |
over | +0.009 |
back | +0.009 |
Spot | +0.009 |
from | +0.009 |
Neg logit contributions | |
inations | -0.014 |
rd | -0.014 |
iden | -0.014 |
nexpected | -0.014 |
iest | -0.014 |
Token | wings |
Token ID | 12098 |
Feature activation | -3.51e-03 |
Pos logit contributions | |
inations | +0.064 |
rd | +0.064 |
iden | +0.063 |
nexpected | +0.062 |
very | +0.062 |
Neg logit contributions | |
Blue | -0.040 |
over | -0.036 |
back | -0.035 |
from | -0.034 |
Spot | -0.033 |
Token | . |
Token ID | 13 |
Feature activation | -1.00e-03 |
Pos logit contributions | |
iest | +0.007 |
rd | +0.007 |
inations | +0.007 |
arest | +0.007 |
nexpected | +0.007 |
Neg logit contributions | |
Blue | -0.005 |
Sunny | -0.004 |
Spot | -0.004 |
isk | -0.004 |
ety | -0.004 |
Token | He |
Token ID | 679 |
Feature activation | -5.41e-03 |
Pos logit contributions | |
inations | +0.002 |
rd | +0.002 |
arest | +0.002 |
iden | +0.002 |
iest | +0.002 |
Neg logit contributions | |
Blue | -0.001 |
over | -0.001 |
Spot | -0.001 |
back | -0.001 |
from | -0.001 |
Token | watched |
Token ID | 7342 |
Feature activation | -1.83e-02 |
Pos logit contributions | |
inations | +0.010 |
iden | +0.010 |
rd | +0.010 |
nexpected | +0.010 |
very | +0.010 |
Neg logit contributions | |
Blue | -0.009 |
over | -0.008 |
back | -0.008 |
from | -0.007 |
Spot | -0.007 |
Token | in |
Token ID | 287 |
Feature activation | -8.87e-04 |
Pos logit contributions | |
nexpected | +0.036 |
very | +0.036 |
inations | +0.035 |
iest | +0.035 |
iden | +0.034 |
Neg logit contributions | |
ety | -0.026 |
Blue | -0.024 |
isk | -0.024 |
blown | -0.022 |
Spot | -0.021 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
rd | +0.001 |
inations | +0.001 |
iden | +0.001 |
nexpected | +0.001 |
very | +0.001 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
Spot | -0.002 |
from | -0.002 |
Token | ement |
Token ID | 972 |
Feature activation | -4.09e-02 |
Pos logit contributions | |
inations | +0.116 |
iden | +0.112 |
oc | +0.107 |
rd | +0.101 |
arest | +0.100 |
Neg logit contributions | |
back | -0.302 |
over | -0.300 |
from | -0.297 |
Blue | -0.293 |
up | -0.279 |
Token | as |
Token ID | 355 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
inations | +0.105 |
rd | +0.105 |
nexpected | +0.103 |
iden | +0.103 |
very | +0.103 |
Neg logit contributions | |
Blue | -0.039 |
over | -0.033 |
back | -0.030 |
Spot | -0.030 |
Sunny | -0.029 |
Token | it |
Token ID | 340 |
Feature activation | -1.36e-02 |
Pos logit contributions | |
Blue | +0.033 |
over | +0.030 |
back | +0.029 |
from | +0.028 |
Spot | +0.028 |
Neg logit contributions | |
inations | -0.046 |
rd | -0.046 |
iden | -0.045 |
nexpected | -0.045 |
very | -0.044 |
Token | flew |
Token ID | 13112 |
Feature activation | +4.78e-02 |
Pos logit contributions | |
inations | +0.031 |
rd | +0.031 |
nexpected | +0.031 |
iden | +0.030 |
very | +0.030 |
Neg logit contributions | |
Blue | -0.017 |
over | -0.015 |
back | -0.014 |
Spot | -0.014 |
from | -0.014 |
Token | away |
Token ID | 1497 |
Feature activation | +2.28e-02 |
Pos logit contributions | |
Blue | +0.056 |
isk | +0.053 |
Sunny | +0.049 |
ety | +0.049 |
Spot | +0.048 |
Neg logit contributions | |
very | -0.101 |
iest | -0.099 |
iden | -0.097 |
nexpected | -0.096 |
rd | -0.095 |
Token | and |
Token ID | 290 |
Feature activation | +3.67e-02 |
Pos logit contributions | |
inations | +0.042 |
rd | +0.042 |
iden | +0.042 |
very | +0.042 |
nexpected | +0.041 |
Neg logit contributions | |
Blue | -0.027 |
over | -0.024 |
back | -0.023 |
Spot | -0.023 |
from | -0.022 |
Token | his |
Token ID | 465 |
Feature activation | +4.22e-02 |
Pos logit contributions | |
Blue | +0.047 |
over | +0.043 |
back | +0.040 |
Spot | +0.040 |
from | +0.039 |
Neg logit contributions | |
rd | -0.082 |
inations | -0.081 |
iden | -0.081 |
nexpected | -0.080 |
very | -0.080 |
Token | wife |
Token ID | 3656 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
Blue | +0.068 |
over | +0.064 |
back | +0.061 |
from | +0.060 |
Spot | +0.059 |
Neg logit contributions | |
inations | -0.081 |
rd | -0.081 |
iden | -0.079 |
nexpected | -0.079 |
very | -0.078 |
Token | stared |
Token ID | 18484 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
rd | +0.042 |
inations | +0.041 |
nexpected | +0.041 |
iden | +0.041 |
arest | +0.041 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.024 |
back | -0.023 |
from | -0.022 |
Spot | -0.021 |
Token | in |
Token ID | 287 |
Feature activation | -1.60e-02 |
Pos logit contributions | |
rd | +0.054 |
inations | +0.054 |
iden | +0.053 |
nexpected | +0.053 |
very | +0.052 |
Neg logit contributions | |
Blue | -0.027 |
over | -0.024 |
back | -0.022 |
Spot | -0.022 |
from | -0.022 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
inations | +0.023 |
rd | +0.022 |
iden | +0.021 |
arest | +0.021 |
iest | +0.021 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.035 |
back | -0.034 |
from | -0.034 |
Spot | -0.033 |
Token | ement |
Token ID | 972 |
Feature activation | -5.22e-02 |
Pos logit contributions | |
inations | +0.114 |
iden | +0.110 |
oc | +0.100 |
arest | +0.098 |
rd | +0.094 |
Neg logit contributions | |
back | -0.310 |
over | -0.300 |
from | -0.295 |
Blue | -0.291 |
too | -0.278 |
Token | . |
Token ID | 13 |
Feature activation | -1.64e-02 |
Pos logit contributions | |
inations | +0.133 |
rd | +0.133 |
iden | +0.132 |
nexpected | +0.130 |
arest | +0.129 |
Neg logit contributions | |
Blue | -0.051 |
over | -0.048 |
from | -0.045 |
back | -0.045 |
Spot | -0.041 |
Token | Their |
Token ID | 5334 |
Feature activation | +1.60e-02 |
Pos logit contributions | |
rd | +0.038 |
iden | +0.038 |
inations | +0.037 |
very | +0.037 |
nexpected | +0.037 |
Neg logit contributions | |
Blue | -0.019 |
over | -0.017 |
back | -0.017 |
from | -0.016 |
Spot | -0.016 |
Token | house |
Token ID | 2156 |
Feature activation | -5.58e-03 |
Pos logit contributions | |
Blue | +0.023 |
back | +0.023 |
over | +0.023 |
from | +0.021 |
Spot | +0.021 |
Neg logit contributions | |
inations | -0.033 |
rd | -0.032 |
iest | -0.032 |
arest | -0.032 |
iden | -0.032 |
Token | looked |
Token ID | 3114 |
Feature activation | -1.46e-02 |
Pos logit contributions | |
inations | +0.011 |
rd | +0.011 |
arest | +0.011 |
iden | +0.011 |
very | +0.011 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.008 |
from | -0.008 |
Spot | -0.007 |
Token | hill |
Token ID | 12788 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
rd | +0.002 |
inations | +0.002 |
iden | +0.002 |
nexpected | +0.002 |
very | +0.002 |
Neg logit contributions | |
Blue | -0.001 |
over | -0.001 |
back | -0.001 |
from | -0.001 |
Spot | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | -8.82e-03 |
Pos logit contributions | |
Blue | +0.034 |
Sunny | +0.030 |
ety | +0.030 |
Wh | +0.029 |
isk | +0.029 |
Neg logit contributions | |
iest | -0.068 |
rd | -0.068 |
very | -0.068 |
iden | -0.067 |
inations | -0.066 |
Token | everyone |
Token ID | 2506 |
Feature activation | +4.48e-03 |
Pos logit contributions | |
inations | +0.016 |
iden | +0.016 |
nexpected | +0.016 |
very | +0.016 |
rd | +0.016 |
Neg logit contributions | |
Blue | -0.014 |
back | -0.012 |
ety | -0.012 |
over | -0.012 |
Spot | -0.012 |
Token | watched |
Token ID | 7342 |
Feature activation | -3.25e-02 |
Pos logit contributions | |
Blue | +0.008 |
over | +0.008 |
back | +0.007 |
from | +0.007 |
Spot | +0.007 |
Neg logit contributions | |
rd | -0.008 |
inations | -0.008 |
iden | -0.008 |
nexpected | -0.008 |
arest | -0.008 |
Token | in |
Token ID | 287 |
Feature activation | +5.41e-04 |
Pos logit contributions | |
inations | +0.067 |
nexpected | +0.066 |
rd | +0.065 |
very | +0.064 |
arest | +0.064 |
Neg logit contributions | |
Blue | -0.044 |
ety | -0.042 |
isk | -0.040 |
over | -0.038 |
Spot | -0.038 |
Token | amaz |
Token ID | 40642 |
Feature activation | -1.18e-01 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
ety | +0.001 |
back | +0.001 |
Spot | +0.001 |
Neg logit contributions | |
rd | -0.001 |
inations | -0.001 |
nexpected | -0.001 |
iden | -0.001 |
very | -0.001 |
Token | ement |
Token ID | 972 |
Feature activation | -4.57e-02 |
Pos logit contributions | |
iden | +0.120 |
inations | +0.118 |
oc | +0.108 |
rd | +0.103 |
arest | +0.098 |
Neg logit contributions | |
back | -0.300 |
over | -0.292 |
Blue | -0.291 |
from | -0.286 |
up | -0.272 |
Token | . |
Token ID | 13 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
inations | +0.108 |
rd | +0.108 |
iden | +0.105 |
nexpected | +0.105 |
very | +0.104 |
Neg logit contributions | |
Blue | -0.054 |
over | -0.048 |
back | -0.045 |
Spot | -0.044 |
from | -0.043 |
Token | The |
Token ID | 383 |
Feature activation | +1.75e-02 |
Pos logit contributions | |
rd | +0.053 |
inations | +0.053 |
iden | +0.053 |
nexpected | +0.053 |
very | +0.052 |
Neg logit contributions | |
Blue | -0.032 |
over | -0.030 |
back | -0.029 |
from | -0.028 |
barking | -0.027 |
Token | rock |
Token ID | 3881 |
Feature activation | +3.60e-03 |
