Token | |
Token ID | 220 |
Feature activation | +7.32e-04 |
Pos logit contributions | |
ited | +0.009 |
saw | +0.008 |
heard | +0.007 |
Hey | +0.007 |
year | +0.007 |
Neg logit contributions | |
anks | -0.012 |
balls | -0.012 |
hyd | -0.011 |
heal | -0.011 |
rid | -0.011 |
Token |
|
Token ID | 198 |
Feature activation | +1.79e-02 |
Pos logit contributions | |
ited | +0.001 |
saw | +0.001 |
heard | +0.001 |
two | +0.000 |
little | +0.000 |
Neg logit contributions | |
hyd | -0.002 |
anks | -0.002 |
balls | -0.002 |
rid | -0.002 |
bleed | -0.002 |
Token |
|
Token ID | 198 |
Feature activation | +9.74e-03 |
Pos logit contributions | |
ited | +0.028 |
saw | +0.022 |
heard | +0.021 |
Hey | +0.020 |
year | +0.019 |
Neg logit contributions | |
anks | -0.034 |
balls | -0.034 |
heal | -0.031 |
hyd | -0.031 |
rid | -0.031 |
Token | Le |
Token ID | 3123 |
Feature activation | +2.69e-02 |
Pos logit contributions | |
ited | +0.016 |
saw | +0.013 |
heard | +0.012 |
little | +0.011 |
year | +0.011 |
Neg logit contributions | |
balls | -0.017 |
anks | -0.017 |
hyd | -0.016 |
heal | -0.016 |
fly | -0.016 |
Token | ah |
Token ID | 993 |
Feature activation | +3.68e-02 |
Pos logit contributions | |
ited | +0.063 |
saw | +0.058 |
heard | +0.056 |
Hey | +0.055 |
little | +0.054 |
Neg logit contributions | |
anks | -0.027 |
balls | -0.025 |
fly | -0.025 |
hyd | -0.025 |
ometer | -0.024 |
Token | replied |
Token ID | 8712 |
Feature activation | +1.04e-01 |
Pos logit contributions | |
ited | +0.042 |
Hey | +0.032 |
grown | +0.030 |
saw | +0.028 |
heard | +0.028 |
Neg logit contributions | |
anks | -0.076 |
balls | -0.075 |
hyd | -0.075 |
heal | -0.074 |
ometer | -0.073 |
Token | , |
Token ID | 11 |
Feature activation | +4.20e-02 |
Pos logit contributions | |
ited | +0.043 |
saw | +0.041 |
heard | +0.033 |
little | +0.023 |
two | +0.020 |
Neg logit contributions | |
anks | -0.309 |
hyd | -0.301 |
balls | -0.296 |
rid | -0.294 |
bleed | -0.291 |
Token | " |
Token ID | 366 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
saw | +0.052 |
ited | +0.051 |
heard | +0.048 |
little | +0.044 |
two | +0.043 |
Neg logit contributions | |
anks | -0.092 |
hyd | -0.085 |
rid | -0.085 |
balls | -0.084 |
bleed | -0.083 |
Token | I |
Token ID | 40 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
anks | +0.036 |
balls | +0.034 |
hyd | +0.033 |
rid | +0.033 |
heal | +0.033 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.019 |
heard | -0.018 |
Hey | -0.016 |
year | -0.015 |
Token | have |
Token ID | 423 |
Feature activation | +1.19e-02 |
Pos logit contributions | |
anks | +0.020 |
hyd | +0.019 |
rid | +0.019 |
heal | +0.019 |
balls | +0.019 |
Neg logit contributions | |
ited | -0.019 |
saw | -0.016 |
heard | -0.016 |
Hey | -0.015 |
year | -0.015 |
Token | a |
Token ID | 257 |
Feature activation | +1.27e-02 |
Pos logit contributions | |
ited | +0.017 |
saw | +0.013 |
Hey | +0.013 |
heard | +0.012 |
year | +0.012 |
Neg logit contributions | |
rid | -0.023 |
anks | -0.023 |
hyd | -0.023 |
balls | -0.023 |
heal | -0.023 |
Token | moon |
Token ID | 8824 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
ited | +0.016 |
saw | +0.014 |
heard | +0.013 |
Hey | +0.012 |
did | +0.012 |
Neg logit contributions | |
anks | -0.017 |
balls | -0.016 |
hyd | -0.016 |
heal | -0.016 |
rid | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | -1.57e-02 |
Pos logit contributions | |
anks | +0.029 |
balls | +0.028 |
rid | +0.028 |
bleed | +0.027 |
hyd | +0.027 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.020 |
heard | -0.019 |
two | -0.017 |
little | -0.016 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.77e-02 |
Pos logit contributions | |
anks | +0.034 |
balls | +0.033 |
hyd | +0.032 |
rid | +0.032 |
heal | +0.032 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.018 |
heard | -0.017 |
little | -0.015 |
two | -0.015 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.11e-02 |
Pos logit contributions | |
ited | +0.126 |
saw | +0.097 |
heard | +0.093 |
conflict | +0.086 |
noticed | +0.085 |
Neg logit contributions | |
anks | -0.146 |
balls | -0.145 |
hyd | -0.139 |
rid | -0.137 |
heal | -0.136 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.34e-02 |
Pos logit contributions | |
ited | +0.102 |
Hey | +0.071 |
heard | +0.070 |
saw | +0.068 |
Fine | +0.058 |
Neg logit contributions | |
balls | -0.209 |
anks | -0.205 |
rid | -0.198 |
ectar | -0.196 |
heal | -0.194 |
Token | a |
Token ID | 257 |
Feature activation | +1.04e-01 |
Pos logit contributions | |
ited | +0.103 |
saw | +0.074 |
Hey | +0.070 |
conflict | +0.068 |
heard | +0.067 |
Neg logit contributions | |
heal | -0.149 |
balls | -0.148 |
hyd | -0.145 |
bleed | -0.145 |
rid | -0.143 |
Token | time |
Token ID | 640 |
Feature activation | +7.24e-02 |
Pos logit contributions | |
ited | +0.080 |
saw | +0.062 |
Hey | +0.060 |
heard | +0.058 |
two | +0.054 |
Neg logit contributions | |
sing | -0.244 |
balls | -0.242 |
hyd | -0.241 |
heal | -0.232 |
anks | -0.231 |
Token | , |
Token ID | 11 |
Feature activation | +8.58e-02 |
Pos logit contributions | |
ited | +0.079 |
saw | +0.054 |
heard | +0.052 |
Hey | +0.045 |
noticed | +0.041 |
Neg logit contributions | |
balls | -0.175 |
anks | -0.170 |
hyd | -0.163 |
heal | -0.161 |
sing | -0.161 |
Token | there |
Token ID | 612 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
ited | +0.140 |
oph | +0.108 |
aman | +0.097 |
amusement | +0.095 |
café | +0.089 |
Neg logit contributions | |
balls | -0.163 |
fly | -0.155 |
sing | -0.154 |
ectar | -0.151 |
hyd | -0.150 |
Token | was |
Token ID | 373 |
Feature activation | +7.25e-02 |
Pos logit contributions | |
ited | +0.144 |
vious | +0.091 |
Exc | +0.091 |
Next | +0.090 |
conflict | +0.089 |
Neg logit contributions | |
balls | -0.152 |
heal | -0.147 |
hyd | -0.144 |
fly | -0.141 |
cr | -0.141 |
Token | a |
Token ID | 257 |
Feature activation | +6.22e-02 |
Pos logit contributions | |
ited | +0.106 |
saw | +0.081 |
heard | +0.078 |
sighed | +0.071 |
Hey | +0.070 |
Neg logit contributions | |
balls | -0.142 |
hyd | -0.140 |
anks | -0.139 |
ectar | -0.137 |
rid | -0.136 |
Token | park |
Token ID | 3952 |
Feature activation | +4.00e-03 |
Pos logit contributions | |
ited | +0.015 |
heard | +0.011 |
frowned | +0.010 |
ised | +0.010 |
sighed | +0.010 |
Neg logit contributions | |
heal | -0.019 |
float | -0.018 |
hyd | -0.018 |
balls | -0.018 |
drown | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
ited | +0.005 |
saw | +0.004 |
heard | +0.004 |
Hey | +0.004 |
two | +0.004 |
Neg logit contributions | |
anks | -0.009 |
balls | -0.008 |
hyd | -0.008 |
heal | -0.008 |
rid | -0.008 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.37e-02 |
Pos logit contributions | |
anks | +0.039 |
balls | +0.038 |
hyd | +0.037 |
rid | +0.037 |
heal | +0.036 |
Neg logit contributions | |
ited | -0.018 |
saw | -0.017 |
heard | -0.016 |
two | -0.014 |
little | -0.014 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.83e-02 |
Pos logit contributions | |
ited | +0.124 |
saw | +0.094 |
heard | +0.090 |
conflict | +0.085 |
noticed | +0.084 |
Neg logit contributions | |
anks | -0.134 |
balls | -0.133 |
hyd | -0.128 |
heal | -0.125 |
rid | -0.125 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.94e-02 |
Pos logit contributions | |
ited | +0.095 |
heard | +0.065 |
saw | +0.065 |
Hey | +0.065 |
year | +0.052 |
Neg logit contributions | |
anks | -0.206 |
balls | -0.206 |
rid | -0.196 |
heal | -0.195 |
ometer | -0.194 |
Token | a |
Token ID | 257 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
ited | +0.083 |
saw | +0.062 |
heard | +0.057 |
Hey | +0.055 |
year | +0.051 |
Neg logit contributions | |
balls | -0.153 |
anks | -0.152 |
heal | -0.152 |
hyd | -0.152 |
bleed | -0.149 |
Token | time |
Token ID | 640 |
Feature activation | +7.12e-02 |
Pos logit contributions | |
ited | +0.076 |
saw | +0.056 |
Hey | +0.053 |
heard | +0.052 |
two | +0.047 |
Neg logit contributions | |
hyd | -0.249 |
sing | -0.247 |
balls | -0.247 |
anks | -0.241 |
heal | -0.241 |
Token | , |
Token ID | 11 |
Feature activation | +8.43e-02 |
Pos logit contributions | |
ited | +0.070 |
saw | +0.049 |
heard | +0.047 |
Hey | +0.040 |
noticed | +0.035 |
Neg logit contributions | |
balls | -0.177 |
anks | -0.175 |
hyd | -0.169 |
heal | -0.166 |
sing | -0.164 |
Token | there |
Token ID | 612 |
Feature activation | +8.22e-02 |
Pos logit contributions | |
ited | +0.126 |
oph | +0.089 |
aman | +0.079 |
Hey | +0.075 |
sighed | +0.074 |
Neg logit contributions | |
balls | -0.171 |
fly | -0.166 |
ectar | -0.162 |
hyd | -0.162 |
heal | -0.162 |
Token | was |
Token ID | 373 |
Feature activation | +6.65e-02 |
Pos logit contributions | |
ited | +0.125 |
conflict | +0.076 |
Hey | +0.076 |
enough | +0.076 |
vious | +0.076 |
Neg logit contributions | |
balls | -0.158 |
heal | -0.155 |
hyd | -0.154 |
fly | -0.149 |
anks | -0.148 |
Token | a |
Token ID | 257 |
Feature activation | +6.04e-02 |
Pos logit contributions | |
ited | +0.094 |
saw | +0.072 |
heard | +0.069 |
Hey | +0.062 |
sighed | +0.062 |
Neg logit contributions | |
balls | -0.131 |
hyd | -0.131 |
anks | -0.129 |
rid | -0.127 |
uggets | -0.127 |
Token | again |
Token ID | 757 |
Feature activation | +9.32e-03 |
Pos logit contributions | |
ited | +0.035 |
saw | +0.033 |
heard | +0.032 |
year | +0.028 |
little | +0.028 |
Neg logit contributions | |
anks | -0.048 |
balls | -0.046 |
hyd | -0.045 |
rid | -0.044 |
bleed | -0.044 |
Token | . |
Token ID | 13 |
Feature activation | -2.57e-02 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.015 |
Hey | +0.015 |
heard | +0.014 |
year | +0.014 |
Neg logit contributions | |
anks | -0.014 |
hyd | -0.014 |
balls | -0.014 |
rid | -0.013 |
heal | -0.013 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +6.39e-02 |
Pos logit contributions | |
anks | +0.057 |
balls | +0.056 |
hyd | +0.055 |
rid | +0.054 |
heal | +0.054 |
Neg logit contributions | |
saw | -0.029 |
ited | -0.028 |
heard | -0.026 |
two | -0.024 |
little | -0.024 |
Token | Once |
Token ID | 7454 |
Feature activation | +7.92e-02 |
Pos logit contributions | |
ited | +0.071 |
saw | +0.055 |
heard | +0.051 |
noticed | +0.043 |
little | +0.042 |
Neg logit contributions | |
anks | -0.152 |
balls | -0.149 |
hyd | -0.145 |
rid | -0.142 |
heal | -0.142 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.72e-02 |
Pos logit contributions | |
ited | +0.083 |
saw | +0.057 |
heard | +0.057 |
Hey | +0.055 |
year | +0.046 |
Neg logit contributions | |
anks | -0.190 |
balls | -0.189 |
hyd | -0.180 |
rid | -0.179 |
heal | -0.179 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.071 |
saw | +0.056 |
heard | +0.051 |
Hey | +0.045 |
year | +0.044 |
Neg logit contributions | |
anks | -0.158 |
balls | -0.157 |
hyd | -0.155 |
heal | -0.153 |
rid | -0.152 |
Token | time |
Token ID | 640 |
Feature activation | +6.83e-02 |
Pos logit contributions | |
ited | +0.068 |
saw | +0.047 |
heard | +0.042 |
Hey | +0.038 |
two | +0.033 |
Neg logit contributions | |
balls | -0.267 |
hyd | -0.266 |
anks | -0.265 |
sing | -0.260 |
heal | -0.259 |
Token | there |
Token ID | 612 |
Feature activation | +7.26e-02 |
Pos logit contributions | |
ited | +0.053 |
saw | +0.040 |
heard | +0.037 |
Hey | +0.028 |
year | +0.026 |
Neg logit contributions | |
anks | -0.181 |
balls | -0.179 |
hyd | -0.175 |
rid | -0.171 |
heal | -0.171 |
Token | was |
Token ID | 373 |
Feature activation | +7.19e-02 |
Pos logit contributions | |
ited | +0.096 |
enough | +0.059 |
heard | +0.058 |
conflict | +0.057 |
saw | +0.057 |
Neg logit contributions | |
balls | -0.153 |
hyd | -0.150 |
heal | -0.149 |
anks | -0.146 |
fly | -0.146 |
Token | a |
Token ID | 257 |
Feature activation | +6.07e-02 |
Pos logit contributions | |
ited | +0.071 |
saw | +0.058 |
heard | +0.054 |
Hey | +0.044 |
year | +0.044 |
Neg logit contributions | |
anks | -0.176 |
balls | -0.173 |
hyd | -0.170 |
rid | -0.167 |
heal | -0.166 |
Token | tiny |
Token ID | 7009 |
Feature activation | +4.49e-02 |
Pos logit contributions | |
saw | +0.039 |
ited | +0.037 |
little | +0.032 |
heard | +0.031 |
year | +0.024 |
Neg logit contributions | |
anks | -0.174 |
balls | -0.164 |
rid | -0.162 |
hyd | -0.160 |
bleed | -0.159 |
Token | smile |
Token ID | 8212 |
Feature activation | +1.19e-02 |
Pos logit contributions | |
ited | +0.025 |
saw | +0.024 |
heard | +0.022 |
little | +0.020 |
two | +0.019 |
Neg logit contributions | |
anks | -0.050 |
balls | -0.048 |
hyd | -0.047 |
rid | -0.046 |
ometers | -0.045 |
Token | . |
Token ID | 13 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.016 |
heard | +0.015 |
two | +0.014 |
year | +0.014 |
Neg logit contributions | |
anks | -0.023 |
balls | -0.022 |
hyd | -0.022 |
rid | -0.022 |
heal | -0.021 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.40e-02 |
Pos logit contributions | |
anks | +0.029 |
rid | +0.028 |
balls | +0.028 |
hyd | +0.027 |
bleed | +0.027 |
Neg logit contributions | |
saw | -0.010 |
heard | -0.008 |
two | -0.008 |
little | -0.008 |
ited | -0.008 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.32e-02 |
Pos logit contributions | |
ited | +0.129 |
saw | +0.101 |
heard | +0.096 |
Hey | +0.088 |
noticed | +0.087 |
Neg logit contributions | |
anks | -0.165 |
balls | -0.163 |
hyd | -0.158 |
rid | -0.155 |
heal | -0.155 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.30e-02 |
Pos logit contributions | |
ited | +0.098 |
heard | +0.067 |
saw | +0.066 |
Hey | +0.065 |
year | +0.052 |
Neg logit contributions | |
balls | -0.221 |
anks | -0.219 |
rid | -0.209 |
heal | -0.208 |
hyd | -0.208 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.094 |
saw | +0.068 |
heard | +0.062 |
Hey | +0.062 |
conflict | +0.058 |
Neg logit contributions | |
balls | -0.156 |
heal | -0.156 |
hyd | -0.155 |
anks | -0.154 |
bleed | -0.152 |
Token | time |
Token ID | 640 |
Feature activation | +7.46e-02 |
Pos logit contributions | |
ited | +0.072 |
saw | +0.052 |
heard | +0.047 |
Hey | +0.046 |
two | +0.039 |
Neg logit contributions | |
hyd | -0.257 |
balls | -0.255 |
sing | -0.250 |
anks | -0.249 |
heal | -0.249 |
Token | , |
Token ID | 11 |
Feature activation | +8.80e-02 |
Pos logit contributions | |
ited | +0.069 |
saw | +0.050 |
heard | +0.047 |
Hey | +0.039 |
noticed | +0.035 |
Neg logit contributions | |
balls | -0.188 |
anks | -0.188 |
hyd | -0.182 |
heal | -0.177 |
rid | -0.176 |
Token | in |
Token ID | 287 |
Feature activation | +7.18e-02 |
Pos logit contributions | |
ited | +0.155 |
oph | +0.117 |
aman | +0.106 |
heard | +0.102 |
amusement | +0.102 |
Neg logit contributions | |
balls | -0.157 |
sing | -0.150 |
hyd | -0.149 |
fly | -0.148 |
heal | -0.147 |
Token | a |
Token ID | 257 |
Feature activation | +7.01e-02 |
Pos logit contributions | |
ited | +0.099 |
saw | +0.075 |
heard | +0.073 |
Hey | +0.067 |
noticed | +0.066 |
Neg logit contributions | |
hyd | -0.148 |
heal | -0.144 |
balls | -0.144 |
rid | -0.140 |
bleed | -0.139 |
Token | small |
Token ID | 1402 |
Feature activation | +5.26e-02 |
Pos logit contributions | |
ited | +0.098 |
Hey | +0.072 |
heard | +0.072 |
enough | +0.068 |
saw | +0.067 |
Neg logit contributions | |
hyd | -0.141 |
heal | -0.135 |
balls | -0.130 |
sing | -0.129 |
drown | -0.129 |
Token | safe |
Token ID | 3338 |
Feature activation | -2.25e-02 |
Pos logit contributions | |
anks | +0.116 |
balls | +0.111 |
hyd | +0.107 |
rid | +0.104 |
heal | +0.102 |
Neg logit contributions | |
ited | -0.107 |
saw | -0.102 |
heard | -0.097 |
little | -0.087 |
two | -0.087 |
Token | . |
Token ID | 13 |
Feature activation | -2.21e-02 |
Pos logit contributions | |
anks | +0.045 |
balls | +0.044 |
hyd | +0.041 |
ophone | +0.041 |
rid | +0.041 |
Neg logit contributions | |
ited | -0.030 |
saw | -0.029 |
heard | -0.028 |
two | -0.025 |
year | -0.025 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.29e-02 |
Pos logit contributions | |
anks | +0.057 |
balls | +0.057 |
hyd | +0.056 |
rid | +0.056 |
bleed | +0.055 |
Neg logit contributions | |
saw | -0.017 |
ited | -0.014 |
little | -0.014 |
heard | -0.014 |
two | -0.013 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.97e-02 |
Pos logit contributions | |
ited | +0.121 |
saw | +0.092 |
heard | +0.088 |
conflict | +0.083 |
noticed | +0.082 |
Neg logit contributions | |
anks | -0.134 |
balls | -0.132 |
hyd | -0.128 |
rid | -0.125 |
heal | -0.124 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.97e-02 |
Pos logit contributions | |
ited | +0.100 |
Hey | +0.072 |
heard | +0.068 |
saw | +0.066 |
Fine | +0.058 |
Neg logit contributions | |
balls | -0.203 |
anks | -0.203 |
rid | -0.194 |
ectar | -0.194 |
ometer | -0.192 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.093 |
saw | +0.067 |
Hey | +0.062 |
heard | +0.061 |
conflict | +0.059 |
Neg logit contributions | |
heal | -0.146 |
balls | -0.145 |
hyd | -0.143 |
bleed | -0.143 |
anks | -0.142 |
Token | time |
Token ID | 640 |
Feature activation | +7.28e-02 |
Pos logit contributions | |
ited | +0.077 |
saw | +0.060 |
Hey | +0.058 |
heard | +0.055 |
two | +0.051 |
Neg logit contributions | |
sing | -0.240 |
hyd | -0.239 |
balls | -0.238 |
anks | -0.230 |
heal | -0.230 |
Token | , |
Token ID | 11 |
Feature activation | +8.47e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.051 |
heard | +0.049 |
Hey | +0.042 |
noticed | +0.038 |
Neg logit contributions | |
balls | -0.179 |
anks | -0.176 |
hyd | -0.171 |
heal | -0.168 |
ectar | -0.166 |
Token | there |
Token ID | 612 |
Feature activation | +8.41e-02 |
Pos logit contributions | |
ited | +0.131 |
oph | +0.096 |
aman | +0.087 |
amusement | +0.083 |
Fine | +0.081 |
Neg logit contributions | |
balls | -0.165 |
fly | -0.161 |
hyd | -0.157 |
ectar | -0.157 |
heal | -0.155 |
Token | was |
Token ID | 373 |
Feature activation | +6.77e-02 |
Pos logit contributions | |
ited | +0.136 |
vious | +0.084 |
Exc | +0.083 |
conflict | +0.083 |
Hey | +0.082 |
Neg logit contributions | |
balls | -0.155 |
heal | -0.152 |
hyd | -0.149 |
fly | -0.146 |
cr | -0.143 |
Token | a |
Token ID | 257 |
Feature activation | +5.66e-02 |
Pos logit contributions | |
ited | +0.101 |
saw | +0.077 |
heard | +0.074 |
sighed | +0.068 |
Hey | +0.066 |
Neg logit contributions | |
balls | -0.130 |
hyd | -0.129 |
anks | -0.128 |
ectar | -0.126 |
uggets | -0.126 |
Token |
|
Token ID | 198 |
Feature activation | -1.09e-02 |
Pos logit contributions | |
anks | +0.013 |
hyd | +0.013 |
balls | +0.013 |
rid | +0.013 |
bleed | +0.012 |
Neg logit contributions | |
ited | -0.005 |
saw | -0.005 |
heard | -0.005 |
two | -0.004 |
little | -0.004 |
Token | THE |
Token ID | 10970 |
Feature activation | -2.14e-02 |
Pos logit contributions | |
anks | +0.025 |
balls | +0.025 |
hyd | +0.024 |
rid | +0.024 |
heal | +0.024 |
Neg logit contributions | |
ited | -0.012 |
saw | -0.011 |
heard | -0.010 |
little | -0.008 |
year | -0.008 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +6.92e-02 |
Pos logit contributions | |
anks | +0.054 |
rid | +0.052 |
ectar | +0.052 |
uggets | +0.051 |
hyd | +0.051 |
Neg logit contributions | |
saw | -0.019 |
heard | -0.017 |
did | -0.016 |
thought | -0.015 |
two | -0.014 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.87e-02 |
Pos logit contributions | |
ited | +0.112 |
saw | +0.078 |
heard | +0.076 |
conflict | +0.075 |
enough | +0.073 |
Neg logit contributions | |
anks | -0.130 |
balls | -0.127 |
hyd | -0.125 |
rid | -0.121 |
heal | -0.120 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.03e-02 |
Pos logit contributions | |
ited | +0.107 |
Hey | +0.077 |
heard | +0.074 |
saw | +0.069 |
Fine | +0.067 |
Neg logit contributions | |
balls | -0.191 |
anks | -0.187 |
ometer | -0.183 |
heal | -0.181 |
rid | -0.181 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.106 |
saw | +0.075 |
Hey | +0.072 |
conflict | +0.071 |
oph | +0.068 |
Neg logit contributions | |
heal | -0.140 |
bleed | -0.135 |
balls | -0.135 |
hyd | -0.132 |
rid | -0.132 |
Token | time |
Token ID | 640 |
Feature activation | +7.30e-02 |
Pos logit contributions | |
ited | +0.073 |
saw | +0.056 |
heard | +0.050 |
Hey | +0.047 |
two | +0.042 |
Neg logit contributions | |
hyd | -0.252 |
balls | -0.251 |
sing | -0.249 |
heal | -0.245 |
anks | -0.245 |
Token | there |
Token ID | 612 |
Feature activation | +7.50e-02 |
Pos logit contributions | |
ited | +0.069 |
saw | +0.049 |
heard | +0.047 |
Hey | +0.038 |
two | +0.035 |
Neg logit contributions | |
balls | -0.183 |
anks | -0.182 |
hyd | -0.176 |
heal | -0.173 |
rid | -0.172 |
Token | was |
Token ID | 373 |
Feature activation | +6.35e-02 |
Pos logit contributions | |
ited | +0.111 |
conflict | +0.069 |
Next | +0.067 |
enough | +0.067 |
Exc | +0.066 |
Neg logit contributions | |
balls | -0.144 |
heal | -0.144 |
hyd | -0.142 |
fly | -0.140 |
rid | -0.138 |
Token | a |
Token ID | 257 |
Feature activation | +5.89e-02 |
Pos logit contributions | |
ited | +0.083 |
saw | +0.065 |
heard | +0.062 |
Hey | +0.054 |
year | +0.053 |
Neg logit contributions | |
balls | -0.133 |
hyd | -0.132 |
anks | -0.130 |
heal | -0.129 |
rid | -0.129 |
Token | little |
Token ID | 1310 |
Feature activation | +4.26e-02 |
Pos logit contributions | |
ited | +0.032 |
saw | +0.028 |
heard | +0.022 |
little | +0.018 |
year | +0.015 |
Neg logit contributions | |
anks | -0.173 |
balls | -0.166 |
hyd | -0.163 |
rid | -0.162 |
heal | -0.160 |
Token | together |
Token ID | 1978 |
Feature activation | +1.87e-02 |
Pos logit contributions | |
anks | +0.042 |
drown | +0.039 |
bleed | +0.039 |
hyd | +0.039 |
rid | +0.039 |
Neg logit contributions | |
ited | -0.025 |
saw | -0.024 |
heard | -0.023 |
two | -0.022 |
little | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | -1.14e-02 |
Pos logit contributions | |
ited | +0.028 |
saw | +0.026 |
heard | +0.025 |
year | +0.022 |
two | +0.022 |
Neg logit contributions | |
anks | -0.036 |
balls | -0.035 |
hyd | -0.034 |
rid | -0.034 |
heal | -0.033 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.27e-02 |
Pos logit contributions | |
anks | +0.025 |
balls | +0.025 |
hyd | +0.024 |
rid | +0.024 |
heal | +0.024 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.013 |
heard | -0.012 |
little | -0.010 |
two | -0.010 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.93e-02 |
Pos logit contributions | |
ited | +0.129 |
saw | +0.101 |
heard | +0.096 |
Hey | +0.089 |
noticed | +0.088 |
Neg logit contributions | |
anks | -0.159 |
balls | -0.157 |
hyd | -0.152 |
rid | -0.150 |
heal | -0.150 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.28e-02 |
Pos logit contributions | |
ited | +0.101 |
Hey | +0.073 |
heard | +0.069 |
saw | +0.067 |
Fine | +0.060 |
Neg logit contributions | |
balls | -0.203 |
anks | -0.201 |
ectar | -0.192 |
rid | -0.192 |
heal | -0.190 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.102 |
saw | +0.073 |
Hey | +0.069 |
conflict | +0.066 |
heard | +0.066 |
Neg logit contributions | |
heal | -0.148 |
balls | -0.147 |
hyd | -0.144 |
bleed | -0.143 |
anks | -0.142 |
Token | time |
Token ID | 640 |
Feature activation | +7.01e-02 |
Pos logit contributions | |
