Token | lonely |
Token ID | 21757 |
Feature activation | +7.92e-03 |
Pos logit contributions | |
Ant | +0.018 |
Next | +0.018 |
hes | +0.018 |
kins | +0.017 |
� | +0.017 |
Neg logit contributions | |
goodbye | -0.016 |
away | -0.014 |
back | -0.013 |
attention | -0.013 |
farewell | -0.012 |
Token | three |
Token ID | 1115 |
Feature activation | +7.12e-02 |
Pos logit contributions | |
Ant | +0.009 |
Salt | +0.009 |
Fire | +0.009 |
Next | +0.009 |
� | +0.009 |
Neg logit contributions | |
goodbye | -0.018 |
attention | -0.016 |
away | -0.016 |
ogging | -0.016 |
des | -0.015 |
Token | year |
Token ID | 614 |
Feature activation | +3.01e-02 |
Pos logit contributions | |
Ant | +0.136 |
� | +0.130 |
Next | +0.130 |
Salt | +0.128 |
Fire | +0.128 |
Neg logit contributions | |
goodbye | -0.110 |
away | -0.096 |
ogging | -0.093 |
attention | -0.091 |
ly | -0.090 |
Token | old |
Token ID | 1468 |
Feature activation | +6.98e-02 |
Pos logit contributions | |
Fire | +0.036 |
twins | +0.035 |
Cere | +0.035 |
Next | +0.034 |
brothers | +0.034 |
Neg logit contributions | |
goodbye | -0.058 |
des | -0.056 |
farewell | -0.053 |
contain | -0.053 |
of | -0.052 |
Token | , |
Token ID | 11 |
Feature activation | -7.35e-03 |
Pos logit contributions | |
Ant | +0.141 |
hes | +0.139 |
� | +0.135 |
Next | +0.135 |
Mouse | +0.132 |
Neg logit contributions | |
goodbye | -0.103 |
ly | -0.087 |
away | -0.085 |
back | -0.082 |
attention | -0.082 |
Token | living |
Token ID | 2877 |
Feature activation | +1.16e-01 |
Pos logit contributions | |
goodbye | +0.012 |
away | +0.010 |
attention | +0.010 |
ogging | +0.010 |
des | +0.009 |
Neg logit contributions | |
Ant | -0.014 |
� | -0.013 |
Fire | -0.013 |
Next | -0.013 |
Mouse | -0.013 |
Token | all |
Token ID | 477 |
Feature activation | +4.07e-02 |
Pos logit contributions | |
Ant | +0.221 |
� | +0.217 |
Next | +0.212 |
Mouse | +0.211 |
Salt | +0.207 |
Neg logit contributions | |
away | -0.168 |
goodbye | -0.167 |
back | -0.152 |
attention | -0.149 |
ly | -0.143 |
Token | alone |
Token ID | 3436 |
Feature activation | +3.49e-02 |
Pos logit contributions | |
Ant | +0.081 |
� | +0.077 |
Mouse | +0.075 |
Next | +0.074 |
Salt | +0.074 |
Neg logit contributions | |
goodbye | -0.063 |
attention | -0.054 |
away | -0.053 |
ogging | -0.051 |
des | -0.048 |
Token | in |
Token ID | 287 |
Feature activation | +3.49e-02 |
Pos logit contributions | |
Ant | +0.076 |
Mouse | +0.073 |
Next | +0.072 |
Salt | +0.072 |
� | +0.072 |
Neg logit contributions | |
goodbye | -0.044 |
away | -0.038 |
attention | -0.035 |
ly | -0.034 |
back | -0.033 |
Token | her |
Token ID | 607 |
Feature activation | +2.58e-02 |
Pos logit contributions | |
Ant | +0.065 |
� | +0.062 |
hes | +0.061 |
Mouse | +0.060 |
Next | +0.060 |
Neg logit contributions | |
goodbye | -0.055 |
attention | -0.048 |
away | -0.047 |
ogging | -0.045 |
farewell | -0.043 |
Token | big |
Token ID | 1263 |
Feature activation | -2.32e-02 |
Pos logit contributions | |
Ant | +0.031 |
hes | +0.028 |
Salt | +0.028 |
� | +0.028 |
Next | +0.027 |
Neg logit contributions | |
goodbye | -0.059 |
away | -0.053 |
attention | -0.052 |
ogging | -0.051 |
farewell | -0.050 |
Token | " |
Token ID | 1 |
Feature activation | +4.42e-02 |
Pos logit contributions | |
Ant | +0.008 |
Salt | +0.008 |
Next | +0.008 |
poker | +0.008 |
sibling | +0.008 |
Neg logit contributions | |
goodbye | -0.010 |
away | -0.009 |
ogging | -0.008 |
farewell | -0.008 |
closer | -0.008 |
Token | Hello |
Token ID | 15496 |
Feature activation | +2.29e-02 |
Pos logit contributions | |
brothers | +0.065 |
sisters | +0.061 |
ates | +0.059 |
sibling | +0.058 |
kins | +0.058 |
Neg logit contributions | |
goodbye | -0.079 |
away | -0.076 |
closer | -0.075 |
des | -0.073 |
farewell | -0.071 |
Token | , |
Token ID | 11 |
Feature activation | +2.94e-02 |
Pos logit contributions | |
Ant | +0.053 |
Fire | +0.051 |
� | +0.051 |
Next | +0.051 |
Salt | +0.051 |
Neg logit contributions | |
goodbye | -0.025 |
attention | -0.020 |
away | -0.020 |
ogging | -0.019 |
des | -0.019 |
Token | rh |
Token ID | 9529 |
Feature activation | +7.88e-02 |
Pos logit contributions | |
Ant | +0.039 |
� | +0.039 |
hes | +0.038 |
Next | +0.037 |
Salt | +0.036 |
Neg logit contributions | |
goodbye | -0.062 |
attention | -0.053 |
away | -0.053 |
ogging | -0.052 |
ly | -0.051 |
Token | in |
Token ID | 259 |
Feature activation | +6.67e-02 |
Pos logit contributions | |
Next | +0.141 |
� | +0.133 |
Salt | +0.131 |
Cere | +0.129 |
hes | +0.129 |
Neg logit contributions | |
away | -0.126 |
goodbye | -0.110 |
ogging | -0.106 |
ly | -0.105 |
of | -0.103 |
Token | oc |
Token ID | 420 |
Feature activation | +1.15e-01 |
Pos logit contributions | |
Next | +0.094 |
hes | +0.093 |
Ant | +0.092 |
Salt | +0.090 |
� | +0.089 |
Neg logit contributions | |
goodbye | -0.127 |
away | -0.119 |
ly | -0.112 |
ogging | -0.109 |
attention | -0.108 |
Token | eros |
Token ID | 27498 |
Feature activation | +5.50e-02 |
Pos logit contributions | |
Ant | +0.152 |
Next | +0.142 |
hes | +0.142 |
Salt | +0.140 |
� | +0.139 |
Neg logit contributions | |
goodbye | -0.239 |
away | -0.214 |
attention | -0.211 |
ogging | -0.207 |
farewell | -0.203 |
Token | !" |
Token ID | 2474 |
Feature activation | +4.29e-03 |
Pos logit contributions | |
Ant | +0.131 |
Fire | +0.124 |
Next | +0.124 |
Salt | +0.124 |
� | +0.124 |
Neg logit contributions | |
goodbye | -0.057 |
attention | -0.045 |
away | -0.045 |
ogging | -0.044 |
farewell | -0.041 |
Token | Anna |
Token ID | 11735 |
Feature activation | -3.00e-02 |
Pos logit contributions | |
Ant | +0.006 |
� | +0.006 |
hes | +0.006 |
Mouse | +0.006 |
Salt | +0.006 |
Neg logit contributions | |
goodbye | -0.009 |
attention | -0.007 |
away | -0.007 |
ogging | -0.007 |
farewell | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | -1.99e-02 |
Pos logit contributions | |
goodbye | +0.041 |
away | +0.033 |
ogging | +0.033 |
attention | +0.033 |
ly | +0.032 |
Neg logit contributions | |
Ant | -0.062 |
� | -0.060 |
Next | -0.059 |
Salt | -0.058 |
Fire | -0.058 |
Token | Ben |
Token ID | 3932 |
