Token | sad |
Token ID | 6507 |
Feature activation | +2.46e-02 |
Pos logit contributions | |
ends | +0.008 |
counts | +0.008 |
Or | +0.007 |
One | +0.007 |
cannot | +0.007 |
Neg logit contributions | |
tune | -0.010 |
sensation | -0.010 |
dolphin | -0.009 |
shark | -0.009 |
Dolphin | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | +1.30e-02 |
Pos logit contributions | |
tune | +0.063 |
sensation | +0.061 |
dolphin | +0.059 |
shark | +0.057 |
Dolphin | +0.057 |
Neg logit contributions | |
ends | -0.023 |
counts | -0.022 |
One | -0.021 |
Or | -0.021 |
Maybe | -0.016 |
Token | She |
Token ID | 1375 |
Feature activation | -8.47e-03 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.023 |
face | +0.021 |
Sister | +0.020 |
expression | +0.019 |
Neg logit contributions | |
ends | -0.020 |
counts | -0.020 |
One | -0.018 |
ours | -0.018 |
Or | -0.017 |
Token | threw |
Token ID | 9617 |
Feature activation | -6.18e-03 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
One | +0.012 |
Or | +0.012 |
has | +0.010 |
Neg logit contributions | |
sensation | -0.017 |
tune | -0.017 |
dolphin | -0.015 |
shark | -0.015 |
Dolphin | -0.015 |
Token | a |
Token ID | 257 |
Feature activation | +3.75e-03 |
Pos logit contributions | |
counts | +0.011 |
ends | +0.011 |
One | +0.010 |
Or | +0.010 |
Ready | +0.009 |
Neg logit contributions | |
tune | -0.011 |
sensation | -0.011 |
dolphin | -0.010 |
shark | -0.009 |
Dolphin | -0.009 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.98e-01 |
Pos logit contributions | |
sensation | +0.005 |
tune | +0.005 |
urg | +0.004 |
Dolphin | +0.004 |
dolphin | +0.004 |
Neg logit contributions | |
ends | -0.008 |
One | -0.008 |
counts | -0.008 |
Or | -0.007 |
Ready | -0.007 |
Token | rum |
Token ID | 6582 |
Feature activation | +2.16e-02 |
Pos logit contributions | |
tune | +0.320 |
sensation | +0.315 |
dolphin | +0.287 |
shark | +0.273 |
chill | +0.257 |
Neg logit contributions | |
counts | -0.363 |
ends | -0.361 |
One | -0.344 |
Or | -0.332 |
Ready | -0.299 |
Token | and |
Token ID | 290 |
Feature activation | +3.24e-03 |
Pos logit contributions | |
tune | +0.045 |
sensation | +0.045 |
dolphin | +0.041 |
shark | +0.040 |
chill | +0.038 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.029 |
One | -0.028 |
Or | -0.027 |
Ready | -0.024 |
Token | cried |
Token ID | 16896 |
Feature activation | +6.15e-04 |
Pos logit contributions | |
tune | +0.007 |
sensation | +0.006 |
dolphin | +0.006 |
shark | +0.006 |
Dolphin | +0.006 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
Or | -0.004 |
One | -0.004 |
cannot | -0.004 |
Token | loudly |
Token ID | 23112 |
Feature activation | +1.63e-03 |
Pos logit contributions | |
tune | +0.002 |
sensation | +0.002 |
dolphin | +0.002 |
shark | +0.002 |
Dolphin | +0.001 |
Neg logit contributions | |
ends | -0.001 |
Or | -0.001 |
One | -0.001 |
counts | -0.000 |
house | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | +8.73e-03 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
dolphin | +0.004 |
shark | +0.004 |
Dolphin | +0.004 |
Neg logit contributions | |
counts | -0.002 |
ends | -0.002 |
One | -0.001 |
Or | -0.001 |
Ready | -0.001 |
Token | kept |
Token ID | 4030 |
Feature activation | +2.12e-02 |
Pos logit contributions | |
counts | +0.013 |
ends | +0.012 |
One | +0.012 |
Or | +0.011 |
� | +0.010 |
Neg logit contributions | |
sensation | -0.016 |
tune | -0.016 |
dolphin | -0.014 |
shark | -0.014 |
Dolphin | -0.014 |
Token | crying |
Token ID | 13774 |
Feature activation | +4.84e-02 |
Pos logit contributions | |
sensation | +0.045 |
tune | +0.045 |
dolphin | +0.042 |
shark | +0.041 |
Dolphin | +0.040 |
Neg logit contributions | |
ends | -0.026 |
counts | -0.025 |
One | -0.025 |
Or | -0.022 |
here | -0.021 |
Token | and |
Token ID | 290 |
Feature activation | -8.90e-03 |
Pos logit contributions | |
tune | +0.120 |
sensation | +0.117 |
dolphin | +0.114 |
Dolphin | +0.111 |
shark | +0.110 |
Neg logit contributions | |
ends | -0.048 |
counts | -0.046 |
Or | -0.045 |
One | -0.044 |
Maybe | -0.035 |
Token | throwing |
Token ID | 9644 |
Feature activation | +9.41e-03 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
One | +0.012 |
Or | +0.011 |
cannot | +0.010 |
Neg logit contributions | |
tune | -0.018 |
dolphin | -0.017 |
sensation | -0.017 |
shark | -0.017 |
Dolphin | -0.016 |
Token | a |
Token ID | 257 |
Feature activation | -3.50e-03 |
Pos logit contributions | |
tune | +0.020 |
sensation | +0.019 |
dolphin | +0.018 |
shark | +0.017 |
Dolphin | +0.017 |
Neg logit contributions | |
counts | -0.013 |
ends | -0.013 |
One | -0.013 |
Or | -0.012 |
Ready | -0.010 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.97e-01 |
Pos logit contributions | |
One | +0.007 |
ends | +0.007 |
counts | +0.007 |
Or | +0.007 |
house | +0.007 |
Neg logit contributions | |
sensation | -0.004 |
tune | -0.004 |
Dolphin | -0.004 |
dolphin | -0.004 |
streaming | -0.003 |
Token | rum |
Token ID | 6582 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
tune | +0.312 |
sensation | +0.306 |
dolphin | +0.277 |
shark | +0.265 |
chill | +0.248 |
Neg logit contributions | |
counts | -0.368 |
ends | -0.367 |
One | -0.353 |
Or | -0.340 |
Ready | -0.311 |
Token | . |
Token ID | 13 |
Feature activation | +1.16e-02 |
Pos logit contributions | |
tune | +0.031 |
sensation | +0.030 |
dolphin | +0.028 |
shark | +0.027 |
chill | +0.026 |
Neg logit contributions | |
ends | -0.016 |
counts | -0.016 |
One | -0.015 |
Or | -0.014 |
Ready | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | +1.03e-02 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.023 |
dolphin | +0.020 |
streaming | +0.019 |
face | +0.019 |
Neg logit contributions | |
counts | -0.017 |
ends | -0.017 |
One | -0.015 |
Or | -0.015 |
ours | -0.014 |
Token |
|
Token ID | 198 |
Feature activation | +2.06e-02 |
Pos logit contributions | |
tune | +0.019 |
sensation | +0.018 |
dolphin | +0.016 |
face | +0.016 |
streamed | +0.016 |
Neg logit contributions | |
counts | -0.016 |
ends | -0.015 |
ours | -0.014 |
Or | -0.013 |
Ready | -0.013 |
Token | Finally |
Token ID | 11158 |
Feature activation | -2.70e-02 |
Pos logit contributions | |
tune | +0.038 |
sensation | +0.035 |
face | +0.031 |
lull | +0.031 |
streamed | +0.031 |
Neg logit contributions | |
counts | -0.036 |
ends | -0.032 |
ours | -0.030 |
Or | -0.029 |
Ready | -0.028 |
Token | to |
Token ID | 284 |
Feature activation | +1.41e-02 |
Pos logit contributions | |
counts | +0.003 |
ends | +0.003 |
One | +0.003 |
Or | +0.003 |
Ready | +0.002 |
Neg logit contributions | |
sensation | -0.008 |
tune | -0.008 |
dolphin | -0.008 |
shark | -0.008 |
Dolphin | -0.007 |
Token | cry |
Token ID | 3960 |
Feature activation | +3.72e-02 |
Pos logit contributions | |
sensation | +0.027 |
tune | +0.027 |
dolphin | +0.026 |
shark | +0.025 |
Dolphin | +0.024 |
Neg logit contributions | |
counts | -0.021 |
One | -0.020 |
Or | -0.019 |
ends | -0.019 |
� | -0.017 |
Token | and |
Token ID | 290 |
Feature activation | +5.77e-03 |
Pos logit contributions | |
sensation | +0.096 |
tune | +0.092 |
dolphin | +0.091 |
shark | +0.090 |
Dolphin | +0.090 |
Neg logit contributions | |
Or | -0.032 |
ends | -0.032 |
One | -0.029 |
counts | -0.027 |
Maybe | -0.024 |
Token | threw |
Token ID | 9617 |
Feature activation | -1.54e-02 |
Pos logit contributions | |
sensation | +0.012 |
tune | +0.012 |
dolphin | +0.011 |
shark | +0.010 |
Dolphin | +0.010 |
Neg logit contributions | |
counts | -0.008 |
ends | -0.008 |
Or | -0.008 |
One | -0.008 |
Ready | -0.007 |
Token | a |
Token ID | 257 |
Feature activation | -5.15e-04 |
Pos logit contributions | |
counts | +0.024 |
ends | +0.024 |
One | +0.023 |
Or | +0.022 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.029 |
sensation | -0.029 |
dolphin | -0.026 |
shark | -0.025 |
streaming | -0.024 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.96e-01 |
Pos logit contributions | |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
counts | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
sensation | -0.001 |
tune | -0.001 |
urg | -0.001 |
Dolphin | -0.001 |
lee | -0.001 |
Token | rum |
Token ID | 6582 |
Feature activation | +2.29e-02 |
Pos logit contributions | |
tune | +0.316 |
sensation | +0.311 |
dolphin | +0.283 |
shark | +0.270 |
Dolphin | +0.254 |
Neg logit contributions | |
counts | -0.358 |
ends | -0.356 |
One | -0.340 |
Or | -0.329 |
Ready | -0.295 |
Token | . |
Token ID | 13 |
Feature activation | +2.41e-02 |
Pos logit contributions | |
tune | +0.053 |
sensation | +0.053 |
dolphin | +0.049 |
shark | +0.047 |
chill | +0.046 |
Neg logit contributions | |
ends | -0.026 |
counts | -0.026 |
One | -0.024 |
Or | -0.023 |
Ready | -0.020 |
Token | She |
Token ID | 1375 |
Feature activation | -3.69e-03 |
Pos logit contributions | |
tune | +0.052 |
sensation | +0.052 |
dolphin | +0.048 |
shark | +0.047 |
chill | +0.045 |
Neg logit contributions | |
counts | -0.031 |
ends | -0.031 |
One | -0.029 |
Or | -0.028 |
Ready | -0.025 |
Token | didn |
Token ID | 1422 |
Feature activation | +2.97e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.005 |
One | +0.005 |
Or | +0.005 |
cannot | +0.004 |
Neg logit contributions | |
sensation | -0.008 |
tune | -0.007 |
dolphin | -0.007 |
Dolphin | -0.006 |
andel | -0.006 |
Token | 't |
Token ID | 470 |
Feature activation | -3.44e-02 |
Pos logit contributions | |
sensation | +0.059 |
tune | +0.059 |
dolphin | +0.055 |
shark | +0.054 |
streaming | +0.054 |
Neg logit contributions | |
counts | -0.039 |
ends | -0.037 |
One | -0.032 |
Or | -0.031 |
ours | -0.031 |
Token | listen |
Token ID | 6004 |
Feature activation | +1.12e-02 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.023 |
dolphin | +0.021 |
shark | +0.020 |
streaming | +0.020 |
Neg logit contributions | |
counts | -0.012 |
ends | -0.012 |
One | -0.012 |
Or | -0.011 |
Ready | -0.010 |
Token | . |
Token ID | 13 |
Feature activation | +2.62e-03 |
Pos logit contributions | |
tune | +0.022 |
sensation | +0.021 |
dolphin | +0.019 |
shark | +0.018 |
face | +0.018 |
Neg logit contributions | |
counts | -0.017 |
ends | -0.017 |
One | -0.016 |
Or | -0.015 |
Ready | -0.015 |
Token | She |
Token ID | 1375 |
Feature activation | -1.96e-02 |
Pos logit contributions | |
sensation | +0.005 |
tune | +0.005 |
face | +0.005 |
dolphin | +0.005 |
Sister | +0.005 |
Neg logit contributions | |
ends | -0.003 |
counts | -0.003 |
One | -0.003 |
ours | -0.003 |
belongs | -0.003 |
Token | threw |
Token ID | 9617 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
counts | +0.030 |
ends | +0.030 |
One | +0.028 |
Or | +0.028 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.039 |
sensation | -0.038 |
dolphin | -0.034 |
shark | -0.033 |
streaming | -0.032 |
Token | a |
Token ID | 257 |
Feature activation | +8.87e-04 |
Pos logit contributions | |
counts | +0.021 |
ends | +0.020 |
One | +0.020 |
Or | +0.019 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.023 |
sensation | -0.022 |
dolphin | -0.020 |
shark | -0.020 |
chill | -0.019 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.92e-01 |
Pos logit contributions | |
sensation | +0.001 |
tune | +0.001 |
streaming | +0.001 |
urg | +0.001 |
lee | +0.001 |
Neg logit contributions | |
ends | -0.002 |
One | -0.002 |
counts | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | rum |
Token ID | 6582 |
Feature activation | +1.13e-02 |
Pos logit contributions | |
tune | +0.310 |
sensation | +0.306 |
dolphin | +0.277 |
shark | +0.263 |
chill | +0.250 |
Neg logit contributions | |
counts | -0.352 |
ends | -0.351 |
One | -0.336 |
Or | -0.325 |
Ready | -0.292 |
Token | in |
Token ID | 287 |
Feature activation | +2.17e-02 |
Pos logit contributions | |
tune | +0.022 |
sensation | +0.022 |
dolphin | +0.019 |
shark | +0.019 |
streaming | +0.019 |
Neg logit contributions | |
counts | -0.017 |
ends | -0.017 |
One | -0.016 |
Or | -0.015 |
Ready | -0.015 |
Token | the |
Token ID | 262 |
Feature activation | +9.88e-03 |
Pos logit contributions | |
tune | +0.039 |
sensation | +0.038 |
dolphin | +0.035 |
shark | +0.034 |
Dolphin | +0.032 |
Neg logit contributions | |
counts | -0.037 |
One | -0.035 |
ends | -0.035 |
Or | -0.033 |
Ready | -0.031 |
Token | store |
Token ID | 3650 |
Feature activation | -6.35e-02 |
Pos logit contributions | |
tune | +0.017 |
sensation | +0.017 |
ous | +0.016 |
streaming | +0.016 |
urg | +0.015 |
Neg logit contributions | |
One | -0.015 |
counts | -0.015 |
has | -0.015 |
Ready | -0.014 |
ends | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -3.26e-03 |
Pos logit contributions | |
ends | +0.088 |
counts | +0.088 |
One | +0.085 |
Or | +0.079 |
Ready | +0.074 |
Neg logit contributions | |
tune | -0.132 |
sensation | -0.129 |
dolphin | -0.117 |
shark | -0.113 |
face | -0.111 |
Token | She |
Token ID | 1375 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
tune | +0.026 |
sensation | +0.025 |
face | +0.024 |
Sister | +0.023 |
dolphin | +0.022 |
Neg logit contributions | |
ends | -0.016 |
counts | -0.016 |
One | -0.015 |
ours | -0.014 |
Or | -0.013 |
Token | cried |
Token ID | 16896 |
Feature activation | +8.97e-03 |
Pos logit contributions | |
counts | +0.015 |
ends | +0.015 |
One | +0.015 |
Or | +0.014 |
Ready | +0.012 |
Neg logit contributions | |
sensation | -0.021 |
tune | -0.020 |
dolphin | -0.018 |
shark | -0.018 |
streaming | -0.017 |
Token | and |
Token ID | 290 |
Feature activation | +7.99e-03 |
Pos logit contributions | |
sensation | +0.024 |
dolphin | +0.024 |
shark | +0.023 |
tune | +0.023 |
Dolphin | +0.023 |
Neg logit contributions | |
ends | -0.007 |
Or | -0.007 |
One | -0.006 |
counts | -0.006 |
or | -0.004 |
Token | threw |
Token ID | 9617 |
Feature activation | -2.01e-02 |
Pos logit contributions | |
tune | +0.019 |
andel | +0.019 |
sensation | +0.019 |
shark | +0.019 |
Dolphin | +0.019 |
Neg logit contributions | |
counts | -0.007 |
ends | -0.007 |
Or | -0.006 |
One | -0.006 |
she | -0.005 |
Token | a |
Token ID | 257 |
Feature activation | -5.01e-03 |
Pos logit contributions | |
counts | +0.030 |
ends | +0.029 |
One | +0.028 |
Or | +0.027 |
Ready | +0.025 |
Neg logit contributions | |
tune | -0.040 |
sensation | -0.039 |
dolphin | -0.036 |
shark | -0.035 |
Dolphin | -0.034 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
ends | +0.010 |
One | +0.010 |
Or | +0.010 |
counts | +0.009 |
Ready | +0.009 |
Neg logit contributions | |
sensation | -0.007 |
tune | -0.006 |
urg | -0.006 |
Ned | -0.006 |
Dolphin | -0.006 |
Token | rum |
Token ID | 6582 |
Feature activation | +7.50e-03 |
Pos logit contributions | |
tune | +0.305 |
sensation | +0.302 |
dolphin | +0.274 |
shark | +0.261 |
chill | +0.247 |
Neg logit contributions | |
counts | -0.345 |
ends | -0.343 |
One | -0.328 |
Or | -0.317 |
Ready | -0.284 |
Token | . |
Token ID | 13 |
Feature activation | +4.56e-03 |
Pos logit contributions | |
tune | +0.018 |
sensation | +0.017 |
dolphin | +0.016 |
shark | +0.016 |
chill | +0.015 |
Neg logit contributions | |
ends | -0.008 |
counts | -0.008 |
One | -0.008 |
Or | -0.007 |
Ready | -0.007 |
Token | Her |
Token ID | 2332 |
Feature activation | -1.71e-04 |
Pos logit contributions | |
tune | +0.009 |
sensation | +0.009 |
dolphin | +0.008 |
shark | +0.008 |
chill | +0.007 |
Neg logit contributions | |
ends | -0.007 |
counts | -0.007 |
One | -0.007 |
Or | -0.006 |
Ready | -0.006 |
Token | mom |
Token ID | 1995 |
Feature activation | -9.99e-03 |
Pos logit contributions | |
counts | +0.000 |
ends | +0.000 |
One | +0.000 |
Or | +0.000 |
house | +0.000 |
Neg logit contributions | |
tune | -0.000 |
sensation | -0.000 |
dolphin | -0.000 |
Dolphin | -0.000 |
streaming | -0.000 |
Token | tried |
Token ID | 3088 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
ends | +0.014 |
counts | +0.014 |
One | +0.013 |
Or | +0.013 |
house | +0.012 |
Neg logit contributions | |
sensation | -0.021 |
tune | -0.020 |
dolphin | -0.019 |
shark | -0.018 |
Dolphin | -0.018 |
Token | and |
Token ID | 290 |
Feature activation | +7.77e-03 |
Pos logit contributions | |
tune | +0.011 |
sensation | +0.010 |
dolphin | +0.010 |
shark | +0.009 |
chill | +0.009 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
One | -0.005 |
Or | -0.005 |
Ready | -0.004 |
Token | started |
Token ID | 2067 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
tune | +0.016 |
sensation | +0.016 |
dolphin | +0.014 |
Dolphin | +0.014 |
shark | +0.014 |
Neg logit contributions | |
ends | -0.011 |
counts | -0.011 |
Or | -0.010 |
One | -0.009 |
cannot | -0.009 |
Token | to |
Token ID | 284 |
Feature activation | +1.42e-03 |
Pos logit contributions | |
ends | +0.025 |
counts | +0.024 |
One | +0.024 |
Or | +0.021 |
cannot | +0.017 |
Neg logit contributions | |
tune | -0.064 |
sensation | -0.063 |
dolphin | -0.060 |
shark | -0.058 |
Dolphin | -0.058 |
Token | throw |
Token ID | 3714 |
Feature activation | -1.57e-02 |
Pos logit contributions | |
tune | +0.003 |
sensation | +0.003 |
andel | +0.003 |
dolphin | +0.003 |
Dolphin | +0.003 |
Neg logit contributions | |
counts | -0.002 |
One | -0.002 |
Or | -0.002 |
ends | -0.002 |
� | -0.001 |
Token | a |
Token ID | 257 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
counts | +0.025 |
ends | +0.024 |
One | +0.024 |
Or | +0.022 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.030 |
sensation | -0.029 |
dolphin | -0.027 |
shark | -0.026 |
Dolphin | -0.025 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
ends | +0.019 |
One | +0.018 |
counts | +0.018 |
Or | +0.017 |
ours | +0.016 |
Neg logit contributions | |
sensation | -0.016 |
tune | -0.016 |
streaming | -0.015 |
streamed | -0.014 |
tunes | -0.013 |
Token | rum |
Token ID | 6582 |
Feature activation | +7.91e-03 |
Pos logit contributions | |
tune | +0.298 |
sensation | +0.293 |
dolphin | +0.263 |
shark | +0.251 |
chill | +0.237 |
Neg logit contributions | |
counts | -0.354 |
ends | -0.353 |
One | -0.339 |
Or | -0.328 |
Ready | -0.301 |
Token | . |
Token ID | 13 |
Feature activation | +1.32e-02 |
Pos logit contributions | |
tune | +0.018 |
sensation | +0.018 |
dolphin | +0.016 |
shark | +0.016 |
streaming | +0.015 |
Neg logit contributions | |
counts | -0.009 |
ends | -0.009 |
One | -0.009 |
Or | -0.008 |
Ready | -0.008 |
Token | Suddenly |
Token ID | 24975 |
Feature activation | -2.06e-02 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.027 |
dolphin | +0.025 |
shark | +0.024 |
streaming | +0.023 |
Neg logit contributions | |
ends | -0.018 |
counts | -0.018 |
One | -0.017 |
Or | -0.017 |
Ready | -0.015 |
Token | , |
Token ID | 11 |
Feature activation | +3.07e-02 |
Pos logit contributions | |
ends | +0.032 |
counts | +0.031 |
One | +0.030 |
Or | +0.029 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.039 |
sensation | -0.038 |
dolphin | -0.035 |
shark | -0.034 |
streaming | -0.033 |
Token | she |
Token ID | 673 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
tune | +0.080 |
sensation | +0.077 |
dolphin | +0.072 |
shark | +0.070 |
Dolphin | +0.068 |
Neg logit contributions | |
ends | -0.028 |
counts | -0.027 |
One | -0.026 |
Or | -0.024 |
Ready | -0.021 |
Token | her |
Token ID | 607 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
sensation | +0.060 |
tune | +0.059 |
dolphin | +0.057 |
shark | +0.055 |
Dolphin | +0.054 |
Neg logit contributions | |
counts | -0.037 |
ends | -0.034 |
Or | -0.032 |
One | -0.031 |
Ready | -0.028 |
Token | feet |
Token ID | 3625 |
Feature activation | +5.55e-02 |
Pos logit contributions | |
ends | +0.023 |
counts | +0.022 |
Or | +0.020 |
cannot | +0.019 |
One | +0.019 |
Neg logit contributions | |
sensation | -0.022 |
tune | -0.020 |
dolphin | -0.019 |
Dolphin | -0.018 |
streamed | -0.018 |
Token | and |
Token ID | 290 |
Feature activation | -3.37e-03 |
Pos logit contributions | |
tune | +0.124 |
sensation | +0.122 |
dolphin | +0.115 |
shark | +0.112 |
Dolphin | +0.108 |
Neg logit contributions | |
ends | -0.068 |
counts | -0.066 |
One | -0.063 |
Or | -0.060 |
Ready | -0.052 |
Token | threw |
Token ID | 9617 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
counts | +0.005 |
ends | +0.004 |
Or | +0.004 |
One | +0.004 |
cannot | +0.004 |
Neg logit contributions | |
tune | -0.007 |
sensation | -0.007 |
dolphin | -0.006 |
Dolphin | -0.006 |
andel | -0.006 |
Token | a |
Token ID | 257 |
Feature activation | -4.38e-03 |
Pos logit contributions | |
counts | +0.034 |
ends | +0.033 |
One | +0.032 |
Or | +0.030 |
Ready | +0.028 |
Neg logit contributions | |
tune | -0.043 |
sensation | -0.043 |
dolphin | -0.039 |
shark | -0.038 |
Dolphin | -0.036 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
ends | +0.011 |
One | +0.010 |
counts | +0.010 |
Or | +0.010 |
Ready | +0.010 |
Neg logit contributions | |
sensation | -0.004 |
tune | -0.004 |
streaming | -0.003 |
Dolphin | -0.003 |
dolphin | -0.003 |
Token | rum |
Token ID | 6582 |
Feature activation | +3.21e-03 |
Pos logit contributions | |
tune | +0.303 |
sensation | +0.299 |
dolphin | +0.271 |
shark | +0.258 |
potato | +0.243 |
Neg logit contributions | |
counts | -0.347 |
ends | -0.345 |
One | -0.330 |
Or | -0.319 |
Ready | -0.288 |
Token | . |
Token ID | 13 |
Feature activation | +4.85e-03 |
Pos logit contributions | |
tune | +0.007 |
sensation | +0.007 |
dolphin | +0.007 |
shark | +0.007 |
chill | +0.006 |
Neg logit contributions | |
ends | -0.004 |
counts | -0.004 |
One | -0.003 |
Or | -0.003 |
Ready | -0.003 |
Token | She |
Token ID | 1375 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
tune | +0.011 |
sensation | +0.011 |
dolphin | +0.010 |
shark | +0.010 |
chill | +0.010 |
Neg logit contributions | |
ends | -0.006 |
counts | -0.006 |
Or | -0.006 |
One | -0.005 |
Ready | -0.005 |
Token | grabbed |
Token ID | 13176 |
Feature activation | +2.43e-02 |
Pos logit contributions | |
counts | +0.017 |
ends | +0.016 |
One | +0.015 |
Or | +0.015 |
cannot | +0.014 |
Neg logit contributions | |
sensation | -0.025 |
tune | -0.025 |
Dolphin | -0.022 |
dolphin | -0.022 |
andel | -0.022 |
Token | the |
Token ID | 262 |
Feature activation | -1.52e-03 |
Pos logit contributions | |
tune | +0.052 |
sensation | +0.051 |
dolphin | +0.046 |
shark | +0.045 |
lull | +0.044 |
Neg logit contributions | |
counts | -0.033 |
ends | -0.033 |
One | -0.031 |
Or | -0.028 |
Ready | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | -7.26e-03 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.023 |
dolphin | +0.021 |
shark | +0.021 |
streaming | +0.020 |
Neg logit contributions | |
counts | -0.013 |
ends | -0.013 |
One | -0.012 |
Or | -0.011 |
Ready | -0.010 |
Token | He |
Token ID | 679 |
Feature activation | -1.22e-02 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
One | +0.011 |
Or | +0.011 |
Ready | +0.010 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.013 |
dolphin | -0.012 |
shark | -0.012 |
streaming | -0.011 |
Token | kept |
Token ID | 4030 |
Feature activation | +1.89e-02 |
Pos logit contributions | |
counts | +0.018 |
ends | +0.017 |
Or | +0.016 |
One | +0.016 |
has | +0.014 |
Neg logit contributions | |
tune | -0.026 |
sensation | -0.025 |
dolphin | -0.023 |
Dolphin | -0.022 |
shark | -0.022 |
Token | throwing |
Token ID | 9644 |
Feature activation | -1.21e-03 |
Pos logit contributions | |
tune | +0.040 |
sensation | +0.040 |
dolphin | +0.037 |
lull | +0.035 |
shark | +0.035 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.024 |
One | -0.022 |
Or | -0.020 |
ours | -0.019 |
Token | a |
Token ID | 257 |
Feature activation | -5.91e-03 |
Pos logit contributions | |
counts | +0.002 |
ends | +0.002 |
One | +0.002 |
Or | +0.002 |
Ready | +0.002 |
Neg logit contributions | |
tune | -0.002 |
sensation | -0.002 |
dolphin | -0.002 |
shark | -0.002 |
streaming | -0.002 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.88e-01 |