Pos logit contributions | |
Blue | +0.022 |
over | +0.020 |
back | +0.019 |
Spot | +0.019 |
from | +0.019 |
Neg logit contributions | |
inations | -0.040 |
rd | -0.040 |
iden | -0.039 |
nexpected | -0.039 |
very | -0.038 |
Token | was |
Token ID | 373 |
Feature activation | +2.03e-03 |
Pos logit contributions | |
Blue | +0.006 |
over | +0.005 |
back | +0.005 |
Spot | +0.005 |
from | +0.005 |
Neg logit contributions | |
inations | -0.007 |
rd | -0.007 |
iden | -0.007 |
nexpected | -0.007 |
very | -0.007 |
Token | a |
Token ID | 257 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
rd | +0.017 |
iden | +0.016 |
inations | +0.016 |
very | +0.016 |
iest | +0.016 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
blown | -0.006 |
pering | -0.005 |
barking | -0.005 |
Token | snack |
Token ID | 26906 |
Feature activation | +3.49e-03 |
Pos logit contributions | |
rd | +0.060 |
inations | +0.060 |
nexpected | +0.059 |
arest | +0.058 |
iden | +0.057 |
Neg logit contributions | |
Blue | -0.037 |
over | -0.034 |
back | -0.033 |
Spot | -0.032 |
from | -0.031 |
Token | . |
Token ID | 13 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.004 |
barking | +0.004 |
ety | +0.004 |
blown | +0.004 |
Neg logit contributions | |
iest | -0.008 |
rd | -0.008 |
iden | -0.007 |
very | -0.007 |
inations | -0.007 |
Token |
|
Token ID | 198 |
Feature activation | +1.45e-02 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.012 |
back | +0.012 |
from | +0.012 |
blown | +0.011 |
Neg logit contributions | |
iden | -0.022 |
very | -0.022 |
rd | -0.022 |
nexpected | -0.021 |
inations | -0.021 |
Token |
|
Token ID | 198 |
Feature activation | +2.34e-02 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.012 |
Spot | +0.011 |
from | +0.010 |
back | +0.010 |
Neg logit contributions | |
inations | -0.039 |
rd | -0.038 |
nexpected | -0.038 |
iest | -0.037 |
arest | -0.037 |
Token | The |
Token ID | 464 |
Feature activation | +4.95e-02 |
Pos logit contributions | |
Blue | +0.035 |
over | +0.033 |
Spot | +0.031 |
back | +0.030 |
from | +0.030 |
Neg logit contributions | |
inations | -0.049 |
rd | -0.048 |
nexpected | -0.048 |
iden | -0.047 |
iest | -0.047 |
Token | mom |
Token ID | 1995 |
Feature activation | -3.71e-03 |
Pos logit contributions | |
Blue | +0.066 |
over | +0.061 |
back | +0.056 |
from | +0.056 |
Spot | +0.055 |
Neg logit contributions | |
rd | -0.109 |
inations | -0.108 |
iden | -0.108 |
nexpected | -0.106 |
very | -0.105 |
Token | said |
Token ID | 531 |
Feature activation | -1.33e-02 |
Pos logit contributions | |
iden | +0.007 |
very | +0.007 |
inations | +0.006 |
rd | +0.006 |
nexpected | +0.006 |
Neg logit contributions | |
Blue | -0.006 |
back | -0.006 |
over | -0.006 |
barking | -0.006 |
pering | -0.006 |
Token | , |
Token ID | 11 |
Feature activation | +3.43e-02 |
Pos logit contributions | |
inations | +0.030 |
rd | +0.029 |
arest | +0.029 |
iden | +0.029 |
nexpected | +0.029 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.017 |
back | -0.016 |
from | -0.016 |
Spot | -0.016 |
Token | " |
Token ID | 366 |
Feature activation | +3.75e-02 |
Pos logit contributions | |
Blue | +0.027 |
over | +0.024 |
back | +0.022 |
ety | +0.021 |
from | +0.021 |
Neg logit contributions | |
iden | -0.092 |
rd | -0.091 |
very | -0.091 |
inations | -0.090 |
nexpected | -0.090 |
Token | Let |
Token ID | 5756 |
Feature activation | -1.81e-03 |
Pos logit contributions | |
Blue | +0.048 |
over | +0.045 |
back | +0.041 |
ety | +0.040 |
Spot | +0.040 |
Neg logit contributions | |
inations | -0.089 |
rd | -0.087 |
nexpected | -0.084 |
arest | -0.084 |
iest | -0.082 |
Token | The |
Token ID | 383 |
Feature activation | +2.41e-02 |
Pos logit contributions | |
inations | +0.043 |
rd | +0.042 |
nexpected | +0.042 |
iest | +0.041 |
arest | +0.041 |
Neg logit contributions | |
Blue | -0.025 |
over | -0.023 |
Spot | -0.021 |
back | -0.021 |
from | -0.020 |
Token | owner |
Token ID | 4870 |
Feature activation | -2.01e-02 |
Pos logit contributions | |
Blue | +0.026 |
over | +0.023 |
back | +0.022 |
Spot | +0.021 |
from | +0.021 |
Neg logit contributions | |
inations | -0.060 |
rd | -0.060 |
iden | -0.059 |
nexpected | -0.058 |
arest | -0.058 |
Token | was |
Token ID | 373 |
Feature activation | +4.99e-03 |
Pos logit contributions | |
rd | +0.039 |
inations | +0.039 |
iden | +0.038 |
nexpected | +0.038 |
arest | +0.038 |
Neg logit contributions | |
Blue | -0.032 |
over | -0.030 |
back | -0.028 |
from | -0.028 |
Spot | -0.028 |
Token | happy |
Token ID | 3772 |
Feature activation | -1.60e-02 |
Pos logit contributions | |
Blue | +0.007 |
ety | +0.006 |
isk | +0.006 |
Wh | +0.006 |
Spot | +0.005 |
Neg logit contributions | |
rd | -0.010 |
iden | -0.010 |
inations | -0.010 |
very | -0.010 |
arest | -0.010 |
Token | that |
Token ID | 326 |
Feature activation | -5.27e-03 |
Pos logit contributions | |
inations | +0.041 |
rd | +0.041 |
iden | +0.040 |
nexpected | +0.040 |
arest | +0.040 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.014 |
back | -0.013 |
from | -0.013 |
Spot | -0.013 |
Token | the |
Token ID | 262 |
Feature activation | +4.26e-02 |
Pos logit contributions | |
rd | +0.010 |
inations | +0.010 |
iden | +0.010 |
nexpected | +0.010 |
very | +0.010 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
from | -0.007 |
Spot | -0.007 |
Token | cat |
Token ID | 3797 |
Feature activation | +2.38e-03 |
Pos logit contributions | |
Blue | +0.055 |
over | +0.051 |
back | +0.047 |
from | +0.047 |
Spot | +0.046 |
Neg logit contributions | |
rd | -0.095 |
inations | -0.095 |
iden | -0.094 |
nexpected | -0.092 |
very | -0.092 |
Token | was |
Token ID | 373 |
Feature activation | +6.61e-03 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.003 |
back | +0.003 |
Spot | +0.003 |
from | +0.003 |
Neg logit contributions | |
rd | -0.005 |
inations | -0.005 |
iden | -0.004 |
nexpected | -0.004 |
very | -0.004 |
Token | so |
Token ID | 523 |
Feature activation | -3.81e-02 |
Pos logit contributions | |
Blue | +0.009 |
ety | +0.008 |
isk | +0.008 |
Spot | +0.008 |
Wh | +0.008 |
Neg logit contributions | |
inations | -0.014 |
rd | -0.014 |
iest | -0.014 |
iden | -0.013 |
very | -0.013 |
Token | creative |
Token ID | 7325 |
Feature activation | -1.40e-02 |
Pos logit contributions | |
rd | +0.078 |
inations | +0.078 |
iden | +0.076 |
nexpected | +0.076 |
very | +0.075 |
Neg logit contributions | |
Blue | -0.058 |
over | -0.054 |
back | -0.051 |
from | -0.050 |
Spot | -0.050 |
Token | in |
Token ID | 287 |
Feature activation | +1.92e-02 |
Pos logit contributions | |
inations | +0.033 |
rd | +0.033 |
iden | +0.033 |
nexpected | +0.033 |
arest | +0.032 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.014 |
from | -0.013 |
Spot | -0.013 |
Token |
|
Token ID | 198 |
Feature activation | +2.39e-03 |
Pos logit contributions | |
rd | +0.005 |
inations | +0.005 |
iden | +0.005 |
iest | +0.005 |
nexpected | +0.005 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
from | -0.002 |
Spot | -0.002 |
Token | May |
Token ID | 6747 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.004 |
from | +0.003 |
also | +0.003 |
back | +0.003 |
Neg logit contributions | |
inations | -0.005 |
rd | -0.005 |
iden | -0.005 |
iest | -0.005 |
nexpected | -0.004 |
Token | a |
Token ID | 64 |
Feature activation | -1.38e-02 |
Pos logit contributions | |
inations | +0.027 |
rd | +0.027 |
nexpected | +0.027 |
arest | +0.026 |
iest | +0.026 |
Neg logit contributions | |
Blue | -0.030 |
over | -0.029 |
from | -0.029 |
back | -0.028 |
also | -0.027 |
Token | left |
Token ID | 1364 |
Feature activation | +2.82e-02 |
Pos logit contributions | |
inations | +0.028 |
rd | +0.027 |
arest | +0.027 |
iden | +0.027 |
iest | +0.026 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.020 |
back | -0.019 |
from | -0.019 |
Spot | -0.018 |
Token | the |
Token ID | 262 |
Feature activation | -4.16e-03 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.021 |
ety | +0.019 |
Spot | +0.019 |
back | +0.019 |
Neg logit contributions | |
rd | -0.073 |
inations | -0.073 |
iden | -0.072 |
nexpected | -0.072 |
very | -0.072 |
Token | park |
Token ID | 3952 |
Feature activation | +3.65e-02 |
Pos logit contributions | |
inations | +0.009 |
rd | +0.009 |
iden | +0.008 |
nexpected | +0.008 |
arest | +0.008 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.006 |
back | -0.006 |
from | -0.005 |
Spot | -0.005 |
Token | feeling |
Token ID | 4203 |
Feature activation | -1.79e-03 |
Pos logit contributions | |
Blue | +0.041 |
ety | +0.035 |
Spot | +0.034 |
Sunny | +0.033 |
over | +0.033 |
Neg logit contributions | |
rd | -0.087 |
very | -0.082 |
inations | -0.082 |
iden | -0.082 |
iest | -0.081 |
Token | happy |
Token ID | 3772 |
Feature activation | -9.77e-03 |
Pos logit contributions | |
rd | +0.004 |
inations | +0.004 |
iden | +0.004 |
nexpected | +0.004 |
very | +0.004 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
from | -0.002 |
Spot | -0.002 |
Token | and |