Feature activation | -6.07e-02 |
Pos logit contributions | |
goodbye | +0.039 |
attention | +0.033 |
ogging | +0.032 |
away | +0.032 |
farewell | +0.031 |
Neg logit contributions | |
Ant | -0.030 |
� | -0.029 |
hes | -0.028 |
kins | -0.028 |
Fire | -0.028 |
Token | , |
Token ID | 11 |
Feature activation | -9.50e-03 |
Pos logit contributions | |
away | +0.012 |
goodbye | +0.011 |
des | +0.010 |
ogging | +0.010 |
attention | +0.010 |
Neg logit contributions | |
Ant | -0.025 |
Salt | -0.024 |
hes | -0.023 |
� | -0.023 |
Will | -0.023 |
Token | " |
Token ID | 366 |
Feature activation | +4.90e-02 |
Pos logit contributions | |
goodbye | +0.022 |
back | +0.018 |
farewell | +0.017 |
ogging | +0.016 |
closer | +0.016 |
Neg logit contributions | |
Ant | -0.013 |
Fire | -0.012 |
Mouse | -0.012 |
poker | -0.012 |
Next | -0.012 |
Token | Y |
Token ID | 56 |
Feature activation | +8.35e-03 |
Pos logit contributions | |
poker | +0.067 |
kins | +0.066 |
hes | +0.065 |
Fire | +0.064 |
brothers | +0.064 |
Neg logit contributions | |
goodbye | -0.102 |
des | -0.092 |
away | -0.091 |
ly | -0.087 |
farewell | -0.086 |
Token | ay |
Token ID | 323 |
Feature activation | +3.59e-02 |
Pos logit contributions | |
Next | +0.014 |
Ant | +0.014 |
Salt | +0.014 |
� | +0.014 |
Cere | +0.013 |
Neg logit contributions | |
goodbye | -0.014 |
away | -0.013 |
attention | -0.012 |
contain | -0.012 |
ogging | -0.012 |
Token | , |
Token ID | 11 |
Feature activation | +2.51e-02 |
Pos logit contributions | |
Ant | +0.086 |
Salt | +0.082 |
Fire | +0.082 |
Next | +0.082 |
� | +0.082 |
Neg logit contributions | |
goodbye | -0.036 |
away | -0.029 |
attention | -0.028 |
ogging | -0.027 |
des | -0.026 |
Token | muff |
Token ID | 27563 |
Feature activation | +1.09e-01 |
Pos logit contributions | |
Ant | +0.046 |
� | +0.043 |
Mouse | +0.043 |
Next | +0.042 |
hes | +0.042 |
Neg logit contributions | |
goodbye | -0.044 |
attention | -0.035 |
ogging | -0.033 |
back | -0.033 |
farewell | -0.032 |
Token | ins |
Token ID | 1040 |
Feature activation | +8.69e-02 |
Pos logit contributions | |
Next | +0.157 |
� | +0.153 |
Fire | +0.146 |
Ant | +0.146 |
Brush | +0.144 |
Neg logit contributions | |
goodbye | -0.199 |
away | -0.195 |
ogging | -0.195 |
attention | -0.187 |
of | -0.183 |
Token | . |
Token ID | 13 |
Feature activation | +1.90e-02 |
Pos logit contributions | |
Ant | +0.206 |
� | +0.201 |
Next | +0.201 |
Will | +0.200 |
Really | +0.199 |
Neg logit contributions | |
goodbye | -0.089 |
attention | -0.073 |
away | -0.067 |
ogging | -0.064 |
ly | -0.063 |
Token | I |
Token ID | 314 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
kins | +0.034 |
web | +0.031 |
oyd | +0.031 |
Ant | +0.031 |
lf | +0.030 |
Neg logit contributions | |
goodbye | -0.034 |
farewell | -0.025 |
attention | -0.024 |
bye | -0.024 |
des | -0.024 |
Token | love |
Token ID | 1842 |
Feature activation | +2.91e-02 |
Pos logit contributions | |
goodbye | +0.014 |
away | +0.012 |
attention | +0.012 |
des | +0.012 |
farewell | +0.011 |
Neg logit contributions | |
Ant | -0.023 |
Next | -0.022 |
Fire | -0.022 |
� | -0.022 |
Salt | -0.022 |
Token | muff |
Token ID | 27563 |
Feature activation | +8.29e-02 |
Pos logit contributions | |
Ant | +0.054 |
� | +0.054 |
Mouse | +0.053 |
oyd | +0.052 |
hes | +0.051 |
Neg logit contributions | |
goodbye | -0.045 |
attention | -0.039 |
back | -0.036 |
ogging | -0.035 |
away | -0.035 |
Token | was |
Token ID | 373 |
Feature activation | +2.60e-02 |
Pos logit contributions | |
ters | +0.050 |
poker | +0.047 |
� | +0.046 |
Ant | +0.044 |
chefs | +0.044 |
Neg logit contributions | |
goodbye | -0.048 |
back | -0.039 |
ly | -0.039 |
away | -0.039 |
des | -0.038 |
Token | three |
Token ID | 1115 |
Feature activation | +7.26e-02 |
Pos logit contributions | |
� | +0.058 |
Ant | +0.056 |
kins | +0.056 |
Really | +0.054 |
Mouse | +0.054 |
Neg logit contributions | |
goodbye | -0.034 |
back | -0.028 |
attention | -0.028 |
closer | -0.025 |
to | -0.024 |
Token | years |
Token ID | 812 |
Feature activation | -1.91e-03 |
Pos logit contributions | |
Ant | +0.115 |
� | +0.109 |
Next | +0.108 |
Salt | +0.107 |
Fire | +0.106 |
Neg logit contributions | |
goodbye | -0.134 |
away | -0.118 |
ogging | -0.117 |
attention | -0.116 |
ly | -0.112 |
Token | old |
Token ID | 1468 |
Feature activation | +6.91e-02 |
Pos logit contributions | |
goodbye | +0.003 |
away | +0.003 |
des | +0.003 |
attention | +0.003 |
farewell | +0.003 |
Neg logit contributions | |
Fire | -0.003 |
Ant | -0.003 |
Next | -0.003 |
Salt | -0.003 |
brothers | -0.003 |
Token | and |
Token ID | 290 |
Feature activation | +3.80e-02 |
Pos logit contributions | |
hes | +0.169 |
Ant | +0.168 |
� | +0.164 |
Next | +0.163 |
Bees | +0.159 |
Neg logit contributions | |
goodbye | -0.067 |
ly | -0.054 |
away | -0.051 |
back | -0.050 |
attention | -0.050 |
Token | living |
Token ID | 2877 |
Feature activation | +1.08e-01 |
Pos logit contributions | |
ters | +0.081 |
� | +0.077 |
chicks | +0.077 |
arians | +0.073 |
ities | +0.073 |
Neg logit contributions | |
goodbye | -0.056 |
back | -0.042 |
attention | -0.038 |
to | -0.037 |
hy | -0.036 |
Token | near |
Token ID | 1474 |
Feature activation | +1.86e-02 |
Pos logit contributions | |
� | +0.187 |
Ant | +0.185 |
Mouse | +0.174 |
Next | +0.172 |
Really | +0.170 |
Neg logit contributions | |
away | -0.176 |
goodbye | -0.174 |
attention | -0.171 |
back | -0.160 |
ogging | -0.155 |
Token | the |
Token ID | 262 |
Feature activation | -1.68e-02 |
Pos logit contributions | |
Ant | +0.039 |
� | +0.038 |
Mouse | +0.036 |
Really | +0.036 |
Next | +0.035 |
Neg logit contributions | |
goodbye | -0.026 |
attention | -0.023 |
away | -0.022 |
ogging | -0.021 |
back | -0.020 |
Token | sea |
Token ID | 5417 |
Feature activation | +1.65e-02 |
Pos logit contributions | |
goodbye | +0.031 |
away | +0.027 |
attention | +0.027 |
ogging | +0.027 |
farewell | +0.026 |
Neg logit contributions | |
Ant | -0.026 |
Next | -0.025 |
Salt | -0.024 |
Fire | -0.024 |
� | -0.024 |
Token | . |
Token ID | 13 |
Feature activation | -1.08e-02 |
Pos logit contributions | |