Pos logit contributions | |
ends | +0.009 |
counts | +0.009 |
One | +0.009 |
Or | +0.009 |
Ready | +0.008 |
Neg logit contributions | |
sensation | -0.011 |
tune | -0.010 |
streaming | -0.009 |
dolphin | -0.009 |
Dolphin | -0.009 |
Token | rum |
Token ID | 6582 |
Feature activation | +2.82e-03 |
Pos logit contributions | |
tune | +0.293 |
sensation | +0.289 |
dolphin | +0.257 |
shark | +0.245 |
streaming | +0.231 |
Neg logit contributions | |
ends | -0.356 |
counts | -0.356 |
One | -0.342 |
Or | -0.331 |
Ready | -0.306 |
Token | . |
Token ID | 13 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
sensation | +0.006 |
tune | +0.006 |
dolphin | +0.005 |
shark | +0.005 |
face | +0.005 |
Neg logit contributions | |
counts | -0.004 |
ends | -0.004 |
One | -0.004 |
Or | -0.003 |
Ready | -0.003 |
Token | His |
Token ID | 2399 |
Feature activation | +6.34e-03 |
Pos logit contributions | |
ends | +0.026 |
counts | +0.025 |
One | +0.024 |
Or | +0.023 |
Ready | +0.022 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.027 |
dolphin | -0.024 |
shark | -0.023 |
streaming | -0.022 |
Token | mom |
Token ID | 1995 |
Feature activation | -2.04e-02 |
Pos logit contributions | |
tune | +0.012 |
sensation | +0.012 |
dolphin | +0.011 |
streaming | +0.011 |
Dolphin | +0.010 |
Neg logit contributions | |
ends | -0.009 |
counts | -0.009 |
Or | -0.009 |
One | -0.008 |
cannot | -0.008 |
Token | was |
Token ID | 373 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
counts | +0.030 |
ends | +0.030 |
Or | +0.028 |
One | +0.028 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.041 |
sensation | -0.041 |
dolphin | -0.037 |
shark | -0.036 |
Dolphin | -0.035 |
Token | and |
Token ID | 290 |
Feature activation | -1.84e-02 |
Pos logit contributions | |
counts | +0.007 |
ends | +0.007 |
One | +0.007 |
Or | +0.006 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.013 |
chill | -0.012 |
Token | so |
Token ID | 523 |
Feature activation | +2.04e-04 |
Pos logit contributions | |
counts | +0.027 |
ends | +0.027 |
One | +0.026 |
Or | +0.024 |
cannot | +0.023 |
Neg logit contributions | |
tune | -0.037 |
sensation | -0.036 |
dolphin | -0.033 |
shark | -0.033 |
Dolphin | -0.031 |
Token | she |
Token ID | 673 |
Feature activation | -1.64e-02 |
Pos logit contributions | |
tune | +0.001 |
sensation | +0.001 |
dolphin | +0.000 |
shark | +0.000 |
potato | +0.000 |
Neg logit contributions | |
counts | -0.000 |
ends | -0.000 |
One | -0.000 |
Or | -0.000 |
Ready | -0.000 |
Token | threw |
Token ID | 9617 |
Feature activation | -1.77e-02 |
Pos logit contributions | |
One | +0.022 |
counts | +0.021 |
ends | +0.020 |
Or | +0.019 |
� | +0.018 |
Neg logit contributions | |
sensation | -0.036 |
tune | -0.034 |
andel | -0.030 |
ulp | -0.030 |
streaming | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -4.02e-03 |
Pos logit contributions | |
counts | +0.026 |
ends | +0.026 |
One | +0.025 |
Or | +0.023 |
Ready | +0.021 |
Neg logit contributions | |
tune | -0.036 |
sensation | -0.035 |
dolphin | -0.032 |
shark | -0.031 |
streaming | -0.030 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.86e-01 |
Pos logit contributions | |
ends | +0.005 |
One | +0.005 |
Or | +0.005 |
counts | +0.005 |
ours | +0.005 |
Neg logit contributions | |
sensation | -0.008 |
tune | -0.007 |
urg | -0.007 |
streaming | -0.007 |
lee | -0.007 |
Token | rum |
Token ID | 6582 |
Feature activation | +1.30e-02 |
Pos logit contributions | |
tune | +0.304 |
sensation | +0.302 |
dolphin | +0.274 |
shark | +0.260 |
chill | +0.247 |
Neg logit contributions | |
counts | -0.337 |
ends | -0.334 |
One | -0.318 |
Or | -0.306 |
Ready | -0.271 |
Token | when |
Token ID | 618 |
Feature activation | +1.20e-02 |
Pos logit contributions | |
tune | +0.029 |
sensation | +0.029 |
dolphin | +0.027 |
shark | +0.026 |
chill | +0.025 |
Neg logit contributions | |
ends | -0.016 |
counts | -0.016 |
One | -0.015 |
Or | -0.014 |
Ready | -0.012 |
Token | her |
Token ID | 607 |
Feature activation | -5.82e-03 |
Pos logit contributions | |
tune | +0.027 |
sensation | +0.027 |
dolphin | +0.025 |
shark | +0.024 |
streaming | +0.024 |
Neg logit contributions | |
ends | -0.014 |
counts | -0.014 |
One | -0.013 |
Or | -0.013 |
Ready | -0.011 |
Token | parents |
Token ID | 3397 |
Feature activation | +4.15e-02 |
Pos logit contributions | |
ends | +0.007 |
counts | +0.007 |
Or | +0.007 |
One | +0.007 |
cannot | +0.007 |
Neg logit contributions | |
sensation | -0.012 |
tune | -0.012 |
lull | -0.010 |
hing | -0.010 |
streaming | -0.010 |
Token | said |
Token ID | 531 |
Feature activation | -3.75e-02 |
Pos logit contributions | |
sensation | +0.089 |
tune | +0.089 |
dolphin | +0.077 |
shark | +0.076 |
lull | +0.073 |
Neg logit contributions | |
ends | -0.056 |
counts | -0.056 |
One | -0.053 |
Or | -0.050 |
� | -0.045 |
Token | surprise |
Token ID | 5975 |
Feature activation | +2.12e-03 |
Pos logit contributions | |
ends | +0.015 |
counts | +0.014 |
One | +0.014 |
Or | +0.013 |
Ready | +0.012 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.012 |
potato | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
tune | +0.005 |
sensation | +0.005 |
dolphin | +0.004 |
shark | +0.004 |
face | +0.004 |
Neg logit contributions | |
ends | -0.003 |
counts | -0.003 |
One | -0.003 |
Or | -0.002 |
Ready | -0.002 |
Token | He |
Token ID | 679 |
Feature activation | -1.38e-02 |
Pos logit contributions | |
ends | +0.038 |
counts | +0.038 |
One | +0.036 |
ours | +0.033 |
Or | +0.033 |
Neg logit contributions | |
sensation | -0.030 |
tune | -0.029 |
shark | -0.026 |
face | -0.025 |
dolphin | -0.025 |
Token | threw |
Token ID | 9617 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
counts | +0.021 |
ends | +0.020 |
One | +0.020 |
Or | +0.019 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.028 |
sensation | -0.027 |
dolphin | -0.024 |
shark | -0.023 |
streaming | -0.023 |
Token | a |
Token ID | 257 |
Feature activation | +6.48e-04 |
Pos logit contributions | |
counts | +0.026 |
ends | +0.026 |
One | +0.025 |
Or | +0.024 |
Ready | +0.022 |
Neg logit contributions | |
tune | -0.026 |
sensation | -0.026 |
dolphin | -0.023 |
shark | -0.022 |
streaming | -0.021 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.82e-01 |
Pos logit contributions | |
sensation | +0.001 |
tune | +0.001 |
streaming | +0.001 |
Dolphin | +0.001 |
dolphin | +0.001 |
Neg logit contributions | |
ends | -0.001 |
One | -0.001 |
counts | -0.001 |
Or | -0.001 |
Ready | -0.001 |
Token | rum |
Token ID | 6582 |
Feature activation | +3.95e-03 |
Pos logit contributions | |
tune | +0.279 |
sensation | +0.274 |
dolphin | +0.242 |
shark | +0.231 |
streaming | +0.218 |
Neg logit contributions | |
ends | -0.350 |
counts | -0.350 |
One | -0.337 |
Or | -0.325 |
Ready | -0.304 |
Token | . |
Token ID | 13 |
Feature activation | -2.09e-02 |
Pos logit contributions | |
tune | +0.008 |
sensation | +0.008 |
dolphin | +0.008 |
shark | +0.007 |
streaming | +0.007 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
One | -0.005 |
Or | -0.005 |
Ready | -0.004 |
Token | He |
Token ID | 679 |
Feature activation | -1.13e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.034 |
One | +0.032 |
Or | +0.031 |
Ready | +0.029 |
Neg logit contributions | |
tune | -0.039 |
sensation | -0.038 |
dolphin | -0.035 |
shark | -0.034 |
streaming | -0.032 |
Token | kicked |
Token ID | 12165 |
Feature activation | +2.84e-03 |
Pos logit contributions | |
counts | +0.016 |
ends | +0.016 |
One | +0.015 |
Or | +0.014 |
cannot | +0.013 |
Neg logit contributions | |
tune | -0.024 |
sensation | -0.023 |
dolphin | -0.021 |
Dolphin | -0.020 |
streaming | -0.020 |
Token | and |
Token ID | 290 |
Feature activation | +8.85e-03 |
Pos logit contributions | |
tune | +0.005 |
sensation | +0.005 |
dolphin | +0.004 |
shark | +0.004 |
streaming | +0.004 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
One | -0.005 |
Or | -0.005 |
Ready | -0.004 |
Token | stubborn |
Token ID | 24444 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
ends | +0.045 |
counts | +0.045 |
One | +0.043 |
Or | +0.041 |
Ready | +0.038 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.056 |
dolphin | -0.050 |
shark | -0.049 |
streaming | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | -2.98e-02 |
Pos logit contributions | |
ends | +0.018 |
counts | +0.017 |
One | +0.017 |
Or | +0.016 |
Ready | +0.016 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.027 |
dolphin | -0.024 |
shark | -0.023 |
face | -0.023 |
Token | She |
Token ID | 1375 |
Feature activation | -3.48e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.033 |
One | +0.031 |
ours | +0.027 |
Or | +0.026 |
Neg logit contributions | |
sensation | -0.067 |
tune | -0.067 |
dolphin | -0.061 |
face | -0.059 |
shark | -0.059 |
Token | throws |
Token ID | 12542 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
ends | +0.043 |
One | +0.043 |
counts | +0.039 |
Or | +0.038 |
ours | +0.037 |
Neg logit contributions | |
sensation | -0.069 |
tune | -0.067 |
face | -0.066 |
dolphin | -0.065 |
shark | -0.064 |
Token | a |
Token ID | 257 |
Feature activation | +1.09e-04 |
Pos logit contributions | |
counts | +0.020 |
ends | +0.020 |
One | +0.019 |
Or | +0.018 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.024 |
sensation | -0.023 |
dolphin | -0.021 |
shark | -0.020 |
streaming | -0.019 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.68e-01 |
Pos logit contributions | |
sensation | +0.000 |
tune | +0.000 |
streaming | +0.000 |
dolphin | +0.000 |
Dolphin | +0.000 |
Neg logit contributions | |
ends | -0.000 |
One | -0.000 |
counts | -0.000 |
Or | -0.000 |
Ready | -0.000 |
Token | rum |
Token ID | 6582 |
Feature activation | -8.92e-03 |
Pos logit contributions | |
tune | +0.256 |
sensation | +0.250 |
dolphin | +0.221 |
shark | +0.211 |
streaming | +0.199 |
Neg logit contributions | |
ends | -0.327 |
counts | -0.326 |
One | -0.314 |
Or | -0.304 |
Ready | -0.286 |
Token | . |
Token ID | 13 |
Feature activation | -3.18e-02 |
Pos logit contributions | |
counts | +0.013 |
One | +0.013 |
ends | +0.013 |
Ready | +0.012 |
Or | +0.012 |
Neg logit contributions | |
sensation | -0.017 |
tune | -0.017 |
dolphin | -0.015 |
face | -0.015 |
streaming | -0.015 |
Token | She |
Token ID | 1375 |
Feature activation | -3.37e-02 |
Pos logit contributions | |
ends | +0.036 |
counts | +0.036 |
One | +0.033 |
Or | +0.029 |
ours | +0.028 |
Neg logit contributions | |
sensation | -0.072 |
tune | -0.072 |
dolphin | -0.066 |
shark | -0.064 |
face | -0.063 |
Token | kicks |
Token ID | 17364 |
Feature activation | -3.17e-03 |
Pos logit contributions | |
ends | +0.043 |
One | +0.043 |
counts | +0.039 |
Or | +0.038 |
Ready | +0.036 |
Neg logit contributions | |
sensation | -0.068 |
tune | -0.067 |
shark | -0.065 |
dolphin | -0.065 |
face | -0.063 |
Token | and |
Token ID | 290 |
Feature activation | +9.20e-03 |
Pos logit contributions | |
counts | +0.005 |
ends | +0.005 |
One | +0.005 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.006 |
sensation | -0.006 |
dolphin | -0.005 |
shark | -0.005 |
streaming | -0.005 |
Token | vegetables |
Token ID | 13701 |
Feature activation | +3.81e-02 |
Pos logit contributions | |
tune | +0.035 |
sensation | +0.035 |
dolphin | +0.030 |
shark | +0.029 |
streaming | +0.029 |
Neg logit contributions | |
counts | -0.038 |
ends | -0.038 |
One | -0.037 |
Or | -0.036 |
Ready | -0.033 |
Token | laughed |
Token ID | 13818 |
Feature activation | +2.12e-02 |
Pos logit contributions | |
tune | +0.072 |
sensation | +0.070 |
dolphin | +0.064 |
shark | +0.061 |
streaming | +0.059 |
Neg logit contributions | |
counts | -0.060 |
ends | -0.060 |
One | -0.057 |
Or | -0.055 |
Ready | -0.050 |
Token | at |
Token ID | 379 |
Feature activation | -6.02e-03 |
Pos logit contributions | |
sensation | +0.049 |
tune | +0.049 |
dolphin | +0.045 |
shark | +0.044 |
chill | +0.043 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.023 |
One | -0.021 |
Or | -0.021 |
Ready | -0.017 |
Token | the |
Token ID | 262 |
Feature activation | -2.78e-03 |
Pos logit contributions | |
counts | +0.008 |
One | +0.008 |
ends | +0.008 |
Or | +0.008 |
Ready | +0.007 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.012 |
chill | -0.010 |
dolphin | -0.010 |
shark | -0.010 |
Token | boast |
Token ID | 28020 |
Feature activation | +6.47e-02 |
Pos logit contributions | |
counts | +0.005 |
One | +0.005 |
ends | +0.005 |
Or | +0.004 |
Ready | +0.004 |
Neg logit contributions | |
tune | -0.004 |
sensation | -0.004 |
streaming | -0.004 |
streamed | -0.004 |
dolphin | -0.004 |
Token | ful |
Token ID | 913 |
Feature activation | +1.68e-01 |
Pos logit contributions | |
tune | +0.106 |
sensation | +0.104 |
dolphin | +0.095 |
shark | +0.094 |
chill | +0.092 |
Neg logit contributions | |
ends | -0.114 |
counts | -0.108 |
One | -0.104 |
Or | -0.103 |
one | -0.090 |
Token | onion |
Token ID | 21670 |
Feature activation | +6.84e-02 |
Pos logit contributions | |
sensation | +0.308 |
tune | +0.303 |
streaming | +0.277 |
streamed | +0.271 |
dolphin | +0.270 |
Neg logit contributions | |
counts | -0.253 |
ends | -0.239 |
One | -0.239 |
Or | -0.232 |
has | -0.207 |
Token | . |
Token ID | 13 |
Feature activation | +7.01e-03 |
Pos logit contributions | |
tune | +0.158 |
sensation | +0.158 |
dolphin | +0.147 |
shark | +0.144 |
Dolphin | +0.143 |
Neg logit contributions | |
counts | -0.071 |
ends | -0.066 |
Or | -0.063 |
One | -0.062 |
house | -0.053 |
Token | One |
Token ID | 1881 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.013 |
dolphin | +0.012 |
shark | +0.011 |
Dolphin | +0.011 |
Neg logit contributions | |
ends | -0.011 |
counts | -0.011 |
One | -0.010 |
Or | -0.010 |
ours | -0.009 |
Token | of |
Token ID | 286 |
Feature activation | +8.25e-02 |
Pos logit contributions | |
counts | +0.153 |
One | +0.151 |
ends | +0.136 |
Or | +0.132 |
� | +0.132 |
Neg logit contributions | |
tune | -0.187 |
sensation | -0.176 |
streaming | -0.164 |
dolphin | -0.159 |
streamed | -0.156 |
Token | them |
Token ID | 606 |
Feature activation | +4.14e-02 |
Pos logit contributions | |
tune | +0.165 |
sensation | +0.157 |
chill | +0.138 |
dolphin | +0.136 |
shark | +0.135 |
Neg logit contributions | |
One | -0.127 |
counts | -0.116 |
ends | -0.108 |
Or | -0.100 |
� | -0.096 |
Token | out |
Token ID | 503 |
Feature activation | +3.38e-02 |
Pos logit contributions | |
ends | +0.016 |
counts | +0.015 |
Or | +0.014 |
One | +0.013 |
Ready | +0.013 |
Neg logit contributions | |
sensation | -0.026 |
tune | -0.025 |
dolphin | -0.023 |
shark | -0.022 |
chill | -0.022 |
Token | loud |
Token ID | 7812 |
Feature activation | +3.84e-02 |
Pos logit contributions | |
sensation | +0.079 |
tune | +0.076 |
dolphin | +0.072 |
Sister | +0.071 |
Dolphin | +0.071 |
Neg logit contributions | |
ends | -0.034 |
One | -0.034 |
counts | -0.033 |
Or | -0.032 |
� | -0.028 |
Token | and |
Token ID | 290 |
Feature activation | +1.67e-02 |
Pos logit contributions | |
sensation | +0.087 |
tune | +0.084 |
dolphin | +0.078 |
streamed | +0.077 |
chill | +0.076 |
Neg logit contributions | |
counts | -0.043 |
ends | -0.037 |
Or | -0.034 |
One | -0.031 |
has | -0.030 |
Token | enjoyed |
Token ID | 8359 |
Feature activation | +2.53e-02 |
Pos logit contributions | |
sensation | +0.032 |
tune | +0.031 |
dolphin | +0.028 |
shark | +0.028 |
potato | +0.027 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.026 |
One | -0.024 |
Or | -0.023 |
Ready | -0.022 |
Token | the |
Token ID | 262 |
Feature activation | +7.03e-03 |
Pos logit contributions | |
tune | +0.047 |
sensation | +0.046 |
dolphin | +0.042 |
shark | +0.040 |
streamed | +0.040 |
Neg logit contributions | |
ends | -0.040 |
counts | -0.039 |
One | -0.038 |
Or | -0.034 |
Ready | -0.033 |
Token | sensation |
Token ID | 18098 |
Feature activation | +1.57e-01 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.013 |
dolphin | +0.012 |
shark | +0.011 |
chill | +0.011 |
Neg logit contributions | |
ends | -0.011 |
counts | -0.011 |
One | -0.011 |
Or | -0.010 |
Ready | -0.009 |
Token | of |
Token ID | 286 |
Feature activation | +5.22e-02 |
Pos logit contributions | |
tune | +0.334 |
sensation | +0.314 |
dolphin | +0.312 |
Dolphin | +0.298 |
shark | +0.298 |
Neg logit contributions | |
counts | -0.200 |
ends | -0.191 |
One | -0.181 |
Or | -0.165 |
� | -0.162 |
Token | being |
Token ID | 852 |
Feature activation | +6.64e-02 |
Pos logit contributions | |
tune | +0.099 |
sensation | +0.093 |
dolphin | +0.090 |
shark | +0.086 |
Dolphin | +0.085 |
Neg logit contributions | |
counts | -0.077 |
One | -0.074 |
ends | -0.071 |
ours | -0.067 |
Ready | -0.066 |
Token | so |
Token ID | 523 |
Feature activation | -3.98e-03 |
Pos logit contributions | |
tune | +0.130 |
sensation | +0.129 |
dolphin | +0.118 |
shark | +0.115 |
potato | +0.109 |
Neg logit contributions | |
ends | -0.101 |
counts | -0.099 |
One | -0.092 |
Or | -0.086 |
Ready | -0.080 |
Token | high |
Token ID | 1029 |
Feature activation | -2.12e-02 |
Pos logit contributions | |
ends | +0.006 |
counts | +0.006 |
One | +0.006 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.008 |
sensation | -0.008 |
dolphin | -0.007 |
shark | -0.007 |
potato | -0.007 |
Token | up |
Token ID | 510 |
Feature activation | -4.47e-02 |
Pos logit contributions | |
counts | +0.026 |
ends | +0.025 |
One | +0.024 |
Or | +0.022 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.046 |
dolphin | -0.043 |
shark | -0.041 |
Dolphin | -0.040 |
Token | his |
Token ID | 465 |
Feature activation | -2.74e-03 |
Pos logit contributions | |
tune | +0.082 |
sensation | +0.078 |
chill | +0.070 |
lull | +0.069 |
dolphin | +0.069 |
Neg logit contributions | |
counts | -0.069 |
One | -0.069 |
ends | -0.067 |
Or | -0.064 |
" | -0.060 |
Token | feet |
Token ID | 3625 |
Feature activation | +5.00e-02 |
Pos logit contributions | |
ends | +0.005 |
counts | +0.005 |
Or | +0.005 |
One | +0.005 |
ours | +0.005 |
Neg logit contributions | |
sensation | -0.004 |
tune | -0.004 |
dolphin | -0.003 |
streamed | -0.003 |
streaming | -0.003 |
Token | and |
Token ID | 290 |
Feature activation | -1.75e-02 |
Pos logit contributions | |
tune | +0.101 |
sensation | +0.098 |
dolphin | +0.088 |
shark | +0.085 |
streaming | +0.084 |
Neg logit contributions | |
counts | -0.074 |
ends | -0.072 |
One | -0.068 |
Or | -0.066 |
Ready | -0.062 |
Token | threw |
Token ID | 9617 |
Feature activation | -2.61e-02 |
Pos logit contributions | |
counts | +0.026 |
ends | +0.025 |
Or | +0.022 |
One | +0.021 |
has | +0.020 |
Neg logit contributions | |
tune | -0.035 |
sensation | -0.034 |
andel | -0.031 |
dolphin | -0.031 |
Dolphin | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -1.20e-02 |
Pos logit contributions | |
counts | +0.044 |
ends | +0.043 |
One | +0.043 |
Or | +0.040 |
Ready | +0.038 |
Neg logit contributions | |
tune | -0.046 |
sensation | -0.045 |
dolphin | -0.041 |
shark | -0.038 |
Dolphin | -0.037 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
ends | +0.019 |
One | +0.018 |
counts | +0.018 |
Or | +0.018 |
Ready | +0.016 |
Neg logit contributions | |
sensation | -0.021 |
tune | -0.021 |
Dolphin | -0.018 |
streaming | -0.018 |
dolphin | -0.018 |
Token | rum |
Token ID | 6582 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
tune | +0.237 |
sensation | +0.232 |
dolphin | +0.206 |
shark | +0.197 |
streaming | +0.185 |
Neg logit contributions | |
ends | -0.292 |
counts | -0.292 |
One | -0.281 |
Or | -0.271 |
Ready | -0.253 |
Token | . |
Token ID | 13 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
ends | +0.020 |
counts | +0.019 |
One | +0.018 |
Or | +0.017 |
Ready | +0.015 |
Neg logit contributions | |
tune | -0.040 |
sensation | -0.040 |
dolphin | -0.037 |
shark | -0.036 |
chill | -0.034 |
Token |
|
Token ID | 198 |
Feature activation | -1.21e-02 |
Pos logit contributions | |
ends | +0.032 |
counts | +0.031 |
One | +0.031 |
Or | +0.030 |
Ready | +0.026 |
Neg logit contributions | |
tune | -0.060 |
sensation | -0.058 |
dolphin | -0.053 |
shark | -0.051 |
streaming | -0.051 |
Token |
|
Token ID | 198 |
Feature activation | +9.75e-05 |
Pos logit contributions | |
counts | +0.020 |
ends | +0.018 |
ours | +0.016 |
belongs | +0.016 |
Or | +0.016 |
Neg logit contributions | |
tune | -0.021 |
sensation | -0.020 |
dolphin | -0.020 |
shark | -0.019 |
Sister | -0.018 |
Token | " |
Token ID | 1 |
Feature activation | -4.54e-02 |
Pos logit contributions | |
tune | +0.000 |
sensation | +0.000 |
dolphin | +0.000 |
shark | +0.000 |
Sister | +0.000 |
Neg logit contributions | |
counts | -0.000 |
ends | -0.000 |
belongs | -0.000 |
ours | -0.000 |
Or | -0.000 |
Token | Suddenly |
Token ID | 38582 |
Feature activation | -1.17e-02 |
Pos logit contributions | |
tune | +0.046 |
sensation | +0.043 |
dolphin | +0.039 |
lull | +0.037 |
Dolphin | +0.035 |
Neg logit contributions | |
counts | -0.046 |
ends | -0.041 |
belongs | -0.038 |
ours | -0.038 |
Or | -0.037 |
Token | , |
Token ID | 11 |
Feature activation | +3.98e-02 |
Pos logit contributions | |
counts | +0.016 |
ends | +0.015 |
One | +0.015 |
Or | +0.014 |
Ready | +0.013 |
Neg logit contributions | |
tune | -0.025 |
sensation | -0.025 |
dolphin | -0.023 |
shark | -0.022 |
streaming | -0.021 |
Token | she |
Token ID | 673 |
Feature activation | +2.23e-02 |
Pos logit contributions | |
tune | +0.088 |
sensation | +0.085 |
dolphin | +0.077 |
shark | +0.075 |
streaming | +0.073 |
Neg logit contributions | |
counts | -0.050 |
ends | -0.050 |
One | -0.048 |
Or | -0.045 |
Ready | -0.042 |
Token | noticed |
Token ID | 6810 |
Feature activation | +4.58e-02 |
Pos logit contributions | |
tune | +0.041 |
sensation | +0.040 |
dolphin | +0.036 |
shark | +0.035 |
streaming | +0.033 |
Neg logit contributions | |
counts | -0.037 |
ends | -0.037 |
One | -0.035 |
Or | -0.034 |
Ready | -0.031 |
Token | a |
Token ID | 257 |
Feature activation | +1.10e-02 |
Pos logit contributions | |
tune | +0.107 |
sensation | +0.103 |
dolphin | +0.094 |
shark | +0.093 |
Dolphin | +0.092 |
Neg logit contributions | |
counts | -0.050 |
One | -0.049 |
ends | -0.048 |
Or | -0.044 |
Ready | -0.039 |
Token | lull |
Token ID | 42858 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
sensation | +0.020 |
tune | +0.019 |
streamed | +0.018 |
streaming | +0.018 |
ened | +0.017 |
Neg logit contributions | |
One | -0.017 |
Or | -0.015 |
ends | -0.015 |
counts | -0.015 |
Ready | -0.015 |
Token | aby |
Token ID | 3930 |
Feature activation | +6.27e-02 |
Pos logit contributions | |
tant | +0.263 |
tune | +0.254 |
lobster | +0.252 |
sensation | +0.249 |
gul | +0.237 |
Neg logit contributions | |
ends | -0.198 |
house | -0.196 |
" | -0.177 |
Or | -0.169 |
ider | -0.166 |
Token | playing |
Token ID | 2712 |
Feature activation | +3.94e-02 |
Pos logit contributions | |
sensation | +0.143 |
tune | +0.142 |
dolphin | +0.131 |
shark | +0.130 |
Dolphin | +0.125 |
Neg logit contributions | |
counts | -0.068 |
ends | -0.067 |
One | -0.061 |
Or | -0.059 |
� | -0.055 |
Token | in |
Token ID | 287 |
Feature activation | +1.99e-02 |
Pos logit contributions | |
tune | +0.086 |
sensation | +0.083 |
dolphin | +0.082 |
Dolphin | +0.079 |
shark | +0.079 |
Neg logit contributions | |
counts | -0.045 |
ends | -0.043 |
One | -0.042 |
Or | -0.041 |
house | -0.037 |
Token | the |
Token ID | 262 |
Feature activation | +1.88e-02 |
Pos logit contributions | |
tune | +0.039 |
sensation | +0.037 |
dolphin | +0.035 |
shark | +0.034 |
Dolphin | +0.033 |
Neg logit contributions | |
counts | -0.030 |
One | -0.030 |
ends | -0.028 |
Or | -0.027 |
Ready | -0.025 |
Token | distance |
Token ID | 5253 |
Feature activation | +3.59e-02 |
Pos logit contributions | |
tune | +0.034 |
sensation | +0.034 |
dolphin | +0.031 |
shark | +0.030 |
streaming | +0.029 |