ited | +0.077 |
saw | +0.059 |
Hey | +0.055 |
heard | +0.054 |
two | +0.049 |
Neg logit contributions | |
balls | -0.244 |
hyd | -0.243 |
sing | -0.243 |
anks | -0.238 |
heal | -0.235 |
Token | , |
Token ID | 11 |
Feature activation | +8.12e-02 |
Pos logit contributions | |
ited | +0.071 |
saw | +0.050 |
heard | +0.048 |
Hey | +0.042 |
noticed | +0.037 |
Neg logit contributions | |
balls | -0.172 |
anks | -0.170 |
hyd | -0.164 |
heal | -0.161 |
sing | -0.159 |
Token | in |
Token ID | 287 |
Feature activation | +6.92e-02 |
Pos logit contributions | |
ited | +0.132 |
oph | +0.103 |
aman | +0.094 |
amusement | +0.091 |
Exc | +0.085 |
Neg logit contributions | |
balls | -0.151 |
fly | -0.144 |
heal | -0.143 |
sing | -0.142 |
hyd | -0.141 |
Token | a |
Token ID | 257 |
Feature activation | +6.85e-02 |
Pos logit contributions | |
ited | +0.104 |
saw | +0.078 |
heard | +0.076 |
Hey | +0.073 |
noticed | +0.071 |
Neg logit contributions | |
hyd | -0.132 |
heal | -0.128 |
balls | -0.128 |
rid | -0.125 |
sing | -0.123 |
Token | big |
Token ID | 1263 |
Feature activation | +5.29e-02 |
Pos logit contributions | |
ited | +0.098 |
Hey | +0.074 |
heard | +0.071 |
enough | +0.069 |
sighed | +0.068 |
Neg logit contributions | |
hyd | -0.132 |
heal | -0.126 |
drown | -0.120 |
balls | -0.120 |
sing | -0.120 |
Token | shore |
Token ID | 15191 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
anks | +0.045 |
balls | +0.044 |
hyd | +0.043 |
rid | +0.042 |
heal | +0.042 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.017 |
heard | -0.016 |
two | -0.014 |
little | -0.013 |
Token | ! |
Token ID | 0 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
anks | +0.019 |
balls | +0.019 |
hyd | +0.018 |
rid | +0.018 |
heal | +0.018 |
Neg logit contributions | |
ited | -0.015 |
saw | -0.014 |
heard | -0.014 |
two | -0.013 |
little | -0.012 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +6.92e-02 |
Pos logit contributions | |
anks | +0.047 |
balls | +0.046 |
hyd | +0.045 |
rid | +0.044 |
heal | +0.044 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.024 |
heard | -0.023 |
year | -0.019 |
Hey | -0.019 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.73e-02 |
Pos logit contributions | |
ited | +0.108 |
saw | +0.078 |
heard | +0.074 |
conflict | +0.073 |
noticed | +0.069 |
Neg logit contributions | |
anks | -0.133 |
balls | -0.132 |
hyd | -0.126 |
rid | -0.123 |
heal | -0.121 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.19e-02 |
Pos logit contributions | |
ited | +0.107 |
Hey | +0.079 |
heard | +0.073 |
saw | +0.069 |
Fine | +0.069 |
Neg logit contributions | |
balls | -0.185 |
anks | -0.184 |
ectar | -0.178 |
rid | -0.178 |
ometer | -0.177 |
Token | a |
Token ID | 257 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
ited | +0.111 |
saw | +0.079 |
Hey | +0.076 |
conflict | +0.075 |
ised | +0.073 |
Neg logit contributions | |
heal | -0.139 |
balls | -0.136 |
bleed | -0.135 |
rid | -0.132 |
hyd | -0.131 |
Token | time |
Token ID | 640 |
Feature activation | +6.98e-02 |
Pos logit contributions | |
ited | +0.078 |
saw | +0.061 |
Hey | +0.056 |
heard | +0.056 |
two | +0.049 |
Neg logit contributions | |
sing | -0.240 |
hyd | -0.240 |
balls | -0.240 |
heal | -0.232 |
anks | -0.230 |
Token | there |
Token ID | 612 |
Feature activation | +7.77e-02 |
Pos logit contributions | |
ited | +0.082 |
saw | +0.056 |
heard | +0.054 |
Hey | +0.048 |
conflict | +0.044 |
Neg logit contributions | |
balls | -0.164 |
anks | -0.157 |
hyd | -0.152 |
aro | -0.151 |
ectar | -0.151 |
Token | lived |
Token ID | 5615 |
Feature activation | +7.33e-02 |
Pos logit contributions | |
ited | +0.129 |
conflict | +0.086 |
enough | +0.086 |
Next | +0.085 |
weeks | +0.082 |
Neg logit contributions | |
heal | -0.132 |
hyd | -0.131 |
balls | -0.129 |
fly | -0.128 |
cr | -0.127 |
Token | a |
Token ID | 257 |
Feature activation | +6.29e-02 |
Pos logit contributions | |
ited | +0.087 |
saw | +0.066 |
heard | +0.062 |
year | +0.054 |
Hey | +0.052 |
Neg logit contributions | |
balls | -0.167 |
anks | -0.166 |
hyd | -0.164 |
rid | -0.162 |
heal | -0.159 |
Token | giant |
Token ID | 6175 |
Feature activation | +4.23e-02 |
Pos logit contributions | |
ited | +0.042 |
saw | +0.037 |
heard | +0.031 |
little | +0.025 |
year | +0.023 |
Neg logit contributions | |
anks | -0.176 |
balls | -0.169 |
hyd | -0.167 |
rid | -0.165 |
heal | -0.163 |
Token | ball |
Token ID | 2613 |
Feature activation | -3.30e-02 |
Pos logit contributions | |
anks | +0.026 |
balls | +0.025 |
hyd | +0.024 |
rid | +0.024 |
heal | +0.023 |
Neg logit contributions | |
ited | -0.026 |
saw | -0.025 |
heard | -0.024 |
two | -0.021 |
little | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -2.34e-02 |
Pos logit contributions | |
anks | +0.066 |
hyd | +0.064 |
balls | +0.063 |
rid | +0.062 |
drown | +0.060 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.044 |
heard | -0.042 |
two | -0.037 |
little | -0.037 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.31e-02 |
Pos logit contributions | |
anks | +0.050 |
balls | +0.049 |
hyd | +0.049 |
rid | +0.048 |
heal | +0.048 |
Neg logit contributions | |
ited | -0.029 |
saw | -0.027 |
heard | -0.025 |
little | -0.023 |
two | -0.023 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.69e-02 |
Pos logit contributions | |
ited | +0.143 |
saw | +0.108 |
heard | +0.102 |
conflict | +0.099 |
noticed | +0.097 |
Neg logit contributions | |
anks | -0.147 |
balls | -0.144 |
heal | -0.139 |
hyd | -0.139 |
rid | -0.137 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.12e-02 |
Pos logit contributions | |
ited | +0.101 |
Hey | +0.073 |
heard | +0.069 |
saw | +0.066 |
Fine | +0.061 |
Neg logit contributions | |
balls | -0.192 |
anks | -0.191 |
ectar | -0.185 |
rid | -0.183 |
ometer | -0.181 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.095 |
saw | +0.069 |
Hey | +0.063 |
heard | +0.062 |
conflict | +0.061 |
Neg logit contributions | |
heal | -0.148 |
balls | -0.148 |
bleed | -0.145 |
anks | -0.145 |
hyd | -0.144 |
Token | time |
Token ID | 640 |
Feature activation | +7.19e-02 |
Pos logit contributions | |
ited | +0.075 |
saw | +0.058 |
heard | +0.053 |
Hey | +0.052 |
two | +0.048 |
Neg logit contributions | |
hyd | -0.243 |
balls | -0.243 |
sing | -0.243 |
anks | -0.239 |
heal | -0.235 |
Token | , |
Token ID | 11 |
Feature activation | +8.46e-02 |
Pos logit contributions | |
ited | +0.073 |
saw | +0.051 |
heard | +0.049 |
Hey | +0.041 |
noticed | +0.038 |
Neg logit contributions | |
balls | -0.176 |
anks | -0.174 |
hyd | -0.167 |
heal | -0.165 |
ectar | -0.164 |
Token | there |
Token ID | 612 |
Feature activation | +8.15e-02 |
Pos logit contributions | |
ited | +0.126 |
oph | +0.095 |
aman | +0.087 |
amusement | +0.083 |
realised | +0.077 |
Neg logit contributions | |
balls | -0.164 |
fly | -0.162 |
heal | -0.160 |
hyd | -0.159 |
ectar | -0.159 |
Token | was |
Token ID | 373 |
Feature activation | +6.80e-02 |
Pos logit contributions | |
ited | +0.127 |
enough | +0.079 |
conflict | +0.078 |
Exc | +0.078 |
weeks | +0.077 |
Neg logit contributions | |
heal | -0.151 |
balls | -0.150 |
hyd | -0.147 |
fly | -0.144 |
anks | -0.144 |
Token | a |
Token ID | 257 |
Feature activation | +5.94e-02 |
Pos logit contributions | |
ited | +0.106 |
saw | +0.079 |
heard | +0.077 |
sighed | +0.072 |
Hey | +0.070 |
Neg logit contributions | |
balls | -0.126 |
anks | -0.125 |
hyd | -0.125 |
ectar | -0.123 |
uggets | -0.123 |
Token | soup |
Token ID | 17141 |
Feature activation | -5.43e-02 |
Pos logit contributions | |
anks | +0.012 |
balls | +0.012 |
hyd | +0.011 |
heal | +0.011 |
ometers | +0.011 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.013 |
heard | -0.013 |
little | -0.012 |
two | -0.012 |
Token | . |
Token ID | 13 |
Feature activation | -1.93e-02 |
Pos logit contributions | |
anks | +0.106 |
balls | +0.101 |
hyd | +0.101 |
rid | +0.098 |
uggets | +0.097 |
Neg logit contributions | |
ited | -0.076 |
saw | -0.073 |
heard | -0.071 |
did | -0.063 |
little | -0.062 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.85e-02 |
Pos logit contributions | |
anks | +0.046 |
balls | +0.045 |
rid | +0.044 |
hyd | +0.044 |
heal | +0.044 |
Neg logit contributions | |
saw | -0.019 |
ited | -0.018 |
heard | -0.017 |
little | -0.015 |
two | -0.015 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.21e-02 |
Pos logit contributions | |
ited | +0.133 |
saw | +0.105 |
heard | +0.099 |
Hey | +0.091 |
noticed | +0.090 |
Neg logit contributions | |
anks | -0.176 |
balls | -0.173 |
hyd | -0.168 |
rid | -0.166 |
heal | -0.165 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.64e-02 |
Pos logit contributions | |
ited | +0.099 |
heard | +0.068 |
saw | +0.067 |
Hey | +0.067 |
year | +0.055 |
Neg logit contributions | |
balls | -0.216 |
anks | -0.215 |
rid | -0.204 |
hyd | -0.202 |
heal | -0.202 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.094 |
saw | +0.069 |
heard | +0.064 |
Hey | +0.060 |
year | +0.058 |
Neg logit contributions | |
balls | -0.168 |
anks | -0.166 |
heal | -0.166 |
hyd | -0.165 |
bleed | -0.162 |
Token | time |
Token ID | 640 |
Feature activation | +7.40e-02 |
Pos logit contributions | |
ited | +0.073 |
saw | +0.053 |
heard | +0.049 |
Hey | +0.047 |
two | +0.041 |
Neg logit contributions | |
balls | -0.254 |
hyd | -0.254 |
sing | -0.249 |
anks | -0.248 |
heal | -0.245 |
Token | , |
Token ID | 11 |
Feature activation | +8.71e-02 |
Pos logit contributions | |
ited | +0.063 |
saw | +0.046 |
heard | +0.042 |
Hey | +0.034 |
year | +0.031 |
Neg logit contributions | |
anks | -0.192 |
balls | -0.192 |
hyd | -0.186 |
heal | -0.182 |
rid | -0.181 |
Token | there |
Token ID | 612 |
Feature activation | +8.62e-02 |
Pos logit contributions | |
ited | +0.142 |
oph | +0.106 |
aman | +0.097 |
heard | +0.093 |
amusement | +0.093 |
Neg logit contributions | |
balls | -0.171 |
hyd | -0.160 |
sing | -0.158 |
heal | -0.155 |
fly | -0.154 |
Token | was |
Token ID | 373 |
Feature activation | +7.76e-02 |
Pos logit contributions | |
ited | +0.124 |
Exc | +0.074 |
enough | +0.072 |
Hey | +0.072 |
heard | +0.071 |
Neg logit contributions | |
balls | -0.177 |
hyd | -0.171 |
heal | -0.170 |
anks | -0.169 |
fly | -0.164 |
Token | a |
Token ID | 257 |
Feature activation | +6.38e-02 |
Pos logit contributions | |
ited | +0.102 |
saw | +0.077 |
heard | +0.075 |
Hey | +0.066 |
year | +0.064 |
Neg logit contributions | |
anks | -0.166 |
balls | -0.166 |
hyd | -0.163 |
heal | -0.158 |
rid | -0.158 |
Token | end |
Token ID | 886 |
Feature activation | +4.54e-02 |
Pos logit contributions | |
ited | +0.096 |
saw | +0.085 |
heard | +0.083 |
Hey | +0.074 |
noticed | +0.073 |
Neg logit contributions | |
balls | -0.094 |
hyd | -0.093 |
anks | -0.091 |
heal | -0.089 |
rid | -0.088 |
Token | . |
Token ID | 13 |
Feature activation | -2.14e-02 |
Pos logit contributions | |
saw | +0.048 |
heard | +0.046 |
time | +0.041 |
year | +0.040 |
little | +0.040 |
Neg logit contributions | |
anks | -0.111 |
uggets | -0.104 |
balls | -0.104 |
rid | -0.102 |
bleed | -0.101 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.31e-02 |
Pos logit contributions | |
anks | +0.047 |
balls | +0.046 |
hyd | +0.045 |
rid | +0.044 |
heal | +0.044 |
Neg logit contributions | |
ited | -0.026 |
saw | -0.023 |
heard | -0.022 |
Hey | -0.019 |
year | -0.019 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.14e-02 |
Pos logit contributions | |
ited | +0.122 |
saw | +0.093 |
heard | +0.089 |
conflict | +0.085 |
noticed | +0.083 |
Neg logit contributions | |
anks | -0.134 |
balls | -0.129 |
hyd | -0.128 |
heal | -0.124 |
rid | -0.124 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.03e-02 |
Pos logit contributions | |
ited | +0.104 |
Hey | +0.077 |
heard | +0.070 |
Fine | +0.067 |
saw | +0.066 |
Neg logit contributions | |
anks | -0.170 |
balls | -0.169 |
ometer | -0.165 |
ophone | -0.164 |
ectar | -0.163 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.104 |
saw | +0.074 |
Hey | +0.070 |
conflict | +0.068 |
heard | +0.067 |
Neg logit contributions | |
heal | -0.141 |
balls | -0.137 |
bleed | -0.136 |
hyd | -0.135 |
anks | -0.133 |
Token | time |
Token ID | 640 |
Feature activation | +6.96e-02 |
Pos logit contributions | |
ited | +0.081 |
saw | +0.064 |
Hey | +0.060 |
heard | +0.058 |
two | +0.054 |
Neg logit contributions | |
hyd | -0.237 |
sing | -0.236 |
balls | -0.232 |
heal | -0.229 |
anks | -0.224 |
Token | , |
Token ID | 11 |
Feature activation | +7.28e-02 |
Pos logit contributions | |
ited | +0.071 |
saw | +0.051 |
heard | +0.048 |
Hey | +0.041 |
noticed | +0.037 |
Neg logit contributions | |
balls | -0.171 |
anks | -0.169 |
hyd | -0.163 |
heal | -0.160 |
sing | -0.158 |
Token | there |
Token ID | 612 |
Feature activation | +7.60e-02 |
Pos logit contributions | |
ited | +0.106 |
oph | +0.080 |
aman | +0.065 |
weeks | +0.064 |
heard | +0.064 |
Neg logit contributions | |
hyd | -0.143 |
heal | -0.142 |
balls | -0.142 |
sing | -0.140 |
fly | -0.139 |
Token | was |
Token ID | 373 |
Feature activation | +6.27e-02 |
Pos logit contributions | |
ited | +0.126 |
conflict | +0.077 |
Exc | +0.077 |
vious | +0.076 |
weeks | +0.076 |
Neg logit contributions | |
heal | -0.138 |
hyd | -0.137 |
balls | -0.136 |
fly | -0.130 |
anks | -0.129 |
Token | a |
Token ID | 257 |
Feature activation | +5.96e-02 |
Pos logit contributions | |
ited | +0.098 |
saw | +0.074 |
heard | +0.072 |
sighed | +0.067 |
Hey | +0.065 |
Neg logit contributions | |
hyd | -0.116 |
balls | -0.114 |
heal | -0.112 |
uggets | -0.111 |
ectar | -0.111 |
Token | â |
Token ID | 22940 |
Feature activation | +1.08e-02 |
Pos logit contributions | |
anks | +0.070 |
balls | +0.068 |
hyd | +0.067 |
rid | +0.066 |
heal | +0.065 |
Neg logit contributions | |
ited | -0.057 |
saw | -0.053 |
heard | -0.051 |
two | -0.048 |
Hey | -0.047 |
Token | € |
Token ID | 26391 |
Feature activation | -8.00e-03 |
Pos logit contributions | |
ited | +0.021 |
saw | +0.021 |
heard | +0.020 |
little | +0.019 |
two | +0.019 |
Neg logit contributions | |
anks | -0.015 |
hyd | -0.014 |
balls | -0.013 |
heal | -0.013 |
rid | -0.013 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +6.56e-02 |
Pos logit contributions | |
anks | +0.016 |
balls | +0.016 |
hyd | +0.015 |
rid | +0.015 |
heal | +0.015 |
Neg logit contributions | |
ited | -0.011 |
saw | -0.010 |
heard | -0.010 |
Hey | -0.009 |
year | -0.009 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.29e-02 |
Pos logit contributions | |
ited | +0.098 |
saw | +0.071 |
heard | +0.068 |
conflict | +0.067 |
enough | +0.062 |
Neg logit contributions | |
anks | -0.128 |
balls | -0.128 |
hyd | -0.122 |
rid | -0.119 |
ectar | -0.117 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.32e-02 |
Pos logit contributions | |
ited | +0.105 |
Hey | +0.078 |
heard | +0.072 |
Fine | +0.070 |
saw | +0.068 |
Neg logit contributions | |
balls | -0.174 |
anks | -0.171 |
ectar | -0.166 |
ometer | -0.165 |
rid | -0.165 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.116 |
saw | +0.083 |
conflict | +0.080 |
Hey | +0.077 |
ised | +0.076 |
Neg logit contributions | |
heal | -0.138 |
balls | -0.136 |
bleed | -0.134 |
rid | -0.132 |
hyd | -0.131 |
Token | time |
Token ID | 640 |
Feature activation | +6.62e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.058 |
heard | +0.053 |
Hey | +0.048 |
two | +0.044 |
Neg logit contributions | |
hyd | -0.248 |
balls | -0.244 |
sing | -0.243 |
heal | -0.236 |
anks | -0.235 |
Token | there |
Token ID | 612 |
Feature activation | +7.45e-02 |
Pos logit contributions | |
ited | +0.062 |
saw | +0.044 |
heard | +0.041 |
Hey | +0.034 |
year | +0.031 |
Neg logit contributions | |
balls | -0.167 |
anks | -0.166 |
hyd | -0.161 |
heal | -0.157 |
rid | -0.157 |
Token | were |
Token ID | 547 |
Feature activation | +6.62e-02 |
Pos logit contributions | |
ited | +0.111 |
enough | +0.076 |
conflict | +0.073 |
weeks | +0.071 |
Next | +0.069 |
Neg logit contributions | |
hyd | -0.139 |
heal | -0.138 |
balls | -0.135 |
fly | -0.134 |
cr | -0.133 |
Token | two |
Token ID | 734 |
Feature activation | +7.38e-02 |
Pos logit contributions | |
ited | +0.066 |
saw | +0.040 |
heard | +0.038 |
sighed | +0.030 |
Hey | +0.029 |
Neg logit contributions | |
balls | -0.164 |
hyd | -0.163 |
ectar | -0.160 |
anks | -0.159 |
uggets | -0.159 |
Token | best |
Token ID | 1266 |
Feature activation | +6.04e-02 |
Pos logit contributions | |
ited | +0.123 |
saw | +0.107 |
heard | +0.104 |
Hey | +0.099 |
did | +0.093 |
Neg logit contributions | |
anks | -0.128 |
balls | -0.128 |
hyd | -0.123 |
rid | -0.118 |
heal | -0.117 |
Token | safe |
Token ID | 3338 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
anks | +0.060 |
balls | +0.058 |
hyd | +0.056 |
rid | +0.055 |
heal | +0.054 |
Neg logit contributions | |
ited | -0.044 |
saw | -0.041 |
heard | -0.039 |
little | -0.035 |
two | -0.035 |
Token | . |
Token ID | 13 |
Feature activation | -1.75e-02 |
Pos logit contributions | |
anks | +0.043 |
balls | +0.041 |
rid | +0.040 |
bleed | +0.040 |
hyd | +0.040 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.027 |
heard | -0.026 |
two | -0.023 |
enough | -0.022 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
anks | +0.034 |
rid | +0.034 |
balls | +0.033 |
hyd | +0.033 |
bleed | +0.033 |
Neg logit contributions | |
saw | -0.024 |
ited | -0.023 |
heard | -0.022 |
little | -0.021 |
two | -0.020 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.12e-02 |
Pos logit contributions | |
ited | +0.141 |
saw | +0.108 |
heard | +0.102 |
conflict | +0.096 |
Hey | +0.096 |
Neg logit contributions | |
anks | -0.157 |
balls | -0.155 |
hyd | -0.149 |
heal | -0.149 |
rid | -0.147 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.15e-02 |
Pos logit contributions | |
ited | +0.104 |
Hey | +0.073 |
heard | +0.070 |
saw | +0.067 |
Fine | +0.061 |
Neg logit contributions | |
balls | -0.206 |
anks | -0.203 |
ometer | -0.194 |
ectar | -0.193 |
rid | -0.192 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.101 |
saw | +0.071 |
Hey | +0.067 |
conflict | +0.066 |
heard | +0.064 |
Neg logit contributions | |
heal | -0.145 |
balls | -0.145 |
bleed | -0.141 |
hyd | -0.141 |
anks | -0.140 |
Token | time |
Token ID | 640 |
Feature activation | +7.23e-02 |
Pos logit contributions | |
ited | +0.078 |
saw | +0.058 |
heard | +0.054 |
Hey | +0.053 |
two | +0.049 |
Neg logit contributions | |
hyd | -0.243 |
balls | -0.242 |
sing | -0.240 |
anks | -0.234 |
heal | -0.233 |
Token | , |
Token ID | 11 |
Feature activation | +8.28e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.052 |
heard | +0.050 |
Hey | +0.043 |
noticed | +0.039 |
Neg logit contributions | |
balls | -0.177 |
anks | -0.175 |
hyd | -0.169 |
heal | -0.166 |
ectar | -0.164 |
Token | there |
Token ID | 612 |
Feature activation | +8.08e-02 |
Pos logit contributions | |
ited | +0.137 |
oph | +0.107 |
aman | +0.097 |
amusement | +0.093 |
realised | +0.092 |
Neg logit contributions | |
balls | -0.152 |
hyd | -0.144 |
heal | -0.144 |
sing | -0.144 |
ectar | -0.142 |
Token | was |
Token ID | 373 |
Feature activation | +7.05e-02 |
Pos logit contributions | |
ited | +0.124 |
enough | +0.077 |
Exc | +0.077 |
conflict | +0.076 |
weeks | +0.075 |
Neg logit contributions | |
balls | -0.153 |
heal | -0.151 |
hyd | -0.150 |
anks | -0.145 |
fly | -0.142 |
Token | a |
Token ID | 257 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
ited | +0.106 |
saw | +0.079 |
heard | +0.077 |
sighed | +0.071 |
Hey | +0.070 |
Neg logit contributions | |
balls | -0.136 |
hyd | -0.134 |
anks | -0.134 |
ectar | -0.132 |
heal | -0.130 |
Token | her |
Token ID | 607 |
Feature activation | +2.83e-02 |
Pos logit contributions | |
ited | +0.039 |
saw | +0.033 |
heard | +0.031 |
Hey | +0.029 |
noticed | +0.027 |
Neg logit contributions | |
rid | -0.054 |
anks | -0.054 |
heal | -0.054 |
balls | -0.054 |
hyd | -0.053 |
Token | . |
Token ID | 13 |
Feature activation | -1.39e-02 |
Pos logit contributions | |
ited | +0.041 |
saw | +0.039 |
heard | +0.037 |
little | +0.035 |
two | +0.034 |
Neg logit contributions | |
anks | -0.055 |
hyd | -0.053 |
balls | -0.053 |
rid | -0.051 |
bleed | -0.051 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.04e-02 |
Pos logit contributions | |
anks | +0.034 |
hyd | +0.034 |
rid | +0.034 |
balls | +0.034 |
bleed | +0.033 |
Neg logit contributions | |
saw | -0.012 |
ited | -0.010 |
heard | -0.010 |
little | -0.010 |
two | -0.010 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.55e-02 |
Pos logit contributions | |
ited | +0.128 |
saw | +0.099 |
heard | +0.094 |
conflict | +0.087 |
noticed | +0.086 |
Neg logit contributions | |
anks | -0.154 |
balls | -0.151 |
hyd | -0.146 |
rid | -0.144 |
heal | -0.144 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.99e-02 |
Pos logit contributions | |
ited | +0.096 |
Hey | +0.067 |
heard | +0.064 |
saw | +0.063 |
Fine | +0.055 |
Neg logit contributions | |
balls | -0.195 |
anks | -0.194 |
rid | -0.186 |
ectar | -0.185 |
heal | -0.184 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.102 |
saw | +0.072 |
Hey | +0.068 |
conflict | +0.068 |
ised | +0.065 |
Neg logit contributions | |
heal | -0.140 |
balls | -0.138 |
bleed | -0.136 |
hyd | -0.135 |
rid | -0.133 |
Token | time |
Token ID | 640 |
Feature activation | +7.18e-02 |
Pos logit contributions | |
ited | +0.075 |
saw | +0.057 |
heard | +0.052 |
Hey | +0.051 |
two | +0.046 |
Neg logit contributions | |
hyd | -0.247 |
balls | -0.245 |
sing | -0.243 |
anks | -0.238 |
heal | -0.238 |
Token | , |
Token ID | 11 |
Feature activation | +7.99e-02 |
Pos logit contributions | |
ited | +0.069 |
saw | +0.049 |
heard | +0.046 |
Hey | +0.039 |
noticed | +0.035 |
Neg logit contributions | |
balls | -0.180 |
anks | -0.179 |
hyd | -0.173 |
heal | -0.169 |
rid | -0.168 |
Token | there |
Token ID | 612 |
Feature activation | +8.09e-02 |
Pos logit contributions | |
ited | +0.128 |
oph | +0.100 |
aman | +0.087 |
amusement | +0.085 |
weeks | +0.081 |
Neg logit contributions | |
balls | -0.152 |
hyd | -0.145 |
fly | -0.145 |
sing | -0.145 |
heal | -0.144 |
Token | was |
Token ID | 373 |
Feature activation | +6.73e-02 |
Pos logit contributions | |
ited | +0.126 |
Exc | +0.078 |
conflict | +0.078 |
enough | +0.077 |
weeks | +0.076 |
Neg logit contributions | |
balls | -0.153 |
heal | -0.151 |
hyd | -0.150 |
fly | -0.145 |
anks | -0.144 |
Token | a |
Token ID | 257 |
Feature activation | +5.95e-02 |
Pos logit contributions | |
ited | +0.105 |
saw | +0.079 |
heard | +0.077 |
sighed | +0.072 |
Hey | +0.070 |
Neg logit contributions | |
balls | -0.124 |
hyd | -0.123 |
anks | -0.121 |
ectar | -0.120 |
uggets | -0.120 |
Token | make |
Token ID | 787 |
Feature activation | +6.82e-04 |
Pos logit contributions | |
balls | +0.040 |
ometers | +0.039 |
anks | +0.039 |
ometer | +0.039 |
ophone | +0.037 |
Neg logit contributions | |
saw | -0.032 |
heard | -0.030 |
ited | -0.029 |
little | -0.028 |
two | -0.027 |
Token | honey |
Token ID | 12498 |
Feature activation | -4.56e-02 |
Pos logit contributions | |
two | +0.001 |
saw | +0.001 |