Ant | +0.040 |
� | +0.039 |
Mouse | +0.039 |
Next | +0.038 |
Salt | +0.037 |
Neg logit contributions | |
goodbye | -0.017 |
attention | -0.014 |
away | -0.013 |
ogging | -0.013 |
ly | -0.012 |
Token | One |
Token ID | 1881 |
Feature activation | +1.29e-02 |
Pos logit contributions | |
goodbye | +0.021 |
ogging | +0.018 |
attention | +0.018 |
away | +0.018 |
farewell | +0.017 |
Neg logit contributions | |
Ant | -0.016 |
� | -0.015 |
hes | -0.015 |
kins | -0.015 |
Mouse | -0.015 |
Token | cold |
Token ID | 4692 |
Feature activation | +1.42e-02 |
Pos logit contributions | |
Ant | +0.068 |
� | +0.067 |
Next | +0.064 |
Mouse | +0.063 |
Fire | +0.063 |
Neg logit contributions | |
goodbye | -0.060 |
attention | -0.053 |
away | -0.049 |
back | -0.047 |
farewell | -0.047 |
Token | and |
Token ID | 290 |
Feature activation | +2.97e-02 |
Pos logit contributions | |
Salt | +0.024 |
Fire | +0.024 |
Next | +0.023 |
hes | +0.023 |
Ant | +0.023 |
Neg logit contributions | |
goodbye | -0.023 |
farewell | -0.020 |
ogging | -0.020 |
des | -0.020 |
away | -0.020 |
Token | heavy |
Token ID | 4334 |
Feature activation | +3.52e-02 |
Pos logit contributions | |
Ant | +0.056 |
� | +0.056 |
Mouse | +0.055 |
Fire | +0.054 |
Next | +0.053 |
Neg logit contributions | |
goodbye | -0.046 |
attention | -0.040 |
back | -0.037 |
closer | -0.036 |
away | -0.035 |
Token | . |
Token ID | 13 |
Feature activation | +6.84e-03 |
Pos logit contributions | |
Ant | +0.071 |
Salt | +0.069 |
Next | +0.068 |
Fire | +0.068 |
� | +0.068 |
Neg logit contributions | |
goodbye | -0.047 |
away | -0.040 |
ogging | -0.039 |
des | -0.038 |
attention | -0.038 |
Token | " |
Token ID | 366 |
Feature activation | +4.35e-02 |
Pos logit contributions | |
Ant | +0.010 |
Mouse | +0.009 |
kins | +0.009 |
poker | +0.009 |
Fire | +0.009 |
Neg logit contributions | |
goodbye | -0.014 |
ogging | -0.012 |
ly | -0.012 |
back | -0.012 |
attention | -0.012 |
Token | These |
Token ID | 4711 |
Feature activation | +1.07e-01 |
Pos logit contributions | |
brothers | +0.060 |
kins | +0.057 |
sisters | +0.056 |
siblings | +0.052 |
twins | +0.050 |
Neg logit contributions | |
away | -0.087 |
des | -0.083 |
closer | -0.080 |
of | -0.080 |
goodbye | -0.079 |
Token | are |
Token ID | 389 |
Feature activation | +7.56e-02 |
Pos logit contributions | |
Ant | +0.183 |
Next | +0.173 |
Salt | +0.172 |
Fire | +0.172 |
� | +0.172 |
Neg logit contributions | |
goodbye | -0.184 |
away | -0.160 |
attention | -0.160 |
ogging | -0.155 |
farewell | -0.149 |
Token | pretty |
Token ID | 2495 |
Feature activation | +5.48e-02 |
Pos logit contributions | |
Ant | +0.152 |
� | +0.150 |
Next | +0.146 |
Mouse | +0.144 |
Fire | +0.143 |
Neg logit contributions | |
goodbye | -0.107 |
attention | -0.094 |
away | -0.086 |
ogging | -0.081 |
back | -0.080 |
Token | !" |
Token ID | 2474 |
Feature activation | +6.04e-03 |
Pos logit contributions | |
Ant | +0.098 |
� | +0.095 |
chefs | +0.093 |
kins | +0.093 |
Next | +0.093 |
Neg logit contributions | |
goodbye | -0.083 |
ogging | -0.074 |
away | -0.070 |
farewell | -0.070 |
attention | -0.069 |
Token | she |
Token ID | 673 |
Feature activation | -9.97e-03 |
Pos logit contributions | |
Ant | +0.008 |
Next | +0.008 |
Salt | +0.008 |
� | +0.008 |
Fire | +0.008 |
Neg logit contributions | |
goodbye | -0.012 |
away | -0.011 |
attention | -0.011 |
ogging | -0.011 |
farewell | -0.010 |
Token | said |
Token ID | 531 |
Feature activation | -6.89e-03 |
Pos logit contributions | |
goodbye | +0.013 |
ly | +0.011 |
attention | +0.010 |
closer | +0.010 |
ogging | +0.010 |
Neg logit contributions | |
� | -0.021 |
Fire | -0.020 |
Ant | -0.020 |
poker | -0.020 |
Next | -0.020 |
Token | He |
Token ID | 679 |
Feature activation | +2.54e-02 |
Pos logit contributions | |
goodbye | +0.011 |
farewell | +0.008 |
away | +0.008 |
ogging | +0.008 |
des | +0.008 |
Neg logit contributions | |
kins | -0.014 |
hes | -0.013 |
ins | -0.013 |
we | -0.013 |
� | -0.013 |
Token | was |
Token ID | 373 |
Feature activation | +1.92e-02 |
Pos logit contributions | |
ters | +0.035 |
poker | +0.034 |
� | +0.034 |
Salt | +0.033 |
chefs | +0.033 |
Neg logit contributions | |
goodbye | -0.049 |
away | -0.043 |
ly | -0.042 |
farewell | -0.042 |
back | -0.042 |
Token | very |
Token ID | 845 |
Feature activation | +1.57e-02 |
Pos logit contributions | |
� | +0.041 |
kins | +0.039 |
Ant | +0.039 |
Really | +0.038 |
Mouse | +0.036 |
Neg logit contributions | |
goodbye | -0.028 |
attention | -0.022 |
back | -0.022 |
farewell | -0.020 |
closer | -0.020 |
Token | poor |
Token ID | 3595 |
Feature activation | +4.16e-02 |
Pos logit contributions | |
Ant | +0.027 |
� | +0.026 |
Next | +0.025 |
Salt | +0.025 |
Fire | +0.025 |
Neg logit contributions | |
goodbye | -0.027 |
attention | -0.023 |
away | -0.023 |
ogging | -0.022 |
farewell | -0.021 |
Token | , |
Token ID | 11 |
Feature activation | -1.37e-02 |
Pos logit contributions | |
Ant | +0.099 |
� | +0.099 |
Mouse | +0.097 |
hes | +0.096 |
Next | +0.096 |
Neg logit contributions | |
goodbye | -0.045 |
attention | -0.035 |
ogging | -0.034 |
away | -0.033 |
ly | -0.031 |
Token | living |
Token ID | 2877 |
Feature activation | +1.07e-01 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.016 |
attention | +0.016 |
ogging | +0.016 |
farewell | +0.015 |
Neg logit contributions | |
Ant | -0.028 |
� | -0.027 |
Mouse | -0.026 |
Next | -0.026 |
hes | -0.026 |
Token | all |
Token ID | 477 |
Feature activation | +3.45e-02 |
Pos logit contributions | |
� | +0.198 |
Ant | +0.192 |
Next | +0.187 |
Really | +0.186 |
Mouse | +0.185 |
Neg logit contributions | |
away | -0.164 |
goodbye | -0.155 |
back | -0.149 |
attention | -0.146 |
home | -0.138 |
Token | alone |
Token ID | 3436 |
Feature activation | +3.17e-02 |
Pos logit contributions | |
Ant | +0.069 |
� | +0.067 |
Mouse | +0.064 |
Really | +0.064 |
Sal | +0.063 |
Neg logit contributions | |
goodbye | -0.053 |
attention | -0.045 |
away | -0.045 |
ogging | -0.042 |
des | -0.040 |
Token | in |
Token ID | 287 |
Feature activation | +3.30e-02 |
Pos logit contributions | |
Ant | +0.073 |
� | +0.071 |