Token ID | 290 |
Feature activation | -7.08e-03 |
Pos logit contributions | |
rd | +0.026 |
inations | +0.026 |
iden | +0.026 |
very | +0.026 |
nexpected | +0.026 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.006 |
Spot | -0.006 |
ety | -0.006 |
Token | proud |
Token ID | 6613 |
Feature activation | -1.68e-02 |
Pos logit contributions | |
inations | +0.014 |
iden | +0.014 |
rd | +0.014 |
nexpected | +0.014 |
arest | +0.014 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.011 |
back | -0.010 |
from | -0.010 |
Spot | -0.010 |
Token | . |
Token ID | 13 |
Feature activation | -1.64e-02 |
Pos logit contributions | |
inations | +0.040 |
rd | +0.039 |
iden | +0.039 |
nexpected | +0.039 |
arest | +0.038 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.020 |
from | -0.019 |
back | -0.018 |
Spot | -0.017 |
Token | we |
Token ID | 356 |
Feature activation | +1.62e-02 |
Pos logit contributions | |
Blue | +0.021 |
over | +0.020 |
back | +0.019 |
Spot | +0.019 |
from | +0.019 |
Neg logit contributions | |
rd | -0.025 |
inations | -0.025 |
iest | -0.024 |
iden | -0.024 |
arest | -0.024 |
Token | pt |
Token ID | 457 |
Feature activation | -2.08e-02 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.024 |
back | +0.023 |
Spot | +0.022 |
from | +0.022 |
Neg logit contributions | |
rd | -0.034 |
inations | -0.034 |
very | -0.033 |
iden | -0.033 |
iest | -0.033 |
Token | . |
Token ID | 13 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
rd | +0.047 |
iest | +0.046 |
inations | +0.046 |
very | +0.045 |
arest | +0.045 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.022 |
ety | -0.021 |
isk | -0.020 |
Sunny | -0.020 |
Token | She |
Token ID | 1375 |
Feature activation | +6.63e-03 |
Pos logit contributions | |
Blue | +0.010 |
over | +0.009 |
back | +0.008 |
from | +0.008 |
barking | +0.007 |
Neg logit contributions | |
rd | -0.033 |
very | -0.032 |
iden | -0.032 |
inations | -0.032 |
nexpected | -0.032 |
Token | said |
Token ID | 531 |
Feature activation | -2.62e-02 |
Pos logit contributions | |
Blue | +0.010 |
over | +0.010 |
back | +0.009 |
Spot | +0.009 |
from | +0.009 |
Neg logit contributions | |
rd | -0.014 |
inations | -0.014 |
iest | -0.013 |
iden | -0.013 |
arest | -0.013 |
Token | , |
Token ID | 11 |
Feature activation | +3.78e-02 |
Pos logit contributions | |
inations | +0.058 |
rd | +0.057 |
iden | +0.056 |
arest | +0.055 |
nexpected | +0.055 |
Neg logit contributions | |
Blue | -0.038 |
over | -0.034 |
back | -0.033 |
Spot | -0.032 |
from | -0.032 |
Token | " |
Token ID | 366 |
Feature activation | +3.12e-02 |
Pos logit contributions | |
Blue | +0.038 |
ety | +0.038 |
blown | +0.034 |
over | +0.034 |
isk | +0.032 |
Neg logit contributions | |
iden | -0.088 |
very | -0.085 |
nexpected | -0.084 |
rd | -0.083 |
inations | -0.082 |
Token | You |
Token ID | 1639 |
Feature activation | -1.21e-02 |
Pos logit contributions | |
Blue | +0.049 |
Spot | +0.044 |
over | +0.044 |
back | +0.042 |
Sunny | +0.042 |
Neg logit contributions | |
inations | -0.066 |
rd | -0.064 |
iden | -0.063 |
iest | -0.062 |
nexpected | -0.061 |
Token | are |
Token ID | 389 |
Feature activation | +1.80e-03 |
Pos logit contributions | |
rd | +0.027 |
inations | +0.026 |
very | +0.026 |
iden | +0.026 |
nexpected | +0.026 |
Neg logit contributions | |
Blue | -0.017 |
over | -0.016 |
back | -0.015 |
from | -0.015 |
Spot | -0.015 |
Token | mean |
Token ID | 1612 |
Feature activation | -4.90e-03 |
Pos logit contributions | |
Blue | +0.002 |
ety | +0.002 |
asleep | +0.002 |
burn | +0.002 |
isk | +0.002 |
Neg logit contributions | |
inations | -0.004 |
prise | -0.004 |
rd | -0.004 |
nexpected | -0.004 |
arest | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +1.40e-02 |
Pos logit contributions | |
rd | +0.011 |
inations | +0.011 |
iden | +0.010 |
nexpected | +0.010 |
very | +0.010 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
back | -0.006 |
Spot | -0.005 |
from | -0.005 |
Token |
|
Token ID | 198 |
Feature activation | +1.53e-02 |
Pos logit contributions | |
Blue | +0.007 |
over | +0.007 |
Spot | +0.006 |
from | +0.006 |
back | +0.006 |
Neg logit contributions | |
inations | -0.020 |
rd | -0.020 |
nexpected | -0.020 |
iden | -0.019 |
iest | -0.019 |
Token | One |
Token ID | 3198 |
Feature activation | -8.70e-03 |
Pos logit contributions | |
Blue | +0.024 |
over | +0.022 |
back | +0.022 |
from | +0.021 |
Spot | +0.021 |
Neg logit contributions | |
rd | -0.030 |
inations | -0.030 |
iden | -0.029 |
nexpected | -0.029 |
very | -0.029 |
Token | day |
Token ID | 1110 |
Feature activation | -3.01e-02 |
Pos logit contributions | |
inations | +0.018 |
nexpected | +0.017 |
rd | +0.017 |
arest | +0.017 |
iden | +0.017 |
Neg logit contributions | |
Blue | -0.013 |
over | -0.013 |
from | -0.012 |
back | -0.012 |
Spot | -0.012 |
Token | , |
Token ID | 11 |
Feature activation | -8.97e-03 |
Pos logit contributions | |
inations | +0.077 |
rd | +0.074 |
arest | +0.072 |
iden | +0.072 |
nexpected | +0.069 |
Neg logit contributions | |
Blue | -0.037 |
over | -0.034 |
from | -0.033 |
Spot | -0.032 |
back | -0.031 |
Token | the |
Token ID | 262 |
Feature activation | +3.80e-02 |
Pos logit contributions | |
rd | +0.024 |
inations | +0.024 |
iden | +0.024 |
nexpected | +0.024 |
very | +0.024 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
back | -0.006 |
from | -0.005 |
Spot | -0.005 |
Token | little |
Token ID | 1310 |
Feature activation | +5.07e-02 |
Pos logit contributions | |
Blue | +0.054 |
over | +0.050 |
Spot | +0.046 |
back | +0.045 |
barking | +0.045 |
Neg logit contributions | |
rd | -0.080 |
iden | -0.079 |
inations | -0.079 |
very | -0.078 |
iest | -0.077 |
Token | girl |
Token ID | 2576 |
Feature activation | -6.76e-03 |
Pos logit contributions | |
Blue | +0.088 |
over | +0.083 |
isk | +0.082 |
Spot | +0.079 |
barking | +0.079 |
Neg logit contributions | |
very | -0.088 |
iden | -0.086 |
iest | -0.085 |
rd | -0.082 |
nexpected | -0.079 |
Token | wasn |
Token ID | 2492 |
Feature activation | +1.58e-02 |
Pos logit contributions | |
rd | +0.014 |
inations | +0.014 |
iden | +0.014 |
nexpected | +0.014 |
very | +0.014 |
Neg logit contributions | |
Blue | -0.010 |
over | -0.009 |
back | -0.008 |
from | -0.008 |
Spot | -0.008 |
Token | 't |
Token ID | 470 |
Feature activation | +2.41e-02 |
Pos logit contributions | |
Blue | +0.026 |
ety | +0.026 |
ter | +0.025 |
isk | +0.024 |
barking | +0.024 |
Neg logit contributions | |
arest | -0.028 |
inations | -0.027 |
rd | -0.026 |
iden | -0.025 |
very | -0.025 |
Token | feeling |
Token ID | 4203 |
Feature activation | -2.63e-03 |
Pos logit contributions | |
Blue | +0.034 |
Spot | +0.029 |
Sunny | +0.028 |
isk | +0.027 |
over | +0.027 |
Neg logit contributions | |
inations | -0.053 |
iden | -0.052 |
very | -0.051 |
iest | -0.050 |
rd | -0.049 |
Token | so |
Token ID | 523 |
Feature activation | -3.38e-02 |
Pos logit contributions | |
inations | +0.006 |
rd | +0.006 |
iest | +0.006 |
iden | +0.006 |
very | +0.006 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.003 |
Spot | -0.003 |
back | -0.003 |
barking | -0.003 |
Token | was |
Token ID | 373 |
Feature activation | +6.67e-03 |
Pos logit contributions | |
rd | +0.012 |
inations | +0.012 |
iden | +0.012 |
arest | +0.012 |
nexpected | +0.012 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
Spot | -0.007 |
from | -0.007 |
Token | about |
Token ID | 546 |
Feature activation | +2.46e-03 |
Pos logit contributions | |
ety | +0.009 |
Blue | +0.009 |
isk | +0.008 |
ter | +0.008 |
fence | +0.007 |
Neg logit contributions | |
inations | -0.013 |
rd | -0.012 |
iden | -0.012 |
arest | -0.012 |
nexpected | -0.012 |
Token | to |
Token ID | 284 |
Feature activation | -8.83e-03 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.002 |
ety | +0.002 |
back | +0.002 |
isk | +0.002 |
Neg logit contributions | |
inations | -0.006 |
iden | -0.006 |
nexpected | -0.006 |
rd | -0.006 |
very | -0.006 |
Token | pinch |
Token ID | 25394 |
Feature activation | -1.83e-02 |
Pos logit contributions | |
rd | +0.019 |
inations | +0.019 |
nexpected | +0.018 |
iden | +0.018 |
very | +0.018 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
back | -0.010 |
ety | -0.010 |
from | -0.010 |
Token | it |
Token ID | 340 |
Feature activation | +2.12e-02 |
Pos logit contributions | |
inations | +0.047 |
rd | +0.047 |
iden | +0.046 |
arest | +0.046 |
nexpected | +0.046 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.017 |
back | -0.015 |
from | -0.015 |
Spot | -0.015 |
Token | hard |
Token ID | 1327 |
Feature activation | +3.37e-03 |
Pos logit contributions | |
Blue | +0.028 |
ety | +0.026 |
isk | +0.025 |
Dove | +0.024 |
Sunny | +0.024 |
Neg logit contributions | |
iest | -0.042 |
rd | -0.041 |
inations | -0.040 |
nexpected | -0.040 |
ises | -0.039 |
Token | , |
Token ID | 11 |
Feature activation | -5.24e-04 |
Pos logit contributions | |
Blue | +0.004 |
ety | +0.003 |
Sunny | +0.003 |
over | +0.003 |
Spot | +0.003 |