little | +0.001 |
heard | +0.001 |
enough | +0.000 |
Neg logit contributions | |
hyd | -0.002 |
anks | -0.002 |
ometer | -0.002 |
ophone | -0.002 |
ometers | -0.002 |
Token | again |
Token ID | 757 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
aro | +0.095 |
ometers | +0.094 |
ometer | +0.091 |
ophone | +0.090 |
anks | +0.089 |
Neg logit contributions | |
heard | -0.053 |
saw | -0.051 |
two | -0.049 |
little | -0.047 |
enough | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | -1.66e-02 |
Pos logit contributions | |
anks | +0.029 |
balls | +0.029 |
ometer | +0.028 |
ophone | +0.028 |
ometers | +0.028 |
Neg logit contributions | |
saw | -0.015 |
heard | -0.015 |
ited | -0.014 |
little | -0.013 |
two | -0.013 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.11e-02 |
Pos logit contributions | |
bleed | +0.046 |
uries | +0.046 |
sing | +0.045 |
rid | +0.045 |
hyd | +0.045 |
Neg logit contributions | |
saw | -0.010 |
two | -0.009 |
little | -0.008 |
heard | -0.006 |
, | -0.006 |
Token | One |
Token ID | 3198 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.111 |
saw | +0.091 |
heard | +0.086 |
Hey | +0.077 |
year | +0.076 |
Neg logit contributions | |
anks | -0.170 |
balls | -0.167 |
hyd | -0.163 |
rid | -0.160 |
heal | -0.159 |
Token | day |
Token ID | 1110 |
Feature activation | +6.62e-02 |
Pos logit contributions | |
ited | +0.155 |
Hey | +0.136 |
saw | +0.126 |
heard | +0.124 |
Exc | +0.124 |
Neg logit contributions | |
hyd | -0.169 |
heal | -0.168 |
sing | -0.164 |
balls | -0.163 |
ectar | -0.162 |
Token | , |
Token ID | 11 |
Feature activation | +7.22e-02 |
Pos logit contributions | |
ited | +0.042 |
saw | +0.033 |
heard | +0.029 |
two | +0.020 |
little | +0.020 |
Neg logit contributions | |
anks | -0.186 |
balls | -0.181 |
hyd | -0.179 |
rid | -0.176 |
heal | -0.175 |
Token | a |
Token ID | 257 |
Feature activation | +7.68e-02 |
Pos logit contributions | |
ited | +0.080 |
saw | +0.062 |
heard | +0.060 |
year | +0.051 |
Hey | +0.050 |
Neg logit contributions | |
balls | -0.168 |
anks | -0.167 |
hyd | -0.164 |
rid | -0.162 |
heal | -0.160 |
Token | little |
Token ID | 1310 |
Feature activation | +5.23e-02 |
Pos logit contributions | |
ited | +0.047 |
saw | +0.042 |
heard | +0.035 |
little | +0.029 |
two | +0.024 |
Neg logit contributions | |
anks | -0.219 |
balls | -0.210 |
hyd | -0.207 |
rid | -0.206 |
heal | -0.202 |
Token | boy |
Token ID | 2933 |
Feature activation | +5.17e-02 |
Pos logit contributions | |
ited | +0.051 |
saw | +0.045 |
heard | +0.041 |
little | +0.035 |
year | +0.034 |
Neg logit contributions | |
anks | -0.129 |
balls | -0.125 |
hyd | -0.123 |
rid | -0.121 |
heal | -0.120 |
Token | slides |
Token ID | 19392 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
balls | +0.071 |
anks | +0.069 |
hyd | +0.068 |
rid | +0.067 |
heal | +0.067 |
Neg logit contributions | |
ited | -0.057 |
saw | -0.044 |
heard | -0.044 |
year | -0.043 |
Hey | -0.040 |
Token | . |
Token ID | 13 |
Feature activation | -2.33e-02 |
Pos logit contributions | |
anks | +0.024 |
balls | +0.023 |
hyd | +0.023 |
rid | +0.022 |
heal | +0.022 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.018 |
heard | -0.017 |
Hey | -0.016 |
year | -0.015 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.88e-02 |
Pos logit contributions | |
anks | +0.051 |
balls | +0.050 |
hyd | +0.049 |
rid | +0.048 |
heal | +0.047 |
Neg logit contributions | |
ited | -0.029 |
saw | -0.026 |
heard | -0.025 |
little | -0.022 |
two | -0.021 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.04e-02 |
Pos logit contributions | |
ited | +0.127 |
saw | +0.098 |
heard | +0.093 |
conflict | +0.086 |
noticed | +0.086 |
Neg logit contributions | |
anks | -0.150 |
balls | -0.149 |
hyd | -0.142 |
rid | -0.140 |
heal | -0.139 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.31e-02 |
Pos logit contributions | |
ited | +0.102 |
Hey | +0.073 |
heard | +0.070 |
saw | +0.069 |
Fine | +0.058 |
Neg logit contributions | |
balls | -0.206 |
anks | -0.205 |
rid | -0.195 |
heal | -0.195 |
ectar | -0.194 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.099 |
saw | +0.070 |
Hey | +0.066 |
heard | +0.064 |
conflict | +0.063 |
Neg logit contributions | |
heal | -0.151 |
balls | -0.150 |
hyd | -0.148 |
bleed | -0.147 |
anks | -0.147 |
Token | time |
Token ID | 640 |
Feature activation | +7.01e-02 |
Pos logit contributions | |
ited | +0.076 |
saw | +0.057 |
Hey | +0.056 |
heard | +0.053 |
two | +0.048 |
Neg logit contributions | |
hyd | -0.242 |
balls | -0.242 |
sing | -0.241 |
anks | -0.235 |
heal | -0.234 |
Token | , |
Token ID | 11 |
Feature activation | +8.36e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.051 |
heard | +0.049 |
Hey | +0.044 |
noticed | +0.038 |
Neg logit contributions | |
balls | -0.171 |
anks | -0.167 |
hyd | -0.161 |
heal | -0.158 |
sing | -0.157 |
Token | there |
Token ID | 612 |
Feature activation | +8.38e-02 |
Pos logit contributions | |
ited | +0.136 |
oph | +0.099 |
amusement | +0.091 |
aman | +0.090 |
Fine | +0.086 |
Neg logit contributions | |
balls | -0.160 |
fly | -0.151 |
ectar | -0.148 |
sing | -0.148 |
hyd | -0.146 |
Token | was |
Token ID | 373 |
Feature activation | +7.26e-02 |
Pos logit contributions | |
ited | +0.144 |
vious | +0.096 |
Hey | +0.092 |
Exc | +0.092 |
conflict | +0.090 |
Neg logit contributions | |
balls | -0.142 |
heal | -0.140 |
hyd | -0.135 |
float | -0.132 |
fly | -0.130 |
Token | a |
Token ID | 257 |
Feature activation | +6.03e-02 |
Pos logit contributions | |
ited | +0.099 |
saw | +0.076 |
heard | +0.074 |
Hey | +0.066 |
sighed | +0.064 |
Neg logit contributions | |
balls | -0.148 |
anks | -0.147 |
hyd | -0.147 |
ectar | -0.143 |
heal | -0.142 |
Token | day |
Token ID | 1110 |
Feature activation | +2.00e-04 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.016 |
heard | +0.016 |
year | +0.014 |
Hey | +0.014 |
Neg logit contributions | |
anks | -0.026 |
balls | -0.025 |
hyd | -0.025 |
rid | -0.024 |
heal | -0.024 |
Token | !" |
Token ID | 2474 |
Feature activation | -9.05e-03 |
Pos logit contributions | |
ited | +0.000 |
saw | +0.000 |
heard | +0.000 |
year | +0.000 |
two | +0.000 |
Neg logit contributions | |
anks | -0.000 |
balls | -0.000 |
hyd | -0.000 |
rid | -0.000 |
bleed | -0.000 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +6.75e-02 |
Pos logit contributions | |
anks | +0.019 |
balls | +0.019 |
hyd | +0.018 |
rid | +0.018 |
heal | +0.018 |
Neg logit contributions | |
ited | -0.012 |
saw | -0.011 |
heard | -0.011 |
two | -0.009 |
little | -0.009 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.28e-02 |
Pos logit contributions | |
ited | +0.095 |
saw | +0.072 |
heard | +0.067 |
conflict | +0.062 |
noticed | +0.060 |
Neg logit contributions | |
anks | -0.140 |
balls | -0.139 |
hyd | -0.135 |
rid | -0.131 |
heal | -0.129 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.97e-02 |
Pos logit contributions | |
ited | +0.090 |
Hey | +0.064 |
heard | +0.062 |
saw | +0.062 |
Fine | +0.051 |
Neg logit contributions | |
anks | -0.189 |
balls | -0.188 |
rid | -0.181 |
ectar | -0.180 |
heal | -0.180 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.102 |
saw | +0.073 |
Hey | +0.068 |
conflict | +0.068 |
heard | +0.065 |
Neg logit contributions | |
heal | -0.139 |
balls | -0.136 |
bleed | -0.135 |
hyd | -0.135 |
rid | -0.132 |
Token | time |
Token ID | 640 |
Feature activation | +6.91e-02 |
Pos logit contributions | |
ited | +0.076 |
saw | +0.059 |
Hey | +0.054 |
heard | +0.053 |
two | +0.047 |
Neg logit contributions | |
hyd | -0.243 |
sing | -0.241 |
balls | -0.240 |
heal | -0.233 |
anks | -0.231 |
Token | there |
Token ID | 612 |
Feature activation | +7.48e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.051 |
heard | +0.049 |
Hey | +0.043 |
noticed | +0.038 |
Neg logit contributions | |
balls | -0.167 |
anks | -0.163 |
hyd | -0.157 |
heal | -0.154 |
sing | -0.154 |
Token | was |
Token ID | 373 |
Feature activation | +6.92e-02 |
Pos logit contributions | |
ited | +0.119 |
enough | +0.079 |
conflict | +0.079 |
Next | +0.077 |
weeks | +0.077 |
Neg logit contributions | |
heal | -0.130 |
hyd | -0.130 |
balls | -0.129 |
fly | -0.128 |
cr | -0.125 |
Token | a |
Token ID | 257 |
Feature activation | +6.05e-02 |
Pos logit contributions | |
ited | +0.085 |
saw | +0.067 |
heard | +0.063 |
Hey | +0.056 |
year | +0.053 |
Neg logit contributions | |
anks | -0.150 |
balls | -0.150 |
hyd | -0.148 |
rid | -0.145 |
heal | -0.144 |
Token | girl |
Token ID | 2576 |
Feature activation | +4.06e-02 |
Pos logit contributions | |
ited | +0.032 |
saw | +0.028 |
heard | +0.022 |
little | +0.018 |
year | +0.015 |
Neg logit contributions | |
anks | -0.178 |
balls | -0.171 |
rid | -0.168 |
hyd | -0.168 |
heal | -0.165 |
Token | garden |
Token ID | 11376 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
ited | +0.000 |
saw | +0.000 |
heard | +0.000 |
Hey | +0.000 |
two | +0.000 |
Neg logit contributions | |
anks | -0.000 |
balls | -0.000 |
hyd | -0.000 |
rid | -0.000 |
heal | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | -1.11e-02 |
Pos logit contributions | |
bleed | +0.031 |
drown | +0.030 |
anks | +0.030 |
ric | +0.030 |
ometers | +0.029 |
Neg logit contributions | |
saw | -0.014 |
heard | -0.013 |
year | -0.012 |
little | -0.012 |
two | -0.012 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +8.60e-02 |
Pos logit contributions | |
anks | +0.027 |
balls | +0.027 |
hyd | +0.027 |
rid | +0.027 |
bleed | +0.027 |
Neg logit contributions | |
saw | -0.009 |
heard | -0.008 |
ited | -0.008 |
two | -0.007 |
little | -0.007 |
Token | Once |
Token ID | 7454 |
Feature activation | +9.22e-02 |
Pos logit contributions | |
ited | +0.132 |
saw | +0.104 |
heard | +0.098 |
Hey | +0.091 |
noticed | +0.089 |
Neg logit contributions | |
anks | -0.167 |
balls | -0.165 |
hyd | -0.160 |
heal | -0.158 |
rid | -0.158 |
Token | upon |
Token ID | 2402 |
Feature activation | +7.06e-02 |
Pos logit contributions | |
ited | +0.105 |
Hey | +0.073 |
heard | +0.071 |
saw | +0.070 |
Fine | +0.058 |
Neg logit contributions | |
balls | -0.211 |
anks | -0.209 |
rid | -0.198 |
ometer | -0.197 |
heal | -0.197 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.092 |
saw | +0.066 |
heard | +0.060 |
Hey | +0.060 |
conflict | +0.058 |
Neg logit contributions | |
balls | -0.150 |
heal | -0.149 |
hyd | -0.148 |
anks | -0.146 |
bleed | -0.145 |
Token | time |
Token ID | 640 |
Feature activation | +7.33e-02 |
Pos logit contributions | |
ited | +0.077 |
saw | +0.058 |
heard | +0.054 |
Hey | +0.053 |
two | +0.047 |
Neg logit contributions | |
hyd | -0.243 |
balls | -0.242 |
sing | -0.241 |
anks | -0.238 |
heal | -0.235 |
Token | , |
Token ID | 11 |
Feature activation | +8.48e-02 |
Pos logit contributions | |
ited | +0.070 |
saw | +0.050 |
heard | +0.048 |
Hey | +0.040 |
noticed | +0.036 |
Neg logit contributions | |
balls | -0.183 |
anks | -0.182 |
hyd | -0.175 |
heal | -0.172 |
rid | -0.170 |
Token | there |
Token ID | 612 |
Feature activation | +8.30e-02 |
Pos logit contributions | |
ited | +0.134 |
oph | +0.102 |
aman | +0.091 |
amusement | +0.088 |
realised | +0.085 |
Neg logit contributions | |
balls | -0.167 |
fly | -0.157 |
hyd | -0.157 |
sing | -0.155 |
heal | -0.155 |
Token | was |
Token ID | 373 |
Feature activation | +7.21e-02 |
Pos logit contributions | |
ited | +0.120 |
Exc | +0.071 |
Hey | +0.071 |
heard | +0.071 |
conflict | +0.070 |
Neg logit contributions | |
balls | -0.168 |
heal | -0.163 |
hyd | -0.161 |
anks | -0.159 |
fly | -0.155 |
Token | a |
Token ID | 257 |
Feature activation | +6.19e-02 |
Pos logit contributions | |
ited | +0.098 |
saw | +0.075 |
heard | +0.072 |
Hey | +0.064 |
sighed | +0.063 |
Neg logit contributions | |
balls | -0.149 |
anks | -0.148 |
hyd | -0.148 |
heal | -0.143 |
rid | -0.143 |
Token | them |
Token ID | 606 |
Feature activation | -2.36e-02 |
Pos logit contributions | |
anks | +0.023 |
balls | +0.022 |
hyd | +0.021 |
ometer | +0.021 |
bleed | +0.021 |
Neg logit contributions | |
ited | -0.012 |
saw | -0.011 |
heard | -0.011 |
two | -0.010 |
little | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
anks | +0.049 |
hyd | +0.047 |
balls | +0.046 |
bleed | +0.046 |
rid | +0.045 |
Neg logit contributions | |
saw | -0.030 |
ited | -0.030 |
heard | -0.028 |
two | -0.026 |
little | -0.025 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.20e-02 |
Pos logit contributions | |
anks | +0.040 |
balls | +0.038 |
hyd | +0.038 |
heal | +0.038 |
bleed | +0.038 |
Neg logit contributions | |
ited | -0.024 |
saw | -0.023 |
heard | -0.021 |
two | -0.019 |
little | -0.019 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.07e-02 |
Pos logit contributions | |
ited | +0.119 |
saw | +0.089 |
heard | +0.086 |
conflict | +0.082 |
noticed | +0.079 |
Neg logit contributions | |
anks | -0.134 |
balls | -0.132 |
hyd | -0.127 |
rid | -0.124 |
heal | -0.123 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.85e-02 |
Pos logit contributions | |
ited | +0.082 |
saw | +0.057 |
heard | +0.057 |
Hey | +0.053 |
year | +0.046 |
Neg logit contributions | |
anks | -0.193 |
balls | -0.193 |
rid | -0.185 |
hyd | -0.184 |
heal | -0.181 |
Token | a |
Token ID | 257 |
Feature activation | +1.01e-01 |
Pos logit contributions | |
ited | +0.096 |
saw | +0.069 |
Hey | +0.065 |
conflict | +0.063 |
heard | +0.063 |
Neg logit contributions | |
heal | -0.139 |
balls | -0.139 |
bleed | -0.136 |
hyd | -0.136 |
rid | -0.134 |
Token | time |
Token ID | 640 |
Feature activation | +7.15e-02 |
Pos logit contributions | |
ited | +0.075 |
saw | +0.057 |
heard | +0.052 |
Hey | +0.052 |
two | +0.045 |
Neg logit contributions | |
hyd | -0.243 |
balls | -0.242 |
sing | -0.242 |
heal | -0.235 |
anks | -0.234 |
Token | there |
Token ID | 612 |
Feature activation | +7.49e-02 |
Pos logit contributions | |
ited | +0.075 |
saw | +0.051 |
heard | +0.050 |
Hey | +0.044 |
noticed | +0.038 |
Neg logit contributions | |
balls | -0.174 |
anks | -0.170 |
hyd | -0.165 |
heal | -0.162 |
sing | -0.161 |
Token | was |
Token ID | 373 |
Feature activation | +6.82e-02 |
Pos logit contributions | |
ited | +0.117 |
enough | +0.078 |
conflict | +0.078 |
Next | +0.075 |
weeks | +0.074 |
Neg logit contributions | |
hyd | -0.134 |
heal | -0.133 |
balls | -0.131 |
fly | -0.131 |
cr | -0.127 |
Token | a |
Token ID | 257 |
Feature activation | +6.10e-02 |
Pos logit contributions | |
ited | +0.085 |
saw | +0.067 |
heard | +0.064 |
Hey | +0.056 |
year | +0.054 |
Neg logit contributions | |
balls | -0.146 |
anks | -0.146 |
hyd | -0.146 |
rid | -0.141 |
heal | -0.141 |
Token | man |
Token ID | 582 |
Feature activation | +3.01e-02 |
Pos logit contributions | |
ited | +0.031 |
saw | +0.026 |
heard | +0.021 |
little | +0.015 |
year | +0.013 |
Neg logit contributions | |
anks | -0.179 |
balls | -0.173 |
hyd | -0.171 |
rid | -0.170 |
heal | -0.168 |
Token | and |
Token ID | 290 |
Feature activation | -6.45e-02 |
Pos logit contributions | |
anks | +0.091 |
balls | +0.088 |
bleed | +0.088 |
drown | +0.086 |
rid | +0.086 |
Neg logit contributions | |
saw | -0.034 |
ited | -0.033 |
heard | -0.031 |
little | -0.028 |
year | -0.028 |
Token | always |
Token ID | 1464 |
Feature activation | -3.74e-02 |
Pos logit contributions | |
anks | +0.140 |
balls | +0.137 |
hyd | +0.135 |
rid | +0.133 |
heal | +0.132 |
Neg logit contributions | |
ited | -0.082 |
saw | -0.070 |
heard | -0.067 |
Hey | -0.060 |
year | -0.059 |
Token | had |
Token ID | 550 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
hyd | +0.082 |
anks | +0.081 |
rid | +0.079 |
heal | +0.079 |
balls | +0.079 |
Neg logit contributions | |
ited | -0.045 |
saw | -0.036 |
heard | -0.034 |
Hey | -0.033 |
year | -0.032 |
Token | plenty |
Token ID | 6088 |
Feature activation | -5.38e-03 |
Pos logit contributions | |
anks | +0.036 |
balls | +0.035 |
hyd | +0.035 |
rid | +0.035 |
heal | +0.034 |
Neg logit contributions | |
ited | -0.023 |
saw | -0.019 |
heard | -0.018 |
Hey | -0.017 |
year | -0.016 |
Token | of |
Token ID | 286 |
Feature activation | -4.44e-02 |
Pos logit contributions | |
anks | +0.014 |
balls | +0.013 |
hyd | +0.013 |
rid | +0.013 |
heal | +0.013 |
Neg logit contributions | |
ited | -0.005 |
saw | -0.004 |
heard | -0.004 |
year | -0.003 |
two | -0.003 |
Token | rad |
Token ID | 2511 |
Feature activation | -1.11e-01 |
Pos logit contributions | |
anks | +0.087 |
balls | +0.085 |
hyd | +0.083 |
rid | +0.080 |
heal | +0.080 |
Neg logit contributions | |
ited | -0.066 |
saw | -0.060 |
heard | -0.057 |
year | -0.051 |
little | -0.051 |
Token | ishes |
Token ID | 5614 |
Feature activation | -8.63e-02 |
Pos logit contributions | |
anks | +0.136 |
sing | +0.133 |
balls | +0.130 |
heal | +0.129 |
bleed | +0.128 |
Neg logit contributions | |
ited | -0.237 |
saw | -0.228 |
heard | -0.212 |
two | -0.204 |
enough | -0.200 |
Token | to |
Token ID | 284 |
Feature activation | -5.73e-02 |
Pos logit contributions | |
anks | +0.183 |
sing | +0.177 |
bleed | +0.175 |
balls | +0.175 |
hyd | +0.171 |
Neg logit contributions | |
ited | -0.102 |
saw | -0.100 |
heard | -0.094 |
two | -0.084 |
enough | -0.082 |
Token | eat |
Token ID | 4483 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
anks | +0.092 |
balls | +0.089 |
ometer | +0.084 |
hyd | +0.082 |
ophone | +0.082 |
Neg logit contributions | |
ited | -0.104 |
saw | -0.101 |
heard | -0.096 |
little | -0.089 |
two | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | -4.91e-02 |
Pos logit contributions | |
anks | +0.038 |
balls | +0.036 |
sing | +0.035 |
bleed | +0.035 |
hyd | +0.035 |
Neg logit contributions | |
ited | -0.025 |
saw | -0.024 |
heard | -0.022 |
little | -0.021 |
year | -0.021 |
Token | The |
Token ID | 383 |
Feature activation | +2.61e-02 |
Pos logit contributions | |
anks | +0.125 |
balls | +0.125 |
bleed | +0.122 |
rid | +0.121 |
hyd | +0.120 |
Neg logit contributions | |
saw | -0.038 |
ited | -0.037 |
heard | -0.032 |
little | -0.031 |
two | -0.031 |
Token | world |
Token ID | 995 |
Feature activation | -2.87e-02 |
Pos logit contributions | |
anks | +0.003 |
balls | +0.003 |
hyd | +0.003 |
rid | +0.003 |
bleed | +0.003 |
Neg logit contributions | |
ited | -0.001 |
saw | -0.001 |
heard | -0.001 |
little | -0.001 |
year | -0.001 |
Token | where |
Token ID | 810 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
anks | +0.061 |
balls | +0.060 |
hyd | +0.059 |
heal | +0.059 |
rid | +0.058 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.031 |
heard | -0.029 |
Hey | -0.026 |
two | -0.026 |
Token | everyone |
Token ID | 2506 |
Feature activation | -3.99e-02 |
Pos logit contributions | |
anks | +0.040 |
balls | +0.039 |
hyd | +0.038 |
rid | +0.037 |
heal | +0.037 |
Neg logit contributions | |
ited | -0.029 |
saw | -0.027 |
heard | -0.026 |
two | -0.023 |
little | -0.023 |
Token | shared |
Token ID | 4888 |
Feature activation | +1.04e-03 |
Pos logit contributions | |
anks | +0.093 |
balls | +0.091 |
hyd | +0.090 |
rid | +0.088 |
heal | +0.087 |
Neg logit contributions | |
ited | -0.043 |
saw | -0.037 |
heard | -0.035 |
year | -0.029 |
Hey | -0.029 |
Token | their |
Token ID | 511 |
Feature activation | -4.48e-02 |
Pos logit contributions | |
ited | +0.001 |
saw | +0.001 |
heard | +0.001 |
Hey | +0.001 |
year | +0.001 |
Neg logit contributions | |
anks | -0.002 |
balls | -0.002 |
hyd | -0.002 |
heal | -0.002 |
rid | -0.002 |
Token | rad |
Token ID | 2511 |
Feature activation | -1.10e-01 |
Pos logit contributions | |
anks | +0.075 |
balls | +0.073 |
hyd | +0.071 |
rid | +0.070 |
heal | +0.069 |
Neg logit contributions | |
ited | -0.079 |
saw | -0.072 |
heard | -0.069 |
Hey | -0.063 |
year | -0.062 |
Token | ishes |
Token ID | 5614 |
Feature activation | -8.31e-02 |
Pos logit contributions | |
anks | +0.136 |
balls | +0.133 |
bleed | +0.126 |
sing | +0.125 |
heal | +0.124 |
Neg logit contributions | |
ited | -0.233 |
saw | -0.224 |
heard | -0.208 |
little | -0.205 |
two | -0.204 |
Token | . |
Token ID | 13 |
Feature activation | -5.11e-02 |
Pos logit contributions | |
anks | +0.193 |
balls | +0.189 |
bleed | +0.185 |
sing | +0.184 |
hyd | +0.182 |
Neg logit contributions | |
ited | -0.084 |
saw | -0.079 |
heard | -0.073 |
two | -0.068 |
little | -0.067 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +4.59e-02 |
Pos logit contributions | |
balls | +0.132 |
anks | +0.132 |
bleed | +0.129 |
rid | +0.129 |
hyd | +0.127 |
Neg logit contributions | |
saw | -0.038 |
ited | -0.034 |
two | -0.033 |
heard | -0.032 |
little | -0.031 |
Token | Once |
Token ID | 7454 |
Feature activation | +6.77e-02 |
Pos logit contributions | |
ited | +0.070 |
saw | +0.055 |
heard | +0.052 |
conflict | +0.047 |
noticed | +0.047 |
Neg logit contributions | |
anks | -0.091 |
balls | -0.088 |
hyd | -0.086 |
rid | -0.084 |
heal | -0.084 |
Token | upon |
Token ID | 2402 |
Feature activation | +4.39e-02 |
Pos logit contributions | |
ited | +0.074 |
heard | +0.051 |
Hey | +0.051 |
saw | +0.050 |
Fine | +0.040 |
Neg logit contributions | |
balls | -0.157 |
anks | -0.156 |
rid | -0.149 |
ectar | -0.149 |
ometer | -0.148 |
Token | other |
Token ID | 584 |
Feature activation | -1.87e-02 |
Pos logit contributions | |
balls | +0.125 |
anks | +0.124 |
hyd | +0.123 |
rid | +0.121 |
heal | +0.119 |
Neg logit contributions | |
ited | -0.120 |
saw | -0.099 |
heard | -0.096 |
Hey | -0.094 |
year | -0.092 |
Token | things |
Token ID | 1243 |
Feature activation | -6.28e-02 |
Pos logit contributions | |
anks | +0.035 |
balls | +0.034 |
hyd | +0.033 |
rid | +0.033 |
heal | +0.032 |
Neg logit contributions | |
ited | -0.030 |
saw | -0.026 |
heard | -0.024 |
Hey | -0.023 |
year | -0.022 |
Token | that |
Token ID | 326 |
Feature activation | -3.78e-02 |
Pos logit contributions | |
anks | +0.133 |
balls | +0.126 |
hyd | +0.124 |
uggets | +0.120 |
sing | +0.120 |
Neg logit contributions | |
ited | -0.082 |
saw | -0.080 |
heard | -0.075 |
did | -0.066 |
little | -0.065 |
Token | might |
Token ID | 1244 |
Feature activation | -5.62e-02 |
Pos logit contributions | |
anks | +0.090 |
balls | +0.087 |
hyd | +0.083 |
bleed | +0.082 |
rid | +0.081 |
Neg logit contributions | |
saw | -0.040 |
ited | -0.038 |
heard | -0.036 |
did | -0.032 |
two | -0.030 |
Token | make |
Token ID | 787 |
Feature activation | -4.20e-02 |
Pos logit contributions | |
anks | +0.101 |
balls | +0.097 |
hyd | +0.092 |
uggets | +0.090 |
rid | +0.089 |
Neg logit contributions | |
ited | -0.093 |
saw | -0.090 |
heard | -0.086 |
little | -0.077 |
two | -0.077 |
Token | it |
Token ID | 340 |
Feature activation | -1.07e-01 |
Pos logit contributions | |
anks | +0.089 |
balls | +0.087 |
hyd | +0.085 |
sing | +0.081 |
ometer | +0.081 |
Neg logit contributions | |
ited | -0.052 |
saw | -0.050 |
heard | -0.048 |
little | -0.044 |
two | -0.044 |
Token | not |
Token ID | 407 |
Feature activation | -6.18e-02 |
Pos logit contributions | |
anks | +0.191 |
balls | +0.187 |
hyd | +0.179 |
uggets | +0.178 |
aro | +0.174 |
Neg logit contributions | |
saw | -0.167 |
ited | -0.166 |
heard | -0.160 |
little | -0.149 |
two | -0.146 |
Token | as |