Mouse | +0.070 |
Sal | +0.070 |
Next | +0.070 |
Neg logit contributions | |
goodbye | -0.036 |
away | -0.029 |
attention | -0.027 |
ly | -0.027 |
back | -0.026 |
Token | a |
Token ID | 257 |
Feature activation | -1.84e-02 |
Pos logit contributions | |
Ant | +0.067 |
Really | +0.067 |
� | +0.067 |
hes | +0.064 |
Sal | +0.064 |
Neg logit contributions | |
goodbye | -0.045 |
attention | -0.038 |
away | -0.036 |
back | -0.035 |
ogging | -0.034 |
Token | small |
Token ID | 1402 |
Feature activation | +2.35e-03 |
Pos logit contributions | |
goodbye | +0.030 |
away | +0.026 |
attention | +0.025 |
ogging | +0.025 |
farewell | +0.024 |
Neg logit contributions | |
Ant | -0.033 |
Next | -0.032 |
hes | -0.031 |
Salt | -0.031 |
� | -0.031 |
Token | � |
Token ID | 95 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
goodbye | +0.012 |
attention | +0.010 |
away | +0.010 |
ogging | +0.010 |
farewell | +0.010 |
Neg logit contributions | |
Ant | -0.011 |
� | -0.010 |
Salt | -0.010 |
Next | -0.010 |
Fire | -0.010 |
Token | € |
Token ID | 26391 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
goodbye | +0.021 |
away | +0.019 |
ogging | +0.018 |
attention | +0.018 |
des | +0.018 |
Neg logit contributions | |
Ant | -0.019 |
Fire | -0.018 |
Next | -0.018 |
hes | -0.017 |
Salt | -0.017 |
Token | � |
Token ID | 129 |
Feature activation | -2.19e-02 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.018 |
attention | +0.018 |
ogging | +0.018 |
farewell | +0.017 |
Neg logit contributions | |
Ant | -0.015 |
Next | -0.014 |
Salt | -0.014 |
Fire | -0.014 |
� | -0.014 |
Token | � |
Token ID | 241 |
Feature activation | +4.27e-02 |
Pos logit contributions | |
goodbye | +0.047 |
attention | +0.043 |
away | +0.042 |
ogging | +0.042 |
des | +0.041 |
Neg logit contributions | |
Next | -0.026 |
Ant | -0.025 |
hes | -0.025 |
Salt | -0.025 |
� | -0.025 |
Token | Hi |
Token ID | 17250 |
Feature activation | +2.63e-02 |
Pos logit contributions | |
Ant | +0.038 |
Fire | +0.034 |
� | +0.033 |
Salt | +0.033 |
hes | +0.033 |
Neg logit contributions | |
goodbye | -0.108 |
away | -0.098 |
attention | -0.097 |
ogging | -0.097 |
des | -0.095 |
Token | rob |
Token ID | 3857 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
Ant | +0.059 |
Next | +0.057 |
Salt | +0.057 |
Fire | +0.057 |
� | +0.056 |
Neg logit contributions | |
goodbye | -0.031 |
away | -0.025 |
attention | -0.024 |
ogging | -0.024 |
farewell | -0.022 |
Token | in |
Token ID | 259 |
Feature activation | +6.73e-02 |
Pos logit contributions | |
� | +0.153 |
Next | +0.153 |
Ant | +0.151 |
Fire | +0.148 |
Salt | +0.146 |
Neg logit contributions | |
goodbye | -0.196 |
ogging | -0.185 |
away | -0.181 |
attention | -0.172 |
ly | -0.171 |
Token | . |
Token ID | 13 |
Feature activation | +1.93e-02 |
Pos logit contributions | |
Ant | +0.169 |
hes | +0.166 |
Next | +0.165 |
Fire | +0.163 |
� | +0.162 |
Neg logit contributions | |
goodbye | -0.060 |
away | -0.044 |
ogging | -0.043 |
attention | -0.039 |
ly | -0.039 |
Token | â |
Token ID | 22940 |
Feature activation | -1.93e-02 |
Pos logit contributions | |
kins | +0.033 |
hes | +0.031 |
Ant | +0.031 |
poker | +0.029 |
oyd | +0.029 |
Neg logit contributions | |
goodbye | -0.035 |
ogging | -0.029 |
farewell | -0.027 |
des | -0.027 |
closer | -0.027 |
Token | € |
Token ID | 26391 |
Feature activation | -8.70e-03 |
Pos logit contributions | |
goodbye | +0.033 |
attention | +0.028 |
away | +0.028 |
ogging | +0.027 |
farewell | +0.026 |
Neg logit contributions | |
Ant | -0.033 |
� | -0.032 |
Next | -0.032 |
Salt | -0.031 |
hes | -0.031 |
Token |
|
Token ID | 198 |
Feature activation | +3.60e-03 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.016 |
attention | +0.015 |
ogging | +0.015 |
farewell | +0.015 |
Neg logit contributions | |
Ant | -0.012 |
Next | -0.011 |
Salt | -0.011 |
Fire | -0.011 |
� | -0.011 |
Token | |
Token ID | 220 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
goodbye | +0.003 |
back | +0.002 |
ogging | +0.002 |
farewell | +0.002 |
closer | +0.002 |
Neg logit contributions | |
kins | -0.003 |
hes | -0.002 |
ities | -0.002 |
poker | -0.002 |
chefs | -0.002 |
Token |
|
Token ID | 198 |
Feature activation | -1.37e-02 |
Pos logit contributions | |
goodbye | +0.028 |
attention | +0.026 |
ogging | +0.025 |
away | +0.024 |
back | +0.023 |
Neg logit contributions | |
Ant | -0.022 |
� | -0.021 |
Fire | -0.020 |
Will | -0.020 |
Salt | -0.020 |
Token |
|
Token ID | 198 |
Feature activation | -9.71e-03 |
Pos logit contributions | |
goodbye | +0.026 |
away | +0.022 |
ogging | +0.022 |
farewell | +0.022 |
attention | +0.022 |
Neg logit contributions | |
Ant | -0.019 |
hes | -0.019 |
� | -0.019 |
Salt | -0.019 |
chefs | -0.019 |
Token | " |
Token ID | 1 |
Feature activation | +4.68e-02 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.017 |
farewell | +0.016 |
closer | +0.016 |
ly | +0.015 |
Neg logit contributions | |
hes | -0.014 |
ities | -0.014 |
Salt | -0.014 |
brothers | -0.014 |
poker | -0.013 |
Token | C |
Token ID | 34 |
Feature activation | +3.32e-03 |
Pos logit contributions | |
brothers | +0.080 |
sibling | +0.076 |
ates | +0.075 |
sisters | +0.072 |
kins | +0.072 |
Neg logit contributions | |
closer | -0.074 |
contain | -0.071 |
des | -0.070 |
of | -0.068 |
away | -0.065 |
Token | oc |
Token ID | 420 |
Feature activation | +1.06e-01 |
Pos logit contributions | |
Ant | +0.005 |
hes | +0.005 |
Fire | +0.005 |
Next | +0.005 |
Will | +0.005 |
Neg logit contributions | |
away | -0.005 |
goodbye | -0.005 |
ogging | -0.005 |
des | -0.005 |
ly | -0.005 |
Token | oa |
Token ID | 12162 |
Feature activation | +1.08e-02 |
Pos logit contributions | |
Ant | +0.123 |
hes | +0.123 |
Next | +0.118 |
� | +0.113 |
Fire | +0.113 |
Neg logit contributions | |
goodbye | -0.227 |
attention | -0.205 |
away | -0.204 |
ly | -0.199 |
ogging | -0.198 |
Token | ? |
Token ID | 30 |
Feature activation | +9.84e-04 |
Pos logit contributions | |
Really | +0.021 |
Ant | +0.020 |
Will | +0.020 |
Next | +0.020 |
hes | +0.019 |
Neg logit contributions | |
goodbye | -0.015 |
ly | -0.013 |