Neg logit contributions | |
rd | -0.008 |
iest | -0.008 |
inations | -0.008 |
very | -0.008 |
nexpected | -0.007 |
Token | when |
Token ID | 618 |
Feature activation | +8.60e-03 |
Pos logit contributions | |
inations | +0.001 |
nexpected | +0.001 |
rd | +0.001 |
very | +0.001 |
iden | +0.001 |
Neg logit contributions | |
Blue | -0.001 |
ety | -0.001 |
isk | -0.001 |
over | -0.000 |
Sunny | -0.000 |
Token | he |
Token ID | 339 |
Feature activation | -9.50e-03 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.012 |
back | +0.012 |
from | +0.011 |
Spot | +0.011 |
Neg logit contributions | |
rd | -0.017 |
inations | -0.017 |
iden | -0.017 |
nexpected | -0.017 |
very | -0.016 |
Token | felt |
Token ID | 2936 |
Feature activation | +3.27e-03 |
Pos logit contributions | |
rd | +0.021 |
inations | +0.020 |
arest | +0.020 |
iden | +0.020 |
iest | +0.020 |
Neg logit contributions | |
Blue | -0.014 |
over | -0.013 |
Spot | -0.012 |
back | -0.012 |
from | -0.012 |
Token | something |
Token ID | 1223 |
Feature activation | -2.34e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.003 |
ety | +0.003 |
back | +0.003 |
Spot | +0.003 |
Neg logit contributions | |
inations | -0.007 |
rd | -0.007 |
nexpected | -0.007 |
iest | -0.007 |
arest | -0.007 |
Token | going |
Token ID | 1016 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
Blue | +0.018 |
ety | +0.017 |
Sunny | +0.016 |
isk | +0.016 |
k | +0.015 |
Neg logit contributions | |
inations | -0.032 |
very | -0.029 |
affles | -0.029 |
prise | -0.029 |
rd | -0.029 |
Token | to |
Token ID | 284 |
Feature activation | +1.87e-02 |
Pos logit contributions | |
Blue | +0.014 |
Sunny | +0.012 |
Spot | +0.011 |
ter | +0.011 |
ety | +0.011 |
Neg logit contributions | |
inations | -0.029 |
iden | -0.028 |
rd | -0.028 |
very | -0.028 |
iest | -0.027 |
Token | sell |
Token ID | 3677 |
Feature activation | -1.81e-03 |
Pos logit contributions | |
Blue | +0.021 |
ety | +0.021 |
over | +0.021 |
ter | +0.020 |
pering | +0.020 |
Neg logit contributions | |
inations | -0.040 |
iden | -0.040 |
rd | -0.039 |
very | -0.039 |
nexpected | -0.038 |
Token | some |
Token ID | 617 |
Feature activation | -5.68e-03 |
Pos logit contributions | |
rd | +0.004 |
nexpected | +0.004 |
inations | +0.004 |
iden | +0.004 |
very | +0.004 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
barking | -0.002 |
back | -0.002 |
ety | -0.002 |
Token | of |
Token ID | 286 |
Feature activation | +1.43e-02 |
Pos logit contributions | |
rd | +0.012 |
inations | +0.012 |
iden | +0.011 |
nexpected | +0.011 |
arest | +0.011 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
Spot | -0.007 |
from | -0.007 |
Token | her |
Token ID | 607 |
Feature activation | +9.49e-06 |
Pos logit contributions | |
over | +0.015 |
Blue | +0.015 |
pering | +0.014 |
back | +0.014 |
ety | +0.014 |
Neg logit contributions | |
inations | -0.031 |
very | -0.031 |
iden | -0.031 |
nexpected | -0.030 |
ocol | -0.030 |
Token | old |
Token ID | 1468 |
Feature activation | +1.89e-02 |
Token | toys |
Token ID | 14958 |
Feature activation | +1.70e-02 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.024 |
back | +0.022 |
Spot | +0.021 |
from | +0.021 |
Neg logit contributions | |
rd | -0.041 |
inations | -0.040 |
iden | -0.040 |
iest | -0.039 |
nexpected | -0.039 |
Token | at |
Token ID | 379 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
Blue | +0.027 |
barking | +0.024 |
Sunny | +0.024 |
Bird | +0.023 |
pering | +0.023 |
Neg logit contributions | |
iest | -0.032 |
rd | -0.031 |
inations | -0.031 |
arest | -0.030 |
iden | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -2.90e-02 |
Pos logit contributions | |
very | +0.020 |
iden | +0.019 |
rd | +0.019 |
prise | +0.019 |
ons | +0.019 |
Neg logit contributions | |
Blue | -0.014 |
ter | -0.013 |
pering | -0.013 |
ters | -0.013 |
barking | -0.013 |
Token | garage |
Token ID | 15591 |
Feature activation | +3.77e-02 |
Pos logit contributions | |
iden | +0.058 |
very | +0.058 |
rd | +0.057 |
inations | +0.057 |
arest | +0.055 |
Neg logit contributions | |
Blue | -0.042 |
over | -0.040 |
ter | -0.038 |
from | -0.038 |
Spot | -0.036 |
Token | help |
Token ID | 1037 |
Feature activation | -2.12e-02 |
Pos logit contributions | |
Blue | +0.014 |
over | +0.013 |
ety | +0.012 |
Spot | +0.012 |
barking | +0.012 |
Neg logit contributions | |
inations | -0.024 |
rd | -0.024 |
arest | -0.023 |
nexpected | -0.023 |
very | -0.023 |
Token | us |
Token ID | 514 |
Feature activation | +2.94e-02 |
Pos logit contributions | |
inations | +0.043 |
rd | +0.043 |
nexpected | +0.042 |
iden | +0.042 |
very | +0.042 |
Neg logit contributions | |
Blue | -0.031 |
over | -0.029 |
back | -0.027 |
Spot | -0.027 |
barking | -0.026 |
Token | find |
Token ID | 1064 |
Feature activation | +7.96e-03 |
Pos logit contributions | |
Blue | +0.035 |
Spot | +0.030 |
barking | +0.030 |
ety | +0.030 |
Wh | +0.029 |
Neg logit contributions | |
inations | -0.069 |
rd | -0.066 |
arest | -0.064 |
iest | -0.064 |
very | -0.063 |
Token | the |
Token ID | 262 |
Feature activation | -2.51e-03 |
Pos logit contributions | |
Blue | +0.009 |
over | +0.008 |
back | +0.008 |
from | +0.007 |
Spot | +0.007 |
Neg logit contributions | |
rd | -0.019 |
inations | -0.019 |
iden | -0.019 |
nexpected | -0.019 |
very | -0.019 |
Token | beach |
Token ID | 10481 |
Feature activation | +1.66e-02 |
Pos logit contributions | |
inations | +0.005 |
rd | +0.005 |
iden | +0.005 |
iest | +0.005 |
nexpected | +0.005 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.003 |
Spot | -0.003 |
barking | -0.003 |
ety | -0.003 |
Token | again |
Token ID | 757 |
Feature activation | +1.88e-02 |
Pos logit contributions | |
Blue | +0.019 |
over | +0.017 |
Spot | +0.016 |
barking | +0.015 |
back | +0.015 |
Neg logit contributions | |
rd | -0.039 |
inations | -0.039 |
iden | -0.038 |
iest | -0.038 |
very | -0.038 |
Token | , |
Token ID | 11 |
Feature activation | -4.54e-03 |
Pos logit contributions | |
Blue | +0.023 |
over | +0.021 |
back | +0.020 |
from | +0.019 |
Spot | +0.019 |
Neg logit contributions | |
rd | -0.044 |
inations | -0.044 |
iden | -0.043 |
nexpected | -0.043 |
very | -0.042 |
Token | no |
Token ID | 645 |
Feature activation | +1.69e-04 |
Pos logit contributions | |
inations | +0.011 |
rd | +0.011 |
iden | +0.011 |
very | +0.010 |
iest | +0.010 |
Neg logit contributions | |
Blue | -0.005 |
ety | -0.004 |
over | -0.004 |
isk | -0.004 |
barking | -0.004 |
Token | matter |
Token ID | 2300 |
Feature activation | -1.45e-02 |
Pos logit contributions | |
Blue | +0.000 |
over | +0.000 |
back | +0.000 |
Spot | +0.000 |
from | +0.000 |
Neg logit contributions | |
inations | -0.000 |
rd | -0.000 |
iden | -0.000 |
nexpected | -0.000 |
iest | -0.000 |
Token | where |
Token ID | 810 |
Feature activation | -2.51e-02 |
Pos logit contributions | |
inations | +0.032 |
nexpected | +0.032 |
rd | +0.032 |
iest | +0.031 |
arest | +0.030 |
Neg logit contributions | |
Blue | -0.017 |
ety | -0.015 |
over | -0.015 |
back | -0.015 |
barking | -0.015 |
Token | we |
Token ID | 356 |
Feature activation | +1.96e-03 |
Pos logit contributions | |
rd | +0.052 |
inations | +0.052 |
iden | +0.051 |
nexpected | +0.050 |
arest | +0.050 |
Neg logit contributions | |
Blue | -0.037 |
over | -0.034 |
back | -0.033 |
Spot | -0.032 |
from | -0.032 |
Token | One |
Token ID | 1881 |
Feature activation | +1.62e-03 |
Pos logit contributions | |
Blue | +0.020 |
from | +0.018 |
over | +0.018 |
ter | +0.018 |
blown | +0.018 |
Neg logit contributions | |
iden | -0.030 |
rd | -0.030 |
very | -0.029 |
inations | -0.029 |
nexpected | -0.029 |
Token | day |
Token ID | 1110 |
Feature activation | -2.28e-02 |
Pos logit contributions | |
Blue | +0.003 |
over | +0.003 |
back | +0.003 |
from | +0.003 |
Spot | +0.003 |
Neg logit contributions | |
inations | -0.003 |
rd | -0.003 |
iden | -0.003 |
nexpected | -0.003 |
arest | -0.003 |
Token | , |
Token ID | 11 |
Feature activation | -3.39e-05 |
Pos logit contributions | |
inations | +0.056 |
rd | +0.055 |
iden | +0.053 |
arest | +0.053 |
very | +0.052 |
Neg logit contributions | |
Blue | -0.029 |
over | -0.027 |
Spot | -0.026 |
from | -0.025 |
back | -0.024 |
Token | Tim |
Token ID | 5045 |
Feature activation | -9.68e-04 |
Pos logit contributions | |
rd | +0.000 |
inations | +0.000 |
iden | +0.000 |
arest | +0.000 |
nexpected | +0.000 |
Neg logit contributions | |
Blue | -0.000 |
over | -0.000 |
Spot | -0.000 |
back | -0.000 |
from | -0.000 |
Token | my |
Token ID | 1820 |
Feature activation | -2.47e-02 |
Pos logit contributions | |
rd | +0.002 |
iden | +0.002 |
inations | +0.002 |
very | +0.002 |
nexpected | +0.002 |
Neg logit contributions | |
Blue | -0.001 |
over | -0.001 |
Sunny | -0.001 |
from | -0.001 |
Spot | -0.001 |
Token | 's |
Token ID | 338 |
Feature activation | +1.70e-02 |
Pos logit contributions | |
inations | +0.056 |
rd | +0.056 |
nexpected | +0.055 |
iden | +0.055 |
arest | +0.055 |
Neg logit contributions | |