Token ID | 355 |
Feature activation | -6.50e-02 |
Pos logit contributions | |
anks | +0.121 |
balls | +0.117 |
hyd | +0.113 |
rid | +0.108 |
ometer | +0.107 |
Neg logit contributions | |
ited | -0.091 |
saw | -0.087 |
heard | -0.083 |
little | -0.076 |
two | -0.074 |
Token | good |
Token ID | 922 |
Feature activation | -5.02e-02 |
Pos logit contributions | |
anks | +0.131 |
balls | +0.127 |
hyd | +0.125 |
rid | +0.122 |
heal | +0.121 |
Neg logit contributions | |
ited | -0.092 |
saw | -0.083 |
heard | -0.079 |
year | -0.070 |
Hey | -0.070 |
Token | . |
Token ID | 13 |
Feature activation | -5.08e-02 |
Pos logit contributions | |
anks | +0.100 |
hyd | +0.099 |
balls | +0.099 |
ophone | +0.097 |
rid | +0.096 |
Neg logit contributions | |
ited | -0.063 |
saw | -0.060 |
heard | -0.057 |
little | -0.054 |
two | -0.054 |
Token | And |
Token ID | 843 |
Feature activation | -2.95e-02 |
Pos logit contributions | |
anks | +0.116 |
balls | +0.114 |
hyd | +0.112 |
rid | +0.111 |
bleed | +0.109 |
Neg logit contributions | |
ited | -0.056 |
saw | -0.053 |
heard | -0.048 |
two | -0.044 |
little | -0.043 |
Token | and |
Token ID | 290 |
Feature activation | -8.42e-02 |
Pos logit contributions | |
anks | +0.120 |
balls | +0.117 |
hyd | +0.114 |
rid | +0.113 |
bleed | +0.109 |
Neg logit contributions | |
ited | -0.105 |
saw | -0.100 |
heard | -0.097 |
little | -0.089 |
two | -0.088 |
Token | it |
Token ID | 340 |
Feature activation | -6.42e-02 |
Pos logit contributions | |
anks | +0.148 |
balls | +0.141 |
hyd | +0.139 |
rid | +0.133 |
heal | +0.130 |
Neg logit contributions | |
ited | -0.137 |
saw | -0.134 |
heard | -0.127 |
two | -0.117 |
little | -0.116 |
Token | will |
Token ID | 481 |
Feature activation | -5.99e-02 |
Pos logit contributions | |
anks | +0.140 |
uggets | +0.139 |
ectar | +0.131 |
hyd | +0.130 |
balls | +0.130 |
Neg logit contributions | |
saw | -0.079 |
ited | -0.078 |
heard | -0.075 |
did | -0.071 |
little | -0.066 |
Token | be |
Token ID | 307 |
Feature activation | -7.47e-02 |
Pos logit contributions | |
anks | +0.112 |
balls | +0.106 |
uggets | +0.102 |
ometer | +0.101 |
ometers | +0.101 |
Neg logit contributions | |
saw | -0.092 |
ited | -0.090 |
heard | -0.087 |
little | -0.082 |
two | -0.082 |
Token | bigger |
Token ID | 5749 |
Feature activation | -4.29e-02 |
Pos logit contributions | |
anks | +0.154 |
balls | +0.150 |
hyd | +0.146 |
rid | +0.142 |
heal | +0.141 |
Neg logit contributions | |
ited | -0.103 |
saw | -0.094 |
heard | -0.089 |
little | -0.079 |
two | -0.079 |
Token | and |
Token ID | 290 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
anks | +0.083 |
balls | +0.081 |
hyd | +0.079 |
rid | +0.078 |
heal | +0.077 |
Neg logit contributions | |
ited | -0.063 |
saw | -0.058 |
heard | -0.055 |
two | -0.050 |
year | -0.049 |
Token | prett |
Token ID | 46442 |
Feature activation | -3.74e-02 |
Pos logit contributions | |
anks | +0.178 |
balls | +0.170 |
hyd | +0.165 |
rid | +0.160 |
heal | +0.158 |
Neg logit contributions | |
ited | -0.188 |
saw | -0.176 |
heard | -0.169 |
two | -0.154 |
little | -0.154 |
Token | ier |
Token ID | 959 |
Feature activation | -5.13e-02 |
Pos logit contributions | |
anks | +0.061 |
balls | +0.058 |
hyd | +0.057 |
rid | +0.056 |
heal | +0.055 |
Neg logit contributions | |
ited | -0.068 |
saw | -0.064 |
heard | -0.061 |
little | -0.055 |
Hey | -0.055 |
Token | than |
Token ID | 621 |
Feature activation | -5.86e-02 |
Pos logit contributions | |
anks | +0.075 |
balls | +0.073 |
hyd | +0.070 |
rid | +0.069 |
bleed | +0.068 |
Neg logit contributions | |
ited | -0.097 |
saw | -0.093 |
heard | -0.090 |
two | -0.084 |
little | -0.084 |
Token | yours |
Token ID | 12431 |
Feature activation | -4.16e-02 |
Pos logit contributions | |
anks | +0.101 |
balls | +0.096 |
hyd | +0.096 |
rid | +0.090 |
bleed | +0.090 |
Neg logit contributions | |
ited | -0.099 |
saw | -0.093 |
heard | -0.090 |
two | -0.083 |
little | -0.082 |
Token | ." |
Token ID | 526 |
Feature activation | -3.06e-02 |
Pos logit contributions | |
anks | +0.072 |
hyd | +0.070 |
balls | +0.069 |
bleed | +0.068 |
sing | +0.068 |
Neg logit contributions | |
saw | -0.064 |
ited | -0.062 |
heard | -0.062 |
two | -0.058 |
did | -0.056 |
Token | John |
Token ID | 1757 |
Feature activation | -8.25e-03 |
Pos logit contributions | |
ited | +0.026 |
conflict | +0.020 |
Hey | +0.019 |
year | +0.019 |
Next | +0.018 |
Neg logit contributions | |
flies | -0.041 |
hyd | -0.040 |
fly | -0.040 |
heal | -0.039 |
balls | -0.039 |
Token | chose |
Token ID | 7690 |
Feature activation | -7.40e-03 |
Pos logit contributions | |
balls | +0.018 |
anks | +0.018 |
hyd | +0.017 |
rid | +0.017 |
heal | +0.017 |
Neg logit contributions | |
ited | -0.009 |
Hey | -0.007 |
conflict | -0.007 |
saw | -0.006 |
year | -0.006 |
Token | some |
Token ID | 617 |
Feature activation | -8.42e-03 |
Pos logit contributions | |
anks | +0.015 |
balls | +0.015 |
rid | +0.015 |
hyd | +0.015 |
heal | +0.014 |
Neg logit contributions | |
ited | -0.010 |
saw | -0.008 |
heard | -0.008 |
Hey | -0.007 |
year | -0.007 |
Token | carrots |
Token ID | 34397 |
Feature activation | -6.15e-02 |
Pos logit contributions | |
anks | +0.015 |
balls | +0.015 |
rid | +0.015 |
hyd | +0.015 |
heal | +0.015 |
Neg logit contributions | |
ited | -0.013 |
saw | -0.012 |
heard | -0.011 |
Hey | -0.010 |
noticed | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | -5.62e-02 |
Pos logit contributions | |
anks | +0.150 |
balls | +0.143 |
hyd | +0.142 |
rid | +0.141 |
bleed | +0.140 |
Neg logit contributions | |
ited | -0.059 |
saw | -0.054 |
heard | -0.050 |
two | -0.042 |
little | -0.041 |
Token | cuc |
Token ID | 38421 |
Feature activation | -1.05e-01 |
Pos logit contributions | |
balls | +0.111 |
anks | +0.111 |
hyd | +0.108 |
rid | +0.107 |
heal | +0.106 |
Neg logit contributions | |
ited | -0.079 |
saw | -0.068 |
heard | -0.066 |
year | -0.060 |
Hey | -0.060 |
Token | umbers |
Token ID | 17024 |
Feature activation | -3.14e-02 |
Pos logit contributions | |
anks | +0.155 |
balls | +0.149 |
hyd | +0.149 |
heal | +0.143 |
rid | +0.142 |
Neg logit contributions | |
ited | -0.205 |
saw | -0.187 |
heard | -0.180 |
two | -0.168 |
Hey | -0.167 |
Token | to |
Token ID | 284 |
Feature activation | -3.57e-02 |
Pos logit contributions | |
anks | +0.063 |
balls | +0.062 |
hyd | +0.061 |
rid | +0.059 |
heal | +0.059 |
Neg logit contributions | |
ited | -0.045 |
saw | -0.039 |
heard | -0.037 |
Hey | -0.034 |
year | -0.033 |
Token | take |
Token ID | 1011 |
Feature activation | -3.20e-03 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.055 |
hyd | +0.054 |
rid | +0.051 |
bleed | +0.051 |
Neg logit contributions | |
ited | -0.065 |
saw | -0.061 |
heard | -0.058 |
two | -0.054 |
little | -0.054 |
Token | home |
Token ID | 1363 |
Feature activation | -1.02e-02 |
Pos logit contributions | |
anks | +0.007 |
balls | +0.007 |
hyd | +0.007 |
rid | +0.006 |
heal | +0.006 |
Neg logit contributions | |
ited | -0.004 |
saw | -0.004 |
heard | -0.004 |
year | -0.003 |
Hey | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | -4.99e-02 |
Pos logit contributions | |
anks | +0.020 |
balls | +0.019 |
hyd | +0.019 |
bleed | +0.019 |
sing | +0.019 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.013 |
heard | -0.013 |
two | -0.012 |
little | -0.012 |
Token | like |
Token ID | 588 |
Feature activation | -2.71e-02 |
Pos logit contributions | |
anks | +0.151 |
balls | +0.145 |
hyd | +0.143 |
rid | +0.139 |
bleed | +0.139 |
Neg logit contributions | |
saw | -0.080 |
ited | -0.080 |
heard | -0.075 |
two | -0.065 |
did | -0.065 |
Token | lettuce |
Token ID | 39406 |
Feature activation | -6.82e-02 |
Pos logit contributions | |
anks | +0.056 |
balls | +0.053 |
hyd | +0.053 |
rid | +0.051 |
bleed | +0.051 |
Neg logit contributions | |
ited | -0.035 |
saw | -0.035 |
heard | -0.033 |
two | -0.030 |
little | -0.030 |
Token | , |
Token ID | 11 |
Feature activation | -5.34e-02 |
Pos logit contributions | |
anks | +0.177 |
balls | +0.171 |
hyd | +0.171 |
rid | +0.169 |
bleed | +0.166 |
Neg logit contributions | |
ited | -0.053 |
saw | -0.049 |
heard | -0.043 |
two | -0.035 |
little | -0.034 |
Token | tomatoes |
Token ID | 23972 |
Feature activation | -5.60e-02 |
Pos logit contributions | |
anks | +0.098 |
balls | +0.093 |
hyd | +0.092 |
rid | +0.089 |
heal | +0.088 |
Neg logit contributions | |
ited | -0.084 |
saw | -0.080 |
heard | -0.076 |
little | -0.070 |
two | -0.070 |
Token | and |
Token ID | 290 |
Feature activation | -7.90e-02 |
Pos logit contributions | |
anks | +0.151 |
balls | +0.146 |
hyd | +0.146 |
rid | +0.144 |
bleed | +0.142 |
Neg logit contributions | |
ited | -0.038 |
saw | -0.034 |
heard | -0.030 |
two | -0.023 |
little | -0.022 |
Token | cuc |
Token ID | 38421 |
Feature activation | -1.05e-01 |
Pos logit contributions | |
anks | +0.132 |
balls | +0.125 |
hyd | +0.123 |
rid | +0.119 |
heal | +0.117 |
Neg logit contributions | |
ited | -0.139 |
saw | -0.132 |
heard | -0.126 |
two | -0.116 |
little | -0.116 |
Token | umbers |
Token ID | 17024 |
Feature activation | -4.81e-02 |
Pos logit contributions | |
anks | +0.154 |
balls | +0.149 |
hyd | +0.148 |
heal | +0.143 |
bleed | +0.141 |
Neg logit contributions | |
ited | -0.205 |
saw | -0.187 |
heard | -0.181 |
two | -0.168 |
Hey | -0.167 |
Token | . |
Token ID | 13 |
Feature activation | -5.88e-02 |
Pos logit contributions | |
anks | +0.093 |
balls | +0.090 |
hyd | +0.088 |
rid | +0.086 |
heal | +0.085 |
Neg logit contributions | |
ited | -0.071 |
saw | -0.065 |
heard | -0.062 |
two | -0.055 |
year | -0.055 |
Token | He |
Token ID | 679 |
Feature activation | -2.29e-02 |
Pos logit contributions | |
anks | +0.124 |
hyd | +0.119 |
balls | +0.118 |
rid | +0.117 |
bleed | +0.117 |
Neg logit contributions | |
ited | -0.073 |
saw | -0.071 |
heard | -0.066 |
two | -0.064 |
little | -0.061 |
Token | smiled |
Token ID | 13541 |
Feature activation | +9.86e-03 |
Pos logit contributions | |
anks | +0.049 |
balls | +0.049 |
hyd | +0.048 |
rid | +0.047 |
heal | +0.047 |
Neg logit contributions | |
ited | -0.028 |
saw | -0.023 |
heard | -0.022 |
Hey | -0.020 |
year | -0.020 |
Token | as |
Token ID | 355 |
Feature activation | -2.38e-03 |
Pos logit contributions | |
ited | +0.009 |
saw | +0.008 |
heard | +0.007 |
Hey | +0.006 |
year | +0.006 |
Neg logit contributions | |
anks | -0.025 |
balls | -0.024 |
hyd | -0.024 |
rid | -0.023 |
heal | -0.023 |
Token | play |
Token ID | 711 |
Feature activation | +1.68e-02 |
Pos logit contributions | |
anks | +0.046 |
balls | +0.045 |
hyd | +0.043 |
rid | +0.042 |
heal | +0.042 |
Neg logit contributions | |
ited | -0.042 |
saw | -0.038 |
heard | -0.037 |
two | -0.033 |
year | -0.033 |
Token | together |
Token ID | 1978 |
Feature activation | -5.14e-03 |
Pos logit contributions | |
ited | +0.030 |
saw | +0.028 |
heard | +0.027 |
little | +0.024 |
two | +0.024 |
Neg logit contributions | |
anks | -0.028 |
balls | -0.027 |
hyd | -0.026 |
rid | -0.025 |
heal | -0.025 |
Token | without |
Token ID | 1231 |
Feature activation | -3.23e-02 |
Pos logit contributions | |
anks | +0.011 |
balls | +0.011 |
hyd | +0.011 |
rid | +0.011 |
heal | +0.011 |
Neg logit contributions | |
ited | -0.006 |
saw | -0.005 |
heard | -0.005 |
Hey | -0.004 |
two | -0.004 |
Token | making |
Token ID | 1642 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
anks | +0.057 |
balls | +0.056 |
rid | +0.056 |
hyd | +0.055 |
heal | +0.055 |
Neg logit contributions | |
ited | -0.053 |
saw | -0.047 |
heard | -0.045 |
Hey | -0.042 |
year | -0.042 |
Token | boring |
Token ID | 14262 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
anks | +0.043 |
balls | +0.042 |
hyd | +0.041 |
rid | +0.039 |
heal | +0.039 |
Neg logit contributions | |
ited | -0.026 |
saw | -0.024 |
heard | -0.023 |
little | -0.021 |
two | -0.021 |
Token | or |
Token ID | 393 |
Feature activation | -1.05e-01 |
Pos logit contributions | |
anks | +0.066 |
balls | +0.065 |
hyd | +0.064 |
rid | +0.063 |
heal | +0.061 |
Neg logit contributions | |
ited | -0.043 |
saw | -0.041 |
heard | -0.039 |
little | -0.035 |
two | -0.034 |
Token | loud |
Token ID | 7812 |
Feature activation | -4.31e-02 |
Pos logit contributions | |
anks | +0.193 |
balls | +0.186 |
hyd | +0.183 |
rid | +0.177 |
heal | +0.176 |
Neg logit contributions | |
ited | -0.166 |
saw | -0.153 |
heard | -0.147 |
little | -0.131 |
two | -0.131 |
Token | sounds |
Token ID | 5238 |
Feature activation | -2.54e-02 |
Pos logit contributions | |
anks | +0.074 |
balls | +0.071 |
hyd | +0.070 |
rid | +0.068 |
heal | +0.067 |
Neg logit contributions | |
ited | -0.072 |
saw | -0.068 |
heard | -0.066 |
little | -0.060 |
two | -0.059 |
Token | that |
Token ID | 326 |
Feature activation | -3.75e-02 |
Pos logit contributions | |
anks | +0.055 |
hyd | +0.052 |
balls | +0.052 |
uggets | +0.051 |
sing | +0.051 |
Neg logit contributions | |
saw | -0.030 |
ited | -0.029 |
heard | -0.028 |
little | -0.027 |
two | -0.025 |
Token | could |
Token ID | 714 |
Feature activation | -5.62e-02 |
Pos logit contributions | |
anks | +0.086 |
balls | +0.082 |
uggets | +0.082 |
hyd | +0.081 |
sing | +0.078 |
Neg logit contributions | |
saw | -0.043 |
heard | -0.040 |
ited | -0.040 |
did | -0.037 |
time | -0.034 |
Token | hurt |
Token ID | 5938 |
Feature activation | -2.79e-02 |
Pos logit contributions | |
anks | +0.099 |
balls | +0.092 |
ometer | +0.091 |
hyd | +0.091 |
uggets | +0.091 |
Neg logit contributions | |
saw | -0.094 |
ited | -0.090 |
heard | -0.090 |
two | -0.082 |
little | -0.082 |
Token | can |
Token ID | 460 |
Feature activation | -5.61e-02 |
Pos logit contributions | |
anks | +0.122 |
balls | +0.115 |
uggets | +0.114 |
hyd | +0.114 |
ectar | +0.114 |
Neg logit contributions | |
saw | -0.066 |
heard | -0.060 |
did | -0.057 |
ited | -0.056 |
two | -0.055 |
Token | make |
Token ID | 787 |
Feature activation | -3.68e-02 |
Pos logit contributions | |
anks | +0.100 |
hyd | +0.093 |
ectar | +0.092 |
uggets | +0.091 |
balls | +0.090 |
Neg logit contributions | |
saw | -0.096 |
heard | -0.090 |
ited | -0.086 |
two | -0.082 |
little | -0.080 |
Token | it |
Token ID | 340 |
Feature activation | -9.97e-02 |
Pos logit contributions | |
anks | +0.079 |
hyd | +0.078 |
balls | +0.078 |
rid | +0.074 |
sing | +0.073 |
Neg logit contributions | |
saw | -0.042 |
ited | -0.041 |
heard | -0.039 |
little | -0.037 |
two | -0.037 |
Token | go |
Token ID | 467 |
Feature activation | -4.43e-02 |
Pos logit contributions | |
anks | +0.192 |
balls | +0.188 |
uggets | +0.186 |
aro | +0.186 |
ectar | +0.181 |
Neg logit contributions | |
saw | -0.143 |
heard | -0.134 |
little | -0.129 |
ited | -0.128 |
two | -0.126 |
Token | faster |
Token ID | 5443 |
Feature activation | -3.14e-02 |
Pos logit contributions | |
anks | +0.089 |
balls | +0.085 |
hyd | +0.081 |
ectar | +0.081 |
uggets | +0.080 |
Neg logit contributions | |
saw | -0.062 |
ited | -0.061 |
heard | -0.058 |
little | -0.054 |
time | -0.053 |
Token | or |
Token ID | 393 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
anks | +0.060 |
balls | +0.058 |
hyd | +0.057 |
rid | +0.056 |
heal | +0.055 |
Neg logit contributions | |
ited | -0.048 |
saw | -0.043 |
heard | -0.042 |
year | -0.037 |
two | -0.037 |
Token | slower |
Token ID | 13611 |
Feature activation | -3.74e-02 |
Pos logit contributions | |
anks | +0.195 |
balls | +0.183 |
hyd | +0.182 |
ectar | +0.176 |
uggets | +0.175 |
Neg logit contributions | |
saw | -0.160 |
ited | -0.158 |
heard | -0.151 |
two | -0.139 |
little | -0.138 |
Token | . |
Token ID | 13 |
Feature activation | -6.15e-02 |
Pos logit contributions | |
anks | +0.068 |
balls | +0.066 |
hyd | +0.065 |
rid | +0.063 |
heal | +0.062 |
Neg logit contributions | |
ited | -0.059 |
saw | -0.055 |
heard | -0.053 |
two | -0.048 |
little | -0.048 |
Token | Sam |
Token ID | 3409 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
anks | +0.128 |
balls | +0.126 |
hyd | +0.124 |
rid | +0.122 |
bleed | +0.119 |
Neg logit contributions | |
ited | -0.076 |
saw | -0.074 |
heard | -0.069 |
two | -0.068 |
Hey | -0.065 |
Token | is |
Token ID | 318 |
Feature activation | -1.90e-03 |
Pos logit contributions | |
ited | +0.013 |
saw | +0.011 |
heard | +0.010 |
year | +0.010 |
Hey | +0.009 |
Neg logit contributions | |
anks | -0.022 |
balls | -0.021 |
hyd | -0.021 |
heal | -0.021 |
rid | -0.021 |
Token | very |
Token ID | 845 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
anks | +0.004 |
balls | +0.004 |
hyd | +0.004 |
rid | +0.004 |
heal | +0.004 |
Neg logit contributions | |
ited | -0.003 |
saw | -0.002 |
heard | -0.002 |
year | -0.002 |
two | -0.002 |
Token | , |
Token ID | 11 |
Feature activation | +1.87e-03 |
Pos logit contributions | |
anks | +0.019 |
balls | +0.019 |
hyd | +0.018 |
heal | +0.018 |
rid | +0.018 |
Neg logit contributions | |
ited | -0.028 |
saw | -0.025 |
heard | -0.024 |
Hey | -0.023 |
year | -0.023 |
Token | we |
Token ID | 356 |
Feature activation | -3.10e-02 |
Pos logit contributions | |
ited | +0.003 |
saw | +0.002 |
heard | +0.002 |
year | +0.002 |
Hey | +0.002 |
Neg logit contributions | |
anks | -0.004 |
balls | -0.004 |
hyd | -0.004 |
rid | -0.004 |
heal | -0.004 |
Token | found |
Token ID | 1043 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
anks | +0.056 |
hyd | +0.055 |
rid | +0.055 |
balls | +0.055 |
heal | +0.054 |
Neg logit contributions | |
ited | -0.049 |
saw | -0.041 |
heard | -0.039 |
year | -0.037 |
Hey | -0.037 |
Token | a |
Token ID | 257 |
Feature activation | -2.40e-03 |
Pos logit contributions | |
ited | +0.031 |
saw | +0.027 |
heard | +0.025 |
two | +0.022 |
little | +0.021 |
Neg logit contributions | |
anks | -0.075 |
balls | -0.073 |
hyd | -0.071 |
heal | -0.070 |
rid | -0.070 |
Token | rare |
Token ID | 4071 |
Feature activation | -1.71e-02 |
Pos logit contributions | |
anks | +0.005 |
balls | +0.005 |
hyd | +0.005 |
rid | +0.005 |
heal | +0.004 |
Neg logit contributions | |
ited | -0.003 |
saw | -0.003 |
heard | -0.003 |
little | -0.003 |
year | -0.003 |
Token | rad |
Token ID | 2511 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
anks | +0.029 |
balls | +0.028 |
hyd | +0.027 |
rid | +0.026 |
heal | +0.026 |
Neg logit contributions | |
ited | -0.030 |
saw | -0.028 |
heard | -0.027 |
year | -0.024 |
little | -0.024 |
Token | ish |
Token ID | 680 |
Feature activation | -6.49e-02 |
Pos logit contributions | |
anks | +0.132 |
balls | +0.125 |
sing | +0.118 |
bleed | +0.118 |
heal | +0.118 |
Neg logit contributions | |
ited | -0.223 |
saw | -0.215 |
heard | -0.200 |
two | -0.192 |
little | -0.190 |
Token | !" |
Token ID | 2474 |
Feature activation | -2.80e-02 |
Pos logit contributions | |
anks | +0.145 |
bleed | +0.137 |
heal | +0.133 |
rid | +0.133 |
sing | +0.133 |
Neg logit contributions | |
saw | -0.074 |
ited | -0.070 |
heard | -0.067 |
two | -0.061 |
year | -0.060 |
Token | they |
Token ID | 484 |
Feature activation | -6.82e-03 |
Pos logit contributions | |
anks | +0.060 |
balls | +0.058 |
hyd | +0.057 |
rid | +0.056 |
heal | +0.056 |
Neg logit contributions | |
ited | -0.036 |
saw | -0.033 |
heard | -0.031 |
two | -0.028 |
little | -0.027 |
Token | shouted |
Token ID | 17293 |
Feature activation | +3.83e-02 |
Pos logit contributions | |
anks | +0.016 |
balls | +0.015 |
rid | +0.015 |
hyd | +0.015 |
heal | +0.015 |
Neg logit contributions | |
ited | -0.007 |
saw | -0.006 |
heard | -0.005 |
Hey | -0.005 |
year | -0.005 |
Token | , |
Token ID | 11 |
Feature activation | +8.11e-04 |
Pos logit contributions | |
saw | +0.031 |
little | +0.030 |
heard | +0.029 |
two | +0.028 |
, | +0.027 |
Neg logit contributions | |
anks | -0.096 |
bleed | -0.091 |
ometers | -0.090 |
affect | -0.090 |
ophone | -0.089 |
Token | some |
Token ID | 617 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
anks | +0.028 |
balls | +0.027 |
hyd | +0.027 |
rid | +0.026 |
heal | +0.026 |
Neg logit contributions | |
ited | -0.018 |
saw | -0.016 |
heard | -0.015 |
two | -0.013 |
Hey | -0.013 |
Token | buttons |
Token ID | 12163 |
Feature activation | -5.39e-02 |
Pos logit contributions | |
anks | +0.041 |
balls | +0.039 |
hyd | +0.039 |
rid | +0.038 |
heal | +0.037 |
Neg logit contributions | |
ited | -0.031 |
saw | -0.028 |
heard | -0.027 |
two | -0.024 |
year | -0.024 |
Token | or |
Token ID | 393 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
anks | +0.120 |
balls | +0.117 |
hyd | +0.115 |
rid | +0.113 |
heal | +0.113 |
Neg logit contributions | |
ited | -0.066 |
saw | -0.056 |
heard | -0.053 |
year | -0.047 |
Hey | -0.046 |
Token | rib |
Token ID | 12183 |
Feature activation | -5.92e-02 |
Pos logit contributions | |
anks | +0.175 |
balls | +0.170 |
hyd | +0.167 |
rid | +0.162 |
heal | +0.160 |
Neg logit contributions | |
ited | -0.171 |
saw | -0.157 |
heard | -0.151 |
two | -0.137 |
year | -0.136 |
Token | bons |
Token ID | 23461 |
Feature activation | -4.79e-02 |
Pos logit contributions | |
anks | +0.067 |
balls | +0.065 |
hyd | +0.063 |
rid | +0.060 |
heal | +0.059 |
Neg logit contributions | |
ited | -0.135 |
saw | -0.127 |
heard | -0.123 |
two | -0.115 |
Hey | -0.115 |
Token | or |
Token ID | 393 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
anks | +0.089 |
balls | +0.087 |
hyd | +0.085 |
rid | +0.083 |
heal | +0.083 |
Neg logit contributions | |
ited | -0.077 |
saw | -0.067 |
heard | -0.065 |
Hey | -0.059 |
year | -0.059 |
Token | patches |
Token ID | 16082 |
Feature activation | -5.52e-02 |
Pos logit contributions | |
anks | +0.179 |
balls | +0.174 |
hyd | +0.170 |
rid | +0.166 |
heal | +0.165 |
Neg logit contributions | |
ited | -0.176 |
saw | -0.161 |
heard | -0.154 |
year | -0.140 |
Hey | -0.140 |
Token | . |
Token ID | 13 |
Feature activation | -7.67e-02 |
Pos logit contributions | |
anks | +0.103 |
balls | +0.100 |
hyd | +0.098 |
rid | +0.096 |
heal | +0.094 |
Neg logit contributions | |
ited | -0.085 |
saw | -0.078 |
heard | -0.075 |
two | -0.068 |
year | -0.067 |
Token | Would |
Token ID | 10928 |
Feature activation | -2.57e-03 |
Pos logit contributions | |
anks | +0.123 |
balls | +0.121 |
rid | +0.117 |
bleed | +0.114 |
hyd | +0.110 |
Neg logit contributions | |
saw | -0.126 |
two | -0.125 |
heard | -0.119 |
ited | -0.119 |
little | -0.118 |
Token | you |
Token ID | 345 |
Feature activation | -3.65e-02 |
Pos logit contributions | |
anks | +0.004 |
hyd | +0.004 |
bleed | +0.004 |
rid | +0.004 |
balls | +0.004 |
Neg logit contributions | |
saw | -0.004 |
ited | -0.004 |
heard | -0.004 |
two | -0.004 |
little | -0.004 |
Token | like |
Token ID | 588 |
Feature activation | -2.66e-03 |
Pos logit contributions | |
anks | +0.064 |
heal | +0.061 |
rid | +0.061 |
balls | +0.060 |
hyd | +0.060 |
Neg logit contributions | |
ited | -0.062 |
saw | -0.053 |
heard | -0.051 |
Hey | -0.050 |
year | -0.048 |
Token | okay |
Token ID | 8788 |
Feature activation | -8.30e-03 |
Pos logit contributions | |
anks | +0.098 |
balls | +0.097 |
rid | +0.096 |
hyd | +0.095 |
bleed | +0.091 |
Neg logit contributions | |
ited | -0.110 |
saw | -0.097 |
heard | -0.094 |
Hey | -0.092 |
year | -0.092 |
Token | , |