attention | -0.013 |
back | -0.012 |
away | -0.012 |
Token | What |
Token ID | 1867 |
Feature activation | +5.75e-02 |
Pos logit contributions | |
kins | +0.002 |
oyd | +0.002 |
iflower | +0.002 |
magnets | +0.002 |
kin | +0.002 |
Neg logit contributions | |
goodbye | -0.002 |
closer | -0.001 |
ur | -0.001 |
hi | -0.001 |
to | -0.001 |
Token | 's |
Token ID | 338 |
Feature activation | +6.79e-02 |
Pos logit contributions | |
� | +0.109 |
Ant | +0.106 |
Mouse | +0.104 |
Really | +0.102 |
Next | +0.102 |
Neg logit contributions | |
goodbye | -0.086 |
attention | -0.072 |
ogging | -0.070 |
des | -0.069 |
away | -0.067 |
Token | that |
Token ID | 326 |
Feature activation | +5.15e-02 |
Pos logit contributions | |
Ant | +0.146 |
Mouse | +0.145 |
sibling | +0.143 |
� | +0.142 |
Really | +0.139 |
Neg logit contributions | |
goodbye | -0.081 |
attention | -0.078 |
back | -0.067 |
away | -0.064 |
closer | -0.064 |
Token | mom |
Token ID | 1995 |
Feature activation | +8.43e-03 |
Pos logit contributions | |
Ant | +0.038 |
� | +0.037 |
hes | +0.035 |
Mouse | +0.035 |
Salt | +0.034 |
Neg logit contributions | |
goodbye | -0.045 |
attention | -0.040 |
back | -0.039 |
ogging | -0.036 |
farewell | -0.036 |
Token | said |
Token ID | 531 |
Feature activation | -1.70e-02 |
Pos logit contributions | |
Ant | +0.018 |
chefs | +0.017 |
poker | +0.017 |
Blocks | +0.017 |
Fire | +0.017 |
Neg logit contributions | |
goodbye | -0.011 |
ly | -0.010 |
of | -0.009 |
back | -0.008 |
closer | -0.008 |
Token | , |
Token ID | 11 |
Feature activation | -7.79e-03 |
Pos logit contributions | |
goodbye | +0.016 |
away | +0.012 |
attention | +0.012 |
ogging | +0.012 |
farewell | +0.011 |
Neg logit contributions | |
Ant | -0.042 |
Next | -0.040 |
Salt | -0.040 |
� | -0.040 |
Fire | -0.040 |
Token | " |
Token ID | 366 |
Feature activation | +4.80e-02 |
Pos logit contributions | |
goodbye | +0.016 |
back | +0.012 |
ly | +0.012 |
farewell | +0.012 |
to | +0.012 |
Neg logit contributions | |
Ant | -0.013 |
Mouse | -0.012 |
poker | -0.012 |
kins | -0.012 |
ates | -0.011 |
Token | C |
Token ID | 34 |
Feature activation | +9.59e-03 |
Pos logit contributions | |
kins | +0.058 |
sisters | +0.058 |
brothers | +0.057 |
siblings | +0.054 |
sibling | +0.053 |
Neg logit contributions | |
of | -0.095 |
des | -0.092 |
away | -0.087 |
land | -0.087 |
goodbye | -0.086 |
Token | oc |
Token ID | 420 |
Feature activation | +1.05e-01 |
Pos logit contributions | |
Ant | +0.016 |
chefs | +0.015 |
kins | +0.015 |
hes | +0.015 |
sisters | +0.015 |
Neg logit contributions | |
away | -0.015 |
ogging | -0.015 |
goodbye | -0.014 |
des | -0.014 |
ly | -0.013 |
Token | oa |
Token ID | 12162 |
Feature activation | +1.43e-02 |
Pos logit contributions | |
Ant | +0.113 |
hes | +0.109 |
I | +0.109 |
Anything | +0.104 |
Next | +0.104 |
Neg logit contributions | |
goodbye | -0.227 |
attention | -0.216 |
ly | -0.211 |
away | -0.205 |
ogging | -0.203 |
Token | is |
Token ID | 318 |
Feature activation | +3.13e-02 |
Pos logit contributions | |
Really | +0.027 |
Ant | +0.025 |
Someone | +0.024 |
Next | +0.024 |
Mouse | +0.024 |
Neg logit contributions | |
goodbye | -0.021 |
ly | -0.019 |
attention | -0.019 |
! | -0.018 |
back | -0.017 |
Token | a |
Token ID | 257 |
Feature activation | +2.28e-03 |
Pos logit contributions | |
Really | +0.059 |
Ant | +0.058 |
� | +0.057 |
Mouse | +0.056 |
Anything | +0.055 |
Neg logit contributions | |
goodbye | -0.046 |
attention | -0.043 |
back | -0.042 |
to | -0.039 |
closer | -0.037 |
Token | strong |
Token ID | 1913 |
Feature activation | -7.05e-03 |
Pos logit contributions | |
Ant | +0.004 |
Next | +0.004 |
kins | +0.004 |
� | +0.004 |
hes | +0.004 |
Neg logit contributions | |
goodbye | -0.004 |
attention | -0.003 |
away | -0.003 |
back | -0.003 |
farewell | -0.003 |
Token | drink |
Token ID | 4144 |
Feature activation | +1.62e-02 |
Pos logit contributions | |
goodbye | +0.012 |
attention | +0.010 |
away | +0.010 |
ly | +0.010 |
ogging | +0.010 |
Neg logit contributions | |
Ant | -0.012 |
hes | -0.012 |
Next | -0.012 |
� | -0.012 |
kins | -0.011 |
Token | of |
Token ID | 286 |
Feature activation | +7.02e-02 |
Pos logit contributions | |
hes | +0.047 |
kins | +0.045 |
I | +0.045 |
Ant | +0.045 |
Salt | +0.044 |
Neg logit contributions | |
goodbye | -0.056 |
attention | -0.050 |
farewell | -0.049 |
away | -0.049 |
ogging | -0.049 |
Token | le |
Token ID | 443 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
Ant | +0.102 |
� | +0.096 |
hes | +0.091 |
Next | +0.091 |
Salt | +0.091 |
Neg logit contributions | |
goodbye | -0.141 |
attention | -0.129 |
away | -0.126 |
ly | -0.119 |
des | -0.117 |
Token | mons |
Token ID | 11567 |
Feature activation | +2.36e-02 |
Pos logit contributions | |
Ant | +0.034 |
Next | +0.032 |
Will | +0.031 |
� | +0.030 |
Fire | +0.030 |
Neg logit contributions | |
goodbye | -0.069 |
attention | -0.065 |
ogging | -0.065 |
away | -0.064 |
ly | -0.063 |
Token | . |
Token ID | 13 |
Feature activation | -7.06e-03 |
Pos logit contributions | |
Ant | +0.048 |
Next | +0.046 |
Fire | +0.046 |
� | +0.046 |
Salt | +0.046 |
Neg logit contributions | |
goodbye | -0.032 |
attention | -0.027 |
away | -0.027 |
ogging | -0.026 |
des | -0.025 |
Token | " |
Token ID | 366 |
Feature activation | +4.00e-02 |
Pos logit contributions | |
goodbye | +0.013 |
ly | +0.011 |
back | +0.011 |
farewell | +0.009 |
to | +0.009 |
Neg logit contributions | |
Ant | -0.013 |
Mouse | -0.011 |
oyd | -0.011 |
kins | -0.011 |
ates | -0.011 |
Token | These |
Token ID | 4711 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
kins | +0.056 |
brothers | +0.054 |
poker | +0.051 |
pap | +0.049 |
sisters | +0.048 |
Neg logit contributions | |
away | -0.077 |
des | -0.074 |
of | -0.072 |
closer | -0.069 |
goodbye | -0.067 |
Token | le |
Token ID | 443 |
Feature activation | +4.48e-02 |
Pos logit contributions | |
Ant | +0.161 |
Next | +0.153 |
Will | +0.150 |
hes | +0.150 |
Fire | +0.150 |
Neg logit contributions | |
goodbye | -0.182 |
attention | -0.172 |
away | -0.162 |
ogging | -0.158 |
ly | -0.156 |
Token | mons |