Blue | -0.031 |
over | -0.029 |
back | -0.027 |
from | -0.026 |
Spot | -0.026 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.024 |
Sunny | +0.022 |
barking | +0.022 |
ety | +0.021 |
Neg logit contributions | |
iden | -0.033 |
rd | -0.033 |
inations | -0.032 |
iest | -0.032 |
nexpected | -0.032 |
Token | asked |
Token ID | 1965 |
Feature activation | +1.54e-02 |
Pos logit contributions | |
iden | +0.021 |
inations | +0.021 |
very | +0.021 |
rd | +0.020 |
nexpected | +0.020 |
Neg logit contributions | |
Blue | -0.015 |
Sunny | -0.013 |
back | -0.013 |
over | -0.013 |
Spot | -0.013 |
Token | him |
Token ID | 683 |
Feature activation | +2.98e-02 |
Pos logit contributions | |
Blue | +0.026 |
over | +0.024 |
blown | +0.023 |
back | +0.023 |
Sunny | +0.023 |
Neg logit contributions | |
rd | -0.027 |
nexpected | -0.027 |
iden | -0.027 |
inations | -0.026 |
very | -0.026 |
Token | if |
Token ID | 611 |
Feature activation | +8.99e-03 |
Pos logit contributions | |
Blue | +0.030 |
blown | +0.025 |
over | +0.024 |
barking | +0.024 |
ety | +0.024 |
Neg logit contributions | |
rd | -0.072 |
iden | -0.072 |
inations | -0.071 |
nexpected | -0.070 |
iest | -0.070 |
Token | he |
Token ID | 339 |
Feature activation | -9.27e-03 |
Pos logit contributions | |
Blue | +0.015 |
Spot | +0.013 |
over | +0.013 |
back | +0.013 |
from | +0.013 |
Neg logit contributions | |
inations | -0.018 |
rd | -0.018 |
arest | -0.017 |
iden | -0.017 |
iest | -0.017 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.41e-02 |
Pos logit contributions | |
Blue | +0.019 |
over | +0.018 |
back | +0.016 |
from | +0.016 |
Spot | +0.016 |
Neg logit contributions | |
rd | -0.028 |
iden | -0.028 |
inations | -0.028 |
nexpected | -0.027 |
iest | -0.027 |
Token | gave |
Token ID | 2921 |
Feature activation | +9.47e-03 |
Pos logit contributions | |
iden | +0.026 |
very | +0.025 |
inations | +0.025 |
rd | +0.024 |
nexpected | +0.024 |
Neg logit contributions | |
Blue | -0.024 |
Sunny | -0.022 |
pering | -0.021 |
back | -0.021 |
over | -0.021 |
Token | him |
Token ID | 683 |
Feature activation | +1.14e-02 |
Pos logit contributions | |
Blue | +0.017 |
pering | +0.016 |
barking | +0.016 |
over | +0.016 |
Sunny | +0.016 |
Neg logit contributions | |
nexpected | -0.014 |
very | -0.014 |
iden | -0.013 |
rd | -0.013 |
inations | -0.013 |
Token | some |
Token ID | 617 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
Blue | +0.009 |
over | +0.008 |
barking | +0.007 |
back | +0.007 |
from | +0.007 |
Neg logit contributions | |
rd | -0.030 |
inations | -0.030 |
iden | -0.030 |
nexpected | -0.030 |
very | -0.029 |
Token | candy |
Token ID | 18550 |
Feature activation | +1.62e-03 |
Pos logit contributions | |
rd | +0.026 |
inations | +0.025 |
iden | +0.025 |
nexpected | +0.025 |
very | +0.025 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.014 |
from | -0.014 |
back | -0.014 |
Spot | -0.013 |
Token | to |
Token ID | 284 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
Blue | +0.002 |
barking | +0.002 |
Bird | +0.002 |
pering | +0.002 |
over | +0.002 |
Neg logit contributions | |
iest | -0.004 |
very | -0.004 |
rd | -0.004 |
iden | -0.004 |
inations | -0.003 |
Token | eat |
Token ID | 4483 |
Feature activation | +8.46e-04 |
Pos logit contributions | |
rd | +0.031 |
inations | +0.031 |
iden | +0.031 |
nexpected | +0.030 |
arest | +0.030 |
Neg logit contributions | |
Blue | -0.017 |
over | -0.015 |
back | -0.014 |
Spot | -0.014 |
from | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | +7.76e-03 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
from | +0.001 |
Neg logit contributions | |
rd | -0.002 |
inations | -0.002 |
iden | -0.002 |
nexpected | -0.002 |
very | -0.002 |
Token | Tim |
Token ID | 5045 |
Feature activation | -1.54e-02 |
Pos logit contributions | |
back | +0.009 |
Blue | +0.009 |
blown | +0.009 |
over | +0.009 |
faster | +0.009 |
Neg logit contributions | |
iden | -0.017 |
very | -0.016 |
rd | -0.016 |
nexpected | -0.016 |
mmm | -0.016 |
Token | my |
Token ID | 1820 |
Feature activation | -3.73e-02 |
Pos logit contributions | |
rd | +0.040 |
inations | +0.040 |
iden | +0.039 |
nexpected | +0.039 |
very | +0.039 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.012 |
back | -0.011 |
from | -0.011 |
Spot | -0.011 |
Token | was |
Token ID | 373 |
Feature activation | +7.87e-03 |
Pos logit contributions | |
inations | +0.074 |
rd | +0.073 |
arest | +0.072 |
iest | +0.072 |
iden | +0.072 |
Neg logit contributions | |
Blue | -0.059 |
over | -0.055 |
back | -0.052 |
Spot | -0.052 |
from | -0.051 |
Token | her |
Token ID | 607 |
Feature activation | -2.30e-02 |
Pos logit contributions | |
inations | +0.019 |
rd | +0.018 |
nexpected | +0.018 |
iden | +0.018 |
very | +0.018 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
ety | -0.006 |
back | -0.006 |
from | -0.006 |
Token | mouth |
Token ID | 5422 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
inations | +0.049 |
rd | +0.049 |
nexpected | +0.049 |
iden | +0.049 |
arest | +0.048 |
Neg logit contributions | |
Blue | -0.032 |
back | -0.030 |
over | -0.030 |
from | -0.028 |
Spot | -0.028 |
Token | ! |
Token ID | 0 |
Feature activation | -1.51e-02 |
Pos logit contributions | |
rd | +0.029 |
iest | +0.028 |
inations | +0.028 |
very | +0.028 |
arest | +0.027 |
Neg logit contributions | |
Blue | -0.018 |
Spot | -0.015 |
ety | -0.015 |
Wh | -0.015 |
over | -0.014 |
Token | She |
Token ID | 1375 |
Feature activation | -6.40e-03 |
Pos logit contributions | |
rd | +0.035 |
inations | +0.035 |
iden | +0.035 |
nexpected | +0.034 |
very | +0.034 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.016 |
back | -0.015 |
from | -0.015 |
Spot | -0.015 |
Token | was |
Token ID | 373 |
Feature activation | -2.65e-03 |
Pos logit contributions | |
rd | +0.013 |
inations | +0.013 |
arest | +0.013 |
iden | +0.013 |
iest | +0.013 |
Neg logit contributions | |
Blue | -0.010 |
over | -0.009 |
back | -0.009 |
from | -0.009 |
Spot | -0.008 |
Token | so |
Token ID | 523 |
Feature activation | -4.55e-02 |
Pos logit contributions | |
rd | +0.006 |
inations | +0.006 |
iden | +0.006 |
very | +0.006 |
nexpected | +0.006 |
Neg logit contributions | |
Blue | -0.003 |
ety | -0.002 |
Spot | -0.002 |
over | -0.002 |
isk | -0.002 |
Token | excited |
Token ID | 6568 |
Feature activation | -1.97e-02 |
Pos logit contributions | |
rd | +0.099 |
inations | +0.098 |
iden | +0.098 |
nexpected | +0.098 |
iest | +0.095 |
Neg logit contributions | |
Blue | -0.061 |
over | -0.059 |
back | -0.056 |
from | -0.055 |
Spot | -0.053 |
Token | to |
Token ID | 284 |
Feature activation | -4.22e-03 |
Pos logit contributions | |
rd | +0.047 |
inations | +0.046 |
iden | +0.046 |
very | +0.045 |
nexpected | +0.045 |
Neg logit contributions | |
Blue | -0.023 |
over | -0.020 |
back | -0.019 |
Spot | -0.019 |
from | -0.018 |
Token | show |
Token ID | 905 |
Feature activation | -2.49e-02 |
Pos logit contributions | |
rd | +0.010 |
inations | +0.009 |
iden | +0.009 |
nexpected | +0.009 |
very | +0.009 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.005 |
back | -0.005 |
from | -0.005 |
Spot | -0.005 |
Token | her |
Token ID | 607 |
Feature activation | +1.37e-02 |
Pos logit contributions | |
rd | +0.065 |
inations | +0.065 |
iden | +0.064 |
nexpected | +0.063 |
arest | +0.063 |
Neg logit contributions | |
Blue | -0.024 |
over | -0.022 |
back | -0.020 |
from | -0.019 |
Spot | -0.019 |
Token | mom |
Token ID | 1995 |
Feature activation | -7.63e-03 |
Pos logit contributions | |
Blue | +0.020 |
over | +0.019 |
back | +0.018 |
from | +0.018 |
Spot | +0.017 |
Neg logit contributions | |
rd | -0.029 |
arest | -0.029 |
inations | -0.029 |
nexpected | -0.028 |
very | -0.028 |
Token | are |
Token ID | 389 |
Feature activation | +6.96e-04 |
Pos logit contributions | |
inations | +0.085 |
rd | +0.085 |
very | +0.083 |
iden | +0.083 |
nexpected | +0.083 |
Neg logit contributions | |
Blue | -0.058 |
over | -0.053 |
from | -0.051 |
Spot | -0.050 |
back | -0.050 |
Token | you |
Token ID | 345 |
Feature activation | -3.03e-02 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
from | +0.001 |
Neg logit contributions | |
rd | -0.001 |
inations | -0.001 |
iden | -0.001 |
nexpected | -0.001 |
very | -0.001 |
Token | doing |
Token ID | 1804 |
Feature activation | -5.25e-02 |
Pos logit contributions | |
iden | +0.053 |
inations | +0.052 |
very | +0.052 |
rd | +0.050 |
nexpected | +0.050 |
Neg logit contributions | |
Blue | -0.054 |
Spot | -0.048 |
Sunny | -0.048 |
ety | -0.047 |
over | -0.047 |
Token | , |
Token ID | 11 |
Feature activation | -1.67e-02 |
Pos logit contributions | |
rd | +0.111 |
inations | +0.110 |
iden | +0.109 |
very | +0.108 |
nexpected | +0.108 |
Neg logit contributions | |
Blue | -0.073 |
over | -0.066 |
back | -0.063 |
Spot | -0.062 |
ety | -0.062 |
Token | Bear |
Token ID | 14732 |
Feature activation | -3.28e-02 |
Pos logit contributions | |
iden | +0.036 |
very | +0.035 |
inations | +0.034 |
rd | +0.034 |