Neg logit contributions | |
One | -0.030 |
counts | -0.029 |
ends | -0.027 |
Or | -0.026 |
Ready | -0.026 |
Token | being |
Token ID | 852 |
Feature activation | +1.80e-02 |
Pos logit contributions | |
sensation | +0.040 |
tune | +0.039 |
shark | +0.038 |
dolphin | +0.037 |
Dolphin | +0.035 |
Neg logit contributions | |
counts | -0.027 |
One | -0.027 |
Or | -0.025 |
ends | -0.025 |
ours | -0.022 |
Token | careless |
Token ID | 36138 |
Feature activation | +1.86e-03 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.036 |
dolphin | +0.033 |
shark | +0.032 |
streaming | +0.030 |
Neg logit contributions | |
ends | -0.026 |
counts | -0.026 |
One | -0.024 |
Or | -0.023 |
Ready | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -7.91e-04 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
dolphin | +0.004 |
shark | +0.004 |
streaming | +0.004 |
Neg logit contributions | |
counts | -0.002 |
ends | -0.002 |
One | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | The |
Token ID | 383 |
Feature activation | -4.18e-04 |
Pos logit contributions | |
ends | +0.001 |
counts | +0.001 |
One | +0.001 |
ours | +0.001 |
Or | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
shark | -0.001 |
Dolphin | -0.001 |
Token | sorrow |
Token ID | 24140 |
Feature activation | +9.79e-02 |
Pos logit contributions | |
counts | +0.001 |
One | +0.001 |
ends | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
streaming | -0.001 |
streamed | -0.001 |
Token | ful |
Token ID | 913 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
tune | +0.217 |
sensation | +0.214 |
dolphin | +0.193 |
shark | +0.187 |
Dolphin | +0.185 |
Neg logit contributions | |
ends | -0.122 |
counts | -0.119 |
Or | -0.107 |
has | -0.103 |
One | -0.103 |
Token | little |
Token ID | 1310 |
Feature activation | +3.63e-02 |
Pos logit contributions | |
sensation | +0.252 |
tune | +0.251 |
dolphin | +0.224 |
streaming | +0.219 |
Dolphin | +0.213 |
Neg logit contributions | |
counts | -0.256 |
One | -0.241 |
Or | -0.240 |
ends | -0.240 |
has | -0.221 |
Token | boy |
Token ID | 2933 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
tune | +0.057 |
sensation | +0.056 |
dolphin | +0.049 |
streaming | +0.047 |
streamed | +0.046 |
Neg logit contributions | |
counts | -0.066 |
One | -0.062 |
ends | -0.062 |
Or | -0.061 |
Ready | -0.057 |
Token | started |
Token ID | 2067 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
sensation | +0.064 |
tune | +0.064 |
dolphin | +0.059 |
Dolphin | +0.056 |
shark | +0.056 |
Neg logit contributions | |
counts | -0.043 |
ends | -0.041 |
Or | -0.039 |
One | -0.038 |
has | -0.036 |
Token | to |
Token ID | 284 |
Feature activation | +6.23e-03 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
One | +0.011 |
Or | +0.011 |
� | +0.010 |
Neg logit contributions | |
sensation | -0.026 |
tune | -0.026 |
dolphin | -0.025 |
shark | -0.024 |
Dolphin | -0.024 |
Token | cry |
Token ID | 3960 |
Feature activation | +3.30e-02 |
Pos logit contributions | |
tune | +0.011 |
sensation | +0.011 |
dolphin | +0.011 |
shark | +0.010 |
Dolphin | +0.010 |
Neg logit contributions | |
counts | -0.009 |
One | -0.009 |
ends | -0.009 |
Or | -0.009 |
� | -0.008 |
Token | being |
Token ID | 852 |
Feature activation | +1.80e-02 |
Pos logit contributions | |
sensation | +0.040 |
tune | +0.039 |
shark | +0.038 |
dolphin | +0.037 |
Dolphin | +0.035 |
Neg logit contributions | |
counts | -0.027 |
One | -0.027 |
Or | -0.025 |
ends | -0.025 |
ours | -0.022 |
Token | careless |
Token ID | 36138 |
Feature activation | +1.86e-03 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.036 |
dolphin | +0.033 |
shark | +0.032 |
streaming | +0.030 |
Neg logit contributions | |
ends | -0.026 |
counts | -0.026 |
One | -0.024 |
Or | -0.023 |
Ready | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -7.91e-04 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
dolphin | +0.004 |
shark | +0.004 |
streaming | +0.004 |
Neg logit contributions | |
counts | -0.002 |
ends | -0.002 |
One | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | The |
Token ID | 383 |
Feature activation | -4.18e-04 |
Pos logit contributions | |
ends | +0.001 |
counts | +0.001 |
One | +0.001 |
ours | +0.001 |
Or | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
shark | -0.001 |
Dolphin | -0.001 |
Token | sorrow |
Token ID | 24140 |
Feature activation | +9.79e-02 |
Pos logit contributions | |
counts | +0.001 |
One | +0.001 |
ends | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
streaming | -0.001 |
streamed | -0.001 |
Token | ful |
Token ID | 913 |
Feature activation | +1.53e-01 |
Pos logit contributions | |
tune | +0.217 |
sensation | +0.214 |
dolphin | +0.193 |
shark | +0.187 |
Dolphin | +0.185 |
Neg logit contributions | |
ends | -0.122 |
counts | -0.119 |
Or | -0.107 |
has | -0.103 |
One | -0.103 |
Token | little |
Token ID | 1310 |
Feature activation | +3.63e-02 |
Pos logit contributions | |
sensation | +0.252 |
tune | +0.251 |
dolphin | +0.224 |
streaming | +0.219 |
Dolphin | +0.213 |
Neg logit contributions | |
counts | -0.256 |
One | -0.241 |
Or | -0.240 |
ends | -0.240 |
has | -0.221 |
Token | boy |
Token ID | 2933 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
tune | +0.057 |
sensation | +0.056 |
dolphin | +0.049 |
streaming | +0.047 |
streamed | +0.046 |
Neg logit contributions | |
counts | -0.066 |
One | -0.062 |
ends | -0.062 |
Or | -0.061 |
Ready | -0.057 |
Token | started |
Token ID | 2067 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
sensation | +0.064 |
tune | +0.064 |
dolphin | +0.059 |
Dolphin | +0.056 |
shark | +0.056 |
Neg logit contributions | |
counts | -0.043 |
ends | -0.041 |
Or | -0.039 |
One | -0.038 |
has | -0.036 |
Token | to |
Token ID | 284 |
Feature activation | +6.23e-03 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
One | +0.011 |
Or | +0.011 |
� | +0.010 |
Neg logit contributions | |
sensation | -0.026 |
tune | -0.026 |
dolphin | -0.025 |
shark | -0.024 |
Dolphin | -0.024 |
Token | cry |
Token ID | 3960 |
Feature activation | +3.30e-02 |
Pos logit contributions | |
tune | +0.011 |
sensation | +0.011 |
dolphin | +0.011 |
shark | +0.010 |
Dolphin | +0.010 |
Neg logit contributions | |
counts | -0.009 |
One | -0.009 |
ends | -0.009 |
Or | -0.009 |
� | -0.008 |
Token | and |
Token ID | 290 |
Feature activation | +3.74e-03 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
One | +0.005 |
Or | +0.005 |
Ready | +0.004 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.011 |
dolphin | -0.010 |
shark | -0.010 |
streamed | -0.010 |
Token | returned |
Token ID | 4504 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
tune | +0.007 |
sensation | +0.007 |
dolphin | +0.006 |
shark | +0.006 |
streaming | +0.006 |
Neg logit contributions | |
counts | -0.006 |
ends | -0.006 |
One | -0.005 |
Or | -0.005 |
Ready | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | +3.03e-02 |
Pos logit contributions | |
counts | +0.033 |
ends | +0.032 |
One | +0.031 |
Or | +0.028 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.042 |
sensation | -0.042 |
dolphin | -0.037 |
shark | -0.036 |
streamed | -0.035 |
Token | his |
Token ID | 465 |
Feature activation | +2.52e-02 |
Pos logit contributions | |
sensation | +0.051 |
tune | +0.051 |
streamed | +0.045 |
dolphin | +0.045 |
streaming | +0.044 |
Neg logit contributions | |
counts | -0.053 |
One | -0.051 |
ends | -0.050 |
Ready | -0.047 |
Or | -0.046 |
Token | rest |
Token ID | 1334 |
Feature activation | +8.01e-02 |
Pos logit contributions | |
tune | +0.043 |
sensation | +0.043 |
dolphin | +0.038 |
shark | +0.036 |
streaming | +0.036 |
Neg logit contributions | |
counts | -0.044 |
ends | -0.043 |
One | -0.042 |
Or | -0.040 |
Ready | -0.038 |
Token | ful |
Token ID | 913 |
Feature activation | +1.51e-01 |
Pos logit contributions | |
tune | +0.192 |
sensation | +0.180 |
dolphin | +0.176 |
shark | +0.171 |
Dolphin | +0.169 |
Neg logit contributions | |
ends | -0.080 |
counts | -0.077 |
One | -0.072 |
Or | -0.061 |
house | -0.057 |
Token | spot |
Token ID | 4136 |
Feature activation | -5.53e-03 |
Pos logit contributions | |
tune | +0.282 |
sensation | +0.281 |
dolphin | +0.260 |
shark | +0.248 |
Dolphin | +0.244 |
Neg logit contributions | |
counts | -0.234 |
One | -0.221 |
ends | -0.217 |
Or | -0.205 |
house | -0.193 |
Token | beneath |
Token ID | 11061 |
Feature activation | +2.75e-02 |
Pos logit contributions | |
ends | +0.007 |
counts | +0.006 |
One | +0.006 |
Or | +0.006 |
house | +0.005 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.012 |
dolphin | -0.011 |
shark | -0.011 |
Dolphin | -0.011 |
Token | the |
Token ID | 262 |
Feature activation | +2.83e-02 |
Pos logit contributions | |
tune | +0.051 |
sensation | +0.050 |
dolphin | +0.046 |
chill | +0.043 |
shark | +0.042 |
Neg logit contributions | |
One | -0.041 |
counts | -0.041 |
ends | -0.036 |
Or | -0.035 |
� | -0.035 |
Token | tree |
Token ID | 5509 |
Feature activation | -2.23e-02 |
Pos logit contributions | |
tune | +0.057 |
sensation | +0.057 |
dolphin | +0.052 |
shark | +0.050 |
streaming | +0.048 |
Neg logit contributions | |
counts | -0.041 |
ends | -0.041 |
One | -0.039 |
Or | -0.037 |
Ready | -0.034 |
Token | with |
Token ID | 351 |
Feature activation | +1.92e-02 |
Pos logit contributions | |
counts | +0.026 |
ends | +0.026 |
One | +0.024 |
Or | +0.023 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.051 |
sensation | -0.051 |
dolphin | -0.047 |
shark | -0.046 |
streaming | -0.044 |
Token | , |
Token ID | 11 |
Feature activation | +4.52e-02 |
Pos logit contributions | |
tune | +0.030 |
sensation | +0.030 |
dolphin | +0.027 |
shark | +0.026 |
potato | +0.025 |
Neg logit contributions | |
counts | -0.025 |
ends | -0.024 |
One | -0.024 |
Or | -0.022 |
Ready | -0.021 |
Token | he |
Token ID | 339 |
Feature activation | +3.65e-02 |
Pos logit contributions | |
tune | +0.098 |
sensation | +0.096 |
dolphin | +0.093 |
potato | +0.087 |
shark | +0.086 |
Neg logit contributions | |
counts | -0.058 |
ends | -0.055 |
One | -0.053 |
Or | -0.050 |
Ready | -0.046 |
Token | remembered |
Token ID | 12086 |
Feature activation | +3.16e-02 |
Pos logit contributions | |
sensation | +0.065 |
tune | +0.065 |
dolphin | +0.058 |
shark | +0.056 |
potato | +0.053 |
Neg logit contributions | |
counts | -0.063 |
ends | -0.061 |
One | -0.060 |
Or | -0.057 |
Ready | -0.052 |
Token | the |
Token ID | 262 |
Feature activation | +1.55e-02 |
Pos logit contributions | |
tune | +0.070 |
sensation | +0.068 |
streamed | +0.059 |
andel | +0.059 |
dolphin | +0.059 |
Neg logit contributions | |
counts | -0.039 |
One | -0.039 |
ends | -0.036 |
Or | -0.032 |
� | -0.031 |
Token | sorrow |
Token ID | 24140 |
Feature activation | +1.04e-01 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.028 |
dolphin | +0.025 |
shark | +0.024 |
streaming | +0.024 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.025 |
One | -0.025 |
Or | -0.023 |
Ready | -0.022 |
Token | ful |
Token ID | 913 |
Feature activation | +1.50e-01 |
Pos logit contributions | |
tune | +0.221 |
sensation | +0.219 |
dolphin | +0.197 |
shark | +0.193 |
Dolphin | +0.187 |
Neg logit contributions | |
ends | -0.137 |
counts | -0.130 |
One | -0.120 |
Or | -0.119 |
� | -0.115 |
Token | look |
Token ID | 804 |
Feature activation | +2.36e-02 |
Pos logit contributions | |
tune | +0.279 |
sensation | +0.267 |
dolphin | +0.251 |
Dolphin | +0.241 |
streaming | +0.239 |
Neg logit contributions | |
counts | -0.241 |
ends | -0.223 |
One | -0.219 |
Or | -0.215 |
Ready | -0.191 |
Token | the |
Token ID | 262 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
tune | +0.050 |
sensation | +0.049 |
dolphin | +0.045 |
shark | +0.043 |
Dolphin | +0.042 |
Neg logit contributions | |
counts | -0.032 |
ends | -0.032 |
One | -0.030 |
Or | -0.028 |
Ready | -0.025 |
Token | older |
Token ID | 4697 |
Feature activation | -1.93e-02 |
Pos logit contributions | |
counts | +0.023 |
ends | +0.023 |
One | +0.022 |
Or | +0.021 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.028 |
sensation | -0.028 |
dolphin | -0.026 |
shark | -0.025 |
streaming | -0.024 |
Token | man |
Token ID | 582 |
Feature activation | -4.99e-03 |
Pos logit contributions | |
ends | +0.037 |
counts | +0.036 |
One | +0.035 |
Or | +0.034 |
Ready | +0.032 |
Neg logit contributions | |
tune | -0.031 |
sensation | -0.030 |
dolphin | -0.026 |
shark | -0.025 |
chill | -0.024 |
Token | had |
Token ID | 550 |
Feature activation | +1.49e-04 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
One | +0.006 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.011 |
sensation | -0.011 |
dolphin | -0.010 |
shark | -0.010 |
chill | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | +1.77e-02 |
Pos logit contributions | |
tune | +0.026 |
sensation | +0.024 |
dolphin | +0.024 |
Dolphin | +0.023 |
shark | +0.023 |
Neg logit contributions | |
One | -0.020 |
counts | -0.020 |
ends | -0.018 |
Or | -0.018 |
Ready | -0.017 |
Token | the |
Token ID | 262 |
Feature activation | +1.45e-03 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.035 |
dolphin | +0.033 |
Dolphin | +0.031 |
andel | +0.031 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.023 |
One | -0.022 |
Or | -0.021 |
has | -0.021 |
Token | lightning |
Token ID | 14357 |
Feature activation | +5.58e-02 |
Pos logit contributions | |
tune | +0.003 |
sensation | +0.003 |
Dolphin | +0.002 |
dolphin | +0.002 |
streaming | +0.002 |
Neg logit contributions | |
counts | -0.002 |
One | -0.002 |
ends | -0.002 |
Or | -0.002 |
house | -0.002 |
Token | struck |
Token ID | 7425 |
Feature activation | -1.44e-02 |
Pos logit contributions | |
tune | +0.102 |
sensation | +0.102 |
Dolphin | +0.101 |
dolphin | +0.100 |
lobster | +0.097 |
Neg logit contributions | |
counts | -0.082 |
ends | -0.074 |
One | -0.073 |
house | -0.072 |
Or | -0.070 |
Token | the |
Token ID | 262 |
Feature activation | +3.46e-03 |
Pos logit contributions | |
counts | +0.019 |
One | +0.019 |
ends | +0.018 |
Or | +0.017 |
Ready | +0.016 |
Neg logit contributions | |
tune | -0.029 |
sensation | -0.028 |
dolphin | -0.028 |
shark | -0.026 |
Dolphin | -0.026 |
Token | hyd |
Token ID | 7409 |
Feature activation | +1.49e-01 |
Pos logit contributions | |
tune | +0.006 |
sensation | +0.006 |
streaming | +0.005 |
dolphin | +0.005 |
streamed | +0.005 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
One | -0.005 |
Or | -0.005 |
Ready | -0.005 |
Token | rant |
Token ID | 5250 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
tune | +0.271 |
sensation | +0.268 |
dolphin | +0.264 |
shark | +0.249 |
Dolphin | +0.243 |
Neg logit contributions | |
counts | -0.232 |
ends | -0.229 |
One | -0.218 |
Or | -0.207 |
belongs | -0.196 |
Token | . |
Token ID | 13 |
Feature activation | +9.53e-03 |
Pos logit contributions | |
tune | +0.050 |
sensation | +0.050 |
dolphin | +0.047 |
Dolphin | +0.046 |
Sister | +0.045 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.022 |
One | -0.022 |
Or | -0.022 |
house | -0.021 |
Token | Suddenly |
Token ID | 24975 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
tune | +0.019 |
sensation | +0.019 |
dolphin | +0.018 |
shark | +0.017 |
streaming | +0.016 |
Neg logit contributions | |
counts | -0.014 |
One | -0.014 |
ends | -0.014 |
Or | -0.013 |
Ready | -0.012 |
Token | water |
Token ID | 1660 |
Feature activation | +6.07e-03 |
Pos logit contributions | |
counts | +0.025 |
ends | +0.024 |
One | +0.023 |
Or | +0.022 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.041 |
sensation | -0.041 |
dolphin | -0.037 |
shark | -0.035 |
streaming | -0.035 |
Token | was |
Token ID | 373 |
Feature activation | -1.32e-02 |
Pos logit contributions | |
tune | +0.014 |
sensation | +0.013 |
manner | +0.012 |
toddler | +0.012 |
tunes | +0.012 |
Neg logit contributions | |
counts | -0.007 |
ends | -0.006 |
has | -0.006 |
cannot | -0.006 |
One | -0.006 |
Token | They |
Token ID | 1119 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
counts | +0.047 |
ends | +0.047 |
One | +0.042 |
belongs | +0.040 |
ours | +0.040 |
Neg logit contributions | |
tune | -0.050 |
sensation | -0.047 |
dolphin | -0.047 |
shark | -0.045 |
Sister | -0.045 |
Token | are |
Token ID | 389 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
ends | +0.029 |
counts | +0.028 |
One | +0.027 |
Or | +0.026 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.040 |
sensation | -0.039 |
dolphin | -0.036 |
shark | -0.035 |
chill | -0.033 |
Token | powerful |
Token ID | 3665 |
Feature activation | -4.27e-02 |
Pos logit contributions | |
ends | +0.037 |
counts | +0.035 |
One | +0.033 |
Or | +0.032 |
Ready | +0.029 |
Neg logit contributions | |
tune | -0.043 |
sensation | -0.043 |
dolphin | -0.038 |
shark | -0.037 |
chill | -0.036 |
Token | . |
Token ID | 13 |
Feature activation | -4.27e-02 |
Pos logit contributions | |
ends | +0.060 |
One | +0.060 |
counts | +0.057 |
Or | +0.054 |
Ready | +0.052 |
Neg logit contributions | |
tune | -0.089 |
sensation | -0.086 |
dolphin | -0.080 |
shark | -0.079 |
tunes | -0.074 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.35e-02 |
Pos logit contributions | |
ends | +0.065 |
counts | +0.065 |
One | +0.057 |
ours | +0.054 |
belongs | +0.054 |
Neg logit contributions | |
tune | -0.077 |
sensation | -0.074 |
dolphin | -0.072 |
shark | -0.069 |
Sister | -0.068 |
Token | One |
Token ID | 3198 |
Feature activation | -1.58e-01 |
Pos logit contributions | |
ends | +0.060 |
counts | +0.056 |
Or | +0.052 |
One | +0.052 |
ours | +0.049 |
Neg logit contributions | |
sensation | -0.089 |
tune | -0.087 |
dolphin | -0.082 |
shark | -0.082 |
Dolphin | -0.077 |
Token | day |
Token ID | 1110 |
Feature activation | +1.78e-02 |
Pos logit contributions | |
counts | +0.271 |
One | +0.269 |
ends | +0.258 |
Or | +0.236 |
has | +0.218 |
Neg logit contributions | |
tune | -0.285 |
sensation | -0.272 |
streaming | -0.248 |
dolphin | -0.247 |
shark | -0.240 |
Token | , |
Token ID | 11 |
Feature activation | -3.64e-03 |
Pos logit contributions | |
sensation | +0.028 |
tune | +0.028 |
dolphin | +0.027 |
shark | +0.027 |
streamed | +0.025 |
Neg logit contributions | |
counts | -0.030 |
ends | -0.030 |
One | -0.029 |
Or | -0.028 |
Ready | -0.027 |
Token | Tom |
Token ID | 4186 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
Or | +0.005 |
One | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.006 |
tune | -0.006 |
dolphin | -0.006 |
shark | -0.006 |
chill | -0.005 |
Token | and |
Token ID | 290 |
Feature activation | +2.20e-02 |
Pos logit contributions | |
ends | +0.055 |
One | +0.051 |
counts | +0.048 |
Or | +0.048 |
Ready | +0.048 |
Neg logit contributions | |
sensation | -0.057 |
tune | -0.055 |
dolphin | -0.053 |
shark | -0.052 |
potato | -0.049 |
Token | Lily |
Token ID | 20037 |
Feature activation | -4.54e-02 |
Pos logit contributions | |
tune | +0.035 |
sensation | +0.035 |
dolphin | +0.030 |
shark | +0.029 |
streaming | +0.028 |
Neg logit contributions | |
counts | -0.041 |
ends | -0.041 |
One | -0.040 |
Or | -0.039 |
ours | -0.035 |
Token | They |
Token ID | 1119 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
counts | +0.048 |
ends | +0.047 |
belongs | +0.041 |
ours | +0.041 |
One | +0.040 |
Neg logit contributions | |
dolphin | -0.039 |
tune | -0.037 |
Dolphin | -0.037 |
Sister | -0.037 |
shark | -0.035 |
Token | are |
Token ID | 389 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
counts | +0.020 |
ends | +0.020 |
One | +0.019 |
Or | +0.018 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.033 |
sensation | -0.032 |
dolphin | -0.030 |
shark | -0.029 |
streaming | -0.028 |
Token | careful |
Token ID | 8161 |
Feature activation | +2.10e-03 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.032 |
One | +0.031 |
Or | +0.029 |
ours | +0.027 |
Neg logit contributions | |
tune | -0.040 |
sensation | -0.040 |
dolphin | -0.035 |
shark | -0.034 |
tant | -0.033 |
Token | . |
Token ID | 13 |
Feature activation | -3.80e-02 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
face | +0.004 |
dolphin | +0.004 |
shark | +0.003 |
Neg logit contributions | |
counts | -0.003 |
One | -0.003 |
ends | -0.003 |
Ready | -0.003 |
Or | -0.003 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.59e-02 |
Pos logit contributions | |
counts | +0.060 |
ends | +0.058 |
One | +0.052 |
belongs | +0.050 |
ours | +0.050 |
Neg logit contributions | |
tune | -0.066 |
dolphin | -0.064 |
sensation | -0.063 |
Dolphin | -0.060 |
Sister | -0.060 |
Token | One |
Token ID | 3198 |
Feature activation | -1.58e-01 |
Pos logit contributions | |
ends | +0.060 |
counts | +0.057 |
One | +0.053 |
Or | +0.052 |
ours | +0.048 |
Neg logit contributions | |
sensation | -0.097 |
tune | -0.096 |
dolphin | -0.091 |
shark | -0.090 |
Dolphin | -0.085 |
Token | morning |
Token ID | 3329 |
Feature activation | -7.51e-04 |
Pos logit contributions | |
One | +0.257 |
counts | +0.254 |
ends | +0.237 |
Or | +0.215 |
has | +0.207 |
Neg logit contributions | |
tune | -0.301 |
sensation | -0.283 |
streaming | -0.269 |
dolphin | -0.258 |
streamed | -0.252 |
Token | , |
Token ID | 11 |
Feature activation | -3.37e-03 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
shark | -0.001 |
potato | -0.001 |
Token | a |
Token ID | 257 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.005 |
One | +0.005 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.006 |
tune | -0.006 |
dolphin | -0.005 |
shark | -0.005 |
face | -0.005 |
Token | little |
Token ID | 1310 |
Feature activation | +2.99e-02 |
Pos logit contributions | |
One | +0.031 |
ends | +0.030 |
Or | +0.029 |
counts | +0.027 |
has | +0.026 |
Neg logit contributions | |
sensation | -0.030 |
tune | -0.028 |
streaming | -0.027 |
ened | -0.025 |
streamed | -0.025 |
Token | ot |
Token ID | 30972 |
Feature activation | +3.43e-03 |
Pos logit contributions | |
sensation | +0.054 |
tune | +0.052 |
streaming | +0.045 |
streamed | +0.044 |
lull | +0.043 |
Neg logit contributions | |
One | -0.048 |
counts | -0.048 |
ends | -0.046 |
Or | -0.045 |
house | -0.041 |
Token | shoes |
Token ID | 10012 |
Feature activation | -3.97e-02 |
Pos logit contributions | |
counts | +0.029 |
ends | +0.029 |
One | +0.027 |
Or | +0.027 |
Ready | +0.025 |
Neg logit contributions | |
tune | -0.030 |
sensation | -0.030 |
dolphin | -0.026 |
streaming | -0.025 |
shark | -0.025 |
Token | are |
Token ID | 389 |
Feature activation | -4.69e-02 |
Pos logit contributions | |
ends | +0.068 |
counts | +0.068 |
One | +0.066 |
Or | +0.063 |
Ready | +0.060 |
Neg logit contributions | |
tune | -0.068 |
sensation | -0.067 |
dolphin | -0.060 |
shark | -0.058 |
streaming | -0.055 |
Token | better |
Token ID | 1365 |
Feature activation | -5.08e-02 |
Pos logit contributions | |
counts | +0.066 |
ends | +0.065 |
One | +0.061 |
Or | +0.061 |
Ready | +0.055 |
Neg logit contributions | |
tune | -0.097 |
sensation | -0.096 |
dolphin | -0.088 |
shark | -0.085 |
streaming | -0.083 |
Token | . |
Token ID | 13 |
Feature activation | -3.97e-02 |
Pos logit contributions | |
Ready | +0.075 |
counts | +0.075 |
One | +0.075 |
Or | +0.069 |
ends | +0.069 |
Neg logit contributions | |
tune | -0.097 |
sensation | -0.089 |
dolphin | -0.085 |
shark | -0.082 |
face | -0.080 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.49e-02 |
Pos logit contributions | |
counts | +0.069 |
ends | +0.067 |
One | +0.059 |
belongs | +0.059 |
ours | +0.059 |
Neg logit contributions | |
tune | -0.058 |
dolphin | -0.057 |
Sister | -0.056 |
Dolphin | -0.055 |
sensation | -0.054 |
Token | One |
Token ID | 3198 |
Feature activation | -1.58e-01 |
Pos logit contributions | |
ends | +0.064 |
counts | +0.059 |
Or | +0.056 |
One | +0.055 |
ours | +0.054 |
Neg logit contributions | |
sensation | -0.088 |
tune | -0.085 |