Token ID | 11567 |
Feature activation | +4.52e-02 |
Pos logit contributions | |
Will | +0.050 |
Ant | +0.050 |
Next | +0.047 |
Fire | +0.046 |
Bear | +0.045 |
Neg logit contributions | |
attention | -0.090 |
ogging | -0.089 |
oot | -0.088 |
goodbye | -0.087 |
ly | -0.084 |
Token | are |
Token ID | 389 |
Feature activation | +6.41e-02 |
Pos logit contributions | |
Ant | +0.092 |
Will | +0.087 |
Fire | +0.086 |
Really | +0.086 |
Next | +0.086 |
Neg logit contributions | |
goodbye | -0.061 |
attention | -0.055 |
ly | -0.051 |
ogging | -0.051 |
away | -0.051 |
Token | sour |
Token ID | 11348 |
Feature activation | +3.54e-02 |
Pos logit contributions | |
Ant | +0.133 |
� | +0.130 |
Next | +0.128 |
Mouse | +0.128 |
oyd | +0.128 |
Neg logit contributions | |
goodbye | -0.080 |
attention | -0.076 |
back | -0.067 |
away | -0.063 |
ly | -0.060 |
Token | ," |
Token ID | 553 |
Feature activation | +2.39e-02 |
Pos logit contributions | |
Ant | +0.076 |
Next | +0.072 |
Fire | +0.072 |
Salt | +0.072 |
� | +0.072 |
Neg logit contributions | |
goodbye | -0.045 |
away | -0.037 |
attention | -0.037 |
ogging | -0.036 |
farewell | -0.034 |
Token | bitter |
Token ID | 12922 |
Feature activation | +2.96e-02 |
Pos logit contributions | |
� | +0.099 |
Ant | +0.094 |
Mouse | +0.092 |
Next | +0.089 |
Bear | +0.088 |
Neg logit contributions | |
goodbye | -0.071 |
back | -0.063 |
attention | -0.062 |
closer | -0.055 |
away | -0.051 |
Token | . |
Token ID | 13 |
Feature activation | +1.72e-02 |
Pos logit contributions | |
Ant | +0.055 |
Salt | +0.054 |
Next | +0.054 |
hes | +0.053 |
Fire | +0.053 |
Neg logit contributions | |
goodbye | -0.044 |
away | -0.038 |
attention | -0.036 |
ogging | -0.036 |
farewell | -0.036 |
Token |
|
Token ID | 198 |
Feature activation | +6.92e-03 |
Pos logit contributions | |
Ant | +0.030 |
kins | +0.029 |
� | +0.029 |
sisters | +0.028 |
sibling | +0.028 |
Neg logit contributions | |
goodbye | -0.028 |
ogging | -0.022 |
away | -0.022 |
farewell | -0.021 |
owl | -0.021 |
Token |
|
Token ID | 198 |
Feature activation | +6.09e-03 |
Pos logit contributions | |
Ant | +0.009 |
� | +0.008 |
Salt | +0.008 |
Next | +0.008 |
Fire | +0.008 |
Neg logit contributions | |
goodbye | -0.014 |
away | -0.013 |
attention | -0.013 |
ogging | -0.013 |
farewell | -0.012 |
Token | " |
Token ID | 1 |
Feature activation | +4.47e-02 |
Pos logit contributions | |
Ant | +0.008 |
Salt | +0.007 |
Next | +0.007 |
Fire | +0.007 |
� | +0.007 |
Neg logit contributions | |
goodbye | -0.013 |
away | -0.012 |
ogging | -0.012 |
attention | -0.011 |
des | -0.011 |
Token | These |
Token ID | 4711 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
brothers | +0.071 |
sisters | +0.067 |
poker | +0.066 |
sibling | +0.066 |
ates | +0.065 |
Neg logit contributions | |
des | -0.064 |
closer | -0.063 |
goodbye | -0.062 |
away | -0.061 |
of | -0.061 |
Token | flowers |
Token ID | 12734 |
Feature activation | +4.34e-02 |
Pos logit contributions | |
Ant | +0.196 |
Next | +0.191 |
� | +0.188 |
Will | +0.188 |
hes | +0.187 |
Neg logit contributions | |
goodbye | -0.152 |
attention | -0.138 |
away | -0.126 |
ly | -0.120 |
back | -0.117 |
Token | are |
Token ID | 389 |
Feature activation | +5.36e-02 |
Pos logit contributions | |
Ant | +0.082 |
Next | +0.079 |
Fire | +0.078 |
� | +0.078 |
Salt | +0.077 |
Neg logit contributions | |
goodbye | -0.066 |
attention | -0.057 |
away | -0.056 |
ogging | -0.054 |
ly | -0.052 |
Token | not |
Token ID | 407 |
Feature activation | -1.94e-02 |
Pos logit contributions | |
Ant | +0.103 |
� | +0.102 |
Next | +0.099 |
Mouse | +0.098 |
Will | +0.096 |
Neg logit contributions | |
goodbye | -0.080 |
attention | -0.071 |
away | -0.064 |
back | -0.064 |
ly | -0.060 |
Token | for |
Token ID | 329 |
Feature activation | +5.39e-02 |
Pos logit contributions | |
goodbye | +0.028 |
attention | +0.025 |
away | +0.024 |
ly | +0.022 |
back | +0.022 |
Neg logit contributions | |
Ant | -0.038 |
� | -0.038 |
Next | -0.036 |
Mouse | -0.036 |
Will | -0.036 |
Token | eating |
Token ID | 6600 |
Feature activation | +4.23e-02 |
Pos logit contributions | |
Ant | +0.092 |
Mouse | +0.087 |
� | +0.087 |
hes | +0.086 |
Next | +0.085 |
Neg logit contributions | |
goodbye | -0.092 |
attention | -0.082 |
ogging | -0.078 |
away | -0.078 |
ly | -0.074 |
Token | old |
Token ID | 1468 |
Feature activation | +4.70e-02 |
Pos logit contributions | |
Ant | +0.029 |
� | +0.027 |
Next | +0.027 |
Salt | +0.027 |
Fire | +0.027 |
Neg logit contributions | |
goodbye | -0.033 |
attention | -0.029 |
away | -0.029 |
ogging | -0.028 |
farewell | -0.027 |
Token | photo |
Token ID | 4590 |
Feature activation | +1.81e-02 |
Pos logit contributions | |
Ant | +0.070 |
� | +0.065 |
Next | +0.064 |
Mouse | +0.063 |
Salt | +0.063 |
Neg logit contributions | |
goodbye | -0.095 |
attention | -0.085 |
away | -0.080 |
ogging | -0.080 |
ly | -0.078 |
Token | albums |
Token ID | 16788 |
Feature activation | -5.66e-03 |
Pos logit contributions | |
Ant | +0.039 |
Next | +0.037 |
� | +0.037 |
Salt | +0.037 |
Fire | +0.037 |
Neg logit contributions | |
goodbye | -0.023 |
attention | -0.019 |
away | -0.019 |
ogging | -0.018 |
farewell | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
goodbye | +0.007 |
away | +0.006 |
attention | +0.006 |
ogging | +0.006 |
farewell | +0.006 |
Neg logit contributions | |
Ant | -0.012 |
Next | -0.011 |
Salt | -0.011 |
� | -0.011 |
Fire | -0.011 |
Token | " |
Token ID | 366 |
Feature activation | +4.19e-02 |
Pos logit contributions | |
goodbye | +0.023 |
ly | +0.019 |
farewell | +0.018 |
back | +0.018 |
ogging | +0.018 |
Neg logit contributions | |
Ant | -0.015 |
kins | -0.014 |
Mouse | -0.013 |
oyd | -0.012 |
Bear | -0.012 |
Token | These |
Token ID | 4711 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
kins | +0.058 |
poker | +0.052 |
brothers | +0.049 |
Salt | +0.047 |
Ant | +0.046 |
Neg logit contributions | |
goodbye | -0.087 |
away | -0.085 |
des | -0.083 |
of | -0.080 |
farewell | -0.078 |
Token | are |
Token ID | 389 |
Feature activation | +7.44e-02 |
Pos logit contributions | |
Ant | +0.176 |
� | +0.172 |