Token ID | 11 |
Feature activation | -3.23e-02 |
Pos logit contributions | |
anks | +0.024 |
balls | +0.023 |
rid | +0.023 |
hyd | +0.023 |
bleed | +0.022 |
Neg logit contributions | |
ited | -0.004 |
saw | -0.004 |
heard | -0.003 |
little | -0.003 |
two | -0.002 |
Token | Anna |
Token ID | 11735 |
Feature activation | -9.42e-03 |
Pos logit contributions | |
balls | +0.058 |
anks | +0.058 |
hyd | +0.056 |
heal | +0.055 |
rid | +0.054 |
Neg logit contributions | |
ited | -0.053 |
saw | -0.045 |
heard | -0.044 |
year | -0.040 |
Hey | -0.039 |
Token | ? |
Token ID | 30 |
Feature activation | -7.65e-02 |
Pos logit contributions | |
anks | +0.018 |
balls | +0.017 |
hyd | +0.017 |
rid | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.015 |
saw | -0.013 |
heard | -0.013 |
year | -0.011 |
Hey | -0.011 |
Token | Does |
Token ID | 8314 |
Feature activation | -4.58e-03 |
Pos logit contributions | |
anks | +0.142 |
balls | +0.133 |
rid | +0.131 |
heal | +0.125 |
hyd | +0.125 |
Neg logit contributions | |
saw | -0.118 |
ited | -0.109 |
heard | -0.108 |
Hey | -0.104 |
two | -0.104 |
Token | it |
Token ID | 340 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.010 |
hyd | +0.010 |
balls | +0.009 |
rid | +0.009 |
sing | +0.009 |
Neg logit contributions | |
saw | -0.005 |
ited | -0.005 |
heard | -0.005 |
little | -0.005 |
two | -0.005 |
Token | hurt |
Token ID | 5938 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
anks | +0.191 |
balls | +0.185 |
hyd | +0.184 |
uggets | +0.182 |
aro | +0.177 |
Neg logit contributions | |
saw | -0.160 |
ited | -0.153 |
heard | -0.152 |
did | -0.141 |
little | -0.137 |
Token | a |
Token ID | 257 |
Feature activation | -2.83e-02 |
Pos logit contributions | |
anks | +0.066 |
balls | +0.064 |
hyd | +0.062 |
rid | +0.060 |
heal | +0.059 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.043 |
heard | -0.041 |
little | -0.036 |
two | -0.036 |
Token | lot |
Token ID | 1256 |
Feature activation | -4.32e-02 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.054 |
hyd | +0.053 |
ometers | +0.053 |
rid | +0.051 |
Neg logit contributions | |
ited | -0.039 |
saw | -0.039 |
heard | -0.037 |
little | -0.035 |
year | -0.034 |
Token | ?" |
Token ID | 1701 |
Feature activation | -2.50e-02 |
Pos logit contributions | |
anks | +0.075 |
heal | +0.072 |
hyd | +0.071 |
balls | +0.071 |
rid | +0.070 |
Neg logit contributions | |
ited | -0.075 |
saw | -0.066 |
heard | -0.063 |
noticed | -0.057 |
Hey | -0.057 |
Token |
|
Token ID | 198 |
Feature activation | -2.32e-03 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.056 |
hyd | +0.056 |
rid | +0.054 |
heal | +0.054 |
Neg logit contributions | |
ited | -0.028 |
saw | -0.026 |
heard | -0.024 |
two | -0.021 |
little | -0.021 |
Token | not |
Token ID | 407 |
Feature activation | -4.18e-02 |
Pos logit contributions | |
anks | +0.080 |
balls | +0.078 |
hyd | +0.076 |
rid | +0.075 |
heal | +0.074 |
Neg logit contributions | |
ited | -0.059 |
saw | -0.054 |
heard | -0.052 |
two | -0.046 |
little | -0.046 |
Token | something |
Token ID | 1223 |
Feature activation | -4.50e-02 |
Pos logit contributions | |
anks | +0.090 |
balls | +0.087 |
hyd | +0.085 |
rid | +0.082 |
heal | +0.081 |
Neg logit contributions | |
ited | -0.051 |
saw | -0.050 |
heard | -0.047 |
little | -0.045 |
two | -0.044 |
Token | you |
Token ID | 345 |
Feature activation | -5.73e-02 |
Pos logit contributions | |
anks | +0.110 |
hyd | +0.103 |
balls | +0.102 |
sing | +0.101 |
bleed | +0.101 |
Neg logit contributions | |
saw | -0.045 |
little | -0.042 |
heard | -0.039 |
ited | -0.039 |
time | -0.038 |
Token | can |
Token ID | 460 |
Feature activation | -6.63e-02 |
Pos logit contributions | |
anks | +0.106 |
balls | +0.100 |
heal | +0.098 |
hyd | +0.098 |
rid | +0.098 |
Neg logit contributions | |
ited | -0.092 |
saw | -0.078 |
heard | -0.076 |
year | -0.075 |
Hey | -0.072 |
Token | buy |
Token ID | 2822 |
Feature activation | -2.36e-02 |
Pos logit contributions | |
anks | +0.124 |
hyd | +0.121 |
balls | +0.120 |
rid | +0.115 |
ometers | +0.114 |
Neg logit contributions | |
ited | -0.097 |
saw | -0.096 |
heard | -0.091 |
two | -0.084 |
little | -0.082 |
Token | or |
Token ID | 393 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.044 |
balls | +0.042 |
hyd | +0.042 |
rid | +0.040 |
heal | +0.040 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.034 |
heard | -0.033 |
two | -0.030 |
little | -0.029 |
Token | take |
Token ID | 1011 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
anks | +0.167 |
balls | +0.162 |
hyd | +0.159 |
rid | +0.155 |
heal | +0.154 |
Neg logit contributions | |
ited | -0.187 |
saw | -0.170 |
heard | -0.164 |
year | -0.151 |
Hey | -0.150 |
Token | . |
Token ID | 13 |
Feature activation | -5.91e-02 |
Pos logit contributions | |
anks | +0.023 |
balls | +0.023 |
rid | +0.022 |
hyd | +0.022 |
heal | +0.022 |
Neg logit contributions | |
ited | -0.018 |
saw | -0.015 |
heard | -0.015 |
Hey | -0.013 |
year | -0.013 |
Token | Love |
Token ID | 5896 |
Feature activation | +1.51e-02 |
Pos logit contributions | |
anks | +0.118 |
balls | +0.118 |
rid | +0.112 |
hyd | +0.111 |
heal | +0.110 |
Neg logit contributions | |
saw | -0.075 |
ited | -0.074 |
two | -0.071 |
heard | -0.070 |
Hey | -0.068 |
Token | is |
Token ID | 318 |
Feature activation | -2.70e-02 |
Pos logit contributions | |
ited | +0.016 |
saw | +0.013 |
heard | +0.012 |
Hey | +0.010 |
year | +0.010 |
Neg logit contributions | |
anks | -0.037 |
balls | -0.036 |
hyd | -0.035 |
heal | -0.035 |
rid | -0.035 |
Token | something |
Token ID | 1223 |
Feature activation | -2.72e-02 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.056 |
hyd | +0.056 |
bleed | +0.054 |
rid | +0.054 |
Neg logit contributions | |
ited | -0.032 |
saw | -0.031 |
heard | -0.029 |
two | -0.027 |
little | -0.027 |
Token | more |
Token ID | 517 |
Feature activation | -3.26e-02 |
Pos logit contributions | |
anks | +0.080 |
balls | +0.077 |
hyd | +0.076 |
rid | +0.073 |
heal | +0.073 |
Neg logit contributions | |
ited | -0.067 |
saw | -0.062 |
heard | -0.060 |
two | -0.054 |
little | -0.054 |
Token | important |
Token ID | 1593 |
Feature activation | -1.21e-02 |
Pos logit contributions | |
anks | +0.063 |
balls | +0.062 |
hyd | +0.059 |
bleed | +0.058 |
sing | +0.057 |
Neg logit contributions | |
ited | -0.047 |
saw | -0.046 |
heard | -0.043 |
little | -0.042 |
two | -0.040 |
Token | than |
Token ID | 621 |
Feature activation | -4.54e-02 |
Pos logit contributions | |
anks | +0.018 |
balls | +0.017 |
hyd | +0.017 |
rid | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.024 |
saw | -0.022 |
heard | -0.022 |
year | -0.020 |
two | -0.020 |
Token | being |
Token ID | 852 |
Feature activation | -5.91e-02 |
Pos logit contributions | |
anks | +0.082 |
balls | +0.079 |
hyd | +0.078 |
rid | +0.074 |
bleed | +0.074 |
Neg logit contributions | |
ited | -0.072 |
saw | -0.068 |
heard | -0.065 |
two | -0.060 |
little | -0.060 |
Token | clever |
Token ID | 14169 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
anks | +0.120 |
balls | +0.117 |
hyd | +0.115 |
rid | +0.114 |
heal | +0.112 |
Neg logit contributions | |
ited | -0.082 |
saw | -0.073 |
heard | -0.069 |
year | -0.062 |
Hey | -0.062 |
Token | or |
Token ID | 393 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.040 |
balls | +0.039 |
hyd | +0.038 |
rid | +0.038 |
heal | +0.038 |
Neg logit contributions | |
ited | -0.024 |
saw | -0.020 |
heard | -0.018 |
Hey | -0.016 |
year | -0.016 |
Token | playing |
Token ID | 2712 |
Feature activation | -3.00e-02 |
Pos logit contributions | |
anks | +0.193 |
balls | +0.186 |
hyd | +0.182 |
rid | +0.176 |
heal | +0.175 |
Neg logit contributions | |
ited | -0.159 |
saw | -0.148 |
heard | -0.141 |
two | -0.128 |
little | -0.128 |
Token | tricks |
Token ID | 15910 |
Feature activation | -6.37e-02 |
Pos logit contributions | |
anks | +0.055 |
balls | +0.053 |
hyd | +0.052 |
rid | +0.051 |
bleed | +0.050 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.044 |
heard | -0.042 |
little | -0.039 |
two | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | -5.48e-02 |
Pos logit contributions | |
anks | +0.132 |
hyd | +0.129 |
balls | +0.128 |
rid | +0.125 |
bleed | +0.124 |
Neg logit contributions | |
ited | -0.081 |
saw | -0.076 |
heard | -0.073 |
two | -0.066 |
little | -0.064 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +3.91e-02 |
Pos logit contributions | |
balls | +0.156 |
anks | +0.150 |
rid | +0.147 |
bleed | +0.147 |
sing | +0.143 |
Neg logit contributions | |
saw | -0.028 |
year | -0.026 |
two | -0.025 |
Once | -0.025 |
heard | -0.023 |
Token | Once |
Token ID | 7454 |
Feature activation | +6.14e-02 |
Pos logit contributions | |
ited | +0.054 |
saw | +0.044 |
heard | +0.041 |
noticed | +0.036 |
little | +0.036 |
Neg logit contributions | |
anks | -0.083 |
balls | -0.080 |
hyd | -0.078 |
rid | -0.077 |
heal | -0.077 |
Token | we |
Token ID | 356 |
Feature activation | -5.59e-02 |
Pos logit contributions | |
anks | +0.137 |
balls | +0.134 |
hyd | +0.131 |
rid | +0.129 |
heal | +0.129 |
Neg logit contributions | |
ited | -0.127 |
saw | -0.113 |
heard | -0.108 |
Hey | -0.099 |
year | -0.099 |
Token | can |
Token ID | 460 |
Feature activation | -4.86e-02 |
Pos logit contributions | |
anks | +0.116 |
balls | +0.113 |
hyd | +0.110 |
rid | +0.107 |
heal | +0.105 |
Neg logit contributions | |
ited | -0.076 |
saw | -0.071 |
heard | -0.067 |
two | -0.059 |
did | -0.059 |
Token | make |
Token ID | 787 |
Feature activation | -2.32e-02 |
Pos logit contributions | |
anks | +0.091 |
balls | +0.090 |
hyd | +0.088 |
rid | +0.085 |
ectar | +0.082 |
Neg logit contributions | |
ited | -0.074 |
saw | -0.071 |
heard | -0.067 |
two | -0.062 |
little | -0.060 |
Token | some |
Token ID | 617 |
Feature activation | -2.71e-02 |
Pos logit contributions | |
anks | +0.051 |
hyd | +0.050 |
balls | +0.050 |
sing | +0.048 |
fly | +0.048 |
Neg logit contributions | |
saw | -0.025 |
two | -0.024 |
ited | -0.024 |
heard | -0.023 |
little | -0.023 |
Token | cards |
Token ID | 4116 |
Feature activation | -8.55e-02 |
Pos logit contributions | |
anks | +0.055 |
balls | +0.054 |
hyd | +0.053 |
rid | +0.052 |
heal | +0.051 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.034 |
heard | -0.032 |
little | -0.029 |
year | -0.029 |
Token | or |
Token ID | 393 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.170 |
balls | +0.164 |
hyd | +0.163 |
bleed | +0.159 |
sing | +0.159 |
Neg logit contributions | |
saw | -0.112 |
ited | -0.110 |
heard | -0.106 |
two | -0.098 |
little | -0.095 |
Token | drawings |
Token ID | 23388 |
Feature activation | -5.85e-02 |
Pos logit contributions | |
anks | +0.174 |
balls | +0.170 |
heal | +0.167 |
hyd | +0.166 |
rid | +0.165 |
Neg logit contributions | |
ited | -0.182 |
saw | -0.156 |
heard | -0.150 |
year | -0.142 |
Hey | -0.142 |
Token | to |
Token ID | 284 |
Feature activation | -5.96e-02 |
Pos logit contributions | |
anks | +0.104 |
balls | +0.101 |
hyd | +0.101 |
rid | +0.098 |
sing | +0.096 |
Neg logit contributions | |
ited | -0.091 |
saw | -0.087 |
heard | -0.084 |
two | -0.077 |
little | -0.076 |
Token | cheer |
Token ID | 14042 |
Feature activation | -3.23e-02 |
Pos logit contributions | |
anks | +0.096 |
balls | +0.093 |
hyd | +0.091 |
rid | +0.087 |
bleed | +0.085 |
Neg logit contributions | |
ited | -0.107 |
saw | -0.102 |
heard | -0.098 |
little | -0.089 |
two | -0.089 |
Token | up |
Token ID | 510 |
Feature activation | -2.73e-02 |
Pos logit contributions | |
anks | +0.068 |
hyd | +0.066 |
balls | +0.066 |
bleed | +0.064 |
rid | +0.064 |
Neg logit contributions | |
ited | -0.039 |
saw | -0.038 |
heard | -0.036 |
little | -0.033 |
two | -0.033 |
Token | the |
Token ID | 262 |
Feature activation | -1.77e-02 |
Pos logit contributions | |
anks | +0.053 |
balls | +0.051 |
hyd | +0.051 |
rid | +0.049 |
heal | +0.049 |
Neg logit contributions | |
ited | -0.041 |
saw | -0.037 |
heard | -0.035 |
Hey | -0.031 |
year | -0.031 |
Token | Please |
Token ID | 5492 |
Feature activation | -2.54e-02 |
Pos logit contributions | |
anks | +0.093 |
hyd | +0.093 |
balls | +0.092 |
rid | +0.091 |
heal | +0.090 |
Neg logit contributions | |
ited | -0.057 |
saw | -0.051 |
heard | -0.048 |
little | -0.042 |
year | -0.042 |
Token | help |
Token ID | 1037 |
Feature activation | +2.22e-02 |
Pos logit contributions | |
anks | +0.051 |
balls | +0.049 |
hyd | +0.048 |
rid | +0.047 |
bleed | +0.047 |
Neg logit contributions | |
ited | -0.035 |
saw | -0.033 |
heard | -0.032 |
little | -0.029 |
two | -0.028 |
Token | me |
Token ID | 502 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
ited | +0.032 |
saw | +0.031 |
heard | +0.030 |
little | +0.028 |
two | +0.028 |
Neg logit contributions | |
anks | -0.043 |
balls | -0.041 |
hyd | -0.040 |
bleed | -0.039 |
rid | -0.039 |
Token | shake |
Token ID | 13279 |
Feature activation | -2.22e-03 |
Pos logit contributions | |
anks | +0.063 |
balls | +0.061 |
ometers | +0.059 |
ometer | +0.059 |
hyd | +0.058 |
Neg logit contributions | |
saw | -0.044 |
ited | -0.043 |
heard | -0.042 |
little | -0.041 |
two | -0.041 |
Token | this |
Token ID | 428 |
Feature activation | -1.65e-02 |
Pos logit contributions | |
anks | +0.005 |
balls | +0.005 |
hyd | +0.005 |
rid | +0.004 |
heal | +0.004 |
Neg logit contributions | |
ited | -0.003 |
saw | -0.003 |
heard | -0.002 |
Hey | -0.002 |
year | -0.002 |
Token | rad |
Token ID | 2511 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.029 |
balls | +0.028 |
hyd | +0.028 |
rid | +0.027 |
heal | +0.027 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.025 |
heard | -0.024 |
Hey | -0.021 |
year | -0.021 |
Token | ish |
Token ID | 680 |
Feature activation | -5.92e-02 |
Pos logit contributions | |
anks | +0.124 |
balls | +0.120 |
hyd | +0.111 |
sing | +0.109 |
heal | +0.109 |
Neg logit contributions | |
ited | -0.229 |
saw | -0.217 |
heard | -0.206 |
two | -0.195 |
noticed | -0.191 |
Token | , |
Token ID | 11 |
Feature activation | -3.49e-02 |
Pos logit contributions | |
anks | +0.112 |
balls | +0.107 |
hyd | +0.105 |
rid | +0.103 |
bleed | +0.101 |
Neg logit contributions | |
ited | -0.089 |
saw | -0.083 |
heard | -0.080 |
two | -0.074 |
little | -0.072 |
Token | it |
Token ID | 340 |
Feature activation | -4.07e-02 |
Pos logit contributions | |
anks | +0.063 |
balls | +0.059 |
hyd | +0.058 |
rid | +0.056 |
bleed | +0.056 |
Neg logit contributions | |
ited | -0.055 |
saw | -0.055 |
heard | -0.052 |
little | -0.049 |
two | -0.049 |
Token | 's |
Token ID | 338 |
Feature activation | -2.29e-02 |
Pos logit contributions | |
anks | +0.089 |
uggets | +0.085 |
balls | +0.083 |
ectar | +0.080 |
aro | +0.079 |
Neg logit contributions | |
saw | -0.053 |
ited | -0.053 |
heard | -0.049 |
did | -0.046 |
two | -0.044 |
Token | too |
Token ID | 1165 |
Feature activation | -3.97e-02 |
Pos logit contributions | |
anks | +0.048 |
balls | +0.046 |
hyd | +0.045 |
rid | +0.044 |
heal | +0.043 |
Neg logit contributions | |
ited | -0.031 |
saw | -0.028 |
heard | -0.027 |
two | -0.024 |
little | -0.024 |
Token | things |
Token ID | 1243 |
Feature activation | -4.72e-02 |
Pos logit contributions | |
anks | +0.034 |
balls | +0.033 |
hyd | +0.032 |
rid | +0.032 |
heal | +0.031 |
Neg logit contributions | |
ited | -0.036 |
saw | -0.033 |
heard | -0.032 |
year | -0.029 |
Hey | -0.029 |
Token | that |
Token ID | 326 |
Feature activation | -3.53e-02 |
Pos logit contributions | |
anks | +0.119 |
balls | +0.116 |
hyd | +0.115 |
sing | +0.114 |
rid | +0.112 |
Neg logit contributions | |
saw | -0.037 |
ited | -0.034 |
heard | -0.032 |
little | -0.029 |
two | -0.028 |
Token | could |
Token ID | 714 |
Feature activation | -5.08e-02 |
Pos logit contributions | |
anks | +0.074 |
balls | +0.073 |
uggets | +0.071 |
hyd | +0.070 |
sing | +0.067 |
Neg logit contributions | |
saw | -0.048 |
heard | -0.043 |
did | -0.043 |
ited | -0.041 |
two | -0.039 |
Token | cause |
Token ID | 2728 |
Feature activation | -4.14e-02 |
Pos logit contributions | |
anks | +0.096 |
balls | +0.092 |
ophone | +0.090 |
hyd | +0.090 |
ometer | +0.089 |
Neg logit contributions | |
saw | -0.076 |
ited | -0.074 |
heard | -0.072 |
two | -0.068 |
little | -0.067 |
Token | harm |
Token ID | 4419 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
anks | +0.076 |
rid | +0.073 |
balls | +0.072 |
heal | +0.072 |
hyd | +0.071 |
Neg logit contributions | |
ited | -0.067 |
saw | -0.058 |
heard | -0.055 |
Hey | -0.051 |
year | -0.051 |
Token | or |
Token ID | 393 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
anks | +0.037 |
balls | +0.037 |
hyd | +0.036 |
rid | +0.035 |
sing | +0.034 |
Neg logit contributions | |
ited | -0.023 |
saw | -0.022 |
heard | -0.021 |
little | -0.019 |
two | -0.019 |
Token | trouble |
Token ID | 5876 |
Feature activation | -4.08e-02 |
Pos logit contributions | |
anks | +0.176 |
balls | +0.170 |
hyd | +0.166 |
rid | +0.158 |
heal | +0.155 |
Neg logit contributions | |
ited | -0.170 |
saw | -0.165 |
heard | -0.158 |
little | -0.144 |
two | -0.144 |
Token | . |
Token ID | 13 |
Feature activation | -4.00e-02 |
Pos logit contributions | |
anks | +0.078 |
balls | +0.077 |
hyd | +0.076 |
sing | +0.073 |
uggets | +0.073 |
Neg logit contributions | |
saw | -0.055 |
ited | -0.054 |
heard | -0.053 |
two | -0.048 |
little | -0.048 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +7.21e-02 |
Pos logit contributions | |
balls | +0.093 |
anks | +0.091 |
rid | +0.089 |
hyd | +0.088 |
bleed | +0.087 |
Neg logit contributions | |
saw | -0.040 |
ited | -0.040 |
heard | -0.037 |
little | -0.036 |
two | -0.034 |
Token | Once |
Token ID | 7454 |
Feature activation | +8.31e-02 |
Pos logit contributions | |
ited | +0.116 |
saw | +0.091 |
heard | +0.086 |
noticed | +0.079 |
Hey | +0.079 |
Neg logit contributions | |
anks | -0.137 |
balls | -0.134 |
hyd | -0.129 |
heal | -0.129 |
rid | -0.128 |
Token | upon |
Token ID | 2402 |
Feature activation | +6.34e-02 |
Pos logit contributions | |
ited | +0.093 |
Hey | +0.068 |
heard | +0.065 |
saw | +0.062 |
Fine | +0.054 |
Neg logit contributions | |
anks | -0.188 |
balls | -0.188 |
rid | -0.179 |
heal | -0.179 |
ometer | -0.178 |
Token | make |
Token ID | 787 |
Feature activation | -8.51e-03 |
Pos logit contributions | |
anks | +0.039 |
hyd | +0.038 |
heal | +0.038 |
rid | +0.038 |
balls | +0.037 |
Neg logit contributions | |
ited | -0.048 |
saw | -0.042 |
heard | -0.041 |
year | -0.039 |
Hey | -0.039 |
Token | sure |
Token ID | 1654 |
Feature activation | +7.06e-03 |
Pos logit contributions | |
anks | +0.018 |
balls | +0.018 |
hyd | +0.018 |
rid | +0.017 |
heal | +0.017 |
Neg logit contributions | |
ited | -0.010 |
saw | -0.010 |
heard | -0.009 |
little | -0.008 |
two | -0.008 |
Token | nothing |
Token ID | 2147 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
ited | +0.010 |
saw | +0.009 |
heard | +0.008 |
Hey | +0.008 |
year | +0.008 |
Neg logit contributions | |
anks | -0.014 |
balls | -0.014 |
hyd | -0.014 |
rid | -0.013 |
heal | -0.013 |
Token | would |
Token ID | 561 |
Feature activation | -5.27e-02 |
Pos logit contributions | |
anks | +0.027 |
balls | +0.026 |
hyd | +0.026 |
rid | +0.026 |
heal | +0.026 |
Neg logit contributions | |
ited | -0.017 |
saw | -0.015 |
heard | -0.014 |
Hey | -0.013 |
year | -0.013 |
Token | make |
Token ID | 787 |
Feature activation | -3.17e-02 |
Pos logit contributions | |
anks | +0.080 |
balls | +0.077 |
hyd | +0.076 |
rid | +0.074 |
heal | +0.074 |
Neg logit contributions | |
ited | -0.100 |
saw | -0.092 |
heard | -0.089 |
Hey | -0.082 |
year | -0.082 |
Token | it |
Token ID | 340 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
anks | +0.073 |
balls | +0.073 |
hyd | +0.071 |
sing | +0.069 |
rid | +0.068 |
Neg logit contributions | |
ited | -0.032 |
saw | -0.031 |
heard | -0.029 |
little | -0.027 |
two | -0.027 |
Token | slip |
Token ID | 13819 |
Feature activation | -3.54e-02 |
Pos logit contributions | |
anks | +0.177 |
balls | +0.176 |
hyd | +0.168 |
uggets | +0.163 |
aro | +0.163 |
Neg logit contributions | |
ited | -0.167 |
saw | -0.164 |
heard | -0.158 |
little | -0.147 |
two | -0.143 |
Token | away |
Token ID | 1497 |
Feature activation | -2.50e-02 |
Pos logit contributions | |
anks | +0.070 |
balls | +0.068 |
hyd | +0.067 |
rid | +0.065 |
heal | +0.064 |
Neg logit contributions | |
ited | -0.051 |
saw | -0.047 |
heard | -0.045 |
two | -0.039 |
year | -0.039 |
Token | again |
Token ID | 757 |
Feature activation | -2.28e-02 |
Pos logit contributions | |
anks | +0.047 |
balls | +0.046 |
hyd | +0.045 |
rid | +0.044 |
heal | +0.043 |
Neg logit contributions | |
ited | -0.038 |
saw | -0.035 |
heard | -0.033 |
two | -0.030 |
little | -0.030 |
Token | . |
Token ID | 13 |
Feature activation | -4.43e-02 |
Pos logit contributions | |
anks | +0.039 |
balls | +0.038 |
hyd | +0.038 |
rid | +0.037 |
heal | +0.037 |
Neg logit contributions | |
ited | -0.039 |
saw | -0.035 |
heard | -0.034 |
Hey | -0.031 |
year | -0.030 |
Token | Then |
Token ID | 3244 |
Feature activation | +1.55e-02 |
Pos logit contributions | |
anks | +0.097 |
balls | +0.095 |
hyd | +0.094 |
rid | +0.092 |
bleed | +0.091 |
Neg logit contributions | |
ited | -0.053 |
saw | -0.049 |
heard | -0.046 |
two | -0.042 |
little | -0.041 |
Token | hopefully |
Token ID | 11481 |
Feature activation | -2.26e-02 |
Pos logit contributions | |
anks | +0.141 |
balls | +0.133 |
hyd | +0.125 |
ophone | +0.123 |
ectar | +0.123 |
Neg logit contributions | |
saw | -0.127 |
ited | -0.123 |
heard | -0.118 |
little | -0.109 |
two | -0.108 |
Token | you |
Token ID | 345 |
Feature activation | -5.45e-02 |
Pos logit contributions | |
anks | +0.039 |
balls | +0.038 |
hyd | +0.037 |
rid | +0.036 |
heal | +0.035 |
Neg logit contributions | |
ited | -0.038 |
saw | -0.036 |
heard | -0.034 |
little | -0.031 |
two | -0.031 |
Token | will |
Token ID | 481 |
Feature activation | -2.91e-02 |
Pos logit contributions | |
anks | +0.101 |
balls | +0.098 |
hyd | +0.095 |
rid | +0.093 |
heal | +0.091 |
Neg logit contributions | |
ited | -0.085 |
saw | -0.080 |
heard | -0.077 |
two | -0.069 |
little | -0.068 |
Token | be |
Token ID | 307 |
Feature activation | -5.86e-02 |
Pos logit contributions | |
anks | +0.053 |
balls | +0.052 |
hyd | +0.048 |
rid | +0.048 |
ometers | +0.047 |
Neg logit contributions | |
ited | -0.045 |
saw | -0.044 |
heard | -0.043 |
two | -0.039 |
little | -0.039 |
Token | healthy |
Token ID | 5448 |
Feature activation | -3.17e-02 |
Pos logit contributions | |
anks | +0.110 |
balls | +0.107 |
hyd | +0.103 |
rid | +0.100 |
heal | +0.100 |
Neg logit contributions | |
ited | -0.092 |
saw | -0.085 |
heard | -0.082 |
two | -0.074 |
little | -0.073 |
Token | and |
Token ID | 290 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
anks | +0.055 |
hyd | +0.054 |
balls | +0.053 |
heal | +0.052 |
rid | +0.051 |
Neg logit contributions | |
ited | -0.055 |
saw | -0.048 |
heard | -0.046 |
Hey | -0.044 |
did | -0.041 |
Token | back |
Token ID | 736 |
Feature activation | -3.42e-02 |
Pos logit contributions | |
anks | +0.178 |