dolphin | -0.082 |
shark | -0.082 |
Dolphin | -0.078 |
Token | day |
Token ID | 1110 |
Feature activation | +1.87e-02 |
Pos logit contributions | |
One | +0.262 |
counts | +0.260 |
ends | +0.242 |
Or | +0.223 |
has | +0.209 |
Neg logit contributions | |
tune | -0.293 |
sensation | -0.278 |
streaming | -0.259 |
dolphin | -0.254 |
shark | -0.246 |
Token | , |
Token ID | 11 |
Feature activation | -4.10e-03 |
Pos logit contributions | |
sensation | +0.031 |
tune | +0.031 |
dolphin | +0.030 |
shark | +0.029 |
streamed | +0.028 |
Neg logit contributions | |
counts | -0.031 |
ends | -0.030 |
One | -0.029 |
Or | -0.029 |
Ready | -0.027 |
Token | a |
Token ID | 257 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
counts | +0.007 |
ends | +0.006 |
Or | +0.006 |
One | +0.006 |
Ready | +0.006 |
Neg logit contributions | |
sensation | -0.007 |
tune | -0.007 |
dolphin | -0.007 |
shark | -0.007 |
Dolphin | -0.006 |
Token | little |
Token ID | 1310 |
Feature activation | +3.13e-02 |
Pos logit contributions | |
One | +0.032 |
ends | +0.032 |
Or | +0.030 |
counts | +0.028 |
has | +0.027 |
Neg logit contributions | |
sensation | -0.029 |
tune | -0.026 |
streaming | -0.026 |
ened | -0.023 |
streamed | -0.023 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.31e-02 |
Pos logit contributions | |
sensation | +0.056 |
tune | +0.054 |
streaming | +0.048 |
streamed | +0.046 |
lull | +0.045 |
Neg logit contributions | |
One | -0.050 |
counts | -0.050 |
ends | -0.049 |
Or | -0.047 |
house | -0.043 |
Token | and |
Token ID | 290 |
Feature activation | -1.98e-02 |
Pos logit contributions | |
ends | +0.082 |
counts | +0.080 |
One | +0.077 |
Or | +0.073 |
Ready | +0.067 |
Neg logit contributions | |
tune | -0.124 |
sensation | -0.122 |
dolphin | -0.112 |
shark | -0.109 |
face | -0.105 |
Token | to |
Token ID | 284 |
Feature activation | +7.01e-03 |
Pos logit contributions | |
counts | +0.027 |
ends | +0.026 |
One | +0.025 |
Or | +0.025 |
Ready | +0.022 |
Neg logit contributions | |
sensation | -0.042 |
tune | -0.042 |
dolphin | -0.039 |
shark | -0.037 |
Dolphin | -0.036 |
Token | share |
Token ID | 2648 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
sensation | +0.014 |
tune | +0.014 |
dolphin | +0.013 |
shark | +0.013 |
streamed | +0.013 |
Neg logit contributions | |
One | -0.010 |
counts | -0.010 |
Or | -0.009 |
ends | -0.009 |
ours | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | -3.11e-02 |
Pos logit contributions | |
ends | +0.019 |
counts | +0.019 |
One | +0.018 |
Or | +0.017 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.024 |
sensation | -0.023 |
dolphin | -0.021 |
face | -0.021 |
shark | -0.021 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.20e-02 |
Pos logit contributions | |
counts | +0.056 |
ends | +0.053 |
belongs | +0.048 |
ours | +0.046 |
One | +0.045 |
Neg logit contributions | |
tune | -0.044 |
dolphin | -0.043 |
Sister | -0.043 |
Dolphin | -0.042 |
sensation | -0.041 |
Token | One |
Token ID | 3198 |
Feature activation | -1.57e-01 |
Pos logit contributions | |
ends | +0.055 |
counts | +0.052 |
One | +0.048 |
Or | +0.048 |
ours | +0.044 |
Neg logit contributions | |
sensation | -0.089 |
tune | -0.088 |
dolphin | -0.082 |
shark | -0.081 |
streamed | -0.077 |
Token | day |
Token ID | 1110 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
One | +0.251 |
counts | +0.248 |
ends | +0.230 |
Or | +0.211 |
has | +0.205 |
Neg logit contributions | |
tune | -0.305 |
sensation | -0.286 |
streaming | -0.274 |
dolphin | -0.264 |
streamed | -0.256 |
Token | , |
Token ID | 11 |
Feature activation | -2.12e-03 |
Pos logit contributions | |
tune | +0.025 |
sensation | +0.025 |
dolphin | +0.024 |
shark | +0.024 |
streamed | +0.023 |
Neg logit contributions | |
counts | -0.024 |
ends | -0.024 |
One | -0.022 |
Or | -0.022 |
Ready | -0.021 |
Token | a |
Token ID | 257 |
Feature activation | -1.76e-02 |
Pos logit contributions | |
counts | +0.003 |
ends | +0.003 |
Or | +0.003 |
One | +0.003 |
Ready | +0.003 |
Neg logit contributions | |
sensation | -0.004 |
tune | -0.004 |
dolphin | -0.003 |
shark | -0.003 |
chill | -0.003 |
Token | little |
Token ID | 1310 |
Feature activation | +3.04e-02 |
Pos logit contributions | |
One | +0.032 |
ends | +0.032 |
Or | +0.030 |
counts | +0.029 |
has | +0.027 |
Neg logit contributions | |
sensation | -0.030 |
tune | -0.028 |
streaming | -0.027 |
ened | -0.025 |
streamed | -0.025 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.11e-02 |
Pos logit contributions | |
sensation | +0.055 |
tune | +0.054 |
streaming | +0.047 |
streamed | +0.046 |
lull | +0.045 |
Neg logit contributions | |
One | -0.048 |
counts | -0.048 |
ends | -0.046 |
Or | -0.045 |
house | -0.042 |
Token | not |
Token ID | 407 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
tune | +0.018 |
sensation | +0.018 |
dolphin | +0.017 |
shark | +0.016 |
face | +0.016 |
Neg logit contributions | |
counts | -0.010 |
ends | -0.009 |
Or | -0.009 |
One | -0.009 |
Ready | -0.008 |
Token | bother |
Token ID | 11393 |
Feature activation | -2.27e-03 |
Pos logit contributions | |
counts | +0.015 |
ends | +0.014 |
One | +0.014 |
Or | +0.013 |
Ready | +0.012 |
Neg logit contributions | |
sensation | -0.022 |
tune | -0.022 |
dolphin | -0.020 |
shark | -0.020 |
Dolphin | -0.019 |
Token | anyone |
Token ID | 2687 |
Feature activation | -2.53e-02 |
Pos logit contributions | |
counts | +0.003 |
One | +0.003 |
ends | +0.003 |
Or | +0.003 |
ours | +0.003 |
Neg logit contributions | |
sensation | -0.005 |
tune | -0.005 |
grin | -0.004 |
streamed | -0.004 |
chill | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | -2.92e-02 |
Pos logit contributions | |
ends | +0.035 |
counts | +0.034 |
One | +0.034 |
Or | +0.032 |
Ready | +0.031 |
Neg logit contributions | |
tune | -0.053 |
sensation | -0.050 |
dolphin | -0.047 |
shark | -0.046 |
face | -0.044 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.48e-02 |
Pos logit contributions | |
counts | +0.045 |
ends | +0.044 |
One | +0.039 |
belongs | +0.038 |
Or | +0.037 |
Neg logit contributions | |
tune | -0.053 |
sensation | -0.051 |
dolphin | -0.049 |
shark | -0.047 |
Sister | -0.047 |
Token | One |
Token ID | 3198 |
Feature activation | -1.57e-01 |
Pos logit contributions | |
ends | +0.056 |
counts | +0.054 |
One | +0.050 |
Or | +0.048 |
ours | +0.043 |
Neg logit contributions | |
sensation | -0.098 |
tune | -0.098 |
dolphin | -0.091 |
shark | -0.089 |
Dolphin | -0.085 |
Token | day |
Token ID | 1110 |
Feature activation | +1.86e-02 |
Pos logit contributions | |
One | +0.245 |
counts | +0.241 |
ends | +0.222 |
Or | +0.198 |
has | +0.195 |
Neg logit contributions | |
tune | -0.306 |
sensation | -0.287 |
streaming | -0.279 |
dolphin | -0.266 |
streamed | -0.260 |
Token | , |
Token ID | 11 |
Feature activation | -1.65e-03 |
Pos logit contributions | |
tune | +0.038 |
sensation | +0.037 |
dolphin | +0.034 |
shark | +0.033 |
streaming | +0.032 |
Neg logit contributions | |
counts | -0.027 |
ends | -0.026 |
One | -0.025 |
Or | -0.024 |
Ready | -0.022 |
Token | a |
Token ID | 257 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
counts | +0.002 |
ends | +0.002 |
One | +0.002 |
Or | +0.002 |
Ready | +0.002 |
Neg logit contributions | |
tune | -0.003 |
sensation | -0.003 |
dolphin | -0.003 |
shark | -0.003 |
streamed | -0.003 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
One | +0.032 |
ends | +0.032 |
Or | +0.030 |
counts | +0.028 |
has | +0.028 |
Neg logit contributions | |
sensation | -0.029 |
tune | -0.026 |
streaming | -0.026 |
ened | -0.024 |
streamed | -0.024 |
Token | named |
Token ID | 3706 |
Feature activation | -5.12e-02 |
Pos logit contributions | |
tune | +0.025 |
sensation | +0.024 |
dolphin | +0.022 |
shark | +0.022 |
potato | +0.020 |
Neg logit contributions | |
ends | -0.022 |
counts | -0.021 |
One | -0.021 |
Or | -0.020 |
Ready | -0.019 |
Token | They |
Token ID | 1119 |
Feature activation | -2.06e-02 |
Pos logit contributions | |
counts | +0.064 |
ends | +0.063 |
One | +0.058 |
Or | +0.054 |
ours | +0.052 |
Neg logit contributions | |
tune | -0.094 |
sensation | -0.091 |
dolphin | -0.088 |
shark | -0.085 |
Dolphin | -0.082 |
Token | are |
Token ID | 389 |
Feature activation | -2.57e-02 |
Pos logit contributions | |
ends | +0.025 |
counts | +0.025 |
One | +0.024 |
Or | +0.023 |
Ready | +0.020 |
Neg logit contributions | |
tune | -0.046 |
sensation | -0.045 |
dolphin | -0.042 |
shark | -0.040 |
chill | -0.039 |
Token | lucky |
Token ID | 9670 |
Feature activation | -2.33e-02 |
Pos logit contributions | |
ends | +0.043 |
counts | +0.042 |
One | +0.041 |
Or | +0.039 |
Ready | +0.036 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.046 |
dolphin | -0.041 |
shark | -0.039 |
streaming | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | -4.57e-02 |
Pos logit contributions | |
ends | +0.031 |
counts | +0.030 |
One | +0.029 |
Or | +0.027 |
Ready | +0.025 |
Neg logit contributions | |
tune | -0.050 |
sensation | -0.049 |
dolphin | -0.046 |
shark | -0.046 |
potato | -0.043 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -5.07e-02 |
Pos logit contributions | |
counts | +0.063 |
ends | +0.063 |
One | +0.057 |
Or | +0.053 |
ours | +0.052 |
Neg logit contributions | |
tune | -0.091 |
sensation | -0.088 |
dolphin | -0.087 |
shark | -0.083 |
Dolphin | -0.080 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ends | +0.072 |
counts | +0.066 |
Or | +0.062 |
One | +0.061 |
ours | +0.059 |
Neg logit contributions | |
sensation | -0.102 |
tune | -0.099 |
dolphin | -0.096 |
shark | -0.095 |
Dolphin | -0.090 |
Token | day |
Token ID | 1110 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
One | +0.251 |
counts | +0.248 |
ends | +0.233 |
Or | +0.208 |
has | +0.204 |
Neg logit contributions | |
tune | -0.298 |
sensation | -0.288 |
streaming | -0.266 |
dolphin | -0.256 |
streamed | -0.254 |
Token | , |
Token ID | 11 |
Feature activation | -8.90e-03 |
Pos logit contributions | |
dolphin | +0.015 |
shark | +0.014 |
sensation | +0.014 |
tune | +0.014 |
streamed | +0.013 |
Neg logit contributions | |
counts | -0.019 |
ends | -0.019 |
Or | -0.018 |
One | -0.018 |
Ready | -0.017 |
Token | Tom |
Token ID | 4186 |
Feature activation | -3.41e-02 |
Pos logit contributions | |
counts | +0.015 |
ends | +0.014 |
Or | +0.013 |
One | +0.013 |
Ready | +0.013 |
Neg logit contributions | |
sensation | -0.015 |
tune | -0.015 |
dolphin | -0.014 |
shark | -0.014 |
face | -0.013 |
Token | and |
Token ID | 290 |
Feature activation | +1.44e-02 |
Pos logit contributions | |
ends | +0.048 |
One | +0.043 |
Ready | +0.040 |
Or | +0.040 |
counts | +0.038 |
Neg logit contributions | |
sensation | -0.068 |
tune | -0.066 |
dolphin | -0.066 |
shark | -0.066 |
potato | -0.062 |
Token | his |
Token ID | 465 |
Feature activation | -7.55e-03 |
Pos logit contributions | |
tune | +0.024 |
sensation | +0.023 |
dolphin | +0.021 |
shark | +0.020 |
streaming | +0.019 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.026 |
One | -0.025 |
Or | -0.025 |
Ready | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | -2.67e-02 |
Pos logit contributions | |
ends | +0.006 |
counts | +0.005 |
One | +0.005 |
Or | +0.005 |
someday | +0.004 |
Neg logit contributions | |
tune | -0.017 |
sensation | -0.017 |
dolphin | -0.016 |
shark | -0.016 |
Dolphin | -0.016 |
Token | They |
Token ID | 1119 |
Feature activation | -9.83e-03 |
Pos logit contributions | |
counts | +0.036 |
ends | +0.036 |
One | +0.034 |
Or | +0.032 |
Ready | +0.029 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.055 |
dolphin | -0.050 |
shark | -0.049 |
streaming | -0.047 |
Token | hoped |
Token ID | 10719 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
counts | +0.014 |
One | +0.013 |
ends | +0.013 |
Or | +0.012 |
cannot | +0.011 |
Neg logit contributions | |
tune | -0.021 |
sensation | -0.021 |
dolphin | -0.019 |
shark | -0.018 |
Dolphin | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -3.85e-02 |
Pos logit contributions | |
counts | +0.014 |
ends | +0.013 |
One | +0.012 |
Or | +0.012 |
Ready | +0.011 |
Neg logit contributions | |
tune | -0.026 |
sensation | -0.026 |
dolphin | -0.023 |
streamed | -0.022 |
shark | -0.022 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.75e-02 |
Pos logit contributions | |
counts | +0.057 |
ends | +0.055 |
One | +0.051 |
ours | +0.047 |
Or | +0.047 |
Neg logit contributions | |
tune | -0.074 |
sensation | -0.073 |
dolphin | -0.067 |
shark | -0.064 |
Dolphin | -0.063 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ends | +0.070 |
counts | +0.063 |
Or | +0.060 |
One | +0.060 |
ours | +0.060 |
Neg logit contributions | |
sensation | -0.093 |
tune | -0.088 |
shark | -0.086 |
dolphin | -0.086 |
Dolphin | -0.081 |
Token | day |
Token ID | 1110 |
Feature activation | +1.46e-02 |
Pos logit contributions | |
counts | +0.255 |
One | +0.247 |
ends | +0.231 |
Or | +0.212 |
has | +0.208 |
Neg logit contributions | |
tune | -0.298 |
sensation | -0.278 |
streaming | -0.260 |
dolphin | -0.252 |
streamed | -0.244 |
Token | , |
Token ID | 11 |
Feature activation | -3.98e-03 |
Pos logit contributions | |
sensation | +0.024 |
tune | +0.024 |
dolphin | +0.024 |
shark | +0.023 |
streamed | +0.022 |
Neg logit contributions | |
counts | -0.024 |
ends | -0.024 |
One | -0.023 |
Or | -0.022 |
Ready | -0.021 |
Token | a |
Token ID | 257 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
counts | +0.007 |
ends | +0.006 |
Or | +0.006 |
One | +0.006 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.007 |
tune | -0.007 |
dolphin | -0.006 |
shark | -0.006 |
streamed | -0.006 |
Token | baby |
Token ID | 5156 |
Feature activation | +2.96e-02 |
Pos logit contributions | |
One | +0.035 |
ends | +0.034 |
Or | +0.033 |
counts | +0.032 |
has | +0.030 |
Neg logit contributions | |
sensation | -0.032 |
tune | -0.029 |
streaming | -0.028 |
streamed | -0.026 |
ened | -0.026 |
Token | k |
Token ID | 479 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
tune | +0.058 |
sensation | +0.057 |
dolphin | +0.050 |
shark | +0.049 |
streaming | +0.047 |
Neg logit contributions | |
counts | -0.046 |
ends | -0.044 |
One | -0.043 |
Or | -0.041 |
Ready | -0.038 |
Token | They |
Token ID | 1119 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
ends | +0.039 |
counts | +0.039 |
One | +0.036 |
Or | +0.035 |
Ready | +0.031 |
Neg logit contributions | |
tune | -0.065 |
sensation | -0.065 |
dolphin | -0.059 |
shark | -0.058 |
streaming | -0.055 |
Token | were |
Token ID | 547 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
counts | +0.016 |
ends | +0.016 |
One | +0.015 |
Or | +0.014 |
has | +0.012 |
Neg logit contributions | |
tune | -0.021 |
sensation | -0.021 |
shark | -0.019 |
dolphin | -0.019 |
potato | -0.018 |
Token | mighty |
Token ID | 18680 |
Feature activation | -5.54e-02 |
Pos logit contributions | |
ends | +0.019 |
counts | +0.019 |
One | +0.018 |
Or | +0.018 |
Ready | +0.016 |
Neg logit contributions | |
tune | -0.022 |
sensation | -0.022 |
dolphin | -0.020 |
shark | -0.019 |
streaming | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -4.15e-02 |
Pos logit contributions | |
ends | +0.095 |
One | +0.094 |
counts | +0.092 |
Or | +0.088 |
Ready | +0.084 |
Neg logit contributions | |
tune | -0.097 |
sensation | -0.092 |
dolphin | -0.086 |
shark | -0.083 |
face | -0.078 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.54e-02 |
Pos logit contributions | |
ends | +0.050 |
counts | +0.050 |
One | +0.047 |
Or | +0.044 |
Ready | +0.040 |
Neg logit contributions | |
tune | -0.095 |
sensation | -0.093 |
dolphin | -0.086 |
shark | -0.083 |
streaming | -0.081 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ends | +0.057 |
counts | +0.054 |
One | +0.050 |
Or | +0.049 |
ours | +0.044 |
Neg logit contributions | |
sensation | -0.099 |
tune | -0.098 |
dolphin | -0.092 |
shark | -0.091 |
streamed | -0.086 |
Token | day |
Token ID | 1110 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
One | +0.241 |
counts | +0.240 |
ends | +0.224 |
Or | +0.202 |
has | +0.196 |
Neg logit contributions | |
tune | -0.308 |
sensation | -0.295 |
streaming | -0.276 |
dolphin | -0.264 |
streamed | -0.260 |
Token | , |
Token ID | 11 |
Feature activation | -7.36e-03 |
Pos logit contributions | |
sensation | +0.017 |
tune | +0.017 |
dolphin | +0.016 |
shark | +0.016 |
streamed | +0.015 |
Neg logit contributions | |
counts | -0.017 |
ends | -0.017 |
Or | -0.016 |
One | -0.016 |
Ready | -0.015 |
Token | a |
Token ID | 257 |
Feature activation | -1.86e-02 |
Pos logit contributions | |
counts | +0.012 |
ends | +0.012 |
Or | +0.011 |
One | +0.011 |
Ready | +0.010 |
Neg logit contributions | |
sensation | -0.013 |
tune | -0.012 |
dolphin | -0.011 |
shark | -0.011 |
streamed | -0.011 |
Token | boy |
Token ID | 2933 |
Feature activation | +6.62e-03 |
Pos logit contributions | |
One | +0.032 |
ends | +0.031 |
Or | +0.030 |
counts | +0.029 |
Ready | +0.027 |
Neg logit contributions | |
sensation | -0.034 |
tune | -0.031 |
streaming | -0.031 |
ened | -0.029 |
streamed | -0.029 |
Token | and |
Token ID | 290 |
Feature activation | +3.84e-04 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.012 |
dolphin | +0.011 |
shark | +0.011 |
streaming | +0.010 |
Neg logit contributions | |
ends | -0.010 |
counts | -0.010 |
One | -0.010 |
Or | -0.009 |
Ready | -0.009 |
Token | the |
Token ID | 262 |
Feature activation | -1.39e-02 |
Pos logit contributions | |
tune | +0.000 |
sensation | +0.000 |
dolphin | +0.000 |
shark | +0.000 |
streaming | +0.000 |
Neg logit contributions | |
counts | -0.000 |
ends | -0.000 |
One | -0.000 |
Or | -0.000 |
Ready | -0.000 |
Token | needle |
Token ID | 17598 |
Feature activation | +3.17e-02 |
Pos logit contributions | |
ends | +0.025 |
counts | +0.024 |
One | +0.023 |
Or | +0.022 |
Ready | +0.021 |
Neg logit contributions | |
tune | -0.024 |
sensation | -0.024 |
dolphin | -0.021 |
shark | -0.020 |
potato | -0.019 |
Token | was |
Token ID | 373 |
Feature activation | -3.34e-02 |
Pos logit contributions | |
tune | +0.060 |
sensation | +0.059 |
dolphin | +0.054 |
shark | +0.052 |
streaming | +0.050 |
Neg logit contributions | |
ends | -0.049 |
counts | -0.049 |
One | -0.047 |
Or | -0.045 |
Ready | -0.041 |
Token | . |
Token ID | 13 |
Feature activation | -3.78e-02 |
Pos logit contributions | |
ends | +0.048 |
counts | +0.047 |
One | +0.045 |
Or | +0.043 |
Ready | +0.040 |
Neg logit contributions | |
tune | -0.068 |
sensation | -0.068 |
dolphin | -0.062 |
shark | -0.061 |
streaming | -0.059 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -5.02e-02 |
Pos logit contributions | |
counts | +0.053 |
ends | +0.053 |
One | +0.050 |
Or | +0.046 |
Ready | +0.043 |
Neg logit contributions | |
tune | -0.076 |
sensation | -0.076 |
dolphin | -0.070 |
shark | -0.068 |
Dolphin | -0.064 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ends | +0.066 |
counts | +0.061 |
One | +0.058 |
Or | +0.057 |
ours | +0.052 |
Neg logit contributions | |
sensation | -0.107 |
tune | -0.105 |
dolphin | -0.099 |
shark | -0.098 |
streamed | -0.093 |
Token | day |
Token ID | 1110 |
Feature activation | +1.32e-02 |
Pos logit contributions | |
counts | +0.240 |
One | +0.237 |
ends | +0.214 |
Or | +0.200 |
has | +0.199 |
Neg logit contributions | |
tune | -0.311 |
sensation | -0.290 |
streaming | -0.275 |
dolphin | -0.263 |
tant | -0.259 |
Token | , |
Token ID | 11 |
Feature activation | -7.89e-03 |
Pos logit contributions | |
dolphin | +0.019 |
sensation | +0.019 |
shark | +0.019 |
tune | +0.019 |
streamed | +0.017 |
Neg logit contributions | |
counts | -0.024 |
ends | -0.023 |
One | -0.022 |
Or | -0.022 |
Ready | -0.021 |
Token | a |
Token ID | 257 |
Feature activation | -1.88e-02 |
Pos logit contributions | |
counts | +0.013 |
ends | +0.012 |
One | +0.012 |
Or | +0.012 |
Ready | +0.011 |
Neg logit contributions | |
sensation | -0.014 |
tune | -0.013 |
dolphin | -0.013 |
shark | -0.012 |
face | -0.011 |
Token | boy |
Token ID | 2933 |
Feature activation | +8.90e-03 |
Pos logit contributions | |
One | +0.035 |
ends | +0.033 |
Or | +0.033 |
counts | +0.031 |
Ready | +0.029 |
Neg logit contributions | |
sensation | -0.032 |
tune | -0.029 |
streaming | -0.029 |
streamed | -0.027 |
ened | -0.026 |
Token | named |
Token ID | 3706 |
Feature activation | -4.84e-02 |
Pos logit contributions | |
tune | +0.016 |
sensation | +0.015 |
dolphin | +0.014 |
shark | +0.014 |
streamed | +0.013 |
Neg logit contributions | |
ends | -0.015 |
counts | -0.014 |
One | -0.014 |
Or | -0.013 |
Ready | -0.013 |
Token | y |
Token ID | 331 |
Feature activation | -4.03e-02 |
Pos logit contributions | |
counts | +0.067 |
ends | +0.067 |
One | +0.064 |
Or | +0.061 |
Ready | +0.056 |
Neg logit contributions | |
tune | -0.088 |
sensation | -0.087 |
dolphin | -0.079 |
shark | -0.076 |
streaming | -0.073 |
Token | ummy |
Token ID | 13513 |
Feature activation | -6.95e-02 |
Pos logit contributions | |
ends | +0.099 |
counts | +0.097 |
One | +0.093 |
Or | +0.092 |
Ready | +0.089 |
Neg logit contributions | |
tune | -0.042 |
sensation | -0.042 |
dolphin | -0.034 |
shark | -0.032 |
potato | -0.031 |
Token | too |
Token ID | 1165 |
Feature activation | -6.52e-02 |
Pos logit contributions | |
ends | +0.102 |
One | +0.097 |
counts | +0.095 |
Or | +0.092 |
Ready | +0.088 |
Neg logit contributions | |
tune | -0.143 |
sensation | -0.141 |
dolphin | -0.129 |
shark | -0.127 |
potato | -0.122 |
Token | . |
Token ID | 13 |
Feature activation | -2.66e-02 |
Pos logit contributions | |
ends | +0.097 |
counts | +0.096 |
Or | +0.091 |
One | +0.090 |
Ready | +0.087 |
Neg logit contributions | |
tune | -0.126 |
sensation | -0.121 |
dolphin | -0.117 |
shark | -0.113 |
potato | -0.108 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.55e-02 |
Pos logit contributions | |
counts | +0.042 |
ends | +0.041 |
One | +0.037 |
belongs | +0.034 |
ours | +0.034 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.045 |
dolphin | -0.044 |
shark | -0.042 |
Dolphin | -0.042 |
Token | One |
Token ID | 3198 |
Feature activation | -1.56e-01 |
Pos logit contributions | |
ends | +0.065 |
counts | +0.061 |
Or | +0.057 |
One | +0.056 |
ours | +0.055 |
Neg logit contributions | |
sensation | -0.090 |
tune | -0.087 |
dolphin | -0.084 |
shark | -0.083 |
Dolphin | -0.079 |
Token | day |
Token ID | 1110 |
Feature activation | +1.70e-02 |
Pos logit contributions | |
counts | +0.260 |
One | +0.259 |
ends | +0.245 |
Or | +0.224 |
has | +0.210 |
Neg logit contributions | |
tune | -0.286 |
sensation | -0.272 |
streaming | -0.254 |
dolphin | -0.246 |
streamed | -0.240 |
Token | , |
Token ID | 11 |
Feature activation | -2.45e-03 |
Pos logit contributions | |
tune | +0.034 |
sensation | +0.034 |
dolphin | +0.031 |
shark | +0.030 |
streamed | +0.029 |
Neg logit contributions | |
counts | -0.024 |
ends | -0.024 |
One | -0.023 |
Or | -0.022 |
Ready | -0.020 |
Token | a |
Token ID | 257 |
Feature activation | -1.76e-02 |
Pos logit contributions | |
counts | +0.004 |
ends | +0.004 |
One | +0.003 |
Or | +0.003 |
Ready | +0.003 |
Neg logit contributions | |
sensation | -0.005 |
tune | -0.005 |
dolphin | -0.004 |
shark | -0.004 |
face | -0.004 |
Token | woman |
Token ID | 2415 |
Feature activation | +6.86e-03 |
Pos logit contributions | |
One | +0.033 |
ends | +0.033 |
Or | +0.031 |
counts | +0.030 |
has | +0.029 |
Neg logit contributions | |
sensation | -0.029 |
tune | -0.026 |
streaming | -0.026 |
streamed | -0.023 |
ened | -0.023 |
Token | went |
Token ID | 1816 |
Feature activation | -4.71e-03 |
Pos logit contributions | |
tune | +0.010 |
sensation | +0.009 |
dolphin | +0.009 |
shark | +0.008 |
streamed | +0.008 |
Neg logit contributions | |
ends | -0.014 |
One | -0.013 |
ours | -0.012 |
Ready | -0.012 |
counts | -0.012 |
Token | she |