Fire | +0.171 |
Next | +0.170 |
Salt | +0.167 |
Neg logit contributions | |
goodbye | -0.165 |
ogging | -0.145 |
away | -0.144 |
des | -0.138 |
attention | -0.137 |
Token | pictures |
Token ID | 5986 |
Feature activation | +2.92e-02 |
Pos logit contributions | |
Ant | +0.142 |
� | +0.134 |
Next | +0.134 |
Salt | +0.133 |
Fire | +0.132 |
Neg logit contributions | |
goodbye | -0.114 |
attention | -0.098 |
away | -0.096 |
ogging | -0.092 |
farewell | -0.089 |
Token | of |
Token ID | 286 |
Feature activation | +6.07e-02 |
Pos logit contributions | |
Ant | +0.061 |
Next | +0.058 |
Salt | +0.058 |
� | +0.058 |
Fire | +0.058 |
Neg logit contributions | |
goodbye | -0.038 |
away | -0.032 |
attention | -0.031 |
ogging | -0.031 |
farewell | -0.029 |
Token | when |
Token ID | 618 |
Feature activation | +3.94e-02 |
Pos logit contributions | |
Ant | +0.106 |
� | +0.101 |
hes | +0.100 |
Mouse | +0.100 |
Sal | +0.100 |
Neg logit contributions | |
goodbye | -0.103 |
attention | -0.091 |
away | -0.087 |
back | -0.082 |
farewell | -0.080 |
Token | I |
Token ID | 314 |
Feature activation | -3.01e-02 |
Pos logit contributions | |
Ant | +0.073 |
Mouse | +0.070 |
Sal | +0.068 |
hes | +0.067 |
poker | +0.066 |
Neg logit contributions | |
goodbye | -0.067 |
attention | -0.056 |
away | -0.052 |
back | -0.051 |
ly | -0.050 |
Token | â |
Token ID | 22940 |
Feature activation | -4.43e-03 |
Pos logit contributions | |
Ant | +0.111 |
Next | +0.105 |
Salt | +0.103 |
� | +0.103 |
hes | +0.102 |
Neg logit contributions | |
goodbye | -0.125 |
away | -0.106 |
attention | -0.106 |
ogging | -0.103 |
farewell | -0.101 |
Token | € |
Token ID | 26391 |
Feature activation | +2.32e-03 |
Pos logit contributions | |
goodbye | +0.007 |
away | +0.006 |
attention | +0.006 |
ogging | +0.006 |
des | +0.006 |
Neg logit contributions | |
Ant | -0.008 |
Fire | -0.008 |
Salt | -0.008 |
Next | -0.008 |
hes | -0.007 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.61e-02 |
Pos logit contributions | |
Ant | +0.003 |
Next | +0.003 |
Fire | +0.003 |
Salt | +0.003 |
Mouse | +0.003 |
Neg logit contributions | |
goodbye | -0.005 |
ogging | -0.004 |
away | -0.004 |
attention | -0.004 |
des | -0.004 |
Token | s |
Token ID | 82 |
Feature activation | +7.20e-02 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.017 |
attention | +0.017 |
ogging | +0.016 |
farewell | +0.015 |
Neg logit contributions | |
Ant | -0.035 |
Next | -0.033 |
Salt | -0.033 |
Fire | -0.033 |
� | -0.033 |
Token | a |
Token ID | 257 |
Feature activation | +2.01e-02 |
Pos logit contributions | |
Ant | +0.121 |
Will | +0.115 |
chefs | +0.114 |
Really | +0.114 |
Mouse | +0.113 |
Neg logit contributions | |
goodbye | -0.124 |
attention | -0.108 |
farewell | -0.101 |
back | -0.099 |
land | -0.098 |
Token | rob |
Token ID | 3857 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
Next | +0.036 |
Ant | +0.036 |
hes | +0.035 |
� | +0.035 |
Salt | +0.034 |
Neg logit contributions | |
goodbye | -0.032 |
away | -0.028 |
attention | -0.026 |
farewell | -0.026 |
ogging | -0.025 |
Token | in |
Token ID | 259 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
Next | +0.165 |
� | +0.164 |
Fire | +0.158 |
Ant | +0.157 |
Salt | +0.157 |
Neg logit contributions | |
goodbye | -0.172 |
ogging | -0.167 |
away | -0.161 |
ly | -0.156 |
attention | -0.154 |
Token | . |
Token ID | 13 |
Feature activation | +1.31e-02 |
Pos logit contributions | |
Ant | +0.142 |
hes | +0.138 |
Next | +0.138 |
Really | +0.137 |
� | +0.135 |
Neg logit contributions | |
goodbye | -0.069 |
away | -0.054 |
ly | -0.051 |
ogging | -0.050 |
attention | -0.050 |
Token | It |
Token ID | 632 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
kins | +0.021 |
Ant | +0.020 |
hes | +0.019 |
poker | +0.019 |
oyd | +0.019 |
Neg logit contributions | |
goodbye | -0.024 |
ogging | -0.022 |
des | -0.020 |
farewell | -0.020 |
attention | -0.019 |
Token | â |
Token ID | 22940 |
Feature activation | -2.46e-02 |
Pos logit contributions | |
Ant | +0.050 |
� | +0.050 |
chefs | +0.050 |
ities | +0.049 |
Salt | +0.049 |
Neg logit contributions | |
goodbye | -0.050 |
closer | -0.044 |
back | -0.043 |
ly | -0.041 |
attention | -0.040 |
Token | € |
Token ID | 26391 |
Feature activation | -1.22e-02 |
Pos logit contributions | |
goodbye | +0.037 |
away | +0.032 |
attention | +0.032 |
ogging | +0.032 |
des | +0.030 |
Neg logit contributions | |
Ant | -0.047 |
Fire | -0.044 |
Salt | -0.044 |
Will | -0.044 |
hes | -0.043 |
Token | a |
Token ID | 257 |
Feature activation | +4.71e-03 |
Pos logit contributions | |
Ant | +0.006 |
� | +0.006 |
Mouse | +0.006 |
Fire | +0.006 |
Next | +0.006 |
Neg logit contributions | |
goodbye | -0.005 |
attention | -0.004 |
back | -0.003 |
farewell | -0.003 |
away | -0.003 |
Token | doll |
Token ID | 3654 |
Feature activation | +4.97e-02 |
Pos logit contributions | |
Ant | +0.007 |
Next | +0.007 |
hes | +0.007 |
� | +0.007 |
Salt | +0.007 |
Neg logit contributions | |
goodbye | -0.009 |
attention | -0.007 |
away | -0.007 |
farewell | -0.007 |
ogging | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | +3.99e-03 |
Pos logit contributions | |
Ant | +0.134 |
Next | +0.128 |
Fire | +0.126 |
hes | +0.126 |
Mouse | +0.126 |
Neg logit contributions | |
goodbye | -0.039 |
attention | -0.025 |
away | -0.024 |
ogging | -0.023 |
ly | -0.023 |
Token | and |
Token ID | 290 |
Feature activation | +1.85e-02 |
Pos logit contributions | |
Ant | +0.009 |
� | +0.008 |
Mouse | +0.008 |
Fire | +0.008 |
Really | +0.008 |
Neg logit contributions | |
goodbye | -0.005 |
attention | -0.004 |
back | -0.004 |
farewell | -0.004 |
ly | -0.004 |
Token | a |
Token ID | 257 |
Feature activation | +7.48e-03 |
Pos logit contributions | |
Ant | +0.037 |
� | +0.037 |
Mouse | +0.035 |
Salt | +0.035 |
kins | +0.034 |
Neg logit contributions | |
goodbye | -0.029 |
attention | -0.022 |
ly | -0.021 |
back | -0.021 |
closer | -0.021 |
Token | shelf |
Token ID | 18316 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
Ant | +0.012 |
Next | +0.012 |
� | +0.011 |
hes | +0.011 |
Salt | +0.011 |
Neg logit contributions | |