balls | +0.172 |
hyd | +0.168 |
rid | +0.163 |
heal | +0.161 |
Neg logit contributions | |
ited | -0.171 |
saw | -0.158 |
heard | -0.151 |
two | -0.137 |
little | -0.136 |
Token | to |
Token ID | 284 |
Feature activation | -2.45e-02 |
Pos logit contributions | |
anks | +0.062 |
balls | +0.060 |
hyd | +0.059 |
rid | +0.058 |
heal | +0.057 |
Neg logit contributions | |
ited | -0.054 |
saw | -0.050 |
heard | -0.048 |
two | -0.044 |
little | -0.043 |
Token | your |
Token ID | 534 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
anks | +0.047 |
balls | +0.045 |
ometer | +0.044 |
ometers | +0.043 |
ophone | +0.042 |
Neg logit contributions | |
saw | -0.036 |
ited | -0.036 |
heard | -0.034 |
two | -0.032 |
little | -0.032 |
Token | old |
Token ID | 1468 |
Feature activation | -3.22e-02 |
Pos logit contributions | |
anks | +0.019 |
balls | +0.018 |
hyd | +0.018 |
rid | +0.018 |
bleed | +0.017 |
Neg logit contributions | |
ited | -0.017 |
saw | -0.016 |
heard | -0.015 |
little | -0.014 |
year | -0.014 |
Token | self |
Token ID | 2116 |
Feature activation | -9.10e-03 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.057 |
hyd | +0.056 |
rid | +0.055 |
bleed | +0.053 |
Neg logit contributions | |
ited | -0.050 |
saw | -0.048 |
heard | -0.046 |
little | -0.042 |
two | -0.042 |
Token | real |
Token ID | 1103 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
heal | +0.021 |
rid | +0.020 |
hyd | +0.020 |
balls | +0.020 |
anks | +0.020 |
Neg logit contributions | |
ited | -0.017 |
saw | -0.014 |
heard | -0.013 |
Hey | -0.013 |
year | -0.013 |
Token | wheat |
Token ID | 17135 |
Feature activation | -8.06e-02 |
Pos logit contributions | |
anks | +0.053 |
balls | +0.051 |
hyd | +0.050 |
rid | +0.049 |
bleed | +0.048 |
Neg logit contributions | |
ited | -0.036 |
saw | -0.035 |
heard | -0.033 |
little | -0.030 |
two | -0.030 |
Token | could |
Token ID | 714 |
Feature activation | -5.89e-02 |
Pos logit contributions | |
anks | +0.174 |
balls | +0.168 |
hyd | +0.165 |
rid | +0.162 |
heal | +0.160 |
Neg logit contributions | |
ited | -0.102 |
saw | -0.093 |
heard | -0.088 |
two | -0.076 |
year | -0.076 |
Token | make |
Token ID | 787 |
Feature activation | -4.39e-02 |
Pos logit contributions | |
anks | +0.105 |
balls | +0.100 |
ophone | +0.095 |
hyd | +0.095 |
ometer | +0.094 |
Neg logit contributions | |
ited | -0.094 |
saw | -0.092 |
heard | -0.088 |
two | -0.082 |
little | -0.081 |
Token | good |
Token ID | 922 |
Feature activation | -5.54e-02 |
Pos logit contributions | |
anks | +0.097 |
balls | +0.094 |
hyd | +0.091 |
sing | +0.089 |
bleed | +0.088 |
Neg logit contributions | |
ited | -0.050 |
saw | -0.049 |
heard | -0.046 |
little | -0.044 |
two | -0.043 |
Token | bread |
Token ID | 8509 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
anks | +0.103 |
balls | +0.100 |
hyd | +0.097 |
rid | +0.095 |
heal | +0.094 |
Neg logit contributions | |
ited | -0.087 |
saw | -0.080 |
heard | -0.077 |
two | -0.069 |
little | -0.069 |
Token | . |
Token ID | 13 |
Feature activation | -5.44e-02 |
Pos logit contributions | |
anks | +0.217 |
ophone | +0.205 |
sing | +0.203 |
balls | +0.202 |
bleed | +0.201 |
Neg logit contributions | |
saw | -0.115 |
heard | -0.108 |
ited | -0.108 |
two | -0.102 |
little | -0.100 |
Token | And |
Token ID | 843 |
Feature activation | -3.18e-02 |
Pos logit contributions | |
anks | +0.117 |
balls | +0.115 |
hyd | +0.112 |
rid | +0.111 |
heal | +0.109 |
Neg logit contributions | |
ited | -0.068 |
saw | -0.063 |
heard | -0.059 |
two | -0.053 |
little | -0.053 |
Token | Jack |
Token ID | 3619 |
Feature activation | +8.84e-03 |
Pos logit contributions | |
anks | +0.062 |
balls | +0.061 |
hyd | +0.060 |
rid | +0.059 |
heal | +0.058 |
Neg logit contributions | |
ited | -0.048 |
saw | -0.041 |
heard | -0.039 |
Hey | -0.036 |
year | -0.035 |
Token | loved |
Token ID | 6151 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
ited | +0.010 |
Hey | +0.008 |
year | +0.007 |
conflict | +0.007 |
Next | +0.007 |
Neg logit contributions | |
hyd | -0.019 |
heal | -0.018 |
anks | -0.018 |
fly | -0.018 |
balls | -0.018 |
Token | to |
Token ID | 284 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
ited | +0.026 |
saw | +0.024 |
heard | +0.023 |
little | +0.021 |
two | +0.021 |
Neg logit contributions | |
anks | -0.041 |
balls | -0.039 |
hyd | -0.039 |
rid | -0.038 |
heal | -0.037 |
Token | 's |
Token ID | 338 |
Feature activation | -2.06e-02 |
Pos logit contributions | |
uggets | +0.144 |
anks | +0.141 |
ectar | +0.133 |
balls | +0.132 |
sing | +0.132 |
Neg logit contributions | |
saw | -0.063 |
heard | -0.058 |
did | -0.058 |
ited | -0.053 |
two | -0.052 |
Token | important |
Token ID | 1593 |
Feature activation | -7.19e-03 |
Pos logit contributions | |
anks | +0.038 |
balls | +0.036 |
hyd | +0.035 |
rid | +0.034 |
heal | +0.033 |
Neg logit contributions | |
ited | -0.032 |
saw | -0.031 |
heard | -0.030 |
two | -0.028 |
little | -0.028 |
Token | to |
Token ID | 284 |
Feature activation | -2.85e-02 |
Pos logit contributions | |
anks | +0.015 |
balls | +0.014 |
hyd | +0.014 |
rid | +0.014 |
heal | +0.014 |
Neg logit contributions | |
ited | -0.009 |
saw | -0.009 |
heard | -0.009 |
two | -0.008 |
little | -0.007 |
Token | be |
Token ID | 307 |
Feature activation | -4.64e-02 |
Pos logit contributions | |
anks | +0.049 |
balls | +0.048 |
hyd | +0.047 |
rid | +0.045 |
ometer | +0.044 |
Neg logit contributions | |
ited | -0.047 |
saw | -0.046 |
heard | -0.043 |
two | -0.040 |
little | -0.040 |
Token | quiet |
Token ID | 5897 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
anks | +0.101 |
balls | +0.098 |
hyd | +0.095 |
rid | +0.093 |
heal | +0.092 |
Neg logit contributions | |
ited | -0.058 |
saw | -0.055 |
heard | -0.052 |
two | -0.046 |
little | -0.046 |
Token | and |
Token ID | 290 |
Feature activation | -1.02e-01 |
Pos logit contributions | |
anks | +0.035 |
balls | +0.034 |
hyd | +0.034 |
heal | +0.033 |
rid | +0.033 |
Neg logit contributions | |
ited | -0.029 |
saw | -0.024 |
heard | -0.023 |
Hey | -0.022 |
year | -0.021 |
Token | stay |
Token ID | 2652 |
Feature activation | -3.61e-02 |
Pos logit contributions | |
anks | +0.194 |
balls | +0.187 |
hyd | +0.181 |
rid | +0.176 |
heal | +0.173 |
Neg logit contributions | |
ited | -0.153 |
saw | -0.145 |
heard | -0.138 |
little | -0.124 |
two | -0.123 |
Token | away |
Token ID | 1497 |
Feature activation | -2.65e-02 |
Pos logit contributions | |
anks | +0.068 |
balls | +0.066 |
ometers | +0.061 |
hyd | +0.061 |
ometer | +0.060 |
Neg logit contributions | |
ited | -0.056 |
saw | -0.054 |
heard | -0.051 |
little | -0.047 |
two | -0.047 |
Token | from |
Token ID | 422 |
Feature activation | -1.75e-02 |
Pos logit contributions | |
anks | +0.051 |
balls | +0.050 |
hyd | +0.047 |
ophone | +0.047 |
aro | +0.046 |
Neg logit contributions | |
ited | -0.036 |
saw | -0.036 |
heard | -0.034 |
little | -0.033 |
two | -0.032 |
Token | them |
Token ID | 606 |
Feature activation | -5.19e-02 |
Pos logit contributions | |
anks | +0.034 |
balls | +0.033 |
hyd | +0.033 |
rid | +0.032 |
heal | +0.032 |
Neg logit contributions | |
ited | -0.026 |
saw | -0.023 |
heard | -0.022 |
Hey | -0.020 |
year | -0.020 |
Token | . |
Token ID | 13 |
Feature activation | -5.44e-02 |
Pos logit contributions | |
anks | +0.097 |
balls | +0.095 |
hyd | +0.093 |
rid | +0.090 |
bleed | +0.088 |
Neg logit contributions | |
ited | -0.078 |
saw | -0.074 |
heard | -0.071 |
little | -0.065 |
two | -0.065 |
Token | |
Token ID | 220 |
Feature activation | -1.74e-02 |
Pos logit contributions | |
anks | +0.017 |
balls | +0.017 |
rid | +0.016 |
hyd | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.013 |
saw | -0.013 |
heard | -0.012 |
did | -0.011 |
two | -0.011 |
Token |
|
Token ID | 198 |
Feature activation | +9.32e-03 |
Pos logit contributions | |
hyd | +0.043 |
anks | +0.042 |
balls | +0.041 |
bleed | +0.041 |
rid | +0.041 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.014 |
heard | -0.013 |
two | -0.013 |
little | -0.011 |
Token |
|
Token ID | 198 |
Feature activation | -4.82e-03 |
Pos logit contributions | |
ited | +0.015 |
saw | +0.013 |
heard | +0.012 |
enough | +0.011 |
noticed | +0.011 |
Neg logit contributions | |
balls | -0.017 |
anks | -0.017 |
heal | -0.016 |
rid | -0.016 |
hyd | -0.015 |
Token | Tom |
Token ID | 13787 |
Feature activation | +2.99e-02 |
Pos logit contributions | |
hyd | +0.009 |
balls | +0.009 |
uries | +0.009 |
heal | +0.009 |
anks | +0.009 |
Neg logit contributions | |
ited | -0.006 |
saw | -0.006 |
heard | -0.005 |
enough | -0.004 |
did | -0.004 |
Token | my |
Token ID | 1820 |
Feature activation | +2.95e-02 |
Pos logit contributions | |
ited | +0.042 |
heard | +0.034 |
saw | +0.034 |
Hey | +0.033 |
year | +0.033 |
Neg logit contributions | |
anks | -0.058 |
balls | -0.055 |
ophone | -0.055 |
hyd | -0.055 |
fly | -0.054 |
Token | said |
Token ID | 531 |
Feature activation | +8.82e-02 |
Pos logit contributions | |
ited | +0.027 |
saw | +0.019 |
heard | +0.019 |
year | +0.017 |
Hey | +0.017 |
Neg logit contributions | |
anks | -0.071 |
balls | -0.071 |
hyd | -0.069 |
heal | -0.068 |
rid | -0.068 |
Token | , |
Token ID | 11 |
Feature activation | +1.92e-02 |
Pos logit contributions | |
ited | +0.067 |
saw | +0.055 |
heard | +0.049 |
year | +0.037 |
Hey | +0.037 |
Neg logit contributions | |
anks | -0.235 |
balls | -0.231 |
hyd | -0.228 |
rid | -0.224 |
heal | -0.223 |
Token | � |
Token ID | 6184 |
Feature activation | +1.82e-03 |
Pos logit contributions | |
ited | +0.029 |
saw | +0.026 |
heard | +0.025 |
year | +0.022 |
Hey | +0.022 |
Neg logit contributions | |
anks | -0.037 |
balls | -0.036 |
hyd | -0.035 |
rid | -0.034 |
heal | -0.034 |
Token | � |
Token ID | 95 |
Feature activation | -8.78e-03 |
Pos logit contributions | |
ited | +0.003 |
saw | +0.003 |
Hey | +0.003 |
heard | +0.003 |
did | +0.003 |
Neg logit contributions | |
balls | -0.003 |
anks | -0.003 |
rid | -0.003 |
fly | -0.003 |
bleed | -0.003 |
Token | € |
Token ID | 26391 |
Feature activation | -9.28e-03 |
Pos logit contributions | |
balls | +0.014 |
anks | +0.014 |
rid | +0.013 |
bleed | +0.013 |
fly | +0.013 |
Neg logit contributions | |
ited | -0.016 |
saw | -0.015 |
heard | -0.014 |
Hey | -0.014 |
did | -0.013 |
Token | � |
Token ID | 129 |
Feature activation | +1.24e-03 |
Pos logit contributions | |
anks | +0.015 |
balls | +0.015 |
rid | +0.014 |
hyd | +0.014 |
heal | +0.014 |
Neg logit contributions | |
ited | -0.016 |
saw | -0.016 |
heard | -0.015 |
did | -0.014 |
year | -0.013 |
Token | of |
Token ID | 286 |
Feature activation | +3.73e-03 |
Pos logit contributions | |
ited | +0.036 |
heard | +0.032 |
saw | +0.030 |
Hey | +0.030 |
did | +0.029 |
Neg logit contributions | |
anks | -0.057 |
hyd | -0.057 |
heal | -0.055 |
balls | -0.055 |
ectar | -0.054 |
Token | the |
Token ID | 262 |
Feature activation | -1.13e-03 |
Pos logit contributions | |
ited | +0.006 |
Fine | +0.005 |
saw | +0.005 |
heard | +0.005 |
Hey | +0.005 |
Neg logit contributions | |
rid | -0.007 |
anks | -0.006 |
uggets | -0.006 |
hyd | -0.006 |
heal | -0.006 |
Token | vendors |
Token ID | 17192 |
Feature activation | -1.81e-02 |
Pos logit contributions | |
hyd | +0.002 |
balls | +0.002 |
anks | +0.002 |
rid | +0.002 |
heal | +0.002 |
Neg logit contributions | |
ited | -0.002 |
heard | -0.001 |
saw | -0.001 |
ara | -0.001 |
did | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | -2.24e-02 |
Pos logit contributions | |
hyd | +0.035 |
anks | +0.035 |
balls | +0.035 |
heal | +0.034 |
ometers | +0.033 |
Neg logit contributions | |
ited | -0.025 |
Hey | -0.022 |
heard | -0.019 |
ha | -0.018 |
conflict | -0.018 |
Token | politely |
Token ID | 34313 |
Feature activation | +5.70e-03 |
Pos logit contributions | |
balls | +0.045 |
rid | +0.044 |
hyd | +0.044 |
heal | +0.042 |
anks | +0.042 |
Neg logit contributions | |
ited | -0.030 |
year | -0.022 |
End | -0.021 |
Hey | -0.021 |
time | -0.021 |
Token | said |
Token ID | 531 |
Feature activation | +6.56e-02 |
Pos logit contributions | |
ited | +0.007 |
Hey | +0.006 |
End | +0.006 |
year | +0.006 |
conflict | +0.005 |
Neg logit contributions | |
hyd | -0.011 |
anks | -0.011 |
balls | -0.011 |
heal | -0.010 |
rid | -0.010 |
Token | , |
Token ID | 11 |
Feature activation | +4.40e-03 |
Pos logit contributions | |
ited | +0.096 |
saw | +0.074 |
heard | +0.070 |
Hey | +0.063 |
year | +0.063 |
Neg logit contributions | |
balls | -0.138 |
anks | -0.136 |
heal | -0.131 |
hyd | -0.128 |
rid | -0.128 |
Token | " |
Token ID | 366 |
Feature activation | -2.31e-02 |
Pos logit contributions | |
ited | +0.006 |
saw | +0.006 |
heard | +0.005 |
two | +0.005 |
little | +0.005 |
Neg logit contributions | |
anks | -0.009 |
balls | -0.009 |
hyd | -0.008 |
rid | -0.008 |
heal | -0.008 |
Token | Hello |
Token ID | 15496 |
Feature activation | +2.97e-02 |
Pos logit contributions | |
hyd | +0.043 |
anks | +0.041 |
balls | +0.041 |
cr | +0.040 |
rid | +0.040 |
Neg logit contributions | |
ited | -0.032 |
saw | -0.029 |
heard | -0.027 |
did | -0.026 |
conflict | -0.025 |
Token | , |
Token ID | 11 |
Feature activation | +8.10e-03 |
Pos logit contributions | |
ited | +0.043 |
heard | +0.031 |
sighed | +0.030 |
saw | +0.030 |
End | +0.028 |
Neg logit contributions | |
anks | -0.060 |
balls | -0.057 |
hyd | -0.056 |
flies | -0.054 |
ometers | -0.054 |
Token | could |
Token ID | 714 |
Feature activation | -4.39e-03 |
Pos logit contributions | |
ited | +0.013 |
heard | +0.011 |
saw | +0.010 |
sighed | +0.009 |
ised | +0.009 |
Neg logit contributions | |
anks | -0.014 |
balls | -0.014 |
hyd | -0.014 |
heal | -0.014 |
ometers | -0.013 |
Token | not |
Token ID | 407 |
Feature activation | -3.55e-02 |
Pos logit contributions | |
anks | +0.063 |
balls | +0.061 |
hyd | +0.060 |
rid | +0.058 |
heal | +0.058 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.042 |
heard | -0.040 |
two | -0.036 |
little | -0.036 |
Token | afraid |
Token ID | 7787 |
Feature activation | -8.12e-03 |
Pos logit contributions | |
anks | +0.071 |
balls | +0.069 |
hyd | +0.068 |
rid | +0.066 |
heal | +0.065 |
Neg logit contributions | |
ited | -0.051 |
saw | -0.047 |
heard | -0.044 |
little | -0.039 |
two | -0.039 |
Token | ." |
Token ID | 526 |
Feature activation | -2.65e-02 |
Pos logit contributions | |
anks | +0.014 |
balls | +0.013 |
hyd | +0.013 |
rid | +0.012 |
heal | +0.012 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.013 |
heard | -0.013 |
little | -0.012 |
year | -0.012 |
Token |
|
Token ID | 198 |
Feature activation | +3.39e-04 |
Pos logit contributions | |
anks | +0.070 |
balls | +0.069 |
bleed | +0.067 |
sing | +0.066 |
ophone | +0.066 |
Neg logit contributions | |
saw | -0.021 |
ited | -0.019 |
little | -0.018 |
heard | -0.018 |
two | -0.017 |
Token |
|
Token ID | 198 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
ited | +0.000 |
saw | +0.000 |
heard | +0.000 |
year | +0.000 |
two | +0.000 |
Neg logit contributions | |
anks | -0.001 |
balls | -0.001 |
hyd | -0.001 |
rid | -0.001 |
heal | -0.001 |
Token | The |
Token ID | 464 |
Feature activation | +6.36e-02 |
Pos logit contributions | |
balls | +0.034 |
hyd | +0.034 |
anks | +0.034 |
rid | +0.034 |
heal | +0.034 |
Neg logit contributions | |
ited | -0.017 |
saw | -0.015 |
heard | -0.014 |
Hey | -0.011 |
enough | -0.011 |
Token | little |
Token ID | 1310 |
Feature activation | +3.92e-02 |
Pos logit contributions | |
ited | +0.066 |
Hey | +0.045 |
saw | +0.045 |
aw | +0.044 |
Tom | +0.043 |
Neg logit contributions | |
balls | -0.141 |
heal | -0.140 |
hyd | -0.139 |
rid | -0.138 |
sing | -0.134 |
Token | girl |
Token ID | 2576 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
ited | +0.050 |
heard | +0.039 |
Tom | +0.039 |
saw | +0.039 |
When | +0.039 |
Neg logit contributions | |
hyd | -0.076 |
rid | -0.074 |
heal | -0.069 |
flies | -0.069 |
fly | -0.069 |
Token | realised |
Token ID | 19169 |
Feature activation | +6.86e-02 |
Pos logit contributions | |
ited | +0.012 |
saw | +0.008 |
Hey | +0.008 |
heard | +0.008 |
year | +0.008 |
Neg logit contributions | |
hyd | -0.023 |
balls | -0.023 |
anks | -0.022 |
heal | -0.022 |
rid | -0.022 |
Token | that |
Token ID | 326 |
Feature activation | +3.00e-02 |
Pos logit contributions | |
ited | +0.102 |
saw | +0.082 |
heard | +0.077 |
Hey | +0.074 |
year | +0.067 |
Neg logit contributions | |
heal | -0.133 |
hyd | -0.132 |
anks | -0.131 |
balls | -0.131 |
rid | -0.130 |
Token | if |
Token ID | 611 |
Feature activation | +1.89e-02 |
Pos logit contributions | |
ited | +0.040 |
saw | +0.034 |
heard | +0.032 |
Hey | +0.029 |
year | +0.028 |
Neg logit contributions | |
anks | -0.063 |
balls | -0.062 |
hyd | -0.062 |
rid | -0.061 |
heal | -0.060 |
Token | L |
Token ID | 43 |
Feature activation | +3.68e-02 |
Pos logit contributions | |
ited | +0.001 |
enough | +0.001 |
saw | +0.001 |
heard | +0.001 |
joy | +0.001 |
Neg logit contributions | |
balls | -0.001 |
heal | -0.001 |
hyd | -0.001 |
ophone | -0.001 |
tr | -0.001 |
Token | ily |
Token ID | 813 |
Feature activation | +4.17e-02 |
Pos logit contributions | |
ited | +0.068 |
saw | +0.064 |
heard | +0.060 |
noticed | +0.056 |
arrived | +0.055 |
Neg logit contributions | |
anks | -0.058 |
ectar | -0.055 |
balls | -0.051 |
ophone | -0.051 |
uggets | -0.050 |
Token | 's |
Token ID | 338 |
Feature activation | +6.76e-02 |
Pos logit contributions | |
ited | +0.070 |
grown | +0.060 |
Hey | +0.059 |
joy | +0.056 |
he | +0.056 |
Neg logit contributions | |
hyd | -0.062 |
balls | -0.060 |
gy | -0.059 |
fet | -0.056 |
heal | -0.056 |
Token | mom |
Token ID | 1995 |
Feature activation | +3.23e-02 |
Pos logit contributions | |
ited | +0.136 |
saw | +0.115 |
Hey | +0.115 |
heard | +0.113 |
noticed | +0.103 |
Neg logit contributions | |
balls | -0.101 |
anks | -0.100 |
hyd | -0.094 |
sing | -0.091 |
heal | -0.090 |
Token | my |
Token ID | 1820 |
Feature activation | +2.84e-02 |
Pos logit contributions | |
ited | +0.059 |
Hey | +0.058 |
amusement | +0.055 |
joy | +0.054 |
trouble | +0.054 |
Neg logit contributions | |
heal | -0.036 |
fly | -0.036 |
ophone | -0.035 |
balls | -0.035 |
hyd | -0.034 |
Token | warned |
Token ID | 7728 |
Feature activation | +7.93e-02 |
Pos logit contributions | |
ited | +0.037 |
Hey | +0.032 |
time | +0.030 |
conflict | +0.030 |
Next | +0.029 |
Neg logit contributions | |
hyd | -0.050 |
heal | -0.049 |
balls | -0.049 |
fly | -0.049 |
flies | -0.046 |
Token | her |
Token ID | 607 |
Feature activation | +4.93e-02 |
Pos logit contributions | |
ited | +0.135 |
saw | +0.115 |
heard | +0.104 |
Hey | +0.102 |
noticed | +0.100 |
Neg logit contributions | |
heal | -0.141 |
rid | -0.132 |
anks | -0.132 |
balls | -0.130 |
bleed | -0.128 |
Token | that |
Token ID | 326 |
Feature activation | +4.66e-02 |
Pos logit contributions | |
ited | +0.079 |
saw | +0.070 |
heard | +0.067 |
Hey | +0.061 |
year | +0.060 |
Neg logit contributions | |
anks | -0.092 |
balls | -0.090 |
hyd | -0.088 |
heal | -0.086 |
rid | -0.086 |
Token | it |
Token ID | 340 |
Feature activation | -3.23e-04 |
Pos logit contributions | |
ited | +0.060 |
saw | +0.056 |
heard | +0.053 |
two | +0.047 |
little | +0.047 |
Neg logit contributions | |
anks | -0.099 |
balls | -0.095 |
hyd | -0.093 |
rid | -0.092 |
heal | -0.090 |
Token | might |
Token ID | 1244 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
anks | +0.001 |
balls | +0.001 |
hyd | +0.001 |
rid | +0.001 |
heal | +0.001 |
Neg logit contributions | |
ited | -0.000 |
saw | -0.000 |
heard | -0.000 |
two | -0.000 |
did | -0.000 |
Token | be |
Token ID | 307 |
Feature activation | -4.07e-02 |
Pos logit contributions | |
anks | +0.035 |
balls | +0.034 |
ophone | +0.034 |
uggets | +0.032 |
ometers | +0.032 |
Neg logit contributions | |
saw | -0.027 |
ited | -0.026 |
heard | -0.026 |
time | -0.024 |
two | -0.024 |
Token | Once |
Token ID | 7454 |
Feature activation | +7.05e-02 |
Pos logit contributions | |
ited | +0.073 |
saw | +0.056 |
heard | +0.053 |
conflict | +0.049 |
noticed | +0.048 |
Neg logit contributions | |
anks | -0.094 |
balls | -0.091 |
hyd | -0.089 |
rid | -0.087 |
heal | -0.087 |
Token | upon |
Token ID | 2402 |
Feature activation | +5.11e-02 |
Pos logit contributions | |
ited | +0.079 |
Hey | +0.055 |
heard | +0.053 |
saw | +0.051 |
Fine | +0.045 |
Neg logit contributions | |
balls | -0.161 |
anks | -0.159 |
ometer | -0.153 |
rid | -0.153 |
ectar | -0.152 |
Token | a |
Token ID | 257 |
Feature activation | +8.07e-02 |
Pos logit contributions | |
ited | +0.074 |
saw | +0.054 |
Hey | +0.051 |
conflict | +0.049 |
heard | +0.049 |
Neg logit contributions | |
heal | -0.102 |
balls | -0.101 |
bleed | -0.100 |
hyd | -0.098 |
rid | -0.098 |
Token | time |
Token ID | 640 |
Feature activation | +5.34e-02 |
Pos logit contributions | |
ited | +0.061 |
saw | +0.050 |
Hey | +0.048 |
heard | +0.045 |
two | +0.042 |
Neg logit contributions | |
hyd | -0.188 |
sing | -0.188 |
balls | -0.185 |
heal | -0.181 |
anks | -0.177 |
Token | , |
Token ID | 11 |
Feature activation | +6.42e-02 |
Pos logit contributions | |
ited | +0.050 |
saw | +0.035 |
heard | +0.033 |
Hey | +0.027 |
year | +0.025 |
Neg logit contributions | |
balls | -0.135 |
anks | -0.134 |
hyd | -0.130 |
heal | -0.128 |
rid | -0.127 |
Token | there |
Token ID | 612 |
Feature activation | +6.66e-02 |
Pos logit contributions | |
ited | +0.092 |
oph | +0.068 |
aman | +0.056 |
amusement | +0.056 |
conflict | +0.056 |
Neg logit contributions | |
balls | -0.133 |
fly | -0.129 |
hyd | -0.126 |
sing | -0.125 |
heal | -0.124 |
Token | was |
Token ID | 373 |
Feature activation | +5.29e-02 |
Pos logit contributions | |
ited | +0.108 |
vious | +0.066 |
conflict | +0.066 |
Exc | +0.065 |
Hey | +0.065 |
Neg logit contributions | |
balls | -0.124 |
heal | -0.122 |
hyd | -0.121 |
cr | -0.114 |
fly | -0.114 |
Token | a |
Token ID | 257 |
Feature activation | +4.43e-02 |
Pos logit contributions | |
ited | +0.077 |
saw | +0.058 |
heard | +0.057 |
sighed | +0.051 |
Hey | +0.050 |
Neg logit contributions | |
balls | -0.105 |
hyd | -0.102 |
anks | -0.100 |
uggets | -0.100 |
rid | -0.100 |
Token | little |
Token ID | 1310 |
Feature activation | +2.67e-02 |
Pos logit contributions | |
ited | +0.025 |
saw | +0.019 |
heard | +0.016 |
year | +0.010 |
little | +0.010 |
Neg logit contributions | |
anks | -0.127 |
balls | -0.124 |
hyd | -0.123 |
rid | -0.121 |
heal | -0.120 |
Token | girl |
Token ID | 2576 |
Feature activation | +2.44e-02 |
Pos logit contributions | |
ited | +0.022 |
saw | +0.020 |
heard | +0.017 |
little | +0.015 |
year | +0.014 |
Neg logit contributions | |
anks | -0.071 |
balls | -0.068 |
hyd | -0.066 |
rid | -0.066 |
heal | -0.065 |
Token | named |
Token ID | 3706 |
Feature activation | +3.38e-02 |
Pos logit contributions | |
saw | +0.031 |
ited | +0.029 |
heard | +0.028 |
year | +0.025 |
little | +0.025 |
Neg logit contributions | |
anks | -0.057 |
balls | -0.055 |
rid | -0.051 |
bleed | -0.051 |
ometers | -0.051 |
Token |
|
Token ID | 198 |
Feature activation | -1.60e-03 |
Pos logit contributions | |