Token ID | 673 |
Feature activation | +3.85e-03 |
Pos logit contributions | |
tune | +0.029 |
sensation | +0.029 |
dolphin | +0.026 |
shark | +0.025 |
chill | +0.025 |
Neg logit contributions | |
counts | -0.013 |
ends | -0.013 |
One | -0.011 |
Or | -0.011 |
Ready | -0.009 |
Token | was |
Token ID | 373 |
Feature activation | -3.69e-03 |
Pos logit contributions | |
tune | +0.007 |
sensation | +0.007 |
dolphin | +0.006 |
shark | +0.006 |
streaming | +0.006 |
Neg logit contributions | |
counts | -0.006 |
ends | -0.006 |
One | -0.006 |
Or | -0.006 |
Ready | -0.005 |
Token | smiling |
Token ID | 16755 |
Feature activation | +7.09e-03 |
Pos logit contributions | |
ends | +0.006 |
counts | +0.005 |
One | +0.005 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.007 |
sensation | -0.007 |
dolphin | -0.006 |
shark | -0.006 |
potato | -0.006 |
Token | all |
Token ID | 477 |
Feature activation | -2.48e-02 |
Pos logit contributions | |
sensation | +0.017 |
tune | +0.017 |
chill | +0.015 |
dolphin | +0.015 |
shark | +0.015 |
Neg logit contributions | |
ends | -0.007 |
counts | -0.006 |
One | -0.006 |
Or | -0.006 |
Maybe | -0.005 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -3.92e-02 |
Pos logit contributions | |
counts | +0.042 |
ends | +0.040 |
One | +0.040 |
Or | +0.036 |
ours | +0.034 |
Neg logit contributions | |
tune | -0.044 |
sensation | -0.043 |
dolphin | -0.040 |
shark | -0.039 |
potato | -0.037 |
Token | One |
Token ID | 3198 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
ends | +0.048 |
counts | +0.046 |
One | +0.043 |
Or | +0.042 |
ours | +0.037 |
Neg logit contributions | |
sensation | -0.087 |
tune | -0.087 |
dolphin | -0.080 |
shark | -0.078 |
streaming | -0.075 |
Token | day |
Token ID | 1110 |
Feature activation | +1.37e-02 |
Pos logit contributions | |
One | +0.264 |
counts | +0.262 |
ends | +0.252 |
Or | +0.226 |
� | +0.213 |
Neg logit contributions | |
tune | -0.281 |
sensation | -0.268 |
dolphin | -0.245 |
shark | -0.239 |
streaming | -0.236 |
Token | , |
Token ID | 11 |
Feature activation | -5.39e-04 |
Pos logit contributions | |
tune | +0.029 |
sensation | +0.029 |
dolphin | +0.026 |
shark | +0.026 |
streaming | +0.025 |
Neg logit contributions | |
ends | -0.018 |
counts | -0.018 |
One | -0.017 |
Or | -0.016 |
Ready | -0.015 |
Token | Sally |
Token ID | 25737 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.001 |
sensation | -0.001 |
dolphin | -0.001 |
shark | -0.001 |
streaming | -0.001 |
Token | was |
Token ID | 373 |
Feature activation | -2.44e-03 |
Pos logit contributions | |
ends | +0.018 |
counts | +0.017 |
One | +0.017 |
Or | +0.016 |
Ready | +0.015 |
Neg logit contributions | |
tune | -0.020 |
sensation | -0.020 |
dolphin | -0.018 |
shark | -0.018 |
potato | -0.017 |
Token | outside |
Token ID | 2354 |
Feature activation | -3.74e-03 |
Pos logit contributions | |
counts | +0.004 |
ends | +0.004 |
One | +0.004 |
Or | +0.004 |
Ready | +0.003 |
Neg logit contributions | |
tune | -0.004 |
sensation | -0.004 |
dolphin | -0.004 |
shark | -0.004 |
streaming | -0.004 |
Token | might |
Token ID | 1244 |
Feature activation | -3.59e-02 |
Pos logit contributions | |
tune | +0.025 |
sensation | +0.025 |
dolphin | +0.023 |
shark | +0.023 |
face | +0.022 |
Neg logit contributions | |
ends | -0.012 |
counts | -0.012 |
One | -0.012 |
Or | -0.011 |
Ready | -0.010 |
Token | get |
Token ID | 651 |
Feature activation | -2.55e-02 |
Pos logit contributions | |
ends | +0.056 |
counts | +0.055 |
Or | +0.050 |
One | +0.050 |
Ready | +0.047 |
Neg logit contributions | |
tune | -0.069 |
sensation | -0.067 |
dolphin | -0.060 |
shark | -0.058 |
face | -0.058 |
Token | hurt |
Token ID | 5938 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
counts | +0.041 |
ends | +0.040 |
One | +0.039 |
Or | +0.038 |
Ready | +0.035 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.046 |
dolphin | -0.042 |
shark | -0.041 |
streaming | -0.040 |
Token | . |
Token ID | 13 |
Feature activation | -2.83e-02 |
Pos logit contributions | |
ends | +0.033 |
counts | +0.031 |
One | +0.028 |
Or | +0.026 |
Ready | +0.026 |
Neg logit contributions | |
tune | -0.059 |
sensation | -0.057 |
dolphin | -0.053 |
shark | -0.052 |
potato | -0.050 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.35e-02 |
Pos logit contributions | |
ends | +0.042 |
counts | +0.040 |
ours | +0.035 |
One | +0.034 |
belongs | +0.033 |
Neg logit contributions | |
tune | -0.050 |
sensation | -0.048 |
Sister | -0.047 |
Potato | -0.046 |
Dolphin | -0.045 |
Token | One |
Token ID | 3198 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
ends | +0.057 |
counts | +0.054 |
One | +0.050 |
Or | +0.050 |
ours | +0.045 |
Neg logit contributions | |
sensation | -0.092 |
tune | -0.091 |
dolphin | -0.085 |
shark | -0.084 |
Dolphin | -0.079 |
Token | day |
Token ID | 1110 |
Feature activation | +1.86e-02 |
Pos logit contributions | |
One | +0.258 |
counts | +0.257 |
ends | +0.237 |
Or | +0.219 |
has | +0.205 |
Neg logit contributions | |
tune | -0.288 |
sensation | -0.274 |
streaming | -0.254 |
dolphin | -0.251 |
shark | -0.242 |
Token | , |
Token ID | 11 |
Feature activation | -3.91e-03 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.037 |
dolphin | +0.034 |
shark | +0.033 |
streaming | +0.031 |
Neg logit contributions | |
counts | -0.027 |
ends | -0.027 |
One | -0.025 |
Or | -0.024 |
Ready | -0.022 |
Token | a |
Token ID | 257 |
Feature activation | -1.51e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
One | +0.005 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.008 |
tune | -0.008 |
dolphin | -0.007 |
shark | -0.007 |
chill | -0.006 |
Token | little |
Token ID | 1310 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
One | +0.028 |
ends | +0.028 |
Or | +0.026 |
counts | +0.025 |
has | +0.024 |
Neg logit contributions | |
sensation | -0.025 |
tune | -0.023 |
streaming | -0.023 |
ened | -0.021 |
streamed | -0.021 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
sensation | +0.056 |
tune | +0.054 |
streaming | +0.048 |
streamed | +0.047 |
lull | +0.045 |
Neg logit contributions | |
One | -0.048 |
counts | -0.047 |
ends | -0.045 |
Or | -0.045 |
house | -0.041 |
Token | protect |
Token ID | 1805 |
Feature activation | +2.41e-02 |
Pos logit contributions | |
counts | +0.053 |
ends | +0.049 |
One | +0.045 |
Or | +0.044 |
has | +0.042 |
Neg logit contributions | |
tune | -0.059 |
sensation | -0.058 |
dolphin | -0.051 |
shark | -0.050 |
streaming | -0.049 |
Token | their |
Token ID | 511 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
sensation | +0.048 |
tune | +0.048 |
dolphin | +0.041 |
shark | +0.041 |
chill | +0.041 |
Neg logit contributions | |
counts | -0.035 |
ends | -0.034 |
One | -0.033 |
Or | -0.031 |
ours | -0.030 |
Token | skin |
Token ID | 4168 |
Feature activation | +2.63e-02 |
Pos logit contributions | |
counts | +0.028 |
ends | +0.028 |
One | +0.026 |
Or | +0.025 |
Ready | +0.023 |
Neg logit contributions | |
sensation | -0.031 |
tune | -0.031 |
dolphin | -0.028 |
shark | -0.027 |
streaming | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | -3.52e-02 |
Pos logit contributions | |
tune | +0.062 |
sensation | +0.061 |
dolphin | +0.057 |
shark | +0.055 |
streaming | +0.053 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.028 |
One | -0.027 |
Or | -0.025 |
Ready | -0.022 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.14e-02 |
Pos logit contributions | |
counts | +0.057 |
ends | +0.053 |
One | +0.050 |
belongs | +0.049 |
ours | +0.049 |
Neg logit contributions | |
tune | -0.057 |
sensation | -0.057 |
dolphin | -0.053 |
Sister | -0.052 |
shark | -0.051 |
Token | One |
Token ID | 3198 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
ends | +0.058 |
counts | +0.054 |
Or | +0.050 |
One | +0.050 |
ours | +0.048 |
Neg logit contributions | |
sensation | -0.083 |
tune | -0.081 |
dolphin | -0.077 |
shark | -0.077 |
Dolphin | -0.073 |
Token | day |
Token ID | 1110 |
Feature activation | +1.86e-02 |
Pos logit contributions | |
counts | +0.262 |
One | +0.260 |
ends | +0.248 |
Or | +0.225 |
has | +0.211 |
Neg logit contributions | |
tune | -0.285 |
sensation | -0.271 |
streaming | -0.250 |
dolphin | -0.247 |
shark | -0.240 |
Token | , |
Token ID | 11 |
Feature activation | -4.73e-04 |
Pos logit contributions | |
sensation | +0.031 |
tune | +0.031 |
dolphin | +0.030 |
shark | +0.029 |
streamed | +0.028 |
Neg logit contributions | |
counts | -0.031 |
ends | -0.030 |
One | -0.029 |
Or | -0.029 |
Ready | -0.027 |
Token | Lily |
Token ID | 20037 |
Feature activation | -4.32e-02 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
sensation | -0.001 |
tune | -0.001 |
dolphin | -0.001 |
shark | -0.001 |
chill | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | +2.16e-02 |
Pos logit contributions | |
ends | +0.066 |
counts | +0.063 |
One | +0.062 |
Or | +0.059 |
Ready | +0.056 |
Neg logit contributions | |
tune | -0.084 |
sensation | -0.083 |
dolphin | -0.075 |
shark | -0.072 |
potato | -0.070 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.98e-02 |
Pos logit contributions | |
tune | +0.047 |
sensation | +0.047 |
dolphin | +0.040 |
shark | +0.040 |
chill | +0.039 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.028 |
One | -0.027 |
Or | -0.026 |
ours | -0.024 |
Token | what |
Token ID | 644 |
Feature activation | +2.94e-02 |
Pos logit contributions | |
sensation | +0.015 |
tune | +0.015 |
dolphin | +0.013 |
shark | +0.013 |
streamed | +0.013 |
Neg logit contributions | |
counts | -0.008 |
ends | -0.008 |
One | -0.008 |
Or | -0.008 |
Ready | -0.007 |
Token | you |
Token ID | 345 |
Feature activation | +2.37e-02 |
Pos logit contributions | |
tune | +0.071 |
sensation | +0.070 |
dolphin | +0.067 |
shark | +0.066 |
potato | +0.064 |
Neg logit contributions | |
ends | -0.030 |
counts | -0.027 |
One | -0.027 |
Or | -0.024 |
Ready | -0.022 |
Token | have |
Token ID | 423 |
Feature activation | -3.31e-02 |
Pos logit contributions | |
tune | +0.054 |
sensation | +0.053 |
dolphin | +0.049 |
shark | +0.048 |
streaming | +0.046 |
Neg logit contributions | |
ends | -0.028 |
counts | -0.028 |
One | -0.027 |
Or | -0.025 |
Ready | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | -2.37e-02 |
Pos logit contributions | |
counts | +0.044 |
ends | +0.044 |
One | +0.042 |
Or | +0.040 |
Ready | +0.039 |
Neg logit contributions | |
tune | -0.069 |
sensation | -0.067 |
dolphin | -0.062 |
shark | -0.060 |
face | -0.059 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.20e-02 |
Pos logit contributions | |
ends | +0.035 |
counts | +0.034 |
ours | +0.030 |
One | +0.029 |
belongs | +0.029 |
Neg logit contributions | |
tune | -0.040 |
Potato | -0.039 |
Sister | -0.039 |
Dolphin | -0.039 |
sensation | -0.038 |
Token | One |
Token ID | 3198 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
ends | +0.053 |
counts | +0.051 |
One | +0.047 |
Or | +0.046 |
ours | +0.041 |
Neg logit contributions | |
sensation | -0.091 |
tune | -0.091 |
dolphin | -0.084 |
shark | -0.083 |
Dolphin | -0.079 |
Token | day |
Token ID | 1110 |
Feature activation | +1.84e-02 |
Pos logit contributions | |
One | +0.247 |
counts | +0.244 |
ends | +0.222 |
Or | +0.204 |
has | +0.197 |
Neg logit contributions | |
tune | -0.299 |
sensation | -0.282 |
streaming | -0.269 |
dolphin | -0.260 |
streamed | -0.252 |
Token | , |
Token ID | 11 |
Feature activation | -4.94e-03 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.036 |
dolphin | +0.034 |
shark | +0.033 |
streaming | +0.031 |
Neg logit contributions | |
counts | -0.026 |
ends | -0.026 |
One | -0.025 |
Or | -0.024 |
Ready | -0.022 |
Token | a |
Token ID | 257 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
counts | +0.008 |
ends | +0.007 |
Or | +0.007 |
One | +0.007 |
Ready | +0.006 |
Neg logit contributions | |
sensation | -0.009 |
tune | -0.009 |
dolphin | -0.008 |
shark | -0.008 |
chill | -0.008 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.41e-02 |
Pos logit contributions | |
One | +0.030 |
ends | +0.029 |
Or | +0.028 |
counts | +0.026 |
has | +0.026 |
Neg logit contributions | |
sensation | -0.028 |
tune | -0.026 |
streaming | -0.025 |
ened | -0.023 |
streamed | -0.023 |
Token | and |
Token ID | 290 |
Feature activation | +6.34e-03 |
Pos logit contributions | |
tune | +0.025 |
sensation | +0.025 |
dolphin | +0.023 |
shark | +0.022 |
potato | +0.021 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.023 |
One | -0.022 |
Or | -0.021 |
Ready | -0.020 |
Token | of |
Token ID | 286 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
ends | +0.042 |
counts | +0.042 |
One | +0.040 |
Or | +0.038 |
Ready | +0.036 |
Neg logit contributions | |
tune | -0.044 |
sensation | -0.043 |
dolphin | -0.039 |
shark | -0.038 |
potato | -0.035 |
Token | their |
Token ID | 511 |
Feature activation | +4.74e-04 |
Pos logit contributions | |
tune | +0.022 |
sensation | +0.022 |
dolphin | +0.019 |
shark | +0.019 |
chill | +0.018 |
Neg logit contributions | |
ends | -0.021 |
counts | -0.021 |
One | -0.020 |
Or | -0.019 |
Ready | -0.018 |
Token | work |
Token ID | 670 |
Feature activation | +2.45e-03 |
Pos logit contributions | |
tune | +0.001 |
sensation | +0.001 |
dolphin | +0.001 |
shark | +0.001 |
streaming | +0.001 |
Neg logit contributions | |
counts | -0.001 |
ends | -0.001 |
One | -0.001 |
Or | -0.001 |
Ready | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | -3.09e-02 |
Pos logit contributions | |
tune | +0.006 |
sensation | +0.006 |
dolphin | +0.005 |
shark | +0.005 |
streaming | +0.005 |
Neg logit contributions | |
counts | -0.003 |
ends | -0.003 |
One | -0.003 |
Or | -0.003 |
Ready | -0.002 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.42e-02 |
Pos logit contributions | |
counts | +0.047 |
ends | +0.045 |
One | +0.041 |
Or | +0.038 |
ours | +0.038 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.056 |
dolphin | -0.055 |
shark | -0.052 |
Dolphin | -0.051 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.056 |
counts | +0.053 |
One | +0.049 |
Or | +0.048 |
ours | +0.044 |
Neg logit contributions | |
sensation | -0.096 |
tune | -0.095 |
dolphin | -0.090 |
shark | -0.088 |
streamed | -0.084 |
Token | day |
Token ID | 1110 |
Feature activation | +1.31e-02 |
Pos logit contributions | |
One | +0.242 |
counts | +0.242 |
ends | +0.225 |
Or | +0.203 |
has | +0.197 |
Neg logit contributions | |
tune | -0.301 |
sensation | -0.286 |
streaming | -0.266 |
dolphin | -0.257 |
tant | -0.253 |
Token | , |
Token ID | 11 |
Feature activation | -6.04e-03 |
Pos logit contributions | |
dolphin | +0.020 |
sensation | +0.020 |
shark | +0.019 |
tune | +0.019 |
streamed | +0.018 |
Neg logit contributions | |
counts | -0.023 |
ends | -0.022 |
Or | -0.022 |
One | -0.022 |
Ready | -0.021 |
Token | Lily |
Token ID | 20037 |
Feature activation | -4.63e-02 |
Pos logit contributions | |
counts | +0.010 |
ends | +0.010 |
Or | +0.009 |
One | +0.009 |
Ready | +0.009 |
Neg logit contributions | |
sensation | -0.010 |
tune | -0.010 |
dolphin | -0.009 |
shark | -0.009 |
face | -0.008 |
Token | and |
Token ID | 290 |
Feature activation | +1.58e-02 |
Pos logit contributions | |
ends | +0.076 |
One | +0.071 |
counts | +0.071 |
Or | +0.067 |
Ready | +0.065 |
Neg logit contributions | |
tune | -0.085 |
sensation | -0.084 |
dolphin | -0.077 |
shark | -0.074 |
potato | -0.071 |
Token | Max |
Token ID | 5436 |
Feature activation | +7.49e-03 |
Pos logit contributions | |
tune | +0.031 |
sensation | +0.030 |
dolphin | +0.027 |
shark | +0.026 |
streaming | +0.025 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.024 |
One | -0.023 |
Or | -0.022 |
Ready | -0.021 |
Token | cake |
Token ID | 12187 |
Feature activation | -5.52e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.034 |
One | +0.031 |
Or | +0.030 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.064 |
sensation | -0.064 |
dolphin | -0.059 |
shark | -0.057 |
streaming | -0.055 |
Token | is |
Token ID | 318 |
Feature activation | -7.64e-02 |
Pos logit contributions | |
ends | +0.072 |
counts | +0.070 |
One | +0.070 |
Or | +0.064 |
Ready | +0.062 |
Neg logit contributions | |
tune | -0.118 |
sensation | -0.118 |
dolphin | -0.107 |
shark | -0.104 |
face | -0.103 |
Token | inside |
Token ID | 2641 |
Feature activation | -5.53e-02 |
Pos logit contributions | |
counts | +0.113 |
ends | +0.111 |
One | +0.107 |
Or | +0.102 |
Ready | +0.095 |
Neg logit contributions | |
tune | -0.152 |
sensation | -0.151 |
dolphin | -0.137 |
shark | -0.132 |
potato | -0.127 |
Token | . |
Token ID | 13 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
ends | +0.069 |
counts | +0.069 |
One | +0.065 |
Or | +0.061 |
Ready | +0.055 |
Neg logit contributions | |
tune | -0.123 |
sensation | -0.121 |
dolphin | -0.111 |
shark | -0.108 |
streaming | -0.104 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -3.66e-02 |
Pos logit contributions | |
ends | +0.047 |
counts | +0.046 |
One | +0.042 |
ours | +0.040 |
belongs | +0.039 |
Neg logit contributions | |
tune | -0.045 |
sensation | -0.044 |
dolphin | -0.044 |
Sister | -0.044 |
shark | -0.042 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.050 |
counts | +0.047 |
One | +0.044 |
Or | +0.044 |
ours | +0.041 |
Neg logit contributions | |
sensation | -0.074 |
tune | -0.073 |
dolphin | -0.068 |
shark | -0.068 |
Dolphin | -0.064 |
Token | day |
Token ID | 1110 |
Feature activation | +2.53e-02 |
Pos logit contributions | |
counts | +0.274 |
One | +0.270 |
ends | +0.262 |
Or | +0.244 |
Ready | +0.225 |
Neg logit contributions | |
tune | -0.267 |
sensation | -0.254 |
dolphin | -0.233 |
streaming | -0.225 |
shark | -0.223 |
Token | , |
Token ID | 11 |
Feature activation | -3.97e-03 |
Pos logit contributions | |
tune | +0.051 |
sensation | +0.050 |
dolphin | +0.046 |
shark | +0.045 |
streaming | +0.043 |
Neg logit contributions | |
counts | -0.036 |
ends | -0.036 |
One | -0.034 |
Or | -0.033 |
Ready | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
Or | +0.006 |
One | +0.006 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.007 |
tune | -0.007 |
dolphin | -0.007 |
shark | -0.007 |
face | -0.006 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.42e-02 |
Pos logit contributions | |
One | +0.032 |
ends | +0.032 |
Or | +0.030 |
counts | +0.028 |
has | +0.028 |
Neg logit contributions | |
sensation | -0.028 |
tune | -0.026 |
streaming | -0.025 |
ened | -0.023 |
streamed | -0.023 |
Token | named |
Token ID | 3706 |
Feature activation | -4.59e-02 |
Pos logit contributions | |
tune | +0.026 |
sensation | +0.025 |
dolphin | +0.023 |
shark | +0.022 |
potato | +0.021 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.023 |
One | -0.022 |
Or | -0.021 |
Ready | -0.020 |
Token | They |
Token ID | 1119 |
Feature activation | -9.09e-03 |
Pos logit contributions | |
counts | +0.031 |
ends | +0.030 |
One | +0.028 |
Or | +0.026 |
belongs | +0.025 |
Neg logit contributions | |
tune | -0.041 |
sensation | -0.040 |
dolphin | -0.038 |
shark | -0.036 |
Dolphin | -0.035 |
Token | are |
Token ID | 389 |
Feature activation | -2.40e-02 |
Pos logit contributions | |
ends | +0.012 |
counts | +0.012 |
One | +0.011 |
Or | +0.011 |
Ready | +0.010 |
Neg logit contributions | |
tune | -0.020 |
sensation | -0.019 |
dolphin | -0.018 |
shark | -0.017 |
chill | -0.017 |
Token | happy |
Token ID | 3772 |
Feature activation | -1.58e-02 |
Pos logit contributions | |
counts | +0.041 |
ends | +0.041 |
One | +0.040 |
Or | +0.038 |
Ready | +0.035 |
Neg logit contributions | |
tune | -0.042 |
sensation | -0.041 |
dolphin | -0.037 |
shark | -0.036 |
streaming | -0.034 |
Token | . |
Token ID | 13 |
Feature activation | -2.87e-02 |
Pos logit contributions | |
ends | +0.025 |
One | +0.024 |
counts | +0.024 |
Or | +0.022 |
Ready | +0.021 |
Neg logit contributions | |
tune | -0.031 |
sensation | -0.029 |
dolphin | -0.028 |
shark | -0.028 |
face | -0.025 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.50e-02 |
Pos logit contributions | |
counts | +0.044 |
ends | +0.043 |
One | +0.039 |
belongs | +0.037 |
ours | +0.036 |
Neg logit contributions | |
tune | -0.052 |
dolphin | -0.049 |
sensation | -0.049 |
shark | -0.046 |
Dolphin | -0.046 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.066 |
counts | +0.061 |
Or | +0.057 |
One | +0.057 |
ours | +0.055 |
Neg logit contributions | |
sensation | -0.087 |
tune | -0.085 |
shark | -0.082 |
dolphin | -0.081 |
Dolphin | -0.077 |
Token | day |
Token ID | 1110 |
Feature activation | +1.75e-02 |
Pos logit contributions | |
One | +0.244 |
counts | +0.240 |
ends | +0.221 |
Or | +0.200 |
has | +0.197 |
Neg logit contributions | |
tune | -0.298 |
sensation | -0.285 |
streaming | -0.276 |
streamed | -0.259 |
dolphin | -0.257 |
Token | , |
Token ID | 11 |
Feature activation | -3.75e-03 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.028 |
dolphin | +0.028 |
shark | +0.027 |
streamed | +0.025 |
Neg logit contributions | |
counts | -0.029 |
ends | -0.029 |
One | -0.028 |
Or | -0.028 |
Ready | -0.026 |
Token | a |
Token ID | 257 |
Feature activation | -1.60e-02 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
Or | +0.006 |
One | +0.006 |
Ready | +0.005 |
Neg logit contributions | |
sensation | -0.006 |
tune | -0.006 |
dolphin | -0.006 |
shark | -0.006 |
face | -0.005 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.36e-02 |
Pos logit contributions | |
One | +0.030 |
ends | +0.030 |
Or | +0.028 |
counts | +0.028 |
has | +0.026 |
Neg logit contributions | |
sensation | -0.026 |
tune | -0.024 |
streaming | -0.024 |
streamed | -0.022 |
ened | -0.021 |
Token | named |
Token ID | 3706 |
Feature activation | -4.74e-02 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.022 |
dolphin | +0.021 |
shark | +0.020 |
potato | +0.019 |
Neg logit contributions | |
ends | -0.024 |
One | -0.022 |
counts | -0.022 |
Or | -0.021 |
Ready | -0.021 |
Token | other |
Token ID | 584 |
Feature activation | -6.65e-02 |
Pos logit contributions | |
sensation | +0.132 |
tune | +0.130 |
dolphin | +0.116 |
shark | +0.115 |
streaming | +0.114 |
Neg logit contributions | |
counts | -0.089 |
ends | -0.087 |
One | -0.087 |
Or | -0.079 |
Ready | -0.074 |
Token | and |
Token ID | 290 |
Feature activation | -7.01e-03 |
Pos logit contributions | |
ends | +0.082 |
counts | +0.082 |
One | +0.076 |
Or | +0.073 |
Ready | +0.064 |
Neg logit contributions | |
sensation | -0.148 |
tune | -0.147 |
dolphin | -0.135 |
shark | -0.131 |
streaming | -0.127 |
Token | waited |
Token ID | 13488 |
Feature activation | -3.61e-02 |
Pos logit contributions | |
counts | +0.009 |
ends | +0.008 |
Or | +0.007 |
has | +0.007 |
Maybe | +0.006 |
Neg logit contributions | |
tune | -0.016 |
sensation | -0.015 |
andel | -0.015 |
opher | -0.014 |
Dolphin | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -2.40e-02 |
Pos logit contributions | |
ends | +0.039 |
counts | +0.035 |
One | +0.034 |
Or | +0.033 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.086 |
sensation | -0.085 |
dolphin | -0.080 |
shark | -0.078 |
Dolphin | -0.076 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -3.87e-02 |
Pos logit contributions | |
counts | +0.039 |
ends | +0.037 |
ours | +0.034 |
belongs | +0.034 |
One | +0.034 |
Neg logit contributions | |
tune | -0.040 |
sensation | -0.039 |
dolphin | -0.038 |
Dolphin | -0.036 |
shark | -0.036 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.052 |
counts | +0.049 |
One | +0.045 |
Or | +0.045 |
ours | +0.042 |
Neg logit contributions | |
sensation | -0.080 |
tune | -0.079 |
dolphin | -0.074 |
shark | -0.074 |
Dolphin | -0.070 |
Token | day |
Token ID | 1110 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
counts | +0.247 |
One | +0.245 |
ends | +0.227 |
Or | +0.208 |
has | +0.203 |
Neg logit contributions | |
tune | -0.295 |