mmm | +0.033 |
Neg logit contributions | |
Blue | -0.021 |
over | -0.021 |
blown | -0.020 |
back | -0.019 |
ety | -0.019 |
Token | ?" |
Token ID | 1701 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
iden | +0.052 |
very | +0.051 |
inations | +0.051 |
rd | +0.050 |
iest | +0.049 |
Neg logit contributions | |
Blue | -0.064 |
over | -0.058 |
back | -0.057 |
Spot | -0.057 |
Sunny | -0.055 |
Token | The |
Token ID | 383 |
Feature activation | +3.41e-02 |
Pos logit contributions | |
inations | +0.076 |
rd | +0.076 |
iden | +0.075 |
nexpected | +0.074 |
arest | +0.074 |
Neg logit contributions | |
Blue | -0.040 |
over | -0.036 |
back | -0.034 |
Spot | -0.033 |
from | -0.033 |
Token | Bear |
Token ID | 14732 |
Feature activation | -3.07e-02 |
Pos logit contributions | |
Blue | +0.040 |
over | +0.040 |
from | +0.036 |
ety | +0.035 |
Spot | +0.034 |
Neg logit contributions | |
iden | -0.075 |
very | -0.075 |
rd | -0.073 |
inations | -0.072 |
nexpected | -0.072 |
Token | smiled |
Token ID | 13541 |
Feature activation | -1.58e-03 |
Pos logit contributions | |
rd | +0.047 |
inations | +0.047 |
iden | +0.046 |
nexpected | +0.046 |
very | +0.046 |
Neg logit contributions | |
Blue | -0.061 |
over | -0.057 |
back | -0.055 |
Spot | -0.055 |
from | -0.054 |
Token | and |
Token ID | 290 |
Feature activation | -1.64e-02 |
Pos logit contributions | |
very | +0.004 |
iden | +0.004 |
rd | +0.003 |
ement | +0.003 |
iest | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
isk | -0.002 |
Wh | -0.001 |
blown | -0.001 |
ety | -0.001 |
Token | said |
Token ID | 531 |
Feature activation | -2.61e-02 |
Pos logit contributions | |
iden | +0.026 |
nexpected | +0.025 |
very | +0.025 |
rd | +0.025 |
inations | +0.025 |
Neg logit contributions | |
Blue | -0.031 |
over | -0.029 |
back | -0.028 |
ety | -0.028 |
from | -0.028 |
Token | was |
Token ID | 373 |
Feature activation | +5.59e-03 |
Pos logit contributions | |
inations | +0.062 |
arest | +0.060 |
rd | +0.059 |
affles | +0.058 |
nexpected | +0.058 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.033 |
from | -0.033 |
back | -0.032 |
Spot | -0.031 |
Token | a |
Token ID | 257 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.003 |
back | +0.003 |
Spot | +0.003 |
from | +0.003 |
Neg logit contributions | |
rd | -0.015 |
inations | -0.015 |
iden | -0.015 |
nexpected | -0.015 |
very | -0.015 |
Token | mysterious |
Token ID | 11428 |
Feature activation | -7.81e-03 |
Pos logit contributions | |
rd | +0.030 |
inations | +0.030 |
nexpected | +0.029 |
iest | +0.029 |
arest | +0.028 |
Neg logit contributions | |
Blue | -0.017 |
back | -0.015 |
over | -0.015 |
from | -0.014 |
barking | -0.014 |
Token | creature |
Token ID | 7185 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
inations | +0.017 |
rd | +0.017 |
nexpected | +0.016 |
arest | +0.016 |
iden | +0.016 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
back | -0.010 |
from | -0.010 |
Spot | -0.009 |
Token | with |
Token ID | 351 |
Feature activation | +6.81e-03 |
Pos logit contributions | |
rd | +0.030 |
inations | +0.030 |
iden | +0.029 |
nexpected | +0.029 |
very | +0.029 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.014 |
back | -0.013 |
Spot | -0.013 |
from | -0.013 |
Token | a |
Token ID | 257 |
Feature activation | -2.54e-02 |
Pos logit contributions | |
Blue | +0.007 |
over | +0.006 |
back | +0.006 |
from | +0.006 |
Spot | +0.006 |
Neg logit contributions | |
inations | -0.017 |
rd | -0.017 |
iden | -0.017 |
nexpected | -0.017 |
very | -0.017 |
Token | kind |
Token ID | 1611 |
Feature activation | +7.01e-03 |
Pos logit contributions | |
rd | +0.060 |
inations | +0.059 |
nexpected | +0.058 |
iest | +0.056 |
affles | +0.056 |
Neg logit contributions | |
Blue | -0.034 |
back | -0.030 |
over | -0.030 |
barking | -0.029 |
from | -0.029 |
Token | smile |
Token ID | 8212 |
Feature activation | +1.25e-02 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.011 |
back | +0.010 |
Spot | +0.010 |
ety | +0.010 |
Neg logit contributions | |
rd | -0.013 |
iden | -0.013 |
inations | -0.013 |
very | -0.013 |
iest | -0.013 |
Token | . |
Token ID | 13 |
Feature activation | -9.30e-03 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.013 |
back | +0.013 |
Spot | +0.013 |
Sunny | +0.013 |
Neg logit contributions | |
rd | -0.029 |
iest | -0.028 |
iden | -0.028 |
inations | -0.028 |
very | -0.028 |
Token |
|
Token ID | 198 |
Feature activation | +4.41e-03 |
Pos logit contributions | |
inations | +0.023 |
rd | +0.023 |
nexpected | +0.022 |
iest | +0.022 |
arest | +0.022 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
from | -0.009 |
Spot | -0.009 |
back | -0.009 |
Token |
|
Token ID | 198 |
Feature activation | +9.10e-03 |
Pos logit contributions | |
Blue | +0.003 |
over | +0.003 |
from | +0.002 |
Spot | +0.002 |
also | +0.002 |
Neg logit contributions | |
rd | -0.013 |
inations | -0.013 |
nexpected | -0.013 |
iest | -0.013 |
ocol | -0.012 |
Token | forget |
Token ID | 6044 |
Feature activation | +1.99e-02 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.010 |
Spot | +0.010 |
Sunny | +0.009 |
barking | +0.009 |
Neg logit contributions | |
inations | -0.018 |
nexpected | -0.017 |
rd | -0.017 |
arest | -0.017 |
iest | -0.017 |
Token | about |
Token ID | 546 |
Feature activation | -1.18e-02 |
Pos logit contributions | |
Blue | +0.023 |
ety | +0.020 |
over | +0.020 |
back | +0.019 |
barking | +0.019 |
Neg logit contributions | |
inations | -0.046 |
rd | -0.046 |
nexpected | -0.045 |
iden | -0.045 |
very | -0.044 |
Token | it |
Token ID | 340 |
Feature activation | -2.10e-03 |
Pos logit contributions | |
inations | +0.028 |
nexpected | +0.028 |
rd | +0.028 |
iden | +0.028 |
very | +0.027 |
Neg logit contributions | |
Blue | -0.013 |
over | -0.011 |
back | -0.011 |
ety | -0.011 |
Sunny | -0.010 |
Token | , |
Token ID | 11 |
Feature activation | -1.08e-02 |
Pos logit contributions | |
inations | +0.004 |
rd | +0.004 |
iest | +0.004 |
nexpected | +0.004 |
iden | +0.004 |
Neg logit contributions | |
Blue | -0.003 |
ety | -0.003 |
Sunny | -0.003 |
over | -0.003 |
Spot | -0.003 |
Token | but |
Token ID | 475 |
Feature activation | -2.41e-02 |
Pos logit contributions | |
inations | +0.024 |
iden | +0.024 |
very | +0.023 |
nexpected | +0.023 |
rd | +0.023 |
Neg logit contributions | |
Blue | -0.012 |
ety | -0.011 |
isk | -0.010 |
Sunny | -0.010 |
over | -0.010 |
Token | he |
Token ID | 339 |
Feature activation | -2.73e-02 |
Pos logit contributions | |
rd | +0.049 |
inations | +0.049 |
iden | +0.048 |
nexpected | +0.047 |
very | +0.047 |
Neg logit contributions | |
Blue | -0.037 |
over | -0.034 |
back | -0.033 |
from | -0.032 |
Spot | -0.032 |
Token | just |
Token ID | 655 |
Feature activation | -1.81e-02 |
Pos logit contributions | |
rd | +0.051 |
inations | +0.051 |
iden | +0.050 |
arest | +0.049 |
nexpected | +0.049 |
Neg logit contributions | |
Blue | -0.046 |
over | -0.043 |
back | -0.041 |
from | -0.041 |
Spot | -0.040 |
Token | felt |
Token ID | 2936 |
Feature activation | -6.20e-03 |
Pos logit contributions | |
rd | +0.032 |
inations | +0.031 |
iden | +0.031 |
arest | +0.030 |
very | +0.030 |
Neg logit contributions | |
Blue | -0.033 |
over | -0.031 |
back | -0.030 |
from | -0.029 |
Spot | -0.029 |
Token | worse |
Token ID | 4785 |
Feature activation | +5.01e-03 |
Pos logit contributions | |
inations | +0.014 |
rd | +0.014 |
nexpected | +0.014 |
iden | +0.014 |
iest | +0.014 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
ety | -0.006 |
Spot | -0.006 |
barking | -0.006 |
Token | and |
Token ID | 290 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
Blue | +0.006 |
over | +0.005 |
back | +0.005 |
Spot | +0.005 |
from | +0.005 |
Neg logit contributions | |
rd | -0.012 |
inations | -0.012 |
iden | -0.012 |
nexpected | -0.012 |
very | -0.012 |
Token | worse |
Token ID | 4785 |
Feature activation | -7.67e-04 |
Pos logit contributions | |
inations | +0.033 |
nexpected | +0.033 |
rd | +0.033 |
iden | +0.033 |
iest | +0.032 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.015 |
Spot | -0.015 |
Sunny | -0.015 |
barking | -0.014 |
Token | would |
Token ID | 561 |
Feature activation | +1.99e-02 |
Pos logit contributions | |
rd | +0.004 |
iest | +0.004 |
inations | +0.004 |
iden | +0.003 |
nexpected | +0.003 |
Neg logit contributions | |
Blue | -0.004 |
over | -0.003 |
Spot | -0.003 |
barking | -0.003 |
Sunny | -0.003 |
Token | stop |
Token ID | 2245 |
Feature activation | +2.26e-02 |
Pos logit contributions | |
Blue | +0.026 |
ety | +0.025 |
over | +0.024 |
Spot | +0.023 |
arted | +0.023 |
Neg logit contributions | |
inations | -0.040 |
arest | -0.038 |
iest | -0.038 |
mmm | -0.036 |
inkle | -0.036 |
Token | hurting |
Token ID | 21530 |
Feature activation | -2.55e-04 |
Pos logit contributions | |
Blue | +0.025 |
over | +0.021 |
Spot | +0.021 |
ety | +0.021 |
isk | +0.020 |
Neg logit contributions | |
inations | -0.052 |
very | -0.051 |
rd | -0.050 |
nexpected | -0.050 |