anks | +0.110 |
balls | +0.108 |
hyd | +0.107 |
rid | +0.104 |
heal | +0.103 |
Neg logit contributions | |
ited | -0.054 |
saw | -0.049 |
heard | -0.046 |
two | -0.040 |
Hey | -0.039 |
Token |
|
Token ID | 198 |
Feature activation | -1.33e-02 |
Pos logit contributions | |
anks | +0.003 |
balls | +0.003 |
heal | +0.003 |
hyd | +0.003 |
rid | +0.003 |
Neg logit contributions | |
ited | -0.002 |
saw | -0.002 |
heard | -0.002 |
Hey | -0.002 |
year | -0.002 |
Token | She |
Token ID | 3347 |
Feature activation | +5.92e-03 |
Pos logit contributions | |
balls | +0.024 |
hyd | +0.023 |
heal | +0.023 |
anks | +0.023 |
bleed | +0.023 |
Neg logit contributions | |
ited | -0.021 |
saw | -0.018 |
heard | -0.017 |
enough | -0.016 |
noticed | -0.015 |
Token | waited |
Token ID | 13488 |
Feature activation | +4.60e-02 |
Pos logit contributions | |
ited | +0.009 |
End | +0.006 |
year | +0.006 |
ised | +0.006 |
Today | +0.006 |
Neg logit contributions | |
fet | -0.011 |
hyd | -0.010 |
balls | -0.010 |
ophone | -0.010 |
rid | -0.010 |
Token | for |
Token ID | 329 |
Feature activation | +2.25e-02 |
Pos logit contributions | |
ited | +0.092 |
Hey | +0.077 |
saw | +0.077 |
heard | +0.073 |
year | +0.070 |
Neg logit contributions | |
balls | -0.072 |
anks | -0.069 |
heal | -0.066 |
hyd | -0.066 |
rid | -0.064 |
Token | a |
Token ID | 257 |
Feature activation | +1.60e-02 |
Pos logit contributions | |
ited | +0.030 |
saw | +0.024 |
heard | +0.022 |
Hey | +0.021 |
year | +0.020 |
Neg logit contributions | |
anks | -0.048 |
balls | -0.048 |
rid | -0.048 |
heal | -0.047 |
hyd | -0.046 |
Token | quiet |
Token ID | 5897 |
Feature activation | +2.57e-02 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.016 |
heard | +0.016 |
little | +0.014 |
year | +0.013 |
Neg logit contributions | |
anks | -0.037 |
balls | -0.036 |
hyd | -0.035 |
rid | -0.034 |
heal | -0.034 |
Token | moment |
Token ID | 2589 |
Feature activation | +2.31e-02 |
Pos logit contributions | |
ited | +0.026 |
saw | +0.018 |
Hey | +0.017 |
heard | +0.015 |
Next | +0.015 |
Neg logit contributions | |
heal | -0.060 |
hyd | -0.060 |
balls | -0.059 |
anks | -0.059 |
rid | -0.059 |
Token | . |
Token ID | 13 |
Feature activation | -3.68e-02 |
Pos logit contributions | |
ited | +0.036 |
saw | +0.031 |
heard | +0.029 |
Hey | +0.029 |
noticed | +0.026 |
Neg logit contributions | |
anks | -0.043 |
balls | -0.043 |
hyd | -0.041 |
heal | -0.040 |
rid | -0.040 |
Token | Then |
Token ID | 3244 |
Feature activation | +1.73e-02 |
Pos logit contributions | |
anks | +0.068 |
balls | +0.064 |
hyd | +0.061 |
heal | +0.061 |
ometers | +0.059 |
Neg logit contributions | |
ited | -0.064 |
saw | -0.054 |
heard | -0.052 |
year | -0.049 |
Hey | -0.047 |
Token | she |
Token ID | 673 |
Feature activation | -8.26e-04 |
Pos logit contributions | |
ited | +0.030 |
saw | +0.021 |
year | +0.021 |
heard | +0.021 |
Hey | +0.021 |
Neg logit contributions | |
balls | -0.032 |
anks | -0.031 |
heal | -0.030 |
rid | -0.030 |
hyd | -0.030 |
Token | ?" |
Token ID | 1701 |
Feature activation | +1.82e-02 |
Pos logit contributions | |
anks | +0.114 |
balls | +0.111 |
hyd | +0.109 |
rid | +0.106 |
bleed | +0.104 |
Neg logit contributions | |
ited | -0.081 |
saw | -0.077 |
heard | -0.073 |
two | -0.066 |
little | -0.065 |
Token | asked |
Token ID | 1965 |
Feature activation | +8.27e-02 |
Pos logit contributions | |
ited | +0.025 |
saw | +0.018 |
heard | +0.017 |
conflict | +0.017 |
Hey | +0.016 |
Neg logit contributions | |
heal | -0.035 |
hyd | -0.035 |
balls | -0.034 |
fly | -0.034 |
anks | -0.034 |
Token | Charles |
Token ID | 7516 |
Feature activation | +2.76e-02 |
Pos logit contributions | |
ited | +0.170 |
saw | +0.144 |
heard | +0.137 |
conflict | +0.132 |
Hey | +0.130 |
Neg logit contributions | |
balls | -0.113 |
anks | -0.112 |
heal | -0.108 |
rid | -0.106 |
hyd | -0.105 |
Token | . |
Token ID | 13 |
Feature activation | -9.20e-03 |
Pos logit contributions | |
ited | +0.039 |
saw | +0.037 |
heard | +0.035 |
little | +0.031 |
two | +0.031 |
Neg logit contributions | |
anks | -0.056 |
balls | -0.054 |
hyd | -0.054 |
rid | -0.052 |
heal | -0.051 |
Token |
|
Token ID | 198 |
Feature activation | +1.07e-02 |
Pos logit contributions | |
anks | +0.021 |
balls | +0.020 |
hyd | +0.020 |
rid | +0.020 |
heal | +0.019 |
Neg logit contributions | |
ited | -0.011 |
saw | -0.010 |
heard | -0.009 |
two | -0.008 |
year | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | -3.96e-04 |
Pos logit contributions | |
ited | +0.014 |
saw | +0.012 |
heard | +0.011 |
Hey | +0.010 |
did | +0.009 |
Neg logit contributions | |
anks | -0.023 |
balls | -0.023 |
hyd | -0.022 |
heal | -0.022 |
rid | -0.022 |
Token | " |
Token ID | 1 |
Feature activation | -9.80e-03 |
Pos logit contributions | |
balls | +0.001 |
hyd | +0.001 |
heal | +0.001 |
anks | +0.001 |
fly | +0.001 |
Neg logit contributions | |
ited | -0.001 |
saw | -0.000 |
heard | -0.000 |
did | -0.000 |
conflict | -0.000 |
Token | I |
Token ID | 40 |
Feature activation | -1.13e-02 |
Pos logit contributions | |
anks | +0.021 |
balls | +0.021 |
hyd | +0.020 |
rid | +0.020 |
heal | +0.020 |
Neg logit contributions | |
ited | -0.012 |
saw | -0.011 |
heard | -0.010 |
little | -0.009 |
year | -0.009 |
Token | 'm |
Token ID | 1101 |
Feature activation | +2.00e-03 |
Pos logit contributions | |
hyd | +0.018 |
rid | +0.017 |
anks | +0.017 |
balls | +0.017 |
heal | +0.017 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.017 |
heard | -0.017 |
Hey | -0.017 |
year | -0.016 |
Token | making |
Token ID | 1642 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
saw | +0.003 |
ited | +0.003 |
two | +0.003 |
heard | +0.003 |
little | +0.003 |
Neg logit contributions | |
anks | -0.004 |
balls | -0.004 |
hyd | -0.004 |
heal | -0.004 |
ophone | -0.003 |
Token | a |
Token ID | 257 |
Feature activation | -2.18e-02 |
Pos logit contributions | |
ited | +0.015 |
saw | +0.014 |
heard | +0.013 |
two | +0.012 |
little | +0.012 |
Neg logit contributions | |
anks | -0.034 |
hyd | -0.034 |
balls | -0.034 |
rid | -0.032 |
sing | -0.032 |
Token | lot |
Token ID | 1256 |
Feature activation | +3.48e-03 |
Pos logit contributions | |
ited | +0.003 |
saw | +0.003 |
heard | +0.003 |
little | +0.002 |
year | +0.002 |
Neg logit contributions | |
anks | -0.007 |
balls | -0.007 |
hyd | -0.007 |
rid | -0.007 |
bleed | -0.007 |
Token | . |
Token ID | 13 |
Feature activation | -2.73e-02 |
Pos logit contributions | |
ited | +0.005 |
saw | +0.005 |
heard | +0.005 |
year | +0.004 |
two | +0.004 |
Neg logit contributions | |
anks | -0.007 |
balls | -0.006 |
hyd | -0.006 |
rid | -0.006 |
heal | -0.006 |
Token | Tim |
Token ID | 5045 |
Feature activation | +2.24e-02 |
Pos logit contributions | |
anks | +0.054 |
rid | +0.053 |
balls | +0.053 |
hyd | +0.052 |
bleed | +0.052 |
Neg logit contributions | |
saw | -0.037 |
ited | -0.036 |
heard | -0.033 |
little | -0.032 |
two | -0.032 |
Token | had |
Token ID | 550 |
Feature activation | +2.27e-02 |
Pos logit contributions | |
ited | +0.028 |
Hey | +0.022 |
conflict | +0.021 |
year | +0.021 |
ised | +0.021 |
Neg logit contributions | |
hyd | -0.043 |
balls | -0.042 |
heal | -0.042 |
anks | -0.041 |
ophone | -0.041 |
Token | a |
Token ID | 257 |
Feature activation | +2.02e-02 |
Pos logit contributions | |
ited | +0.032 |
Hey | +0.024 |
saw | +0.023 |
year | +0.022 |
heard | +0.021 |
Neg logit contributions | |
balls | -0.047 |
rid | -0.046 |
anks | -0.046 |
hyd | -0.046 |
heal | -0.044 |
Token | fun |
Token ID | 1257 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
ited | +0.029 |
saw | +0.024 |
heard | +0.023 |
Hey | +0.021 |
year | +0.020 |
Neg logit contributions | |
balls | -0.040 |
anks | -0.040 |
hyd | -0.039 |
rid | -0.038 |
heal | -0.038 |
Token | day |
Token ID | 1110 |
Feature activation | +1.13e-03 |
Pos logit contributions | |
ited | +0.024 |
saw | +0.020 |
heard | +0.020 |
Hey | +0.018 |
two | +0.017 |
Neg logit contributions | |
anks | -0.042 |
balls | -0.041 |
hyd | -0.041 |
rid | -0.039 |
heal | -0.039 |
Token | at |
Token ID | 379 |
Feature activation | +1.14e-02 |
Pos logit contributions | |
ited | +0.001 |
saw | +0.001 |
heard | +0.001 |
two | +0.001 |
year | +0.001 |
Neg logit contributions | |
anks | -0.003 |
balls | -0.003 |
heal | -0.003 |
rid | -0.003 |
bleed | -0.003 |
Token | the |
Token ID | 262 |
Feature activation | +8.70e-03 |
Pos logit contributions | |
ited | +0.012 |
saw | +0.010 |
heard | +0.009 |
Hey | +0.008 |
year | +0.007 |
Neg logit contributions | |
anks | -0.028 |
balls | -0.027 |
hyd | -0.027 |
rid | -0.026 |
heal | -0.026 |
Token | park |
Token ID | 3952 |
Feature activation | -3.49e-04 |
Pos logit contributions | |
ited | +0.013 |
ised | +0.009 |
heard | +0.009 |
sighed | +0.009 |
frowned | +0.009 |
Neg logit contributions | |
heal | -0.017 |
anks | -0.016 |
drown | -0.016 |
float | -0.015 |
hyd | -0.015 |
Token | with |
Token ID | 351 |
Feature activation | +3.28e-04 |
Pos logit contributions | |
anks | +0.001 |
balls | +0.001 |
hyd | +0.001 |
rid | +0.001 |
heal | +0.001 |
Neg logit contributions | |
ited | -0.001 |
saw | -0.001 |
heard | -0.000 |
Hey | -0.000 |
year | -0.000 |
Token | smile |
Token ID | 8212 |
Feature activation | +1.53e-02 |
Pos logit contributions | |
ited | +0.038 |
saw | +0.031 |
heard | +0.030 |
Hey | +0.030 |
year | +0.028 |
Neg logit contributions | |
balls | -0.043 |
hyd | -0.043 |
anks | -0.043 |
heal | -0.042 |
rid | -0.042 |
Token | . |
Token ID | 13 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
ited | +0.028 |
saw | +0.025 |
heard | +0.024 |
Hey | +0.022 |
year | +0.022 |
Neg logit contributions | |
anks | -0.025 |
balls | -0.024 |
hyd | -0.024 |
heal | -0.023 |
rid | -0.023 |
Token | " |
Token ID | 366 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
anks | +0.025 |
balls | +0.024 |
hyd | +0.023 |
rid | +0.023 |
heal | +0.023 |
Neg logit contributions | |
ited | -0.019 |
saw | -0.016 |
heard | -0.015 |
year | -0.014 |
Hey | -0.014 |
Token | If |
Token ID | 1532 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
balls | +0.034 |
anks | +0.034 |
hyd | +0.033 |
rid | +0.033 |
heal | +0.032 |
Neg logit contributions | |
ited | -0.023 |
saw | -0.021 |
heard | -0.020 |
little | -0.017 |
year | -0.017 |
Token | you |
Token ID | 345 |
Feature activation | -2.19e-02 |
Pos logit contributions | |
ited | +0.038 |
saw | +0.034 |
heard | +0.033 |
Hey | +0.031 |
year | +0.031 |
Neg logit contributions | |
anks | -0.021 |
balls | -0.021 |
hyd | -0.020 |
heal | -0.019 |
rid | -0.019 |
Token | guess |
Token ID | 4724 |
Feature activation | +9.09e-03 |
Pos logit contributions | |
anks | +0.036 |
balls | +0.035 |
hyd | +0.035 |
rid | +0.034 |
heal | +0.034 |
Neg logit contributions | |
ited | -0.038 |
saw | -0.035 |
heard | -0.034 |
year | -0.031 |
Hey | -0.031 |
Token | right |
Token ID | 826 |
Feature activation | +1.97e-02 |
Pos logit contributions | |
ited | +0.012 |
saw | +0.011 |
heard | +0.010 |
year | +0.009 |
Hey | +0.009 |
Neg logit contributions | |
anks | -0.019 |
balls | -0.019 |
hyd | -0.018 |
rid | -0.018 |
heal | -0.018 |
Token | , |
Token ID | 11 |
Feature activation | -1.33e-02 |
Pos logit contributions | |
ited | +0.029 |
saw | +0.020 |
Hey | +0.019 |
heard | +0.019 |
oph | +0.019 |
Neg logit contributions | |
balls | -0.040 |
anks | -0.039 |
hyd | -0.039 |
ometer | -0.038 |
rid | -0.038 |
Token | you |
Token ID | 345 |
Feature activation | -1.39e-02 |
Pos logit contributions | |
anks | +0.019 |
balls | +0.019 |
hyd | +0.018 |
heal | +0.018 |
rid | +0.018 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.024 |
heard | -0.023 |
year | -0.022 |
Hey | -0.022 |
Token | get |
Token ID | 651 |
Feature activation | -7.03e-03 |
Pos logit contributions | |
anks | +0.026 |
balls | +0.025 |
hyd | +0.025 |
heal | +0.024 |
rid | +0.024 |
Neg logit contributions | |
ited | -0.022 |
saw | -0.019 |
heard | -0.019 |
Hey | -0.017 |
year | -0.017 |
Token | a |
Token ID | 257 |
Feature activation | +1.53e-03 |
Pos logit contributions | |
anks | +0.014 |
balls | +0.014 |
rid | +0.014 |
hyd | +0.013 |
heal | +0.013 |
Neg logit contributions | |
ited | -0.010 |
saw | -0.009 |
heard | -0.008 |
year | -0.008 |
Hey | -0.008 |
Token | later |
Token ID | 1568 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
anks | +0.033 |
balls | +0.032 |
hyd | +0.032 |
rid | +0.031 |
heal | +0.031 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.024 |
heard | -0.023 |
Hey | -0.021 |
year | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -3.64e-02 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.015 |
Hey | +0.015 |
heard | +0.014 |
conflict | +0.013 |
Neg logit contributions | |
balls | -0.018 |
anks | -0.017 |
heal | -0.017 |
hyd | -0.017 |
rid | -0.016 |
Token | She |
Token ID | 1375 |
Feature activation | +1.50e-03 |
Pos logit contributions | |
anks | +0.079 |
balls | +0.077 |
hyd | +0.076 |
rid | +0.074 |
heal | +0.074 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.041 |
heard | -0.039 |
two | -0.034 |
year | -0.034 |
Token | learned |
Token ID | 4499 |
Feature activation | +4.62e-02 |
Pos logit contributions | |
ited | +0.002 |
Hey | +0.002 |
conflict | +0.002 |
year | +0.002 |
saw | +0.002 |
Neg logit contributions | |
hyd | -0.003 |
balls | -0.003 |
heal | -0.003 |
rid | -0.003 |
bleed | -0.003 |
Token | a |
Token ID | 257 |
Feature activation | +1.03e-02 |
Pos logit contributions | |
ited | +0.049 |
saw | +0.042 |
heard | +0.039 |
Hey | +0.033 |
year | +0.033 |
Neg logit contributions | |
anks | -0.111 |
balls | -0.108 |
hyd | -0.106 |
rid | -0.105 |
heal | -0.104 |
Token | lesson |
Token ID | 11483 |
Feature activation | +2.23e-02 |
Pos logit contributions | |
little | +0.009 |
saw | +0.009 |
heard | +0.008 |
ited | +0.008 |
year | +0.008 |
Neg logit contributions | |
anks | -0.026 |
hyd | -0.025 |
balls | -0.025 |
uggets | -0.024 |
aro | -0.024 |
Token | : |
Token ID | 25 |
Feature activation | -2.16e-02 |
Pos logit contributions | |
ited | +0.042 |
saw | +0.031 |
Hey | +0.031 |
Quick | +0.029 |
Uh | +0.029 |
Neg logit contributions | |
anks | -0.039 |
balls | -0.037 |
heal | -0.035 |
ometers | -0.035 |
hyd | -0.035 |
Token | never |
Token ID | 1239 |
Feature activation | -3.46e-03 |
Pos logit contributions | |
anks | +0.044 |
balls | +0.042 |
hyd | +0.041 |
rid | +0.041 |
bleed | +0.040 |
Neg logit contributions | |
ited | -0.029 |
saw | -0.028 |
heard | -0.026 |
two | -0.024 |
little | -0.024 |
Token | go |
Token ID | 467 |
Feature activation | -4.24e-03 |
Pos logit contributions | |
anks | +0.006 |
balls | +0.006 |
hyd | +0.006 |
rid | +0.006 |
heal | +0.005 |
Neg logit contributions | |
ited | -0.006 |
saw | -0.005 |
heard | -0.005 |
year | -0.005 |
Hey | -0.005 |
Token | on |
Token ID | 319 |
Feature activation | -3.96e-03 |
Pos logit contributions | |
anks | +0.008 |
balls | +0.008 |
hyd | +0.008 |
rid | +0.007 |
heal | +0.007 |
Neg logit contributions | |
ited | -0.007 |
saw | -0.006 |
heard | -0.006 |
Hey | -0.005 |
year | -0.005 |
Token | an |
Token ID | 281 |
Feature activation | +1.22e-02 |
Pos logit contributions | |
anks | +0.009 |
balls | +0.008 |
hyd | +0.008 |
heal | +0.008 |
rid | +0.008 |
Neg logit contributions | |
ited | -0.005 |
saw | -0.004 |
heard | -0.004 |
Hey | -0.004 |
year | -0.003 |
Token | Thank |
Token ID | 6952 |
Feature activation | +6.87e-04 |
Pos logit contributions | |
anks | +0.109 |
rid | +0.103 |
balls | +0.100 |
bleed | +0.100 |
aro | +0.097 |
Neg logit contributions | |
Hi | -0.073 |
saw | -0.072 |
Hey | -0.070 |
two | -0.070 |
did | -0.068 |
Token | you |
Token ID | 345 |
Feature activation | +3.90e-03 |
Pos logit contributions | |
ited | +0.001 |
saw | +0.001 |
two | +0.001 |
heard | +0.001 |
Hey | +0.001 |
Neg logit contributions | |
anks | -0.001 |
sing | -0.001 |
ometer | -0.001 |
hyd | -0.001 |
ometers | -0.001 |
Token | for |
Token ID | 329 |
Feature activation | -6.26e-03 |
Pos logit contributions | |
ited | +0.004 |
saw | +0.004 |
heard | +0.004 |
two | +0.003 |
little | +0.003 |
Neg logit contributions | |
anks | -0.009 |
hyd | -0.009 |
bleed | -0.009 |
rid | -0.009 |
balls | -0.009 |
Token | forgiving |
Token ID | 43486 |
Feature activation | -3.77e-03 |
Pos logit contributions | |
anks | +0.011 |
balls | +0.010 |
hyd | +0.010 |
rid | +0.010 |
heal | +0.010 |
Neg logit contributions | |
ited | -0.011 |
saw | -0.010 |
heard | -0.010 |
two | -0.009 |
little | -0.009 |
Token | me |
Token ID | 502 |
Feature activation | -2.54e-02 |
Pos logit contributions | |
anks | +0.009 |
balls | +0.008 |
hyd | +0.008 |
rid | +0.008 |
heal | +0.008 |
Neg logit contributions | |
ited | -0.005 |
saw | -0.004 |
heard | -0.003 |
year | -0.003 |
noticed | -0.003 |
Token | ." |
Token ID | 526 |
Feature activation | -3.10e-02 |
Pos logit contributions | |
anks | +0.046 |
balls | +0.044 |
bleed | +0.043 |
hyd | +0.042 |
rid | +0.042 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.036 |
heard | -0.036 |
two | -0.035 |
little | -0.034 |
Token |
|
Token ID | 198 |
Feature activation | +4.30e-03 |
Pos logit contributions | |
anks | +0.073 |
hyd | +0.073 |
rid | +0.072 |
affect | +0.072 |
bleed | +0.071 |
Neg logit contributions | |
saw | -0.031 |
two | -0.026 |
heard | -0.026 |
little | -0.025 |
ited | -0.024 |
Token |
|
Token ID | 198 |
Feature activation | -1.01e-02 |
Pos logit contributions | |
ited | +0.005 |
saw | +0.004 |
heard | +0.004 |
year | +0.004 |
two | +0.004 |
Neg logit contributions | |
anks | -0.010 |
balls | -0.010 |
hyd | -0.009 |
rid | -0.009 |
heal | -0.009 |
Token | The |
Token ID | 464 |
Feature activation | +3.47e-02 |
Pos logit contributions | |
anks | +0.021 |
balls | +0.021 |
hyd | +0.020 |
rid | +0.020 |
heal | +0.020 |
Neg logit contributions | |
ited | -0.013 |
saw | -0.012 |
heard | -0.011 |
year | -0.010 |
Hey | -0.010 |
Token | moral |
Token ID | 6573 |
Feature activation | +4.13e-02 |
Pos logit contributions | |
ited | +0.042 |
saw | +0.035 |
heard | +0.033 |
year | +0.028 |
Hey | +0.028 |
Neg logit contributions | |
anks | -0.078 |
balls | -0.078 |
hyd | -0.076 |
rid | -0.074 |
heal | -0.074 |
Token | of |
Token ID | 286 |
Feature activation | +4.38e-02 |
Pos logit contributions | |
ited | +0.059 |
saw | +0.043 |
heard | +0.040 |
noticed | +0.039 |
spotted | +0.034 |
Neg logit contributions | |
anks | -0.087 |
heal | -0.085 |
hyd | -0.084 |
rid | -0.084 |
balls | -0.083 |
Token |
|
Token ID | 198 |
Feature activation | -2.22e-02 |
Pos logit contributions | |
anks | +0.034 |
balls | +0.033 |
hyd | +0.032 |
rid | +0.032 |
heal | +0.032 |
Neg logit contributions | |
ited | -0.019 |
saw | -0.017 |
heard | -0.016 |
Hey | -0.013 |
year | -0.013 |
Token | L |
Token ID | 43 |
Feature activation | +1.38e-02 |
Pos logit contributions | |
balls | +0.047 |
anks | +0.047 |
hyd | +0.046 |
rid | +0.045 |
heal | +0.045 |
Neg logit contributions | |
ited | -0.028 |
saw | -0.024 |
heard | -0.023 |
Hey | -0.020 |
two | -0.019 |
Token | ily |
Token ID | 813 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
ited | +0.025 |
saw | +0.024 |
heard | +0.022 |
noticed | +0.021 |
arrived | +0.021 |
Neg logit contributions | |
anks | -0.021 |
ectar | -0.020 |
ophone | -0.020 |
uggets | -0.019 |
balls | -0.019 |
Token | learned |
Token ID | 4499 |
Feature activation | +4.41e-02 |
Pos logit contributions | |
ited | +0.040 |
grown | +0.033 |
Hey | +0.032 |
year | +0.030 |
time | +0.030 |
Neg logit contributions | |
hyd | -0.059 |
heal | -0.058 |
balls | -0.057 |
fet | -0.055 |
anks | -0.055 |
Token | that |
Token ID | 326 |
Feature activation | +2.66e-02 |
Pos logit contributions | |
ited | +0.068 |
saw | +0.055 |
heard | +0.050 |
Hey | +0.046 |
noticed | +0.044 |
Neg logit contributions | |
anks | -0.090 |
heal | -0.089 |
balls | -0.088 |
rid | -0.086 |
hyd | -0.085 |
Token | it |
Token ID | 340 |
Feature activation | -2.63e-02 |
Pos logit contributions | |
ited | +0.031 |
saw | +0.030 |
heard | +0.028 |
two | +0.026 |
year | +0.025 |
Neg logit contributions | |
anks | -0.059 |
balls | -0.057 |
hyd | -0.055 |
rid | -0.055 |
ophone | -0.054 |
Token | 's |
Token ID | 338 |
Feature activation | -3.32e-03 |
Pos logit contributions | |
anks | +0.058 |
balls | +0.056 |
hyd | +0.056 |
uggets | +0.055 |
sing | +0.054 |
Neg logit contributions | |
saw | -0.030 |
ited | -0.030 |
heard | -0.029 |
two | -0.026 |
did | -0.026 |
Token | important |
Token ID | 1593 |
Feature activation | +2.06e-02 |
Pos logit contributions | |
anks | +0.005 |
hyd | +0.005 |
balls | +0.005 |
heal | +0.005 |
rid | +0.005 |
Neg logit contributions | |
ited | -0.006 |
saw | -0.006 |
heard | -0.005 |
Hey | -0.005 |
year | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | -1.46e-02 |
Pos logit contributions | |
ited | +0.034 |
saw | +0.029 |
heard | +0.028 |
Hey | +0.026 |
year | +0.025 |
Neg logit contributions | |
anks | -0.038 |
balls | -0.038 |
hyd | -0.036 |
rid | -0.036 |
heal | -0.036 |
Token | take |
Token ID | 1011 |
Feature activation | -1.82e-03 |
Pos logit contributions | |
anks | +0.023 |
hyd | +0.022 |
heal | +0.022 |
balls | +0.022 |
rid | +0.022 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.024 |
heard | -0.024 |
Hey | -0.022 |
year | -0.022 |
Token | care |
Token ID | 1337 |
Feature activation | -2.09e-04 |
Pos logit contributions | |
anks | +0.003 |
rid | +0.003 |
hyd | +0.003 |
balls | +0.003 |
heal | +0.003 |
Neg logit contributions | |
ited | -0.003 |
saw | -0.003 |
heard | -0.003 |
Hey | -0.003 |
year | -0.003 |
Token | sat |
Token ID | 3332 |
Feature activation | +3.96e-02 |
Pos logit contributions | |
hyd | +0.017 |
balls | +0.017 |
rid | +0.016 |
anks | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.011 |
Hey | -0.008 |
saw | -0.008 |
year | -0.008 |
heard | -0.008 |
Token | down |
Token ID | 866 |
Feature activation | +2.45e-02 |
Pos logit contributions | |
ited | +0.087 |
heard | +0.073 |
saw | +0.073 |
Hey | +0.073 |
year | +0.070 |
Neg logit contributions | |
hyd | -0.054 |
balls | -0.052 |
anks | -0.051 |
rid | -0.048 |
heal | -0.045 |
Token | to |
Token ID | 284 |
Feature activation | -3.29e-03 |
Pos logit contributions | |
ited | +0.042 |
saw | +0.035 |
heard | +0.034 |
Hey | +0.032 |
year | +0.031 |
Neg logit contributions | |
anks | -0.044 |
hyd | -0.043 |
balls | -0.043 |
rid | -0.042 |
heal | -0.040 |
Token | have |
Token ID | 423 |
Feature activation | -5.27e-03 |
Pos logit contributions | |
anks | +0.005 |
balls | +0.005 |
hyd | +0.005 |
rid | +0.005 |
heal | +0.005 |
Neg logit contributions | |
ited | -0.006 |
saw | -0.005 |
heard | -0.005 |
year | -0.005 |
Hey | -0.005 |
Token | their |
Token ID | 511 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
anks | +0.010 |
hyd | +0.010 |
balls | +0.010 |
rid | +0.010 |
heal | +0.010 |
Neg logit contributions | |
ited | -0.008 |
saw | -0.006 |
heard | -0.006 |
Hey | -0.006 |
year | -0.005 |
Token | picnic |
Token ID | 35715 |
Feature activation | -8.95e-03 |
Pos logit contributions | |
anks | +0.031 |
balls | +0.029 |
hyd | +0.029 |
rid | +0.028 |
heal | +0.028 |
Neg logit contributions | |