sensation | -0.275 |
streaming | -0.260 |
dolphin | -0.254 |
shark | -0.244 |
Token | , |
Token ID | 11 |
Feature activation | -7.94e-04 |
Pos logit contributions | |
tune | +0.038 |
sensation | +0.037 |
dolphin | +0.035 |
shark | +0.034 |
streamed | +0.032 |
Neg logit contributions | |
counts | -0.028 |
ends | -0.028 |
One | -0.027 |
Or | -0.026 |
Ready | -0.024 |
Token | a |
Token ID | 257 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.002 |
sensation | -0.002 |
dolphin | -0.001 |
shark | -0.001 |
streamed | -0.001 |
Token | big |
Token ID | 1263 |
Feature activation | +1.30e-02 |
Pos logit contributions | |
One | +0.030 |
ends | +0.030 |
Or | +0.028 |
counts | +0.027 |
has | +0.026 |
Neg logit contributions | |
sensation | -0.028 |
tune | -0.025 |
streaming | -0.024 |
ened | -0.022 |
streamed | -0.022 |
Token | storm |
Token ID | 6388 |
Feature activation | +5.48e-02 |
Pos logit contributions | |
sensation | +0.026 |
tune | +0.025 |
streaming | +0.023 |
streamed | +0.023 |
lull | +0.021 |
Neg logit contributions | |
counts | -0.018 |
One | -0.017 |
has | -0.017 |
house | -0.016 |
Or | -0.016 |
Token | belong |
Token ID | 5594 |
Feature activation | -2.71e-02 |
Pos logit contributions | |
ends | +0.007 |
counts | +0.007 |
One | +0.007 |
Or | +0.006 |
Ready | +0.006 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.012 |
dolphin | -0.011 |
shark | -0.010 |
streaming | -0.010 |
Token | to |
Token ID | 284 |
Feature activation | -2.14e-02 |
Pos logit contributions | |
One | +0.052 |
ends | +0.050 |
counts | +0.050 |
Or | +0.048 |
Ready | +0.048 |
Neg logit contributions | |
tune | -0.036 |
shark | -0.035 |
sensation | -0.033 |
dolphin | -0.033 |
face | -0.031 |
Token | them |
Token ID | 606 |
Feature activation | -1.54e-02 |
Pos logit contributions | |
counts | +0.032 |
ends | +0.032 |
One | +0.030 |
Or | +0.029 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.042 |
sensation | -0.042 |
dolphin | -0.038 |
shark | -0.037 |
streaming | -0.035 |
Token | . |
Token ID | 13 |
Feature activation | -2.13e-02 |
Pos logit contributions | |
counts | +0.022 |
ends | +0.022 |
One | +0.021 |
Or | +0.019 |
Ready | +0.019 |
Neg logit contributions | |
tune | -0.030 |
sensation | -0.029 |
dolphin | -0.027 |
shark | -0.027 |
face | -0.026 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.23e-02 |
Pos logit contributions | |
counts | +0.039 |
ends | +0.036 |
One | +0.032 |
belongs | +0.032 |
ours | +0.030 |
Neg logit contributions | |
tune | -0.030 |
Sister | -0.029 |
sensation | -0.029 |
dolphin | -0.028 |
Ned | -0.028 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.056 |
counts | +0.054 |
One | +0.050 |
Or | +0.049 |
ours | +0.045 |
Neg logit contributions | |
sensation | -0.088 |
tune | -0.087 |
dolphin | -0.082 |
shark | -0.081 |
Dolphin | -0.077 |
Token | day |
Token ID | 1110 |
Feature activation | +2.07e-02 |
Pos logit contributions | |
One | +0.251 |
counts | +0.246 |
ends | +0.229 |
Or | +0.210 |
has | +0.199 |
Neg logit contributions | |
tune | -0.292 |
sensation | -0.276 |
streaming | -0.260 |
dolphin | -0.252 |
streamed | -0.245 |
Token | , |
Token ID | 11 |
Feature activation | -8.79e-04 |
Pos logit contributions | |
tune | +0.039 |
sensation | +0.039 |
dolphin | +0.036 |
shark | +0.035 |
streamed | +0.034 |
Neg logit contributions | |
counts | -0.031 |
ends | -0.031 |
One | -0.029 |
Or | -0.029 |
Ready | -0.026 |
Token | a |
Token ID | 257 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
Or | +0.001 |
One | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
sensation | -0.002 |
tune | -0.002 |
dolphin | -0.001 |
shark | -0.001 |
face | -0.001 |
Token | little |
Token ID | 1310 |
Feature activation | +3.15e-02 |
Pos logit contributions | |
One | +0.029 |
ends | +0.029 |
Or | +0.027 |
counts | +0.026 |
Ready | +0.025 |
Neg logit contributions | |
sensation | -0.025 |
tune | -0.023 |
streaming | -0.022 |
streamed | -0.020 |
ened | -0.020 |
Token | boy |
Token ID | 2933 |
Feature activation | +1.33e-02 |
Pos logit contributions | |
sensation | +0.057 |
tune | +0.055 |
streaming | +0.048 |
streamed | +0.047 |
lull | +0.046 |
Neg logit contributions | |
One | -0.050 |
counts | -0.050 |
ends | -0.048 |
Or | -0.047 |
house | -0.043 |
Token | love |
Token ID | 1842 |
Feature activation | +4.70e-03 |
Pos logit contributions | |
ends | +0.025 |
counts | +0.025 |
One | +0.024 |
Or | +0.023 |
Ready | +0.021 |
Neg logit contributions | |
tune | -0.034 |
sensation | -0.034 |
dolphin | -0.032 |
shark | -0.030 |
face | -0.030 |
Token | to |
Token ID | 284 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
tune | +0.009 |
sensation | +0.009 |
dolphin | +0.008 |
shark | +0.008 |
streamed | +0.007 |
Neg logit contributions | |
counts | -0.007 |
ends | -0.007 |
One | -0.007 |
Or | -0.007 |
Ready | -0.006 |
Token | paint |
Token ID | 7521 |
Feature activation | -5.22e-02 |
Pos logit contributions | |
counts | +0.027 |
One | +0.026 |
ends | +0.025 |
Or | +0.024 |
Ready | +0.021 |
Neg logit contributions | |
sensation | -0.034 |
tune | -0.034 |
dolphin | -0.030 |
shark | -0.029 |
streaming | -0.029 |
Token | . |
Token ID | 13 |
Feature activation | -3.58e-02 |
Pos logit contributions | |
ends | +0.073 |
counts | +0.070 |
One | +0.068 |
Or | +0.065 |
Ready | +0.061 |
Neg logit contributions | |
tune | -0.108 |
sensation | -0.108 |
dolphin | -0.097 |
shark | -0.094 |
face | -0.092 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -4.35e-02 |
Pos logit contributions | |
ends | +0.053 |
counts | +0.053 |
One | +0.048 |
ours | +0.046 |
Or | +0.045 |
Neg logit contributions | |
tune | -0.066 |
sensation | -0.065 |
dolphin | -0.062 |
shark | -0.059 |
Dolphin | -0.057 |
Token | One |
Token ID | 3198 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
ends | +0.055 |
counts | +0.052 |
One | +0.048 |
Or | +0.047 |
ours | +0.042 |
Neg logit contributions | |
sensation | -0.095 |
tune | -0.094 |
dolphin | -0.089 |
shark | -0.087 |
Dolphin | -0.083 |
Token | day |
Token ID | 1110 |
Feature activation | +1.21e-02 |
Pos logit contributions | |
One | +0.254 |
counts | +0.253 |
ends | +0.242 |
Or | +0.220 |
has | +0.207 |
Neg logit contributions | |
tune | -0.285 |
sensation | -0.271 |
streaming | -0.247 |
dolphin | -0.246 |
shark | -0.243 |
Token | , |
Token ID | 11 |
Feature activation | -4.37e-03 |
Pos logit contributions | |
dolphin | +0.019 |
sensation | +0.019 |
tune | +0.018 |
shark | +0.018 |
streamed | +0.017 |
Neg logit contributions | |
counts | -0.021 |
ends | -0.020 |
Or | -0.020 |
One | -0.020 |
Ready | -0.019 |
Token | a |
Token ID | 257 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
counts | +0.007 |
ends | +0.007 |
One | +0.007 |
Or | +0.006 |
Ready | +0.006 |
Neg logit contributions | |
sensation | -0.008 |
tune | -0.007 |
dolphin | -0.007 |
shark | -0.007 |
face | -0.006 |
Token | boy |
Token ID | 2933 |
Feature activation | +5.99e-03 |
Pos logit contributions | |
One | +0.036 |
ends | +0.035 |
Or | +0.034 |
counts | +0.032 |
has | +0.030 |
Neg logit contributions | |
sensation | -0.035 |
streaming | -0.033 |
tune | -0.032 |
streamed | -0.030 |
ened | -0.030 |
Token | went |
Token ID | 1816 |
Feature activation | -7.93e-03 |
Pos logit contributions | |
tune | +0.010 |
sensation | +0.010 |
dolphin | +0.009 |
shark | +0.009 |
potato | +0.009 |
Neg logit contributions | |
ends | -0.010 |
One | -0.010 |
counts | -0.010 |
Or | -0.009 |
Ready | -0.009 |
Token | a |
Token ID | 257 |
Feature activation | +8.79e-03 |
Pos logit contributions | |
tune | +0.042 |
sensation | +0.040 |
urg | +0.036 |
dolphin | +0.036 |
uc | +0.036 |
Neg logit contributions | |
One | -0.019 |
counts | -0.019 |
ends | -0.017 |
� | -0.015 |
ours | -0.014 |
Token | big |
Token ID | 1263 |
Feature activation | +3.17e-02 |
Pos logit contributions | |
sensation | +0.016 |
tune | +0.015 |
urg | +0.015 |
ened | +0.014 |
tunes | +0.014 |
Neg logit contributions | |
One | -0.013 |
ends | -0.013 |
Or | -0.012 |
� | -0.012 |
counts | -0.011 |
Token | smile |
Token ID | 8212 |
Feature activation | +8.90e-02 |
Pos logit contributions | |
sensation | +0.047 |
tune | +0.046 |
tunes | +0.040 |
urg | +0.039 |
Dolphin | +0.039 |
Neg logit contributions | |
counts | -0.057 |
One | -0.055 |
has | -0.053 |
ends | -0.052 |
Or | -0.052 |
Token | on |
Token ID | 319 |
Feature activation | +3.75e-02 |
Pos logit contributions | |
sensation | +0.219 |
tune | +0.214 |
shark | +0.210 |
dolphin | +0.206 |
tant | +0.205 |
Neg logit contributions | |
ends | -0.062 |
One | -0.055 |
counts | -0.051 |
� | -0.050 |
" | -0.047 |
Token | his |
Token ID | 465 |
Feature activation | +2.71e-02 |
Pos logit contributions | |
sensation | +0.069 |
tune | +0.066 |
Dolphin | +0.064 |
dolphin | +0.064 |
toddler | +0.063 |
Neg logit contributions | |
One | -0.054 |
counts | -0.047 |
" | -0.044 |
ends | -0.043 |
Or | -0.041 |
Token | face |
Token ID | 1986 |
Feature activation | +1.20e-01 |
Pos logit contributions | |
sensation | +0.030 |
Dolphin | +0.028 |
streamed | +0.028 |
Ned | +0.027 |
shark | +0.027 |
Neg logit contributions | |
Or | -0.053 |
ends | -0.052 |
counts | -0.050 |
� | -0.049 |
One | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | +2.49e-02 |
Pos logit contributions | |
sensation | +0.314 |
tune | +0.313 |
dolphin | +0.307 |
shark | +0.305 |
Dolphin | +0.295 |
Neg logit contributions | |
ends | -0.077 |
One | -0.074 |
counts | -0.074 |
� | -0.057 |
Or | -0.056 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.71e-02 |
Pos logit contributions | |
tune | +0.057 |
sensation | +0.056 |
dolphin | +0.052 |
shark | +0.050 |
chill | +0.049 |
Neg logit contributions | |
ends | -0.030 |
counts | -0.030 |
One | -0.029 |
Or | -0.028 |
Ready | -0.025 |
Token | Once |
Token ID | 7454 |
Feature activation | +4.82e-03 |
Pos logit contributions | |
ends | +0.035 |
counts | +0.033 |
One | +0.031 |
Or | +0.031 |
ours | +0.028 |
Neg logit contributions | |
sensation | -0.058 |
tune | -0.058 |
dolphin | -0.054 |
shark | -0.053 |
streamed | -0.051 |
Token | there |
Token ID | 612 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
tune | +0.010 |
sensation | +0.010 |
dolphin | +0.009 |
shark | +0.009 |
streaming | +0.009 |
Neg logit contributions | |
counts | -0.006 |
ends | -0.006 |
One | -0.006 |
Or | -0.006 |
Ready | -0.005 |
Token | was |
Token ID | 373 |
Feature activation | +1.84e-02 |
Pos logit contributions | |
counts | +0.018 |
ends | +0.018 |
One | +0.017 |
Or | +0.016 |
Ready | +0.015 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.026 |
dolphin | -0.024 |
shark | -0.023 |
potato | -0.022 |
Token | with |
Token ID | 351 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
tune | +0.073 |
sensation | +0.071 |
dolphin | +0.065 |
shark | +0.063 |
streaming | +0.061 |
Neg logit contributions | |
ends | -0.046 |
counts | -0.046 |
One | -0.044 |
Or | -0.042 |
Ready | -0.038 |
Token | a |
Token ID | 257 |
Feature activation | +8.05e-03 |
Pos logit contributions | |
tune | +0.036 |
sensation | +0.035 |
dolphin | +0.031 |
shark | +0.031 |
Dolphin | +0.030 |
Neg logit contributions | |
counts | -0.024 |
ends | -0.023 |
One | -0.023 |
Or | -0.021 |
Ready | -0.019 |
Token | smile |
Token ID | 8212 |
Feature activation | +8.86e-02 |
Pos logit contributions | |
sensation | +0.014 |
tune | +0.014 |
streaming | +0.012 |
streamed | +0.012 |
urg | +0.012 |
Neg logit contributions | |
ends | -0.013 |
One | -0.013 |
counts | -0.012 |
Or | -0.012 |
Ready | -0.011 |
Token | on |
Token ID | 319 |
Feature activation | +3.58e-02 |
Pos logit contributions | |
sensation | +0.208 |
tune | +0.206 |
shark | +0.194 |
dolphin | +0.189 |
tant | +0.187 |
Neg logit contributions | |
ends | -0.086 |
counts | -0.082 |
One | -0.075 |
has | -0.066 |
� | -0.065 |
Token | her |
Token ID | 607 |
Feature activation | +7.89e-03 |
Pos logit contributions | |
sensation | +0.084 |
tune | +0.083 |
shark | +0.081 |
dolphin | +0.078 |
Dolphin | +0.076 |
Neg logit contributions | |
One | -0.038 |
counts | -0.035 |
ends | -0.034 |
Or | -0.030 |
She | -0.027 |
Token | face |
Token ID | 1986 |
Feature activation | +1.13e-01 |
Pos logit contributions | |
sensation | +0.009 |
streamed | +0.008 |
tant | +0.008 |
toddler | +0.008 |
Sister | +0.007 |
Neg logit contributions | |
ends | -0.015 |
counts | -0.014 |
Or | -0.014 |
next | -0.014 |
has | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | +1.37e-02 |
Pos logit contributions | |
tune | +0.286 |
sensation | +0.285 |
shark | +0.270 |
dolphin | +0.269 |
Dolphin | +0.258 |
Neg logit contributions | |
counts | -0.096 |
ends | -0.095 |
One | -0.088 |
Or | -0.079 |
� | -0.066 |
Token | What |
Token ID | 1867 |
Feature activation | +2.71e-02 |
Pos logit contributions | |
tune | +0.029 |
sensation | +0.028 |
dolphin | +0.026 |
shark | +0.025 |
streaming | +0.024 |
Neg logit contributions | |
counts | -0.018 |
ends | -0.018 |
One | -0.017 |
Or | -0.016 |
ours | -0.015 |
Token | an |
Token ID | 281 |
Feature activation | -8.50e-03 |
Pos logit contributions | |
tune | +0.051 |
sensation | +0.051 |
dolphin | +0.047 |
shark | +0.045 |
face | +0.043 |
Neg logit contributions | |
ends | -0.041 |
counts | -0.040 |
One | -0.038 |
Or | -0.037 |
Ready | -0.035 |
Token | amazing |
Token ID | 4998 |
Feature activation | +1.24e-02 |
Pos logit contributions | |
ends | +0.017 |
counts | +0.016 |
One | +0.015 |
Or | +0.015 |
Ready | +0.014 |
Neg logit contributions | |
tune | -0.013 |
sensation | -0.013 |
dolphin | -0.012 |
shark | -0.011 |
face | -0.011 |
Token | adventure |
Token ID | 8855 |
Feature activation | +5.43e-03 |
Pos logit contributions | |
sensation | +0.021 |
tune | +0.020 |
streaming | +0.019 |
streamed | +0.019 |
dolphin | +0.017 |
Neg logit contributions | |
counts | -0.021 |
One | -0.020 |
ends | -0.019 |
Or | -0.018 |
house | -0.017 |
Token | stubborn |
Token ID | 24444 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
ends | +0.045 |
counts | +0.045 |
One | +0.043 |
Or | +0.041 |
Ready | +0.038 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.056 |
dolphin | -0.050 |
shark | -0.049 |
streaming | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | -2.98e-02 |
Pos logit contributions | |
ends | +0.018 |
counts | +0.017 |
One | +0.017 |
Or | +0.016 |
Ready | +0.016 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.027 |
dolphin | -0.024 |
shark | -0.023 |
face | -0.023 |
Token | She |
Token ID | 1375 |
Feature activation | -3.48e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.033 |
One | +0.031 |
ours | +0.027 |
Or | +0.026 |
Neg logit contributions | |
sensation | -0.067 |
tune | -0.067 |
dolphin | -0.061 |
face | -0.059 |
shark | -0.059 |
Token | throws |
Token ID | 12542 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
ends | +0.043 |
One | +0.043 |
counts | +0.039 |
Or | +0.038 |
ours | +0.037 |
Neg logit contributions | |
sensation | -0.069 |
tune | -0.067 |
face | -0.066 |
dolphin | -0.065 |
shark | -0.064 |
Token | a |
Token ID | 257 |
Feature activation | +1.09e-04 |
Pos logit contributions | |
counts | +0.020 |
ends | +0.020 |
One | +0.019 |
Or | +0.018 |
Ready | +0.017 |
Neg logit contributions | |
tune | -0.024 |
sensation | -0.023 |
dolphin | -0.021 |
shark | -0.020 |
streaming | -0.019 |
Token | tant |
Token ID | 24246 |
Feature activation | +1.68e-01 |
Pos logit contributions | |
sensation | +0.000 |
tune | +0.000 |
streaming | +0.000 |
dolphin | +0.000 |
Dolphin | +0.000 |
Neg logit contributions | |
ends | -0.000 |
One | -0.000 |
counts | -0.000 |
Or | -0.000 |
Ready | -0.000 |
Token | rum |
Token ID | 6582 |
Feature activation | -8.92e-03 |
Pos logit contributions | |
tune | +0.256 |
sensation | +0.250 |
dolphin | +0.221 |
shark | +0.211 |
streaming | +0.199 |
Neg logit contributions | |
ends | -0.327 |
counts | -0.326 |
One | -0.314 |
Or | -0.304 |
Ready | -0.286 |
Token | . |
Token ID | 13 |
Feature activation | -3.18e-02 |
Pos logit contributions | |
counts | +0.013 |
One | +0.013 |
ends | +0.013 |
Ready | +0.012 |
Or | +0.012 |
Neg logit contributions | |
sensation | -0.017 |
tune | -0.017 |
dolphin | -0.015 |
face | -0.015 |
streaming | -0.015 |
Token | She |
Token ID | 1375 |
Feature activation | -3.37e-02 |
Pos logit contributions | |
ends | +0.036 |
counts | +0.036 |
One | +0.033 |
Or | +0.029 |
ours | +0.028 |
Neg logit contributions | |
sensation | -0.072 |
tune | -0.072 |
dolphin | -0.066 |
shark | -0.064 |
face | -0.063 |
Token | kicks |
Token ID | 17364 |
Feature activation | -3.17e-03 |
Pos logit contributions | |
ends | +0.043 |
One | +0.043 |
counts | +0.039 |
Or | +0.038 |
Ready | +0.036 |
Neg logit contributions | |
sensation | -0.068 |
tune | -0.067 |
shark | -0.065 |
dolphin | -0.065 |
face | -0.063 |
Token | and |
Token ID | 290 |
Feature activation | +9.20e-03 |
Pos logit contributions | |
counts | +0.005 |
ends | +0.005 |
One | +0.005 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.006 |
sensation | -0.006 |
dolphin | -0.005 |
shark | -0.005 |
streaming | -0.005 |
Token | and |
Token ID | 290 |
Feature activation | -9.38e-04 |
Pos logit contributions | |
ends | +0.011 |
counts | +0.011 |
One | +0.011 |
Or | +0.010 |
Ready | +0.009 |
Neg logit contributions | |
tune | -0.019 |
sensation | -0.019 |
dolphin | -0.017 |
shark | -0.017 |
chill | -0.016 |
Token | jump |
Token ID | 4391 |
Feature activation | -1.86e-02 |
Pos logit contributions | |
counts | +0.001 |
ends | +0.001 |
One | +0.001 |
Or | +0.001 |
Ready | +0.001 |
Neg logit contributions | |
tune | -0.002 |
sensation | -0.002 |
dolphin | -0.002 |
shark | -0.002 |
streaming | -0.002 |
Token | in |
Token ID | 287 |
Feature activation | +3.79e-03 |
Pos logit contributions | |
ends | +0.031 |
counts | +0.031 |
One | +0.030 |
Or | +0.028 |
Ready | +0.026 |
Neg logit contributions | |
tune | -0.034 |
sensation | -0.033 |
dolphin | -0.030 |
shark | -0.029 |
streaming | -0.027 |
Token | the |
Token ID | 262 |
Feature activation | -4.89e-03 |
Pos logit contributions | |
sensation | +0.005 |
tune | +0.005 |
dolphin | +0.004 |
shark | +0.004 |
face | +0.004 |
Neg logit contributions | |
ends | -0.009 |
counts | -0.008 |
Or | -0.008 |
One | -0.008 |
Ready | -0.007 |
Token | p |
Token ID | 279 |
Feature activation | +1.56e-02 |
Pos logit contributions | |
counts | +0.008 |
ends | +0.008 |
One | +0.008 |
Or | +0.008 |
Ready | +0.007 |
Neg logit contributions | |
tune | -0.009 |
sensation | -0.008 |
dolphin | -0.008 |
shark | -0.007 |
streaming | -0.007 |
Token | udd |
Token ID | 4185 |
Feature activation | +1.09e-01 |
Pos logit contributions | |
sensation | +0.018 |
tune | +0.018 |
dolphin | +0.016 |
face | +0.016 |
lee | +0.015 |
Neg logit contributions | |
counts | -0.034 |
One | -0.032 |
ends | -0.031 |
Or | -0.031 |
Ready | -0.031 |
Token | les |
Token ID | 829 |
Feature activation | +3.18e-02 |
Pos logit contributions | |
tune | +0.170 |
sensation | +0.165 |
dolphin | +0.143 |
shark | +0.142 |
Dolphin | +0.141 |
Neg logit contributions | |
counts | -0.197 |
One | -0.192 |
ends | -0.186 |
belongs | -0.176 |
ours | -0.168 |
Token | . |
Token ID | 13 |
Feature activation | +1.76e-03 |
Pos logit contributions | |
tune | +0.072 |
sensation | +0.071 |
dolphin | +0.065 |
shark | +0.063 |
streaming | +0.061 |
Neg logit contributions | |
counts | -0.039 |
ends | -0.039 |
One | -0.037 |
Or | -0.035 |
Ready | -0.032 |
Token | One |
Token ID | 1881 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
face | +0.003 |
stomach | +0.003 |
dolphin | +0.003 |
Neg logit contributions | |
ends | -0.002 |
counts | -0.002 |
ours | -0.002 |
Or | -0.002 |
ider | -0.002 |
Token | day |
Token ID | 1110 |
Feature activation | +4.05e-02 |
Pos logit contributions | |
One | +0.162 |
counts | +0.153 |
ends | +0.147 |
Or | +0.137 |
house | +0.128 |
Neg logit contributions | |
tune | -0.191 |
sensation | -0.186 |
streaming | -0.170 |
dolphin | -0.167 |
shark | -0.166 |
Token | , |
Token ID | 11 |
Feature activation | +2.57e-02 |
Pos logit contributions | |
tune | +0.085 |
sensation | +0.084 |
dolphin | +0.077 |
shark | +0.074 |
streaming | +0.071 |
Neg logit contributions | |
ends | -0.056 |
counts | -0.056 |
One | -0.053 |
Or | -0.050 |
Ready | -0.046 |
Token | work |
Token ID | 670 |
Feature activation | +2.07e-02 |
Pos logit contributions | |
sensation | +0.142 |
tune | +0.141 |
dolphin | +0.131 |
shark | +0.125 |
streaming | +0.121 |
Neg logit contributions | |
counts | -0.121 |
ends | -0.115 |
One | -0.113 |
Or | -0.105 |
Ready | -0.096 |
Token | . |
Token ID | 13 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
tune | +0.050 |
sensation | +0.049 |
dolphin | +0.046 |
shark | +0.045 |
potato | +0.043 |
Neg logit contributions | |
counts | -0.022 |
ends | -0.021 |
One | -0.019 |
Or | -0.018 |
Ready | -0.015 |
Token | |
Token ID | 220 |
Feature activation | +9.25e-03 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.028 |
dolphin | +0.025 |
shark | +0.024 |
chill | +0.024 |
Neg logit contributions | |
ends | -0.019 |
counts | -0.018 |
One | -0.018 |
Or | -0.017 |
Ready | -0.016 |
Token |
|
Token ID | 198 |
Feature activation | +1.62e-02 |
Pos logit contributions | |
tune | +0.019 |
sensation | +0.019 |
dolphin | +0.017 |
shark | +0.017 |
Dolphin | +0.016 |
Neg logit contributions | |
counts | -0.012 |
ends | -0.012 |
One | -0.011 |
Or | -0.010 |
ours | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | +2.64e-02 |
Pos logit contributions | |
tune | +0.038 |
sensation | +0.038 |
dolphin | +0.035 |
shark | +0.034 |
streaming | +0.033 |
Neg logit contributions | |
ends | -0.018 |
counts | -0.018 |
One | -0.017 |
Or | -0.016 |
Ready | -0.014 |
Token | Ol |
Token ID | 30098 |
Feature activation | +1.19e-01 |
Pos logit contributions | |
tune | +0.054 |
sensation | +0.053 |
dolphin | +0.048 |
shark | +0.047 |
streaming | +0.045 |
Neg logit contributions | |
ends | -0.038 |
counts | -0.038 |
One | -0.036 |
Or | -0.034 |
Ready | -0.032 |
Token | iver |
Token ID | 1428 |
Feature activation | +5.86e-02 |
Pos logit contributions | |
tune | +0.205 |
dolphin | +0.199 |
sensation | +0.197 |
shark | +0.192 |
Dolphin | +0.177 |
Neg logit contributions | |
counts | -0.201 |
One | -0.199 |
ends | -0.194 |
Or | -0.190 |
Ready | -0.168 |
Token | smiled |
Token ID | 13541 |
Feature activation | -1.93e-02 |
Pos logit contributions | |
sensation | +0.116 |
andel | +0.115 |
tune | +0.115 |
Dolphin | +0.110 |
dolphin | +0.110 |
Neg logit contributions | |
counts | -0.061 |
ends | -0.060 |
has | -0.058 |
house | -0.056 |
� | -0.055 |
Token | and |
Token ID | 290 |
Feature activation | +1.16e-02 |
Pos logit contributions | |
ends | +0.014 |
Or | +0.012 |
One | +0.011 |
counts | +0.011 |
� | +0.010 |
Neg logit contributions | |
sensation | -0.051 |
tune | -0.048 |
shark | -0.046 |
roach | -0.046 |
dolphin | -0.046 |
Token | be |
Token ID | 307 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
sensation | +0.027 |
tune | +0.026 |
andel | +0.025 |
tant | +0.024 |
ascal | +0.024 |
Neg logit contributions | |
counts | -0.013 |
ends | -0.013 |
� | -0.012 |
has | -0.011 |
Or | -0.010 |
Token | amed |
Token ID | 2434 |
Feature activation | -2.94e-03 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.027 |
dolphin | +0.025 |
shark | +0.025 |
potato | +0.023 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.027 |
One | -0.027 |
Or | -0.025 |
Ready | -0.023 |
Token | a |
Token ID | 257 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
ends | +0.024 |
counts | +0.023 |
One | +0.022 |
Or | +0.021 |
Ready | +0.019 |
Neg logit contributions | |
tune | -0.037 |
sensation | -0.037 |
dolphin | -0.033 |
shark | -0.033 |
chill | -0.031 |
Token | few |
Token ID | 1178 |
Feature activation | +5.64e-02 |
Pos logit contributions | |
sensation | +0.038 |
tune | +0.035 |