iest | -0.049 |
Token | soon |
Token ID | 2582 |
Feature activation | +3.12e-02 |
Pos logit contributions | |
rd | +0.001 |
iest | +0.001 |
nexpected | +0.001 |
very | +0.001 |
ises | +0.001 |
Neg logit contributions | |
Blue | -0.000 |
over | -0.000 |
blown | -0.000 |
barking | -0.000 |
isk | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | -1.37e-03 |
Pos logit contributions | |
Blue | +0.046 |
over | +0.041 |
ety | +0.041 |
isk | +0.040 |
Spot | +0.039 |
Neg logit contributions | |
rd | -0.062 |
iest | -0.059 |
inations | -0.059 |
very | -0.057 |
arest | -0.056 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.65e-02 |
Pos logit contributions | |
rd | +0.003 |
iden | +0.003 |
inations | +0.003 |
nexpected | +0.003 |
very | +0.003 |
Neg logit contributions | |
Blue | -0.002 |
over | -0.002 |
back | -0.002 |
from | -0.001 |
barking | -0.001 |
Token | Anna |
Token ID | 31160 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
rd | +0.042 |
iden | +0.041 |
inations | +0.040 |
iest | +0.040 |
very | +0.040 |
Neg logit contributions | |
Blue | -0.050 |
over | -0.047 |
back | -0.046 |
from | -0.045 |
ety | -0.044 |
Token | and |
Token ID | 290 |
Feature activation | +3.84e-02 |
Pos logit contributions | |
rd | +0.026 |
iden | +0.026 |
inations | +0.025 |
nexpected | +0.025 |
very | +0.025 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
back | -0.010 |
barking | -0.010 |
from | -0.010 |
Token | Ben |
Token ID | 3932 |
Feature activation | -2.32e-02 |
Pos logit contributions | |
Blue | +0.038 |
over | +0.038 |
faster | +0.037 |
barking | +0.036 |
isk | +0.036 |
Neg logit contributions | |
iden | -0.086 |
nexpected | -0.085 |
rd | -0.084 |
very | -0.084 |
arest | -0.082 |
Token | are |
Token ID | 389 |
Feature activation | +1.19e-02 |
Pos logit contributions | |
rd | +0.044 |
iden | +0.044 |
nexpected | +0.043 |
inations | +0.043 |
very | +0.042 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.032 |
barking | -0.031 |
back | -0.031 |
from | -0.030 |
Token | twins |
Token ID | 20345 |
Feature activation | +3.05e-03 |
Pos logit contributions | |
Blue | +0.017 |
over | +0.014 |
Spot | +0.013 |
ety | +0.013 |
from | +0.013 |
Neg logit contributions | |
inations | -0.026 |
rd | -0.026 |
iden | -0.025 |
very | -0.025 |
nexpected | -0.024 |
Token | now |
Token ID | 783 |
Feature activation | -4.45e-03 |
Pos logit contributions | |
Blue | +0.001 |
over | +0.001 |
back | +0.001 |
Spot | +0.001 |
barking | +0.001 |
Neg logit contributions | |
inations | -0.001 |
rd | -0.001 |
iden | -0.001 |
nexpected | -0.001 |
iest | -0.001 |
Token | he |
Token ID | 339 |
Feature activation | -7.57e-03 |
Pos logit contributions | |
inations | +0.009 |
rd | +0.009 |
arest | +0.009 |
nexpected | +0.009 |
iden | +0.009 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
back | -0.006 |
from | -0.006 |
Spot | -0.006 |
Token | could |
Token ID | 714 |
Feature activation | -5.87e-03 |
Pos logit contributions | |
rd | +0.016 |
inations | +0.016 |
iden | +0.016 |
nexpected | +0.015 |
arest | +0.015 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.011 |
back | -0.010 |
from | -0.010 |
also | -0.009 |
Token | wear |
Token ID | 5806 |
Feature activation | -6.58e-03 |
Pos logit contributions | |
inations | +0.012 |
rd | +0.012 |
iden | +0.012 |
nexpected | +0.012 |
arest | +0.012 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.008 |
back | -0.007 |
Spot | -0.007 |
from | -0.007 |
Token | his |
Token ID | 465 |
Feature activation | -3.40e-02 |
Pos logit contributions | |
rd | +0.018 |
iden | +0.018 |
arest | +0.018 |
inations | +0.017 |
nexpected | +0.017 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.006 |
back | -0.006 |
from | -0.005 |
Spot | -0.005 |
Token | favorite |
Token ID | 4004 |
Feature activation | -7.53e-02 |
Pos logit contributions | |
rd | +0.068 |
iden | +0.066 |
opter | +0.066 |
very | +0.066 |
nexpected | +0.066 |
Neg logit contributions | |
Blue | -0.054 |
back | -0.053 |
over | -0.049 |
from | -0.046 |
Spot | -0.046 |
Token | shirt |
Token ID | 10147 |
Feature activation | -1.38e-02 |
Pos logit contributions | |
arest | +0.170 |
mmm | +0.170 |
inations | +0.167 |
Jamie | +0.166 |
iden | +0.166 |
Neg logit contributions | |
Blue | -0.102 |
back | -0.099 |
over | -0.096 |
k | -0.088 |
from | -0.085 |
Token | again |
Token ID | 757 |
Feature activation | +1.09e-02 |
Pos logit contributions | |
rd | +0.033 |
inations | +0.033 |
iden | +0.032 |
iest | +0.032 |
very | +0.032 |
Neg logit contributions | |
Blue | -0.015 |
over | -0.013 |
Spot | -0.013 |
back | -0.012 |
barking | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
Blue | +0.011 |
over | +0.011 |
back | +0.010 |
from | +0.010 |
Spot | +0.009 |
Neg logit contributions | |
inations | -0.027 |
nexpected | -0.027 |
iden | -0.027 |
rd | -0.027 |
arest | -0.027 |
Token |
|
Token ID | 198 |
Feature activation | -3.39e-03 |
Pos logit contributions | |
iden | +0.026 |
rd | +0.026 |
very | +0.026 |
inations | +0.025 |
nexpected | +0.025 |
Neg logit contributions | |
Blue | -0.018 |
over | -0.018 |
back | -0.017 |
from | -0.017 |
blown | -0.016 |
Token |
|
Token ID | 198 |
Feature activation | +2.88e-03 |
Pos logit contributions | |
inations | +0.008 |
rd | +0.008 |
iest | +0.007 |
nexpected | +0.007 |
iden | +0.007 |
Neg logit contributions | |
Blue | -0.005 |
over | -0.004 |
Spot | -0.004 |
back | -0.004 |
from | -0.004 |
Token | cross |
Token ID | 3272 |
Feature activation | +1.05e-02 |
Pos logit contributions | |
rd | +0.057 |
inations | +0.057 |
nexpected | +0.055 |
mmm | +0.053 |
arest | +0.053 |
Neg logit contributions | |
Blue | -0.029 |
back | -0.026 |
over | -0.025 |
k | -0.025 |
barking | -0.024 |
Token | was |
Token ID | 373 |
Feature activation | +4.01e-03 |
Pos logit contributions | |
Blue | +0.015 |
over | +0.014 |
back | +0.013 |
from | +0.013 |
Spot | +0.013 |
Neg logit contributions | |
inations | -0.022 |
rd | -0.022 |
iden | -0.021 |
nexpected | -0.021 |
arest | -0.021 |
Token | on |
Token ID | 319 |
Feature activation | +5.92e-03 |
Pos logit contributions | |
Blue | +0.005 |
ety | +0.004 |
Sunny | +0.004 |
Spot | +0.004 |
isk | +0.004 |
Neg logit contributions | |
nexpected | -0.009 |
inations | -0.009 |
iden | -0.009 |
iest | -0.009 |
rd | -0.009 |
Token | her |
Token ID | 607 |
Feature activation | -8.83e-03 |
Pos logit contributions | |
ety | +0.007 |
Blue | +0.006 |
back | +0.006 |
isk | +0.005 |
over | +0.005 |
Neg logit contributions | |
nexpected | -0.013 |
very | -0.013 |
iest | -0.013 |
iden | -0.013 |
rd | -0.013 |
Token | window |
Token ID | 4324 |
Feature activation | -7.90e-03 |
Pos logit contributions | |
rd | +0.020 |
inations | +0.019 |
iden | +0.019 |
nexpected | +0.019 |
arest | +0.019 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
Spot | -0.010 |
from | -0.010 |
ety | -0.010 |
Token | sill |
Token ID | 49276 |
Feature activation | -7.97e-02 |
Pos logit contributions | |
rd | +0.017 |
inations | +0.016 |
iden | +0.016 |
nexpected | +0.016 |
iest | +0.016 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
Spot | -0.010 |
back | -0.010 |
Sunny | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
inations | +0.191 |
rd | +0.190 |
iden | +0.189 |
nexpected | +0.187 |
very | +0.186 |
Neg logit contributions | |
Blue | -0.087 |
over | -0.084 |
back | -0.079 |
from | -0.077 |
Spot | -0.072 |
Token | She |
Token ID | 1375 |
Feature activation | -9.35e-03 |
Pos logit contributions | |
rd | +0.031 |
iden | +0.030 |
inations | +0.030 |
nexpected | +0.030 |
very | +0.030 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.014 |
from | -0.014 |
Spot | -0.014 |
Token | had |
Token ID | 550 |
Feature activation | -9.12e-03 |
Pos logit contributions | |
rd | +0.021 |
inations | +0.021 |
iest | +0.020 |
arest | +0.020 |
iden | +0.020 |
Neg logit contributions | |
over | -0.013 |
Blue | -0.013 |
back | -0.012 |
from | -0.011 |
also | -0.011 |
Token | no |
Token ID | 645 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
inations | +0.021 |
rd | +0.021 |
iden | +0.021 |
nexpected | +0.021 |
very | +0.020 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
Spot | -0.009 |
back | -0.009 |
from | -0.009 |
Token | idea |
Token ID | 2126 |
Feature activation | -8.17e-03 |
Pos logit contributions | |
iden | +0.020 |
rd | +0.020 |
inations | +0.020 |
very | +0.020 |
nexpected | +0.020 |
Neg logit contributions | |
Blue | -0.016 |
over | -0.015 |
back | -0.015 |
from | -0.014 |
Spot | -0.014 |
Token | took |
Token ID | 1718 |
Feature activation | -2.41e-03 |
Pos logit contributions | |
Blue | +0.000 |
over | +0.000 |
back | +0.000 |
from | +0.000 |
Spot | +0.000 |
Neg logit contributions | |
rd | -0.001 |
inations | -0.001 |
iden | -0.001 |
nexpected | -0.001 |
very | -0.001 |
Token | his |
Token ID | 465 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
inations | +0.006 |
rd | +0.006 |
iden | +0.006 |
arest | +0.006 |
very | +0.006 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.002 |