ited | -0.032 |
saw | -0.030 |
heard | -0.028 |
year | -0.026 |
two | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | -3.35e-02 |
Pos logit contributions | |
anks | +0.017 |
balls | +0.016 |
hyd | +0.016 |
rid | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.014 |
saw | -0.013 |
heard | -0.012 |
two | -0.011 |
year | -0.011 |
Token | After |
Token ID | 2293 |
Feature activation | +1.58e-02 |
Pos logit contributions | |
anks | +0.067 |
balls | +0.065 |
hyd | +0.064 |
rid | +0.063 |
heal | +0.062 |
Neg logit contributions | |
ited | -0.048 |
saw | -0.044 |
heard | -0.041 |
two | -0.037 |
little | -0.037 |
Token | they |
Token ID | 484 |
Feature activation | -2.73e-03 |
Pos logit contributions | |
ited | +0.030 |
Hey | +0.022 |
saw | +0.022 |
heard | +0.021 |
Fine | +0.020 |
Neg logit contributions | |
balls | -0.029 |
hyd | -0.026 |
anks | -0.026 |
aro | -0.025 |
heal | -0.025 |
Token | finished |
Token ID | 5201 |
Feature activation | +1.83e-02 |
Pos logit contributions | |
hyd | +0.005 |
rid | +0.005 |
cr | +0.005 |
aro | +0.005 |
heal | +0.005 |
Neg logit contributions | |
ited | -0.003 |
Fine | -0.003 |
Hey | -0.003 |
ised | -0.003 |
Next | -0.003 |
Token | , |
Token ID | 11 |
Feature activation | +8.66e-03 |
Pos logit contributions | |
ited | +0.029 |
saw | +0.024 |
heard | +0.023 |
Hey | +0.022 |
year | +0.021 |
Neg logit contributions | |
anks | -0.035 |
balls | -0.035 |
hyd | -0.034 |
heal | -0.033 |
rid | -0.032 |
Token | in |
Token ID | 287 |
Feature activation | -6.04e-03 |
Pos logit contributions | |
anks | +0.003 |
balls | +0.003 |
hyd | +0.003 |
rid | +0.003 |
heal | +0.003 |
Neg logit contributions | |
ited | -0.002 |
saw | -0.002 |
heard | -0.001 |
two | -0.001 |
year | -0.001 |
Token | a |
Token ID | 257 |
Feature activation | +9.56e-03 |
Pos logit contributions | |
balls | +0.012 |
hyd | +0.012 |
anks | +0.012 |
rid | +0.012 |
heal | +0.012 |
Neg logit contributions | |
ited | -0.008 |
saw | -0.006 |
heard | -0.006 |
Hey | -0.006 |
oph | -0.005 |
Token | big |
Token ID | 1263 |
Feature activation | +4.56e-04 |
Pos logit contributions | |
ited | +0.015 |
Hey | +0.011 |
saw | +0.011 |
heard | +0.011 |
Fine | +0.011 |
Neg logit contributions | |
hyd | -0.017 |
heal | -0.017 |
balls | -0.016 |
rid | -0.016 |
sing | -0.016 |
Token | , |
Token ID | 11 |
Feature activation | -2.14e-02 |
Pos logit contributions | |
ited | +0.001 |
Next | +0.001 |
When | +0.001 |
Fine | +0.001 |
Hey | +0.001 |
Neg logit contributions | |
hyd | -0.001 |
heal | -0.001 |
sing | -0.001 |
fly | -0.001 |
float | -0.001 |
Token | dark |
Token ID | 3223 |
Feature activation | -9.62e-04 |
Pos logit contributions | |
hyd | +0.042 |
balls | +0.041 |
anks | +0.041 |
heal | +0.040 |
rid | +0.040 |
Neg logit contributions | |
ited | -0.031 |
saw | -0.026 |
heard | -0.025 |
year | -0.022 |
Hey | -0.022 |
Token | house |
Token ID | 2156 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
hyd | +0.002 |
fly | +0.002 |
anks | +0.002 |
heal | +0.002 |
balls | +0.002 |
Neg logit contributions | |
ited | -0.002 |
saw | -0.001 |
heard | -0.001 |
Hey | -0.001 |
did | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -2.82e-02 |
Pos logit contributions | |
hyd | +0.032 |
balls | +0.032 |
anks | +0.032 |
heal | +0.030 |
flies | +0.029 |
Neg logit contributions | |
ited | -0.041 |
Hey | -0.033 |
saw | -0.032 |
heard | -0.031 |
oph | -0.031 |
Token | there |
Token ID | 612 |
Feature activation | +5.54e-03 |
Pos logit contributions | |
balls | +0.060 |
hyd | +0.060 |
anks | +0.060 |
heal | +0.057 |
rid | +0.057 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.029 |
heard | -0.028 |
Hey | -0.026 |
year | -0.026 |
Token | lived |
Token ID | 5615 |
Feature activation | +1.87e-02 |
Pos logit contributions | |
ited | +0.007 |
Hey | +0.004 |
conflict | +0.004 |
saw | +0.004 |
Exc | +0.004 |
Neg logit contributions | |
balls | -0.012 |
anks | -0.011 |
heal | -0.011 |
hyd | -0.011 |
flies | -0.011 |
Token | a |
Token ID | 257 |
Feature activation | +1.56e-02 |
Pos logit contributions | |
ited | +0.025 |
saw | +0.018 |
heard | +0.017 |
conflict | +0.015 |
Hey | +0.015 |
Neg logit contributions | |
balls | -0.040 |
anks | -0.039 |
hyd | -0.039 |
rid | -0.039 |
heal | -0.038 |
Token | stubborn |
Token ID | 24444 |
Feature activation | +1.89e-02 |
Pos logit contributions | |
ited | +0.017 |
saw | +0.013 |
heard | +0.013 |
Hey | +0.011 |
two | +0.010 |
Neg logit contributions | |
balls | -0.036 |
hyd | -0.036 |
anks | -0.036 |
heal | -0.035 |
rid | -0.035 |
Token | noticed |
Token ID | 6810 |
Feature activation | +5.62e-02 |
Pos logit contributions | |
balls | +0.014 |
rid | +0.013 |
cr | +0.013 |
hyd | +0.013 |
bleed | +0.013 |
Neg logit contributions | |
ited | -0.006 |
year | -0.004 |
Fine | -0.004 |
Next | -0.004 |
Hey | -0.004 |
Token | the |
Token ID | 262 |
Feature activation | +8.16e-03 |
Pos logit contributions | |
ited | +0.095 |
saw | +0.063 |
year | +0.063 |
heard | +0.059 |
Hey | +0.057 |
Neg logit contributions | |
balls | -0.106 |
rid | -0.104 |
hyd | -0.103 |
heal | -0.100 |
aro | -0.099 |
Token | fan |
Token ID | 4336 |
Feature activation | -5.35e-02 |
Pos logit contributions | |
ited | +0.017 |
saw | +0.015 |
heard | +0.015 |
Hey | +0.014 |
noticed | +0.013 |
Neg logit contributions | |
balls | -0.011 |
anks | -0.011 |
hyd | -0.011 |
rid | -0.011 |
heal | -0.011 |
Token | was |
Token ID | 373 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
anks | +0.108 |
balls | +0.106 |
hyd | +0.105 |
heal | +0.102 |
rid | +0.102 |
Neg logit contributions | |
ited | -0.075 |
saw | -0.065 |
heard | -0.061 |
Hey | -0.057 |
year | -0.055 |
Token | dirty |
Token ID | 11841 |
Feature activation | -4.96e-03 |
Pos logit contributions | |
hyd | +0.021 |
balls | +0.021 |
anks | +0.021 |
rid | +0.021 |
heal | +0.020 |
Neg logit contributions | |
ited | -0.022 |
saw | -0.019 |
heard | -0.018 |
Hey | -0.018 |
year | -0.017 |
Token | with |
Token ID | 351 |
Feature activation | -2.84e-02 |
Pos logit contributions | |
anks | +0.010 |
hyd | +0.010 |
balls | +0.010 |
heal | +0.009 |
rid | +0.009 |
Neg logit contributions | |
ited | -0.007 |
saw | -0.006 |
heard | -0.006 |
Hey | -0.005 |
year | -0.005 |
Token | dust |
Token ID | 8977 |
Feature activation | -2.74e-02 |
Pos logit contributions | |
rid | +0.052 |
hyd | +0.051 |
balls | +0.051 |
anks | +0.050 |
heal | +0.050 |
Neg logit contributions | |
ited | -0.046 |
saw | -0.038 |
heard | -0.037 |
Hey | -0.036 |
year | -0.035 |
Token | and |
Token ID | 290 |
Feature activation | -6.95e-02 |
Pos logit contributions | |
anks | +0.051 |
balls | +0.050 |
hyd | +0.049 |
rid | +0.048 |
heal | +0.048 |
Neg logit contributions | |
ited | -0.043 |
saw | -0.038 |
heard | -0.036 |
Hey | -0.033 |
year | -0.033 |
Token | leaves |
Token ID | 5667 |
Feature activation | -3.88e-02 |
Pos logit contributions | |
balls | +0.126 |
anks | +0.124 |
hyd | +0.123 |
rid | +0.122 |
heal | +0.119 |
Neg logit contributions | |
ited | -0.115 |
saw | -0.095 |
heard | -0.091 |
Hey | -0.090 |
year | -0.088 |
Token | . |
Token ID | 13 |
Feature activation | -4.82e-02 |
Pos logit contributions | |
anks | +0.068 |
balls | +0.067 |
hyd | +0.065 |
rid | +0.064 |
heal | +0.064 |
Neg logit contributions | |
ited | -0.065 |
saw | -0.058 |
heard | -0.055 |
Hey | -0.052 |
year | -0.051 |
Token | Lily |
Token ID | 20037 |
Feature activation | +1.98e-02 |
Pos logit contributions | |
anks | +0.089 |
balls | +0.086 |
hyd | +0.084 |
rid | +0.083 |
heal | +0.082 |
Neg logit contributions | |
ited | -0.075 |
saw | -0.071 |
heard | -0.067 |
little | -0.062 |
two | -0.061 |
Token | isy |
Token ID | 13560 |
Feature activation | -1.86e-02 |
Pos logit contributions | |
anks | +0.017 |
balls | +0.017 |
hyd | +0.016 |
rid | +0.016 |
heal | +0.016 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.019 |
heard | -0.018 |
year | -0.017 |
two | -0.016 |
Token | always |
Token ID | 1464 |
Feature activation | -1.09e-03 |
Pos logit contributions | |
hyd | +0.041 |
balls | +0.040 |
anks | +0.040 |
heal | +0.039 |
rid | +0.039 |
Neg logit contributions | |
ited | -0.021 |
Hey | -0.016 |
saw | -0.014 |
heard | -0.014 |
two | -0.014 |
Token | wanted |
Token ID | 2227 |
Feature activation | +4.02e-02 |
Pos logit contributions | |
hyd | +0.002 |
anks | +0.002 |
balls | +0.002 |
rid | +0.002 |
heal | +0.002 |
Neg logit contributions | |
ited | -0.001 |
saw | -0.001 |
heard | -0.001 |
Hey | -0.001 |
year | -0.001 |
Token | more |
Token ID | 517 |
Feature activation | -9.86e-03 |
Pos logit contributions | |
ited | +0.059 |
saw | +0.052 |
heard | +0.049 |
Hey | +0.045 |
year | +0.044 |
Neg logit contributions | |
anks | -0.080 |
balls | -0.078 |
hyd | -0.076 |
rid | -0.075 |
heal | -0.074 |
Token | sun |
Token ID | 4252 |
Feature activation | -4.83e-02 |
Pos logit contributions | |
anks | +0.021 |
balls | +0.021 |
hyd | +0.020 |
rid | +0.020 |
heal | +0.020 |
Neg logit contributions | |
ited | -0.013 |
saw | -0.011 |
heard | -0.011 |
year | -0.010 |
little | -0.010 |
Token | and |
Token ID | 290 |
Feature activation | -5.76e-02 |
Pos logit contributions | |
drown | +0.097 |
anks | +0.096 |
balls | +0.096 |
hyd | +0.095 |
rid | +0.095 |
Neg logit contributions | |
ited | -0.058 |
saw | -0.058 |
heard | -0.056 |
two | -0.051 |
enough | -0.050 |
Token | more |
Token ID | 517 |
Feature activation | -1.35e-02 |
Pos logit contributions | |
anks | +0.107 |
balls | +0.106 |
hyd | +0.105 |
heal | +0.105 |
rid | +0.105 |
Neg logit contributions | |
ited | -0.091 |
saw | -0.074 |
heard | -0.071 |
Hey | -0.070 |
year | -0.067 |
Token | rain |
Token ID | 6290 |
Feature activation | -4.17e-02 |
Pos logit contributions | |
anks | +0.027 |
balls | +0.027 |
hyd | +0.026 |
rid | +0.025 |
heal | +0.025 |
Neg logit contributions | |
ited | -0.019 |
saw | -0.018 |
heard | -0.017 |
little | -0.016 |
year | -0.016 |
Token | , |
Token ID | 11 |
Feature activation | -2.93e-02 |
Pos logit contributions | |
anks | +0.077 |
hyd | +0.075 |
balls | +0.074 |
rid | +0.072 |
uggets | +0.072 |
Neg logit contributions | |
saw | -0.061 |
ited | -0.061 |
heard | -0.058 |
did | -0.053 |
little | -0.050 |
Token | even |
Token ID | 772 |
Feature activation | -1.18e-02 |
Pos logit contributions | |
anks | +0.066 |
balls | +0.064 |
hyd | +0.063 |
rid | +0.062 |
heal | +0.061 |
Neg logit contributions | |
ited | -0.035 |
saw | -0.031 |
heard | -0.029 |
two | -0.025 |
year | -0.025 |
Token | when |
Token ID | 618 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
anks | +0.025 |
rid | +0.024 |
balls | +0.024 |
hyd | +0.024 |
heal | +0.024 |
Neg logit contributions | |
ited | -0.016 |
saw | -0.013 |
heard | -0.012 |
Hey | -0.011 |
year | -0.011 |
Token | that |
Token ID | 326 |
Feature activation | -2.38e-02 |
Pos logit contributions | |
anks | +0.073 |
balls | +0.070 |
hyd | +0.069 |
uggets | +0.068 |
rid | +0.066 |
Neg logit contributions | |
ited | -0.044 |
saw | -0.044 |
heard | -0.041 |
did | -0.038 |
little | -0.035 |
Token | can |
Token ID | 460 |
Feature activation | -4.08e-02 |
Pos logit contributions | |
anks | +0.050 |
balls | +0.048 |
uggets | +0.047 |
sing | +0.045 |
hyd | +0.045 |
Neg logit contributions | |
saw | -0.033 |
heard | -0.031 |
ited | -0.030 |
did | -0.029 |
two | -0.028 |
Token | 't |
Token ID | 470 |
Feature activation | -3.29e-02 |
Pos logit contributions | |
anks | +0.080 |
uggets | +0.079 |
ectar | +0.075 |
hyd | +0.075 |
balls | +0.074 |
Neg logit contributions | |
saw | -0.061 |
heard | -0.056 |
ited | -0.054 |
two | -0.054 |
little | -0.051 |
Token | be |
Token ID | 307 |
Feature activation | -4.49e-02 |
Pos logit contributions | |
anks | +0.067 |
balls | +0.062 |
uggets | +0.062 |
ometer | +0.060 |
ectar | +0.060 |
Neg logit contributions | |
saw | -0.050 |
heard | -0.047 |
ited | -0.044 |
little | -0.043 |
two | -0.042 |
Token | changed |
Token ID | 3421 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
anks | +0.091 |
balls | +0.089 |
hyd | +0.088 |
rid | +0.086 |
heal | +0.085 |
Neg logit contributions | |
ited | -0.063 |
saw | -0.056 |
heard | -0.053 |
Hey | -0.047 |
year | -0.047 |
Token | or |
Token ID | 393 |
Feature activation | -8.01e-02 |
Pos logit contributions | |
anks | +0.031 |
balls | +0.030 |
hyd | +0.030 |
rid | +0.029 |
heal | +0.029 |
Neg logit contributions | |
ited | -0.025 |
saw | -0.022 |
heard | -0.021 |
Hey | -0.019 |
year | -0.019 |
Token | fixed |
Token ID | 5969 |
Feature activation | +4.93e-03 |
Pos logit contributions | |
anks | +0.146 |
balls | +0.139 |
hyd | +0.136 |
rid | +0.131 |
heal | +0.128 |
Neg logit contributions | |
ited | -0.125 |
saw | -0.121 |
heard | -0.116 |
two | -0.105 |
little | -0.104 |
Token | . |
Token ID | 13 |
Feature activation | -3.60e-02 |
Pos logit contributions | |
ited | +0.007 |
saw | +0.007 |
heard | +0.006 |
two | +0.006 |
little | +0.006 |
Neg logit contributions | |
anks | -0.010 |
balls | -0.009 |
hyd | -0.009 |
rid | -0.009 |
heal | -0.009 |
Token | He |
Token ID | 679 |
Feature activation | -8.11e-03 |
Pos logit contributions | |
balls | +0.076 |
bleed | +0.073 |
hyd | +0.073 |
anks | +0.072 |
rid | +0.072 |
Neg logit contributions | |
ited | -0.041 |
saw | -0.041 |
two | -0.040 |
little | -0.039 |
heard | -0.038 |
Token | took |
Token ID | 1718 |
Feature activation | +2.85e-02 |
Pos logit contributions | |
anks | +0.018 |
hyd | +0.018 |
balls | +0.018 |
heal | +0.017 |
rid | +0.017 |
Neg logit contributions | |
ited | -0.009 |
saw | -0.007 |
heard | -0.007 |
Hey | -0.007 |
year | -0.007 |
Token | this |
Token ID | 428 |
Feature activation | +1.23e-02 |
Pos logit contributions | |
ited | +0.034 |
saw | +0.029 |
heard | +0.027 |
year | +0.023 |
Hey | +0.023 |
Neg logit contributions | |
anks | -0.064 |
balls | -0.063 |
rid | -0.062 |
hyd | -0.062 |
heal | -0.061 |
Token | could |
Token ID | 714 |
Feature activation | -2.48e-02 |
Pos logit contributions | |
anks | +0.054 |
balls | +0.053 |
hyd | +0.052 |
rid | +0.052 |
heal | +0.051 |
Neg logit contributions | |
ited | -0.020 |
saw | -0.015 |
heard | -0.014 |
Hey | -0.013 |
year | -0.012 |
Token | return |
Token ID | 1441 |
Feature activation | +9.71e-04 |
Pos logit contributions | |
anks | +0.052 |
hyd | +0.049 |
balls | +0.049 |
uggets | +0.048 |
ectar | +0.048 |
Neg logit contributions | |
saw | -0.035 |
heard | -0.032 |
little | -0.030 |
ited | -0.030 |
did | -0.030 |
Token | to |
Token ID | 284 |
Feature activation | -2.53e-02 |
Pos logit contributions | |
ited | +0.001 |
saw | +0.001 |
heard | +0.001 |
two | +0.001 |
year | +0.001 |
Neg logit contributions | |
anks | -0.002 |
balls | -0.002 |
hyd | -0.002 |
rid | -0.002 |
heal | -0.002 |
Token | the |
Token ID | 262 |
Feature activation | -1.22e-02 |
Pos logit contributions | |
anks | +0.060 |
balls | +0.058 |
hyd | +0.058 |
rid | +0.057 |
heal | +0.057 |
Neg logit contributions | |
ited | -0.027 |
saw | -0.023 |
heard | -0.022 |
Hey | -0.018 |
year | -0.018 |
Token | zoo |
Token ID | 26626 |
Feature activation | -2.50e-02 |
Pos logit contributions | |
balls | +0.023 |
anks | +0.023 |
hyd | +0.023 |
rid | +0.022 |
heal | +0.022 |
Neg logit contributions | |
ited | -0.019 |
saw | -0.015 |
heard | -0.015 |
Hey | -0.013 |
year | -0.013 |
Token | and |
Token ID | 290 |
Feature activation | -6.84e-02 |
Pos logit contributions | |
anks | +0.046 |
balls | +0.045 |
hyd | +0.044 |
rid | +0.043 |
heal | +0.043 |
Neg logit contributions | |
ited | -0.040 |
saw | -0.035 |
heard | -0.034 |
Hey | -0.031 |
two | -0.030 |
Token | see |
Token ID | 766 |
Feature activation | +1.30e-02 |
Pos logit contributions | |
anks | +0.123 |
balls | +0.119 |
hyd | +0.117 |
rid | +0.113 |
heal | +0.112 |
Neg logit contributions | |
ited | -0.110 |
saw | -0.104 |
heard | -0.099 |
two | -0.089 |
little | -0.089 |
Token | more |
Token ID | 517 |
Feature activation | -3.28e-02 |
Pos logit contributions | |
ited | +0.018 |
saw | +0.015 |
heard | +0.014 |
Hey | +0.013 |
year | +0.013 |
Neg logit contributions | |
anks | -0.027 |
balls | -0.026 |
heal | -0.026 |
rid | -0.026 |
hyd | -0.026 |
Token | animals |
Token ID | 4695 |
Feature activation | -5.67e-02 |
Pos logit contributions | |
anks | +0.065 |
balls | +0.064 |
hyd | +0.063 |
rid | +0.061 |
heal | +0.061 |
Neg logit contributions | |
ited | -0.047 |
saw | -0.042 |
heard | -0.040 |
Hey | -0.036 |
year | -0.036 |
Token | , |
Token ID | 11 |
Feature activation | -4.92e-02 |
Pos logit contributions | |
anks | +0.113 |
balls | +0.109 |
hyd | +0.107 |
rid | +0.105 |
bleed | +0.103 |
Neg logit contributions | |
ited | -0.080 |
saw | -0.076 |
heard | -0.072 |
two | -0.063 |
year | -0.063 |
Token | because |
Token ID | 780 |
Feature activation | -2.67e-02 |
Pos logit contributions | |
balls | +0.089 |
anks | +0.089 |
hyd | +0.088 |
heal | +0.086 |
rid | +0.085 |
Neg logit contributions | |
ited | -0.080 |
saw | -0.065 |
heard | -0.062 |
Hey | -0.061 |
year | -0.059 |
Token | fire |
Token ID | 2046 |
Feature activation | -4.59e-02 |
Pos logit contributions | |
anks | +0.018 |
balls | +0.017 |
hyd | +0.017 |
rid | +0.017 |
heal | +0.017 |
Neg logit contributions | |
ited | -0.015 |
saw | -0.013 |
heard | -0.013 |
Hey | -0.011 |
year | -0.011 |
Token | and |
Token ID | 290 |
Feature activation | -4.82e-02 |
Pos logit contributions | |
anks | +0.099 |
balls | +0.096 |
hyd | +0.094 |
rid | +0.094 |
heal | +0.092 |
Neg logit contributions | |
ited | -0.057 |
saw | -0.052 |
heard | -0.049 |
two | -0.043 |
year | -0.043 |
Token | she |
Token ID | 673 |
Feature activation | -2.33e-02 |
Pos logit contributions | |
hyd | +0.093 |
balls | +0.091 |
heal | +0.091 |
rid | +0.090 |
anks | +0.089 |
Neg logit contributions | |
ited | -0.078 |
Hey | -0.058 |
year | -0.056 |
heard | -0.054 |
saw | -0.053 |
Token | quickly |
Token ID | 2952 |
Feature activation | -2.36e-03 |
Pos logit contributions | |
anks | +0.056 |
balls | +0.055 |
hyd | +0.055 |
heal | +0.054 |
rid | +0.054 |
Neg logit contributions | |
ited | -0.024 |
saw | -0.017 |
heard | -0.016 |
year | -0.015 |
Hey | -0.015 |
Token | put |
Token ID | 1234 |
Feature activation | -2.64e-03 |
Pos logit contributions | |
hyd | +0.005 |
anks | +0.005 |
heal | +0.005 |
bleed | +0.004 |
rid | +0.004 |
Neg logit contributions | |
ited | -0.003 |
Hey | -0.003 |
year | -0.003 |
End | -0.002 |
Next | -0.002 |
Token | it |
Token ID | 340 |
Feature activation | -5.87e-02 |
Pos logit contributions | |
anks | +0.006 |
balls | +0.006 |
hyd | +0.006 |
heal | +0.006 |
sing | +0.006 |
Neg logit contributions | |
ited | -0.003 |
saw | -0.003 |
heard | -0.003 |
two | -0.002 |
little | -0.002 |
Token | out |
Token ID | 503 |
Feature activation | -2.18e-02 |
Pos logit contributions | |
anks | +0.119 |
balls | +0.113 |
hyd | +0.111 |
rid | +0.109 |
heal | +0.108 |
Neg logit contributions | |
ited | -0.082 |
saw | -0.077 |
heard | -0.073 |
two | -0.065 |
little | -0.064 |
Token | . |
Token ID | 13 |
Feature activation | -4.33e-02 |
Pos logit contributions | |
anks | +0.037 |
rid | +0.037 |
balls | +0.037 |
hyd | +0.037 |
heal | +0.036 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.032 |
heard | -0.030 |
Hey | -0.030 |
year | -0.028 |
Token | Tim |
Token ID | 5045 |
Feature activation | +2.07e-02 |
Pos logit contributions | |
anks | +0.087 |
hyd | +0.085 |
rid | +0.084 |
balls | +0.084 |
heal | +0.083 |
Neg logit contributions | |
saw | -0.056 |
ited | -0.055 |
heard | -0.050 |
two | -0.048 |
little | -0.048 |
Token | my |
Token ID | 1820 |
Feature activation | +1.55e-02 |
Pos logit contributions | |
ited | +0.036 |
Hey | +0.034 |
Next | +0.031 |
minutes | +0.031 |
ised | +0.031 |
Neg logit contributions | |
gy | -0.029 |
ophone | -0.028 |
anks | -0.028 |
fly | -0.027 |
hyd | -0.027 |
Token | felt |
Token ID | 2936 |
Feature activation | +2.20e-02 |
Pos logit contributions | |
ited | +0.020 |
Hey | +0.015 |
grown | +0.014 |
year | +0.013 |
Hi | +0.013 |
Neg logit contributions | |
hyd | -0.032 |
balls | -0.032 |
heal | -0.031 |
anks | -0.030 |
rid | -0.030 |
Token | . |
Token ID | 13 |
Feature activation | -4.96e-02 |
Pos logit contributions | |
ited | +0.019 |
saw | +0.017 |
heard | +0.017 |
two | +0.015 |
year | +0.015 |
Neg logit contributions | |
anks | -0.022 |
balls | -0.021 |
hyd | -0.021 |
rid | -0.021 |
heal | -0.020 |
Token | They |
Token ID | 1119 |
Feature activation | -2.05e-02 |
Pos logit contributions | |
anks | +0.082 |
balls | +0.079 |
hyd | +0.077 |
rid | +0.076 |
heal | +0.075 |
Neg logit contributions | |
ited | -0.087 |
saw | -0.083 |
heard | -0.079 |
two | -0.074 |
Hey | -0.073 |
Token | start |
Token ID | 923 |
Feature activation | +3.12e-03 |
Pos logit contributions | |
heal | +0.031 |
anks | +0.031 |
bleed | +0.031 |
hyd | +0.031 |
rid | +0.030 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.031 |
heard | -0.031 |
year | -0.031 |
Hey | -0.030 |
Token | to |
Token ID | 284 |
Feature activation | -2.50e-02 |
Pos logit contributions | |
ited | +0.005 |
saw | +0.004 |
heard | +0.004 |
Hey | +0.004 |
year | +0.004 |
Neg logit contributions | |
anks | -0.006 |
balls | -0.006 |
hyd | -0.006 |
rid | -0.006 |
heal | -0.006 |
Token | kick |
Token ID | 4829 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
anks | +0.036 |
balls | +0.034 |
hyd | +0.034 |
heal | +0.033 |
rid | +0.033 |
Neg logit contributions | |
ited | -0.050 |
saw | -0.046 |
heard | -0.044 |
Hey | -0.042 |
year | -0.041 |
Token | and |
Token ID | 290 |
Feature activation | -6.21e-02 |
Pos logit contributions | |
ited | +0.014 |
saw | +0.012 |
heard | +0.011 |
year | +0.010 |
Hey | +0.010 |
Neg logit contributions | |
anks | -0.021 |
balls | -0.021 |
hyd | -0.021 |
rid | -0.020 |
heal | -0.020 |
Token | scream |
Token ID | 8196 |
Feature activation | +9.42e-03 |
Pos logit contributions | |
anks | +0.112 |
ectar | +0.104 |
hyd | +0.102 |
balls | +0.101 |
ometers | +0.100 |
Neg logit contributions | |
saw | -0.102 |
ited | -0.102 |
heard | -0.098 |
two | -0.090 |
did | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | -4.71e-02 |
Pos logit contributions | |
ited | +0.012 |
saw | +0.012 |
heard | +0.011 |
little | +0.010 |
two | +0.010 |
Neg logit contributions | |
anks | -0.019 |
hyd | -0.018 |
balls | -0.018 |
rid | -0.018 |
bleed | -0.018 |
Token | They |
Token ID | 1119 |
Feature activation | -2.22e-02 |
Pos logit contributions | |
anks | +0.083 |
balls | +0.080 |
hyd | +0.079 |
rid | +0.078 |
heal | +0.077 |
Neg logit contributions | |
ited | -0.076 |
saw | -0.073 |
heard | -0.070 |
two | -0.065 |
little | -0.064 |
Token | say |
Token ID | 910 |
Feature activation | +4.92e-02 |
Pos logit contributions | |
anks | +0.040 |
balls | +0.038 |
hyd | +0.038 |
rid | +0.037 |
heal | +0.036 |
Neg logit contributions | |
ited | -0.037 |
saw | -0.033 |
heard | -0.032 |
year | -0.029 |
Hey | -0.029 |
Token | , |
Token ID | 11 |
Feature activation | -6.37e-03 |
Pos logit contributions | |
ited | +0.060 |
saw | +0.055 |
heard | +0.052 |
two | +0.045 |
little | +0.045 |
Neg logit contributions | |
anks | -0.108 |
balls | -0.105 |
hyd | -0.104 |
rid | -0.102 |
heal | -0.101 |