dolphin | +0.033 |
streaming | +0.033 |
shark | +0.032 |
Neg logit contributions | |
One | -0.032 |
ends | -0.030 |
counts | -0.028 |
Or | -0.027 |
Ready | -0.026 |
Token | days |
Token ID | 1528 |
Feature activation | +3.88e-02 |
Pos logit contributions | |
tune | +0.098 |
sensation | +0.097 |
dolphin | +0.085 |
shark | +0.082 |
chill | +0.079 |
Neg logit contributions | |
ends | -0.098 |
counts | -0.094 |
One | -0.092 |
Or | -0.090 |
Ready | -0.084 |
Token | , |
Token ID | 11 |
Feature activation | +2.06e-02 |
Pos logit contributions | |
tune | +0.077 |
sensation | +0.076 |
dolphin | +0.069 |
shark | +0.067 |
streaming | +0.065 |
Neg logit contributions | |
ends | -0.056 |
counts | -0.056 |
One | -0.054 |
Or | -0.051 |
Ready | -0.047 |
Token | they |
Token ID | 484 |
Feature activation | +2.89e-02 |
Pos logit contributions | |
tune | +0.053 |
sensation | +0.052 |
dolphin | +0.047 |
shark | +0.047 |
chill | +0.046 |
Neg logit contributions | |
ends | -0.019 |
counts | -0.017 |
One | -0.017 |
Or | -0.016 |
house | -0.015 |
Token | got |
Token ID | 1392 |
Feature activation | +2.47e-02 |
Pos logit contributions | |
tune | +0.064 |
sensation | +0.063 |
dolphin | +0.057 |
shark | +0.056 |
Dolphin | +0.053 |
Neg logit contributions | |
counts | -0.038 |
ends | -0.038 |
One | -0.036 |
Or | -0.034 |
house | -0.031 |
Token | the |
Token ID | 262 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
tune | +0.043 |
sensation | +0.043 |
dolphin | +0.039 |
shark | +0.037 |
face | +0.036 |
Neg logit contributions | |
counts | -0.042 |
ends | -0.041 |
One | -0.040 |
Or | -0.038 |
Ready | -0.036 |
Token | book |
Token ID | 1492 |
Feature activation | -1.21e-02 |
Pos logit contributions | |
ends | +0.027 |
counts | +0.026 |
One | +0.025 |
Or | +0.025 |
atters | +0.023 |
Neg logit contributions | |
tune | -0.029 |
sensation | -0.029 |
face | -0.025 |
dolphin | -0.024 |
lull | -0.024 |
Token | back |
Token ID | 736 |
Feature activation | -3.99e-03 |
Pos logit contributions | |
counts | +0.019 |
ends | +0.019 |
One | +0.019 |
Ready | +0.017 |
Or | +0.017 |
Neg logit contributions | |
tune | -0.022 |
sensation | -0.021 |
face | -0.019 |
streaming | -0.018 |
chill | -0.018 |
Token | and |
Token ID | 290 |
Feature activation | +1.15e-03 |
Pos logit contributions | |
counts | +0.006 |
ends | +0.006 |
One | +0.006 |
Or | +0.005 |
Ready | +0.005 |
Neg logit contributions | |
tune | -0.008 |
sensation | -0.008 |
dolphin | -0.007 |
shark | -0.007 |
streaming | -0.007 |
Token | it |
Token ID | 340 |
Feature activation | +2.71e-02 |
Pos logit contributions | |
tune | +0.002 |
sensation | +0.002 |
dolphin | +0.002 |
shark | +0.002 |
Dolphin | +0.002 |
Neg logit contributions | |
counts | -0.002 |
ends | -0.002 |
One | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | night |
Token ID | 1755 |
Feature activation | -8.43e-03 |
Pos logit contributions | |
counts | +0.036 |
ends | +0.035 |
One | +0.034 |
Or | +0.033 |
Ready | +0.031 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.026 |
dolphin | -0.024 |
shark | -0.022 |
potato | -0.021 |
Token | , |
Token ID | 11 |
Feature activation | +8.37e-03 |
Pos logit contributions | |
ends | +0.010 |
counts | +0.009 |
One | +0.009 |
Or | +0.008 |
house | +0.007 |
Neg logit contributions | |
sensation | -0.019 |
tune | -0.019 |
dolphin | -0.018 |
shark | -0.018 |
streamed | -0.017 |
Token | but |
Token ID | 475 |
Feature activation | +2.75e-02 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.013 |
dolphin | +0.013 |
shark | +0.012 |
Dolphin | +0.012 |
Neg logit contributions | |
ends | -0.015 |
counts | -0.015 |
Or | -0.014 |
One | -0.014 |
Ready | -0.012 |
Token | eventually |
Token ID | 4191 |
Feature activation | -4.94e-02 |
Pos logit contributions | |
tune | +0.065 |
sensation | +0.064 |
dolphin | +0.059 |
shark | +0.058 |
potato | +0.056 |
Neg logit contributions | |
ends | -0.031 |
counts | -0.031 |
One | -0.029 |
Or | -0.027 |
Ready | -0.024 |
Token | fell |
Token ID | 3214 |
Feature activation | +1.10e-03 |
Pos logit contributions | |
counts | +0.063 |
ends | +0.062 |
One | +0.058 |
Or | +0.054 |
Ready | +0.049 |
Neg logit contributions | |
tune | -0.111 |
sensation | -0.109 |
dolphin | -0.100 |
shark | -0.097 |
chill | -0.093 |
Token | asleep |
Token ID | 16039 |
Feature activation | +3.97e-02 |
Pos logit contributions | |
tune | +0.002 |
sensation | +0.002 |
dolphin | +0.001 |
shark | +0.001 |
potato | +0.001 |
Neg logit contributions | |
counts | -0.002 |
ends | -0.002 |
One | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | . |
Token ID | 13 |
Feature activation | +2.22e-02 |
Pos logit contributions | |
sensation | +0.092 |
tune | +0.092 |
dolphin | +0.091 |
shark | +0.087 |
potato | +0.087 |
Neg logit contributions | |
ends | -0.041 |
counts | -0.038 |
One | -0.036 |
Or | -0.036 |
house | -0.031 |
Token | The |
Token ID | 383 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
tune | +0.045 |
sensation | +0.045 |
dolphin | +0.040 |
shark | +0.039 |
streaming | +0.038 |
Neg logit contributions | |
counts | -0.032 |
ends | -0.032 |
One | -0.030 |
Or | -0.028 |
Ready | -0.026 |
Token | monster |
Token ID | 9234 |
Feature activation | +4.05e-02 |
Pos logit contributions | |
tune | +0.025 |
sensation | +0.025 |
dolphin | +0.022 |
shark | +0.021 |
streaming | +0.021 |
Neg logit contributions | |
counts | -0.021 |
ends | -0.021 |
One | -0.021 |
Or | -0.019 |
Ready | -0.018 |
Token | came |
Token ID | 1625 |
Feature activation | +1.88e-02 |
Pos logit contributions | |
sensation | +0.079 |
tune | +0.077 |
dolphin | +0.071 |
Dolphin | +0.068 |
potato | +0.068 |
Neg logit contributions | |
counts | -0.062 |
ends | -0.058 |
One | -0.056 |
Or | -0.055 |
house | -0.050 |
Token | back |
Token ID | 736 |
Feature activation | -1.04e-02 |
Pos logit contributions | |
tune | +0.035 |
sensation | +0.034 |
dolphin | +0.031 |
shark | +0.030 |
streaming | +0.029 |
Neg logit contributions | |
counts | -0.031 |
ends | -0.030 |
One | -0.029 |
Or | -0.028 |
Ready | -0.026 |
Token | colors |
Token ID | 7577 |
Feature activation | +8.15e-03 |
Pos logit contributions | |
tune | +0.074 |
sensation | +0.074 |
dolphin | +0.068 |
shark | +0.066 |
Dolphin | +0.063 |
Neg logit contributions | |
counts | -0.065 |
ends | -0.062 |
One | -0.060 |
Or | -0.057 |
Ready | -0.053 |
Token | filled |
Token ID | 5901 |
Feature activation | +1.97e-02 |
Pos logit contributions | |
tune | +0.016 |
sensation | +0.015 |
shark | +0.014 |
dolphin | +0.014 |
Dolphin | +0.014 |
Neg logit contributions | |
counts | -0.012 |
ends | -0.012 |
One | -0.011 |
Or | -0.010 |
has | -0.010 |
Token | the |
Token ID | 262 |
Feature activation | -6.91e-03 |
Pos logit contributions | |
tune | +0.040 |
sensation | +0.039 |
shark | +0.037 |
dolphin | +0.037 |
streaming | +0.035 |
Neg logit contributions | |
One | -0.026 |
counts | -0.026 |
ends | -0.026 |
Or | -0.022 |
ours | -0.020 |
Token | room |
Token ID | 2119 |
Feature activation | +1.91e-03 |
Pos logit contributions | |
ends | +0.012 |
counts | +0.012 |
One | +0.012 |
Or | +0.011 |
Ready | +0.010 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.012 |
dolphin | -0.010 |
shark | -0.010 |
face | -0.010 |
Token | with |
Token ID | 351 |
Feature activation | -9.46e-03 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.004 |
dolphin | +0.003 |
shark | +0.003 |
streaming | +0.003 |
Neg logit contributions | |
ends | -0.003 |
counts | -0.003 |
One | -0.003 |
Or | -0.003 |
Ready | -0.002 |
Token | joy |
Token ID | 8716 |
Feature activation | +4.10e-02 |
Pos logit contributions | |
counts | +0.014 |
ends | +0.014 |
One | +0.013 |
Or | +0.012 |
Ready | +0.012 |
Neg logit contributions | |
tune | -0.018 |
sensation | -0.018 |
shark | -0.016 |
dolphin | -0.016 |
Dolphin | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
tune | +0.099 |
sensation | +0.099 |
shark | +0.095 |
dolphin | +0.095 |
Dolphin | +0.092 |
Neg logit contributions | |
ends | -0.038 |
counts | -0.037 |
One | -0.029 |
has | -0.028 |
Or | -0.027 |
Token |
|
Token ID | 198 |
Feature activation | +1.81e-02 |
Pos logit contributions | |
sensation | +0.038 |
tune | +0.038 |
dolphin | +0.035 |
shark | +0.033 |
Dolphin | +0.033 |
Neg logit contributions | |
counts | -0.027 |
ends | -0.027 |
One | -0.025 |
Or | -0.024 |
ours | -0.022 |
Token |
|
Token ID | 198 |
Feature activation | +2.76e-02 |
Pos logit contributions | |
tune | +0.038 |
sensation | +0.038 |
dolphin | +0.034 |
shark | +0.033 |
streaming | +0.032 |
Neg logit contributions | |
ends | -0.025 |
counts | -0.025 |
One | -0.024 |
Or | -0.022 |
Ready | -0.020 |
Token | Cl |
Token ID | 2601 |
Feature activation | +8.44e-03 |
Pos logit contributions | |
tune | +0.051 |
sensation | +0.049 |
dolphin | +0.045 |
lull | +0.041 |
shark | +0.041 |
Neg logit contributions | |
counts | -0.047 |
ends | -0.043 |
Or | -0.041 |
One | -0.039 |
Ready | -0.038 |
Token | ara |
Token ID | 3301 |
Feature activation | +9.93e-03 |
Pos logit contributions | |
tune | +0.019 |
sensation | +0.019 |
dolphin | +0.017 |
shark | +0.017 |
streamed | +0.016 |
Neg logit contributions | |
counts | -0.010 |
ends | -0.010 |
One | -0.009 |
Or | -0.009 |
Ready | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | +9.15e-03 |
Pos logit contributions | |
counts | +0.009 |
ends | +0.009 |
One | +0.009 |
Or | +0.008 |
house | +0.007 |
Neg logit contributions | |
sensation | -0.020 |
tune | -0.019 |
dolphin | -0.018 |
shark | -0.017 |
chill | -0.017 |
Token | It |
Token ID | 632 |
Feature activation | +3.67e-03 |
Pos logit contributions | |
tune | +0.018 |
sensation | +0.017 |
dolphin | +0.016 |
shark | +0.015 |
potato | +0.015 |
Neg logit contributions | |
counts | -0.014 |
One | -0.014 |
ends | -0.013 |
Or | -0.013 |
Ready | -0.012 |
Token | had |
Token ID | 550 |
Feature activation | +5.71e-03 |
Pos logit contributions | |
tune | +0.008 |
sensation | +0.008 |
dolphin | +0.007 |
shark | +0.007 |
streaming | +0.007 |
Neg logit contributions | |
counts | -0.005 |
ends | -0.005 |
One | -0.005 |
Or | -0.005 |
belongs | -0.004 |
Token | a |
Token ID | 257 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.013 |
dolphin | +0.012 |
shark | +0.012 |
Dolphin | +0.011 |
Neg logit contributions | |
counts | -0.007 |
ends | -0.007 |
One | -0.006 |
Or | -0.006 |
Ready | -0.005 |
Token | very |
Token ID | 845 |
Feature activation | +6.87e-02 |
Pos logit contributions | |
ends | +0.021 |
One | +0.021 |
Or | +0.020 |
counts | +0.020 |
has | +0.018 |
Neg logit contributions | |
sensation | -0.031 |
tune | -0.030 |
streaming | -0.027 |
streamed | -0.027 |
iness | -0.027 |
Token | important |
Token ID | 1593 |
Feature activation | +4.13e-02 |
Pos logit contributions | |
tune | +0.132 |
sensation | +0.131 |
dolphin | +0.115 |
streaming | +0.115 |
streamed | +0.111 |
Neg logit contributions | |
ends | -0.107 |
counts | -0.103 |
One | -0.100 |
Or | -0.094 |
ours | -0.087 |
Token | job |
Token ID | 1693 |
Feature activation | +3.69e-02 |
Pos logit contributions | |
tune | +0.077 |
sensation | +0.077 |
lee | +0.074 |
streaming | +0.072 |
dolphin | +0.072 |
Neg logit contributions | |
counts | -0.058 |
One | -0.055 |
Or | -0.054 |
ends | -0.053 |
Ready | -0.048 |
Token | - |
Token ID | 532 |
Feature activation | -3.02e-02 |
Pos logit contributions | |
tune | +0.080 |
sensation | +0.079 |
dolphin | +0.073 |
shark | +0.070 |
chill | +0.068 |
Neg logit contributions | |
counts | -0.048 |
ends | -0.047 |
One | -0.044 |
Or | -0.042 |
Ready | -0.038 |
Token | to |
Token ID | 284 |
Feature activation | +4.14e-03 |
Pos logit contributions | |
ends | +0.048 |
counts | +0.047 |
One | +0.045 |
Or | +0.043 |
Ready | +0.039 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.055 |
dolphin | -0.051 |
shark | -0.050 |
potato | -0.047 |
Token | power |
Token ID | 1176 |
Feature activation | -6.56e-03 |
Pos logit contributions | |
sensation | +0.008 |
tune | +0.008 |
dolphin | +0.008 |
streaming | +0.007 |
shark | +0.007 |
Neg logit contributions | |
One | -0.006 |
counts | -0.006 |
ends | -0.006 |
Or | -0.006 |
Ready | -0.005 |
Token | a |
Token ID | 257 |
Feature activation | -2.40e-02 |
Pos logit contributions | |
counts | +0.009 |
ends | +0.009 |
One | +0.008 |
Or | +0.008 |
Ready | +0.007 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.012 |
chill | -0.012 |
Token | the |
Token ID | 262 |
Feature activation | -6.55e-03 |
Pos logit contributions | |
tune | +0.010 |
sensation | +0.010 |
dolphin | +0.009 |
shark | +0.009 |
streaming | +0.009 |
Neg logit contributions | |
counts | -0.009 |
ends | -0.009 |
One | -0.009 |
Or | -0.008 |
Ready | -0.008 |
Token | university |
Token ID | 6403 |
Feature activation | -6.13e-02 |
Pos logit contributions | |
ends | +0.009 |
counts | +0.009 |
One | +0.008 |
Or | +0.008 |
ours | +0.007 |
Neg logit contributions | |
tune | -0.013 |
sensation | -0.013 |
dolphin | -0.012 |
face | -0.011 |
shark | -0.011 |
Token | was |
Token ID | 373 |
Feature activation | -2.38e-02 |
Pos logit contributions | |
ends | +0.091 |
counts | +0.086 |
One | +0.084 |
Or | +0.080 |
Ready | +0.077 |
Neg logit contributions | |
tune | -0.123 |
sensation | -0.115 |
dolphin | -0.103 |
shark | -0.100 |
face | -0.098 |
Token | coming |
Token ID | 2406 |
Feature activation | +3.87e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.034 |
One | +0.032 |
Or | +0.031 |
Ready | +0.028 |
Neg logit contributions | |
tune | -0.048 |
sensation | -0.047 |
dolphin | -0.043 |
shark | -0.042 |
streaming | -0.040 |
Token | true |
Token ID | 2081 |
Feature activation | -5.95e-02 |
Pos logit contributions | |
tune | +0.076 |
sensation | +0.073 |
dolphin | +0.066 |
shark | +0.064 |
face | +0.063 |
Neg logit contributions | |
counts | -0.059 |
ends | -0.059 |
One | -0.054 |
Or | -0.053 |
Ready | -0.049 |
Token | . |
Token ID | 13 |
Feature activation | +2.10e-02 |
Pos logit contributions | |
ends | +0.072 |
counts | +0.071 |
One | +0.067 |
Or | +0.064 |
Ready | +0.057 |
Neg logit contributions | |
tune | -0.134 |
sensation | -0.133 |
dolphin | -0.122 |
shark | -0.119 |
streaming | -0.114 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -3.13e-02 |
Pos logit contributions | |
tune | +0.044 |
sensation | +0.043 |
dolphin | +0.040 |
shark | +0.038 |
chill | +0.037 |
Neg logit contributions | |
counts | -0.029 |
ends | -0.029 |
One | -0.028 |
Or | -0.027 |
Ready | -0.024 |
Token | Once |
Token ID | 7454 |
Feature activation | +2.84e-04 |
Pos logit contributions | |
ends | +0.039 |
counts | +0.038 |
One | +0.035 |
Or | +0.034 |
ours | +0.030 |
Neg logit contributions | |
sensation | -0.068 |
tune | -0.068 |
dolphin | -0.063 |
shark | -0.062 |
streamed | -0.059 |
Token | there |
Token ID | 612 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
tune | +0.001 |
sensation | +0.001 |
dolphin | +0.001 |
shark | +0.001 |
potato | +0.001 |
Neg logit contributions | |
counts | -0.000 |
ends | -0.000 |
One | -0.000 |
Or | -0.000 |
Ready | -0.000 |
Token | was |
Token ID | 373 |
Feature activation | +1.28e-02 |
Pos logit contributions | |
counts | +0.024 |
ends | +0.023 |
One | +0.023 |
Or | +0.021 |
Ready | +0.019 |
Neg logit contributions | |
tune | -0.037 |
sensation | -0.035 |
dolphin | -0.033 |
shark | -0.031 |
potato | -0.031 |
Token | a |
Token ID | 257 |
Feature activation | -2.90e-02 |
Pos logit contributions | |
sensation | +0.023 |
tune | +0.023 |
dolphin | +0.021 |
shark | +0.020 |
potato | +0.019 |
Neg logit contributions | |
counts | -0.021 |
ends | -0.021 |
Or | -0.019 |
One | -0.019 |
Ready | -0.018 |
Token | was |
Token ID | 373 |
Feature activation | -1.36e-02 |
Pos logit contributions | |
ends | +0.030 |
counts | +0.030 |
One | +0.029 |
Or | +0.028 |
Ready | +0.026 |
Neg logit contributions | |
tune | -0.033 |
sensation | -0.032 |
dolphin | -0.030 |
shark | -0.028 |
potato | -0.027 |
Token | making |
Token ID | 1642 |
Feature activation | +7.59e-02 |
Pos logit contributions | |
ends | +0.020 |
One | +0.018 |
counts | +0.017 |
Or | +0.016 |
house | +0.015 |
Neg logit contributions | |
tune | -0.028 |
sensation | -0.028 |
dolphin | -0.026 |
shark | -0.025 |
Dolphin | -0.024 |
Token | his |
Token ID | 465 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
tune | +0.147 |
sensation | +0.138 |
streamed | +0.130 |
streaming | +0.126 |
tant | +0.124 |
Neg logit contributions | |
ends | -0.109 |
One | -0.107 |
counts | -0.106 |
Or | -0.098 |
you | -0.090 |
Token | thick |
Token ID | 6546 |
Feature activation | +3.05e-02 |
Pos logit contributions | |
counts | +0.018 |
ends | +0.017 |
One | +0.017 |
Or | +0.016 |
Ready | +0.015 |
Neg logit contributions | |
tune | -0.017 |
sensation | -0.017 |
dolphin | -0.016 |
shark | -0.015 |
streaming | -0.015 |
Token | toast |
Token ID | 27805 |
Feature activation | +3.46e-02 |
Pos logit contributions | |
tune | +0.049 |
sensation | +0.048 |
dolphin | +0.044 |
shark | +0.042 |
Dolphin | +0.042 |
Neg logit contributions | |
counts | -0.056 |
ends | -0.052 |
One | -0.051 |
Or | -0.050 |
house | -0.046 |
Token | in |
Token ID | 287 |
Feature activation | +7.34e-04 |
Pos logit contributions | |
tune | +0.075 |
sensation | +0.074 |
dolphin | +0.068 |
shark | +0.066 |
streaming | +0.064 |
Neg logit contributions | |
ends | -0.044 |
counts | -0.044 |
One | -0.042 |
Or | -0.040 |
Ready | -0.035 |
Token | the |
Token ID | 262 |
Feature activation | -2.01e-03 |
Pos logit contributions | |
tune | +0.001 |
sensation | +0.001 |
dolphin | +0.001 |
shark | +0.001 |
Dolphin | +0.001 |
Neg logit contributions | |
One | -0.001 |
counts | -0.001 |
ends | -0.001 |
Or | -0.001 |
ours | -0.001 |
Token | kitchen |
Token ID | 9592 |
Feature activation | +1.17e-02 |
Pos logit contributions | |
One | +0.003 |
counts | +0.003 |
Or | +0.002 |
ends | +0.002 |
Ready | +0.002 |
Neg logit contributions | |
tune | -0.004 |
sensation | -0.004 |
dolphin | -0.004 |
streaming | -0.004 |
shark | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +9.27e-03 |
Pos logit contributions | |
tune | +0.028 |
sensation | +0.028 |
dolphin | +0.026 |
shark | +0.025 |
streamed | +0.025 |
Neg logit contributions | |
counts | -0.012 |
ends | -0.012 |
One | -0.011 |
Or | -0.011 |
Ready | -0.009 |
Token | She |
Token ID | 1375 |
Feature activation | -1.12e-02 |
Pos logit contributions | |
tune | +0.021 |
sensation | +0.021 |
dolphin | +0.019 |
shark | +0.019 |
Dolphin | +0.018 |
Neg logit contributions | |
counts | -0.011 |
ends | -0.011 |
One | -0.011 |
Or | -0.010 |
Ready | -0.009 |
Token | put |
Token ID | 1234 |
Feature activation | +1.51e-02 |
Pos logit contributions | |
ends | +0.018 |
counts | +0.018 |
One | +0.017 |
Or | +0.017 |
Ready | +0.015 |
Neg logit contributions | |
tune | -0.021 |
sensation | -0.020 |
dolphin | -0.019 |
shark | -0.018 |
streaming | -0.017 |
Token | inside |
Token ID | 2641 |
Feature activation | -3.54e-02 |
Pos logit contributions | |
counts | +0.015 |
ends | +0.015 |
One | +0.014 |
Or | +0.013 |
Ready | +0.012 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.026 |
dolphin | -0.024 |
shark | -0.023 |
streaming | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +2.59e-02 |
Pos logit contributions | |
counts | +0.043 |
ends | +0.041 |
One | +0.040 |
Or | +0.037 |
Ready | +0.033 |
Neg logit contributions | |
tune | -0.080 |
sensation | -0.078 |
dolphin | -0.072 |
shark | -0.070 |
chill | -0.068 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | -2.71e-02 |
Pos logit contributions | |
tune | +0.061 |
sensation | +0.061 |
dolphin | +0.056 |
shark | +0.054 |
chill | +0.053 |
Neg logit contributions | |
counts | -0.029 |
ends | -0.029 |
One | -0.027 |
Or | -0.026 |
Ready | -0.023 |
Token | John |
Token ID | 7554 |
Feature activation | -1.95e-02 |
Pos logit contributions | |
ends | +0.031 |
counts | +0.031 |
One | +0.029 |
Or | +0.027 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.062 |
sensation | -0.062 |
dolphin | -0.057 |
shark | -0.056 |
streaming | -0.054 |
Token | and |
Token ID | 290 |
Feature activation | +1.92e-02 |
Pos logit contributions | |
counts | +0.029 |
ends | +0.029 |
One | +0.027 |
Or | +0.026 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.039 |
sensation | -0.038 |
dolphin | -0.035 |
shark | -0.034 |
streaming | -0.032 |
Token | his |
Token ID | 465 |
Feature activation | +9.44e-04 |
Pos logit contributions | |
sensation | +0.036 |
tune | +0.034 |
andel | +0.029 |
tant | +0.029 |
chill | +0.028 |
Neg logit contributions | |
counts | -0.033 |
One | -0.030 |
Or | -0.029 |
ends | -0.029 |
� | -0.028 |
Token | Dad |
Token ID | 17415 |
Feature activation | +1.36e-02 |
Pos logit contributions | |
sensation | +0.002 |
tune | +0.002 |
lull | +0.002 |
streamed | +0.002 |
streaming | +0.002 |
Neg logit contributions | |
ends | -0.001 |
One | -0.001 |
Or | -0.001 |
counts | -0.001 |
Ready | -0.001 |
Token | went |
Token ID | 1816 |
Feature activation | +4.71e-04 |
Pos logit contributions | |
tune | +0.026 |
sensation | +0.025 |
dolphin | +0.022 |
shark | +0.021 |
chill | +0.021 |
Neg logit contributions | |
counts | -0.022 |
ends | -0.021 |
One | -0.021 |
Or | -0.020 |
Ready | -0.018 |
Token | to |
Token ID | 284 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
tune | +0.001 |
sensation | +0.001 |
dolphin | +0.001 |
shark | +0.001 |
potato | +0.001 |
Neg logit contributions | |
counts | -0.001 |
ends | -0.001 |
One | -0.001 |
Or | -0.001 |
Ready | -0.001 |
Token | the |
Token ID | 262 |
Feature activation | -1.84e-02 |
Pos logit contributions | |
sensation | +0.045 |
tune | +0.044 |
streamed | +0.038 |
streaming | +0.038 |
chill | +0.037 |
Neg logit contributions | |
counts | -0.032 |
One | -0.032 |
ends | -0.030 |
Ready | -0.029 |
Or | -0.029 |
Token | park |
Token ID | 3952 |
Feature activation | -6.78e-02 |
Pos logit contributions | |
One | +0.024 |
counts | +0.023 |
ends | +0.023 |
Or | +0.021 |
Ready | +0.021 |
Neg logit contributions | |
sensation | -0.040 |
tune | -0.040 |
streaming | -0.035 |
streamed | -0.035 |
dolphin | -0.034 |
Token | buttons |
Token ID | 12163 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
ends | +0.053 |
counts | +0.052 |
One | +0.050 |
Or | +0.049 |
Ready | +0.046 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.045 |
dolphin | -0.041 |
shark | -0.038 |
chill | -0.036 |
Token | and |
Token ID | 290 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
counts | +0.038 |
ends | +0.038 |
One | +0.036 |
Or | +0.034 |
Ready | +0.031 |
Neg logit contributions | |
tune | -0.048 |
sensation | -0.047 |
dolphin | -0.042 |
shark | -0.041 |
potato | -0.039 |
Token | levers |
Token ID | 48872 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
counts | +0.027 |
ends | +0.026 |
Or | +0.024 |
One | +0.023 |
Ready | +0.021 |
Neg logit contributions | |
sensation | -0.034 |
tune | -0.034 |
dolphin | -0.030 |
shark | -0.029 |
Dolphin | -0.029 |
Token | did |
Token ID | 750 |
Feature activation | -1.34e-02 |
Pos logit contributions | |
counts | +0.038 |
ends | +0.038 |
One | +0.035 |
Or | +0.034 |
Ready | +0.030 |
Neg logit contributions | |
tune | -0.064 |
sensation | -0.063 |
dolphin | -0.058 |
shark | -0.056 |
potato | -0.054 |
Token | . |
Token ID | 13 |
Feature activation | -6.59e-03 |
Pos logit contributions | |
counts | +0.019 |
ends | +0.019 |
One | +0.018 |
Or | +0.018 |
Ready | +0.017 |
Neg logit contributions | |
sensation | -0.026 |
tune | -0.026 |
face | -0.022 |
dolphin | -0.022 |
shark | -0.022 |
Token | He |
Token ID | 679 |
Feature activation | -7.75e-03 |
Pos logit contributions | |
ends | +0.009 |
counts | +0.009 |
One | +0.009 |
Or | +0.008 |
Ready | +0.007 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.012 |
chill | -0.012 |
Token | said |
Token ID | 531 |
Feature activation | -2.08e-02 |
Pos logit contributions | |