back | -0.002 |
Spot | -0.002 |
from | -0.002 |
Token | paper |
Token ID | 3348 |
Feature activation | -3.08e-03 |
Pos logit contributions | |
rd | +0.023 |
iden | +0.022 |
nexpected | +0.022 |
very | +0.022 |
inations | +0.022 |
Neg logit contributions | |
back | -0.022 |
Blue | -0.021 |
over | -0.021 |
from | -0.020 |
near | -0.019 |
Token | and |
Token ID | 290 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
inations | +0.008 |
rd | +0.008 |
iden | +0.008 |
iest | +0.007 |
nexpected | +0.007 |
Neg logit contributions | |
Blue | -0.003 |
over | -0.003 |
Spot | -0.002 |
Sunny | -0.002 |
back | -0.002 |
Token | cr |
Token ID | 1067 |
Feature activation | -6.10e-02 |
Pos logit contributions | |
rd | +0.028 |
inations | +0.028 |
iden | +0.027 |
arest | +0.027 |
nexpected | +0.027 |
Neg logit contributions | |
Blue | -0.017 |
over | -0.016 |
back | -0.015 |
Spot | -0.015 |
from | -0.015 |
Token | ay |
Token ID | 323 |
Feature activation | -7.85e-02 |
Pos logit contributions | |
rd | +0.042 |
inations | +0.042 |
iden | +0.041 |
nexpected | +0.039 |
very | +0.037 |
Neg logit contributions | |
Blue | -0.173 |
over | -0.168 |
back | -0.163 |
from | -0.162 |
Spot | -0.160 |
Token | ons |
Token ID | 684 |
Feature activation | +2.48e-03 |
Pos logit contributions | |
iden | +0.090 |
rd | +0.090 |
lifetime | +0.088 |
arest | +0.082 |
very | +0.081 |
Neg logit contributions | |
over | -0.187 |
back | -0.179 |
Blue | -0.178 |
and | -0.173 |
from | -0.173 |
Token | . |
Token ID | 13 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
Blue | +0.004 |
ety | +0.003 |
Sunny | +0.003 |
barking | +0.003 |
Bird | +0.003 |
Neg logit contributions | |
iest | -0.005 |
rd | -0.005 |
inations | -0.004 |
arest | -0.004 |
iden | -0.004 |
Token | He |
Token ID | 679 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
Blue | +0.013 |
over | +0.012 |
back | +0.011 |
Spot | +0.011 |
from | +0.011 |
Neg logit contributions | |
inations | -0.029 |
rd | -0.029 |
iden | -0.028 |
nexpected | -0.028 |
arest | -0.028 |
Token | says |
Token ID | 1139 |
Feature activation | -1.62e-02 |
Pos logit contributions | |
inations | +0.020 |
nexpected | +0.020 |
rd | +0.020 |
iden | +0.020 |
very | +0.019 |
Neg logit contributions | |
Blue | -0.020 |
over | -0.018 |
back | -0.018 |
from | -0.017 |
Spot | -0.017 |
Token | he |
Token ID | 339 |
Feature activation | +7.69e-04 |
Pos logit contributions | |
inations | +0.029 |
rd | +0.029 |
iden | +0.028 |
nexpected | +0.028 |
iest | +0.028 |
Neg logit contributions | |
Blue | -0.028 |
over | -0.026 |
back | -0.025 |
from | -0.025 |
Spot | -0.025 |
Token | have |
Token ID | 423 |
Feature activation | +2.13e-03 |
Pos logit contributions | |
over | +0.005 |
Blue | +0.005 |
back | +0.005 |
from | +0.005 |
also | +0.005 |
Neg logit contributions | |
rd | -0.009 |
nexpected | -0.009 |
very | -0.009 |
iden | -0.009 |
inations | -0.009 |
Token | some |
Token ID | 617 |
Feature activation | -2.99e-02 |
Pos logit contributions | |
Blue | +0.002 |
over | +0.002 |
back | +0.001 |
Spot | +0.001 |
from | +0.001 |
Neg logit contributions | |
rd | -0.006 |
inations | -0.006 |
iden | -0.006 |
iest | -0.006 |
nexpected | -0.006 |
Token | paper |
Token ID | 3348 |
Feature activation | -9.03e-03 |
Pos logit contributions | |
rd | +0.064 |
iden | +0.063 |
nexpected | +0.062 |
inations | +0.060 |
opter | +0.060 |
Neg logit contributions | |
over | -0.042 |
from | -0.042 |
Blue | -0.041 |
back | -0.041 |
near | -0.037 |
Token | and |
Token ID | 290 |
Feature activation | -9.02e-03 |
Pos logit contributions | |
rd | +0.024 |
inations | +0.023 |
iden | +0.023 |
nexpected | +0.023 |
very | +0.023 |
Neg logit contributions | |
Blue | -0.008 |
over | -0.007 |
back | -0.007 |
from | -0.007 |
Spot | -0.006 |
Token | cr |
Token ID | 1067 |
Feature activation | -5.86e-02 |
Pos logit contributions | |
rd | +0.019 |
inations | +0.019 |
iden | +0.019 |
nexpected | +0.018 |
iest | +0.018 |
Neg logit contributions | |
Blue | -0.013 |
over | -0.012 |
back | -0.012 |
from | -0.011 |
Spot | -0.011 |
Token | ay |
Token ID | 323 |
Feature activation | -7.90e-02 |
Pos logit contributions | |
rd | +0.045 |
iden | +0.043 |
inations | +0.042 |
nexpected | +0.041 |
very | +0.038 |
Neg logit contributions | |
Blue | -0.162 |
over | -0.159 |
from | -0.154 |
back | -0.153 |
Spot | -0.149 |
Token | ons |
Token ID | 684 |
Feature activation | -5.68e-03 |
Pos logit contributions | |
iden | +0.084 |
rd | +0.082 |
leton | +0.075 |
lifetime | +0.075 |
arest | +0.072 |
Neg logit contributions | |
over | -0.194 |
back | -0.186 |
Blue | -0.186 |
and | -0.180 |
from | -0.179 |
Token | . |
Token ID | 13 |
Feature activation | +1.81e-03 |
Pos logit contributions | |
inations | +0.013 |
rd | +0.013 |
iest | +0.013 |
very | +0.013 |
arest | +0.013 |
Neg logit contributions | |
Blue | -0.006 |
over | -0.005 |
barking | -0.005 |
Spot | -0.005 |
back | -0.005 |
Token | You |
Token ID | 921 |
Feature activation | -2.17e-02 |
Pos logit contributions | |
Blue | +0.003 |
over | +0.003 |
back | +0.002 |
from | +0.002 |
Spot | +0.002 |
Neg logit contributions | |
rd | -0.004 |
inations | -0.004 |
iden | -0.003 |
nexpected | -0.003 |
arest | -0.003 |
Token | can |
Token ID | 460 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
inations | +0.042 |
rd | +0.042 |
iden | +0.041 |
nexpected | +0.040 |
arest | +0.040 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.032 |
back | -0.031 |
Spot | -0.030 |
from | -0.030 |
Token | draw |
Token ID | 3197 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
rd | +0.038 |
inations | +0.038 |
iden | +0.037 |
nexpected | +0.037 |
very | +0.036 |
Neg logit contributions | |
Blue | -0.028 |
over | -0.026 |
back | -0.024 |
from | -0.024 |
Spot | -0.024 |
Token | Tim |
Token ID | 5045 |
Feature activation | -7.64e-03 |
Pos logit contributions | |
Blue | +0.044 |
back | +0.044 |
blown | +0.043 |
from | +0.042 |
over | +0.042 |
Neg logit contributions | |
rd | -0.088 |
iden | -0.086 |
very | -0.084 |
nexpected | -0.083 |
inations | -0.083 |
Token | my |
Token ID | 1820 |
Feature activation | -2.33e-02 |
Pos logit contributions | |
rd | +0.020 |
inations | +0.020 |
iden | +0.020 |
nexpected | +0.020 |
very | +0.019 |
Neg logit contributions | |
Blue | -0.007 |
over | -0.006 |
back | -0.005 |
from | -0.005 |
Spot | -0.005 |
Token | loved |
Token ID | 6151 |
Feature activation | +2.85e-02 |
Pos logit contributions | |
inations | +0.049 |
rd | +0.047 |
arest | +0.047 |
nexpected | +0.047 |
iden | +0.047 |
Neg logit contributions | |
Blue | -0.035 |
over | -0.033 |
back | -0.031 |
Spot | -0.031 |
from | -0.030 |
Token | wearing |
Token ID | 5762 |
Feature activation | +3.44e-03 |
Pos logit contributions | |
Blue | +0.030 |
over | +0.027 |
ter | +0.025 |
ety | +0.025 |
isk | +0.024 |
Neg logit contributions | |
iden | -0.067 |
inations | -0.066 |
very | -0.066 |
nexpected | -0.065 |
rd | -0.065 |
Token | his |
Token ID | 465 |
Feature activation | -3.31e-02 |
Pos logit contributions | |
Blue | +0.004 |
over | +0.004 |
back | +0.003 |
Spot | +0.003 |
from | +0.003 |
Neg logit contributions | |
rd | -0.008 |
inations | -0.008 |
arest | -0.008 |
nexpected | -0.008 |
iden | -0.008 |
Token | favorite |
Token ID | 4004 |
Feature activation | -7.92e-02 |
Pos logit contributions | |
inations | +0.077 |
rd | +0.076 |
nexpected | +0.076 |
iden | +0.075 |
arest | +0.074 |
Neg logit contributions | |
Blue | -0.042 |
back | -0.038 |
over | -0.037 |
Spot | -0.036 |
from | -0.034 |
Token | red |
Token ID | 2266 |
Feature activation | -2.76e-02 |
Pos logit contributions | |
inations | +0.192 |
mmm | +0.189 |
arest | +0.187 |
nexpected | +0.185 |
opter | +0.183 |
Neg logit contributions | |
Blue | -0.094 |
back | -0.094 |
over | -0.087 |
Spot | -0.080 |
k | -0.079 |
Token | cap |
Token ID | 1451 |
Feature activation | -7.82e-03 |
Pos logit contributions | |
inations | +0.052 |
rd | +0.051 |
iden | +0.050 |
nexpected | +0.050 |
arest | +0.050 |
Neg logit contributions | |
Blue | -0.046 |
over | -0.043 |
back | -0.042 |
from | -0.040 |
Spot | -0.040 |
Token | everywhere |
Token ID | 8347 |
Feature activation | -8.13e-03 |
Pos logit contributions | |
rd | +0.017 |
iest | +0.017 |
inations | +0.016 |
very | +0.016 |
iden | +0.016 |
Neg logit contributions | |
Blue | -0.010 |
over | -0.009 |
barking | -0.009 |
Sunny | -0.009 |
pering | -0.009 |
Token | he |
Token ID | 339 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
rd | +0.018 |
inations | +0.018 |
iden | +0.018 |
nexpected | +0.018 |
arest | +0.018 |
Neg logit contributions | |
Blue | -0.011 |
over | -0.010 |
back | -0.009 |
from | -0.009 |
Spot | -0.009 |
Token | went |
Token ID | 1816 |
Feature activation | +5.07e-02 |
Pos logit contributions | |
inations | +0.033 |
rd | +0.033 |
iden | +0.032 |
nexpected | +0.032 |
arest | +0.032 |
Neg logit contributions | |
Blue | -0.012 |
over | -0.011 |
back | -0.011 |
from | -0.010 |
Spot | -0.010 |