ends | +0.013 |
counts | +0.013 |
One | +0.013 |
Or | +0.012 |
Ready | +0.011 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.013 |
dolphin | -0.012 |
shark | -0.012 |
potato | -0.011 |
Token | she |
Token ID | 673 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
ends | +0.025 |
counts | +0.024 |
Or | +0.022 |
One | +0.022 |
Ready | +0.018 |
Neg logit contributions | |
tune | -0.048 |
sensation | -0.048 |
dolphin | -0.042 |
chill | -0.041 |
shark | -0.041 |
Token | could |
Token ID | 714 |
Feature activation | +1.28e-02 |
Pos logit contributions | |
counts | +0.028 |
ends | +0.028 |
One | +0.027 |
Or | +0.026 |
Ready | +0.026 |
Neg logit contributions | |
tune | -0.027 |
sensation | -0.026 |
dolphin | -0.024 |
shark | -0.022 |
face | -0.022 |
Token | help |
Token ID | 1037 |
Feature activation | +4.10e-02 |
Pos logit contributions | |
tune | +0.023 |
sensation | +0.022 |
dolphin | +0.021 |
shark | +0.020 |
face | +0.019 |
Neg logit contributions | |
ends | -0.021 |
counts | -0.021 |
One | -0.020 |
Or | -0.019 |
Ready | -0.018 |
Token | him |
Token ID | 683 |
Feature activation | +1.35e-02 |
Pos logit contributions | |
sensation | +0.085 |
tune | +0.085 |
dolphin | +0.078 |
shark | +0.077 |
streamed | +0.075 |
Neg logit contributions | |
counts | -0.054 |
ends | -0.052 |
One | -0.051 |
Or | -0.051 |
Ready | -0.043 |
Token | We |
Token ID | 775 |
Feature activation | +7.29e-04 |
Pos logit contributions | |
ends | +0.028 |
counts | +0.027 |
One | +0.025 |
Or | +0.022 |
ours | +0.021 |
Neg logit contributions | |
tune | -0.047 |
sensation | -0.045 |
dolphin | -0.044 |
shark | -0.043 |
Dolphin | -0.043 |
Token | will |
Token ID | 481 |
Feature activation | -1.20e-03 |
Pos logit contributions | |
tune | +0.002 |
sensation | +0.002 |
dolphin | +0.001 |
shark | +0.001 |
chill | +0.001 |
Neg logit contributions | |
ends | -0.001 |
counts | -0.001 |
Or | -0.001 |
One | -0.001 |
Ready | -0.001 |
Token | come |
Token ID | 1282 |
Feature activation | -1.68e-02 |
Pos logit contributions | |
ends | +0.002 |
counts | +0.002 |
Or | +0.002 |
One | +0.002 |
Ready | +0.002 |
Neg logit contributions | |
tune | -0.002 |
sensation | -0.002 |
dolphin | -0.002 |
face | -0.002 |
shark | -0.002 |
Token | back |
Token ID | 736 |
Feature activation | -3.63e-02 |
Pos logit contributions | |
counts | +0.033 |
One | +0.030 |
Or | +0.029 |
belongs | +0.029 |
ends | +0.029 |
Neg logit contributions | |
tune | -0.024 |
dolphin | -0.021 |
face | -0.020 |
shark | -0.019 |
Dolphin | -0.019 |
Token | soon |
Token ID | 2582 |
Feature activation | -9.55e-02 |
Pos logit contributions | |
counts | +0.049 |
ends | +0.048 |
One | +0.045 |
Or | +0.045 |
atters | +0.042 |
Neg logit contributions | |
tune | -0.071 |
dolphin | -0.064 |
sensation | -0.062 |
shark | -0.062 |
Dolphin | -0.059 |
Token | . |
Token ID | 13 |
Feature activation | -2.85e-02 |
Pos logit contributions | |
counts | +0.100 |
ends | +0.100 |
One | +0.091 |
Or | +0.090 |
Ready | +0.085 |
Neg logit contributions | |
tune | -0.227 |
sensation | -0.213 |
dolphin | -0.204 |
face | -0.196 |
shark | -0.195 |
Token | Be |
Token ID | 1355 |
Feature activation | +2.75e-02 |
Pos logit contributions | |
ends | +0.036 |
counts | +0.035 |
One | +0.033 |
belongs | +0.028 |
ours | +0.028 |
Neg logit contributions | |
tune | -0.060 |
sensation | -0.056 |
dolphin | -0.055 |
shark | -0.054 |
Dolphin | -0.054 |
Token | a |
Token ID | 257 |
Feature activation | -3.74e-02 |
Pos logit contributions | |
tune | +0.052 |
sensation | +0.052 |
dolphin | +0.047 |
shark | +0.045 |
streaming | +0.043 |
Neg logit contributions | |
ends | -0.043 |
counts | -0.043 |
One | -0.041 |
Or | -0.039 |
Ready | -0.036 |
Token | good |
Token ID | 922 |
Feature activation | +5.51e-02 |
Pos logit contributions | |
ends | +0.065 |
counts | +0.064 |
One | +0.063 |
Or | +0.061 |
Ready | +0.057 |
Neg logit contributions | |
sensation | -0.064 |
tune | -0.063 |
dolphin | -0.055 |
streaming | -0.053 |
shark | -0.053 |
Token | boy |
Token ID | 2933 |
Feature activation | +3.28e-03 |
Pos logit contributions | |
tune | +0.084 |
sensation | +0.083 |
dolphin | +0.073 |
shark | +0.070 |
streaming | +0.066 |
Neg logit contributions | |
ends | -0.107 |
counts | -0.107 |
One | -0.103 |
Or | -0.100 |
Ready | -0.093 |
Token | ." |
Token ID | 526 |
Feature activation | -2.07e-02 |
Pos logit contributions | |
tune | +0.007 |
dolphin | +0.007 |
sensation | +0.007 |
shark | +0.007 |
potato | +0.006 |
Neg logit contributions | |
One | -0.004 |
ends | -0.004 |
counts | -0.003 |
Or | -0.003 |
Ready | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
ends | +0.031 |
One | +0.031 |
counts | +0.030 |
Or | +0.029 |
Ready | +0.028 |
Neg logit contributions | |
sensation | -0.040 |
tune | -0.039 |
dolphin | -0.037 |
shark | -0.035 |
streaming | -0.034 |
Token | Sara |
Token ID | 24799 |
Feature activation | -2.21e-02 |
Pos logit contributions | |
counts | +0.029 |
ends | +0.028 |
Or | +0.025 |
One | +0.025 |
house | +0.024 |
Neg logit contributions | |
sensation | -0.030 |
tune | -0.029 |
dolphin | -0.028 |
Sister | -0.027 |
Ned | -0.027 |
Token | and |
Token ID | 290 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
One | +0.036 |
ends | +0.033 |
ours | +0.032 |
Ready | +0.031 |
Or | +0.030 |
Neg logit contributions | |
tune | -0.037 |
dolphin | -0.035 |
sensation | -0.034 |
face | -0.032 |
potato | -0.032 |
Token | Tom |
Token ID | 4186 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
tune | +0.037 |
sensation | +0.036 |
dolphin | +0.034 |
shark | +0.033 |
streaming | +0.032 |
Neg logit contributions | |
ends | -0.014 |
counts | -0.014 |
One | -0.013 |
Or | -0.012 |
Ready | -0.011 |
Token | pretend |
Token ID | 16614 |
Feature activation | -4.33e-02 |
Pos logit contributions | |
ends | +0.029 |
counts | +0.028 |
One | +0.027 |
Or | +0.026 |
Ready | +0.023 |
Neg logit contributions | |
tune | -0.056 |
sensation | -0.054 |
dolphin | -0.051 |
shark | -0.049 |
face | -0.048 |
Token | the |
Token ID | 262 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
counts | +0.061 |
ends | +0.059 |
One | +0.055 |
Or | +0.054 |
Ready | +0.048 |
Neg logit contributions | |
sensation | -0.089 |
tune | -0.089 |
dolphin | -0.079 |
streaming | -0.076 |
shark | -0.076 |
Token | rug |
Token ID | 14477 |
Feature activation | -2.93e-02 |
Pos logit contributions | |
ends | +0.038 |
counts | +0.037 |
Or | +0.035 |
One | +0.035 |
ours | +0.032 |
Neg logit contributions | |
sensation | -0.034 |
tune | -0.033 |
dolphin | -0.029 |
shark | -0.028 |
chill | -0.027 |
Token | is |
Token ID | 318 |
Feature activation | -5.69e-02 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.032 |
One | +0.031 |
Or | +0.029 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.067 |
sensation | -0.067 |
dolphin | -0.062 |
shark | -0.060 |
streaming | -0.058 |
Token | a |
Token ID | 257 |
Feature activation | -6.19e-02 |
Pos logit contributions | |
ends | +0.081 |
counts | +0.080 |
One | +0.076 |
Or | +0.072 |
Ready | +0.066 |
Neg logit contributions | |
tune | -0.117 |
sensation | -0.115 |
dolphin | -0.104 |
shark | -0.101 |
chill | -0.098 |
Token | map |
Token ID | 3975 |
Feature activation | -7.75e-02 |
Pos logit contributions | |
ends | +0.108 |
counts | +0.108 |
One | +0.104 |
Or | +0.100 |
Ready | +0.093 |
Neg logit contributions | |
tune | -0.106 |
sensation | -0.104 |
dolphin | -0.093 |
shark | -0.089 |
streaming | -0.085 |
Token | of |
Token ID | 286 |
Feature activation | +1.16e-02 |
Pos logit contributions | |
ends | +0.128 |
One | +0.123 |
counts | +0.120 |
Or | +0.116 |
Ready | +0.110 |
Neg logit contributions | |
tune | -0.140 |
dolphin | -0.131 |
sensation | -0.131 |
shark | -0.125 |
face | -0.117 |
Token | bear |
Token ID | 6842 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
ends | +0.022 |
counts | +0.022 |
One | +0.021 |
Or | +0.020 |
Ready | +0.019 |
Neg logit contributions | |
tune | -0.015 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.012 |
potato | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | -2.61e-02 |
Pos logit contributions | |
ends | +0.032 |
counts | +0.031 |
One | +0.030 |
Or | +0.028 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.046 |
sensation | -0.044 |
dolphin | -0.042 |
shark | -0.041 |
potato | -0.040 |
Token | You |
Token ID | 921 |
Feature activation | +4.20e-03 |
Pos logit contributions | |
ends | +0.033 |
One | +0.032 |
counts | +0.030 |
ours | +0.025 |
Or | +0.024 |
Neg logit contributions | |
sensation | -0.055 |
tune | -0.055 |
dolphin | -0.052 |
shark | -0.050 |
Dolphin | -0.049 |
Token | are |
Token ID | 389 |
Feature activation | -3.23e-02 |
Pos logit contributions | |
tune | +0.008 |
sensation | +0.008 |
dolphin | +0.008 |
shark | +0.008 |
face | +0.008 |
Neg logit contributions | |
ends | -0.006 |
One | -0.005 |
counts | -0.005 |
Or | -0.005 |
Ready | -0.005 |
Token | very |
Token ID | 845 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
counts | +0.053 |
One | +0.052 |
ends | +0.051 |
Or | +0.051 |
Maybe | +0.046 |
Neg logit contributions | |
sensation | -0.055 |
tune | -0.055 |
dolphin | -0.051 |
shark | -0.049 |
face | -0.048 |
Token | pretty |
Token ID | 2495 |
Feature activation | -8.40e-02 |
Pos logit contributions | |
counts | +0.048 |
ends | +0.046 |
One | +0.046 |
Or | +0.045 |
Ready | +0.043 |
Neg logit contributions | |
tune | -0.041 |
sensation | -0.041 |
dolphin | -0.038 |
shark | -0.037 |
face | -0.035 |
Token | . |
Token ID | 13 |
Feature activation | -4.30e-02 |
Pos logit contributions | |
ends | +0.127 |
counts | +0.122 |
One | +0.122 |
Or | +0.115 |
Ready | +0.114 |
Neg logit contributions | |
tune | -0.163 |
sensation | -0.156 |
dolphin | -0.151 |
shark | -0.150 |
potato | -0.141 |
Token | But |
Token ID | 887 |
Feature activation | -3.02e-02 |
Pos logit contributions | |
ends | +0.059 |
One | +0.055 |
counts | +0.051 |
ours | +0.046 |
house | +0.041 |
Neg logit contributions | |
tune | -0.085 |
sensation | -0.084 |
dolphin | -0.080 |
Dolphin | -0.077 |
shark | -0.077 |
Token | I |
Token ID | 314 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
ends | +0.032 |
counts | +0.031 |
One | +0.029 |
Or | +0.027 |
atters | +0.023 |
Neg logit contributions | |
tune | -0.072 |
sensation | -0.071 |
dolphin | -0.067 |
shark | -0.065 |
face | -0.063 |
Token | will |
Token ID | 481 |
Feature activation | -3.23e-03 |
Pos logit contributions | |
counts | +0.025 |
ends | +0.025 |
One | +0.024 |
Or | +0.022 |
Ready | +0.022 |
Neg logit contributions | |
sensation | -0.043 |
tune | -0.043 |
dolphin | -0.040 |
shark | -0.038 |
face | -0.037 |
Token | make |
Token ID | 787 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
ends | +0.005 |
counts | +0.005 |
One | +0.005 |
Or | +0.004 |
Ready | +0.004 |
Neg logit contributions | |
tune | -0.006 |
sensation | -0.006 |
dolphin | -0.006 |
face | -0.005 |
shark | -0.005 |
Token | , |
Token ID | 11 |
Feature activation | -2.80e-02 |
Pos logit contributions | |
tune | +0.032 |
sensation | +0.032 |
dolphin | +0.029 |
shark | +0.028 |
streaming | +0.028 |
Neg logit contributions | |
counts | -0.019 |
ends | -0.018 |
One | -0.018 |
Or | -0.016 |
Ready | -0.015 |
Token | I |
Token ID | 314 |
Feature activation | -2.95e-02 |
Pos logit contributions | |
counts | +0.030 |
ends | +0.030 |
One | +0.028 |
Or | +0.027 |
Ready | +0.024 |
Neg logit contributions | |
tune | -0.066 |
sensation | -0.066 |
dolphin | -0.058 |
streaming | -0.058 |
streamed | -0.057 |
Token | want |
Token ID | 765 |
Feature activation | -1.96e-02 |
Pos logit contributions | |
counts | +0.037 |
ends | +0.035 |
One | +0.035 |
Or | +0.033 |
Ready | +0.030 |
Neg logit contributions | |
tune | -0.062 |
sensation | -0.062 |
dolphin | -0.058 |
shark | -0.056 |
chill | -0.054 |
Token | to |
Token ID | 284 |
Feature activation | -6.67e-03 |
Pos logit contributions | |
ends | +0.034 |
counts | +0.031 |
Or | +0.030 |
One | +0.030 |
Ready | +0.027 |
Neg logit contributions | |
tune | -0.035 |
sensation | -0.033 |
dolphin | -0.033 |
shark | -0.032 |
potato | -0.031 |
Token | go |
Token ID | 467 |
Feature activation | -3.10e-02 |
Pos logit contributions | |
ends | +0.011 |
counts | +0.011 |
One | +0.010 |
Or | +0.010 |
Ready | +0.009 |
Neg logit contributions | |
tune | -0.012 |
sensation | -0.012 |
dolphin | -0.011 |
shark | -0.011 |
face | -0.010 |
Token | first |
Token ID | 717 |
Feature activation | -1.12e-01 |
Pos logit contributions | |
counts | +0.048 |
ends | +0.047 |
Or | +0.047 |
One | +0.043 |
Ready | +0.043 |
Neg logit contributions | |
tune | -0.057 |
shark | -0.052 |
dolphin | -0.051 |
sensation | -0.050 |
face | -0.047 |
Token | !" |
Token ID | 2474 |
Feature activation | -2.79e-02 |
Pos logit contributions | |
ends | +0.143 |
Or | +0.127 |
counts | +0.127 |
Ready | +0.121 |
One | +0.113 |
Neg logit contributions | |
tune | -0.244 |
sensation | -0.233 |
shark | -0.218 |
dolphin | -0.218 |
face | -0.206 |
Token | Ben |
Token ID | 3932 |
Feature activation | +4.10e-02 |
Pos logit contributions | |
ends | +0.049 |
counts | +0.048 |
One | +0.046 |
Or | +0.044 |
Ready | +0.042 |
Neg logit contributions | |
tune | -0.048 |
sensation | -0.047 |
dolphin | -0.044 |
shark | -0.042 |
face | -0.040 |
Token | says |
Token ID | 1139 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
tune | +0.087 |
sensation | +0.085 |
dolphin | +0.083 |
shark | +0.081 |
face | +0.079 |
Neg logit contributions | |
ends | -0.051 |
One | -0.050 |
counts | -0.045 |
Or | -0.042 |
ours | -0.041 |
Token | . |
Token ID | 13 |
Feature activation | -8.21e-03 |
Pos logit contributions | |
counts | +0.028 |
ends | +0.027 |
One | +0.027 |
Or | +0.024 |
Ready | +0.023 |
Neg logit contributions | |
tune | -0.046 |
sensation | -0.045 |
dolphin | -0.041 |
shark | -0.040 |
face | -0.039 |
Token | He |
Token ID | 679 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
ends | +0.013 |
counts | +0.013 |
One | +0.012 |
Or | +0.012 |
Ready | +0.011 |
Neg logit contributions | |
tune | -0.016 |
sensation | -0.015 |
dolphin | -0.014 |
shark | -0.013 |
streaming | -0.013 |
Token | the |
Token ID | 262 |
Feature activation | -4.62e-02 |
Pos logit contributions | |
counts | +0.061 |
ends | +0.061 |
One | +0.058 |
Or | +0.056 |
Ready | +0.052 |
Neg logit contributions | |
tune | -0.057 |
sensation | -0.055 |
dolphin | -0.050 |
shark | -0.048 |
streaming | -0.045 |
Token | time |
Token ID | 640 |
Feature activation | +2.86e-03 |
Pos logit contributions | |
ends | +0.081 |
counts | +0.078 |
Or | +0.076 |
One | +0.075 |
Ready | +0.069 |
Neg logit contributions | |
tune | -0.081 |
sensation | -0.079 |
dolphin | -0.071 |
shark | -0.067 |
chill | -0.065 |
Token | . |
Token ID | 13 |
Feature activation | +1.44e-03 |
Pos logit contributions | |
tune | +0.006 |
sensation | +0.006 |
dolphin | +0.005 |
shark | +0.005 |
streaming | +0.005 |
Neg logit contributions | |
counts | -0.004 |
ends | -0.004 |
One | -0.004 |
Or | -0.004 |
Ready | -0.003 |
Token |
|
Token ID | 198 |
Feature activation | +1.83e-03 |
Pos logit contributions | |
sensation | +0.003 |
tune | +0.003 |
dolphin | +0.002 |
face | +0.002 |
Dolphin | +0.002 |
Neg logit contributions | |
ends | -0.002 |
counts | -0.002 |
ours | -0.002 |
One | -0.002 |
Or | -0.002 |
Token |
|
Token ID | 198 |
Feature activation | +1.31e-02 |
Pos logit contributions | |
tune | +0.004 |
sensation | +0.003 |
dolphin | +0.003 |
shark | +0.003 |
face | +0.003 |
Neg logit contributions | |
counts | -0.003 |
ends | -0.003 |
Or | -0.002 |
ours | -0.002 |
Ready | -0.002 |
Token | One |
Token ID | 3198 |
Feature activation | -1.07e-01 |
Pos logit contributions | |
tune | +0.035 |
sensation | +0.033 |
lull | +0.031 |
dolphin | +0.031 |
rhythm | +0.031 |
Neg logit contributions | |
counts | -0.010 |
ours | -0.009 |
ends | -0.008 |
Or | -0.007 |
belongs | -0.007 |
Token | day |
Token ID | 1110 |
Feature activation | +3.85e-02 |
Pos logit contributions | |
One | +0.155 |
counts | +0.148 |
ends | +0.138 |
Or | +0.125 |
� | +0.125 |
Neg logit contributions | |
tune | -0.213 |
sensation | -0.209 |
streaming | -0.199 |
streamed | -0.188 |
tant | -0.188 |
Token | , |
Token ID | 11 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
tune | +0.091 |
sensation | +0.089 |
dolphin | +0.081 |
shark | +0.079 |
chill | +0.076 |
Neg logit contributions | |
ends | -0.047 |
counts | -0.045 |
One | -0.042 |
Or | -0.040 |
Ready | -0.036 |
Token | L |
Token ID | 406 |
Feature activation | +1.64e-02 |
Pos logit contributions | |
tune | +0.072 |
sensation | +0.070 |
shark | +0.064 |
dolphin | +0.064 |
chill | +0.062 |
Neg logit contributions | |
ends | -0.039 |
counts | -0.034 |
One | -0.034 |
Or | -0.032 |
Ready | -0.029 |
Token | ola |
Token ID | 5708 |
Feature activation | +1.78e-02 |
Pos logit contributions | |
tune | +0.024 |
sensation | +0.024 |
dolphin | +0.021 |
shark | +0.020 |
streaming | +0.019 |
Neg logit contributions | |
counts | -0.032 |
ends | -0.032 |
One | -0.031 |
Or | -0.030 |
Ready | -0.028 |
Token | was |
Token ID | 373 |
Feature activation | +4.87e-03 |
Pos logit contributions | |
tune | +0.032 |
sensation | +0.031 |
dolphin | +0.030 |
shark | +0.028 |
potato | +0.027 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.028 |
One | -0.027 |
Or | -0.026 |
Ready | -0.025 |
Token | One |
Token ID | 3198 |
Feature activation | -1.10e-01 |
Pos logit contributions | |
tune | +0.034 |
sensation | +0.033 |
dolphin | +0.032 |
shark | +0.030 |
potato | +0.030 |
Neg logit contributions | |
counts | -0.007 |
ends | -0.007 |
Or | -0.005 |
ours | -0.004 |
Ready | -0.004 |
Token | summer |
Token ID | 3931 |
Feature activation | +3.25e-02 |
Pos logit contributions | |
One | +0.168 |
counts | +0.158 |
ends | +0.157 |
Or | +0.143 |
house | +0.140 |
Neg logit contributions | |
tune | -0.214 |
sensation | -0.212 |
streaming | -0.199 |
streamed | -0.189 |
dolphin | -0.185 |
Token | , |
Token ID | 11 |
Feature activation | +3.04e-02 |
Pos logit contributions | |
tune | +0.056 |
sensation | +0.055 |
dolphin | +0.049 |
shark | +0.047 |
streaming | +0.046 |
Neg logit contributions | |
counts | -0.059 |
ends | -0.055 |
One | -0.055 |
Or | -0.051 |
Ready | -0.049 |
Token | things |
Token ID | 1243 |
Feature activation | +6.00e-03 |
Pos logit contributions | |
tune | +0.053 |
sensation | +0.052 |
streaming | +0.044 |
lull | +0.043 |
chill | +0.043 |
Neg logit contributions | |
One | -0.052 |
ends | -0.051 |
counts | -0.051 |
Or | -0.047 |
ours | -0.046 |
Token | were |
Token ID | 547 |
Feature activation | -7.41e-03 |
Pos logit contributions | |
tune | +0.013 |
sensation | +0.012 |
dolphin | +0.011 |
Dolphin | +0.010 |
shark | +0.010 |
Neg logit contributions | |
counts | -0.009 |
ends | -0.008 |
One | -0.008 |
Or | -0.007 |
belongs | -0.007 |
Token | different |
Token ID | 1180 |
Feature activation | -7.99e-02 |
Pos logit contributions | |
ends | +0.011 |
counts | +0.011 |
One | +0.011 |
Or | +0.010 |
Ready | +0.010 |
Neg logit contributions | |
tune | -0.014 |
sensation | -0.014 |
dolphin | -0.013 |
shark | -0.012 |
streaming | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | +6.30e-03 |
Pos logit contributions | |
counts | +0.105 |
ends | +0.101 |
One | +0.096 |
Or | +0.091 |
Ready | +0.081 |
Neg logit contributions | |
tune | -0.175 |
sensation | -0.170 |
dolphin | -0.159 |
shark | -0.154 |
Dolphin | -0.148 |
Token | Jack |
Token ID | 3619 |
Feature activation | +1.94e-02 |
Pos logit contributions | |
tune | +0.011 |
sensation | +0.010 |
dolphin | +0.009 |
lull | +0.009 |
shark | +0.009 |
Neg logit contributions | |
counts | -0.011 |
One | -0.011 |
Or | -0.011 |
ends | -0.011 |
Ready | -0.010 |
Token | 's |
Token ID | 338 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
tune | +0.040 |
sensation | +0.040 |
dolphin | +0.036 |
shark | +0.035 |
streaming | +0.034 |
Neg logit contributions | |
counts | -0.027 |
ends | -0.027 |
One | -0.026 |
Or | -0.024 |
Ready | -0.022 |
Token | dad |
Token ID | 9955 |
Feature activation | +1.80e-02 |
Pos logit contributions | |
sensation | +0.031 |
tune | +0.030 |
streaming | +0.027 |
streamed | +0.027 |
dolphin | +0.026 |
Neg logit contributions | |
ends | -0.026 |
counts | -0.025 |
Or | -0.024 |
One | -0.024 |
house | -0.022 |
Token | was |
Token ID | 373 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
tune | +0.033 |
sensation | +0.032 |
dolphin | +0.029 |
shark | +0.028 |
potato | +0.027 |
Neg logit contributions | |
ends | -0.029 |
counts | -0.029 |
One | -0.028 |
Or | -0.027 |
Ready | -0.025 |
Token | , |
Token ID | 11 |
Feature activation | +2.46e-02 |
Pos logit contributions | |
tune | +0.077 |
sensation | +0.076 |
dolphin | +0.070 |
shark | +0.067 |
streaming | +0.065 |
Neg logit contributions | |
ends | -0.055 |
counts | -0.055 |
One | -0.052 |
Or | -0.050 |
Ready | -0.046 |
Token | she |
Token ID | 673 |
Feature activation | +1.33e-02 |
Pos logit contributions | |
tune | +0.057 |
sensation | +0.057 |
dolphin | +0.053 |
shark | +0.051 |
potato | +0.050 |
Neg logit contributions | |
counts | -0.029 |
ends | -0.028 |
One | -0.025 |
Or | -0.025 |
Ready | -0.022 |
Token | went |
Token ID | 1816 |
Feature activation | +7.58e-03 |
Pos logit contributions | |
tune | +0.022 |
sensation | +0.021 |
dolphin | +0.020 |
shark | +0.019 |
potato | +0.018 |
Neg logit contributions | |
ends | -0.024 |
counts | -0.024 |
One | -0.023 |
Or | -0.022 |
Ready | -0.021 |
Token | to |
Token ID | 284 |
Feature activation | +2.48e-02 |
Pos logit contributions | |
tune | +0.014 |
sensation | +0.014 |
dolphin | +0.013 |
shark | +0.012 |
potato | +0.012 |
Neg logit contributions | |
counts | -0.012 |
ends | -0.012 |
One | -0.011 |
Or | -0.011 |
Ready | -0.010 |
Token | the |
Token ID | 262 |
Feature activation | -5.74e-03 |
Pos logit contributions | |
sensation | +0.045 |
tune | +0.045 |
dolphin | +0.038 |
shark | +0.038 |
streaming | +0.037 |
Neg logit contributions | |
counts | -0.042 |
One | -0.041 |
ends | -0.041 |
Or | -0.038 |
Ready | -0.036 |
Token | zoo |
Token ID | 26626 |
Feature activation | -7.56e-02 |
Pos logit contributions | |
One | +0.007 |
counts | +0.007 |
ends | +0.007 |
Or | +0.006 |
Ready | +0.006 |
Neg logit contributions | |
sensation | -0.013 |
tune | -0.013 |
streaming | -0.012 |
streamed | -0.012 |
dolphin | -0.011 |
Token | with |
Token ID | 351 |
Feature activation | +1.40e-03 |
Pos logit contributions | |
ends | +0.117 |
counts | +0.116 |
One | +0.110 |
Or | +0.108 |
Ready | +0.104 |
Neg logit contributions | |
tune | -0.144 |
sensation | -0.137 |
dolphin | -0.127 |
shark | -0.122 |
face | -0.120 |
Token | her |
Token ID | 607 |
Feature activation | -7.13e-03 |
Pos logit contributions | |
tune | +0.002 |
sensation | +0.002 |
dolphin | +0.002 |
shark | +0.002 |
face | +0.002 |
Neg logit contributions | |
ends | -0.002 |
counts | -0.002 |
One | -0.002 |
Or | -0.002 |
Ready | -0.002 |
Token | mom |
Token ID | 1995 |
Feature activation | -1.40e-02 |
Pos logit contributions | |
ends | +0.011 |
counts | +0.011 |
One | +0.011 |
Or | +0.010 |
house | +0.009 |
Neg logit contributions | |
sensation | -0.014 |
tune | -0.014 |
streamed | -0.011 |
streaming | -0.011 |
dolphin | -0.011 |
Token | and |
Token ID | 290 |
Feature activation | +7.59e-03 |
Pos logit contributions | |
ends | +0.014 |
counts | +0.014 |
Or | +0.012 |
One | +0.012 |
Ready | +0.011 |
Neg logit contributions | |
tune | -0.034 |
sensation | -0.033 |
dolphin | -0.032 |
potato | -0.031 |
shark | -0.030 |
Token | dad |
Token ID | 9955 |
Feature activation | +1.44e-02 |
Pos logit contributions | |
sensation | +0.012 |
tune | +0.012 |
dolphin | +0.010 |
shark | +0.010 |
streaming | +0.009 |
Neg logit contributions | |
ends | -0.015 |
counts | -0.014 |
One | -0.014 |
Or | -0.014 |
Ready | -0.013 |