goodbye | -0.014 |
attention | -0.012 |
away | -0.011 |
farewell | -0.011 |
back | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | +1.74e-02 |
Pos logit contributions | |
Ant | +0.208 |
Salt | +0.198 |
Fire | +0.198 |
Next | +0.197 |
� | +0.197 |
Neg logit contributions | |
goodbye | -0.139 |
away | -0.117 |
attention | -0.116 |
ogging | -0.114 |
des | -0.109 |
Token | She |
Token ID | 1375 |
Feature activation | +4.04e-02 |
Pos logit contributions | |
Ant | +0.027 |
kins | +0.026 |
sibling | +0.024 |
Mouse | +0.024 |
poker | +0.024 |
Neg logit contributions | |
goodbye | -0.033 |
ogging | -0.028 |
attention | -0.027 |
farewell | -0.026 |
away | -0.026 |
Token | picked |
Token ID | 6497 |
Feature activation | +1.45e-02 |
Pos logit contributions | |
� | +0.077 |
poker | +0.074 |
Fire | +0.072 |
Next | +0.072 |
Salt | +0.071 |
Neg logit contributions | |
goodbye | -0.060 |
ly | -0.049 |
of | -0.049 |
back | -0.049 |
closer | -0.048 |
Token | something |
Token ID | 1223 |
Feature activation | -2.74e-02 |
Pos logit contributions | |
Ant | +0.027 |
Salt | +0.026 |
Next | +0.026 |
hes | +0.026 |
Fire | +0.026 |
Neg logit contributions | |
goodbye | -0.022 |
ogging | -0.019 |
away | -0.019 |
attention | -0.018 |
des | -0.018 |
Token | in |
Token ID | 287 |
Feature activation | +1.49e-02 |
Pos logit contributions | |
goodbye | +0.049 |
away | +0.043 |
attention | +0.042 |
ogging | +0.042 |
farewell | +0.041 |
Neg logit contributions | |
Ant | -0.044 |
Salt | -0.042 |
Next | -0.042 |
hes | -0.042 |
Fire | -0.041 |
Token | pictures |
Token ID | 5986 |
Feature activation | +6.45e-03 |
Pos logit contributions | |
Ant | +0.001 |
Fire | +0.001 |
brothers | +0.001 |
Salt | +0.001 |
Next | +0.001 |
Neg logit contributions | |
goodbye | -0.002 |
ogging | -0.002 |
away | -0.002 |
des | -0.002 |
farewell | -0.002 |
Token | . |
Token ID | 13 |
Feature activation | +1.01e-02 |
Pos logit contributions | |
Ant | +0.012 |
Next | +0.012 |
Salt | +0.012 |
Fire | +0.012 |
� | +0.012 |
Neg logit contributions | |
goodbye | -0.010 |
away | -0.008 |
attention | -0.008 |
ogging | -0.008 |
farewell | -0.008 |
Token |
|
Token ID | 198 |
Feature activation | +7.85e-03 |
Pos logit contributions | |
� | +0.017 |
Ant | +0.016 |
kins | +0.016 |
sisters | +0.016 |
oyd | +0.016 |
Neg logit contributions | |
goodbye | -0.017 |
away | -0.014 |
ogging | -0.013 |
farewell | -0.013 |
back | -0.013 |
Token |
|
Token ID | 198 |
Feature activation | +7.11e-03 |
Pos logit contributions | |
Ant | +0.010 |
Next | +0.009 |
� | +0.009 |
Salt | +0.009 |
Fire | +0.009 |
Neg logit contributions | |
goodbye | -0.017 |
away | -0.015 |
attention | -0.015 |
ogging | -0.015 |
farewell | -0.015 |
Token | " |
Token ID | 1 |
Feature activation | +4.39e-02 |
Pos logit contributions | |
Brush | +0.009 |
� | +0.009 |
Ant | +0.009 |
ities | +0.009 |
poker | +0.009 |
Neg logit contributions | |
goodbye | -0.013 |
away | -0.012 |
ogging | -0.011 |
attention | -0.011 |
of | -0.011 |
Token | These |
Token ID | 4711 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
poker | +0.062 |
ates | +0.062 |
brothers | +0.059 |
sibling | +0.059 |
Cere | +0.059 |
Neg logit contributions | |
away | -0.072 |
of | -0.068 |
closer | -0.068 |
goodbye | -0.067 |
contain | -0.067 |
Token | are |
Token ID | 389 |
Feature activation | +6.69e-02 |
Pos logit contributions | |
Ant | +0.158 |
Fire | +0.154 |
� | +0.153 |
Next | +0.152 |
Salt | +0.150 |
Neg logit contributions | |
goodbye | -0.186 |
ogging | -0.166 |
away | -0.165 |
des | -0.158 |
attention | -0.157 |
Token | terrible |
Token ID | 7818 |
Feature activation | +4.93e-02 |
Pos logit contributions | |
Ant | +0.130 |
� | +0.124 |
Next | +0.123 |
Salt | +0.122 |
Fire | +0.121 |
Neg logit contributions | |
goodbye | -0.100 |
attention | -0.085 |
away | -0.083 |
ogging | -0.080 |
farewell | -0.077 |
Token | ! |
Token ID | 0 |
Feature activation | +1.02e-02 |
Pos logit contributions | |
Ant | +0.095 |
Fire | +0.094 |
Next | +0.093 |
� | +0.093 |
Salt | +0.092 |
Neg logit contributions | |
goodbye | -0.071 |
away | -0.059 |
ogging | -0.059 |
attention | -0.057 |
des | -0.057 |
Token | You |
Token ID | 921 |
Feature activation | +4.11e-03 |
Pos logit contributions | |
Ant | +0.016 |
kins | +0.016 |
hes | +0.015 |
oyd | +0.015 |
iflower | +0.015 |
Neg logit contributions | |
goodbye | -0.018 |
ogging | -0.016 |
ly | -0.015 |
away | -0.015 |
farewell | -0.014 |
Token | ruined |
Token ID | 20484 |
Feature activation | +2.33e-02 |
Pos logit contributions | |
Ant | +0.008 |
� | +0.007 |
Really | +0.007 |
Next | +0.007 |
Salt | +0.007 |
Neg logit contributions | |
goodbye | -0.006 |
away | -0.005 |
attention | -0.005 |
contain | -0.005 |
ly | -0.005 |
Token | ?" |
Token ID | 1701 |
Feature activation | +5.83e-04 |
Pos logit contributions | |
Ant | +0.065 |
Next | +0.060 |
Salt | +0.060 |
Fire | +0.060 |
� | +0.060 |
Neg logit contributions | |
goodbye | -0.058 |
away | -0.051 |
attention | -0.050 |
ogging | -0.050 |
ly | -0.047 |
Token | she |
Token ID | 673 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
Ant | +0.001 |
� | +0.001 |
hes | +0.001 |
poker | +0.001 |
Cere | +0.001 |
Neg logit contributions | |
goodbye | -0.001 |
attention | -0.001 |
ogging | -0.001 |
away | -0.001 |
closer | -0.001 |
Token | asks |
Token ID | 7893 |
Feature activation | +1.49e-02 |
Pos logit contributions | |
goodbye | +0.037 |
attention | +0.034 |
ogging | +0.033 |
away | +0.032 |
ly | +0.032 |
Neg logit contributions | |
Ant | -0.031 |
� | -0.030 |
Salt | -0.028 |
Anything | -0.028 |
kins | -0.028 |
Token | . |
Token ID | 13 |
Feature activation | +1.79e-02 |
Pos logit contributions | |
Ant | +0.028 |
Salt | +0.027 |
Fire | +0.027 |
Next | +0.027 |
kins | +0.026 |
Neg logit contributions | |
goodbye | -0.022 |
away | -0.020 |
des | -0.019 |
ogging | -0.019 |
farewell | -0.018 |
Token | " |
Token ID | 366 |
Feature activation | +4.38e-02 |
Pos logit contributions | |
kins | +0.029 |
Ant | +0.029 |
chefs | +0.028 |
� | +0.028 |
ates | +0.028 |
Neg logit contributions | |
goodbye | -0.030 |
back | -0.025 |
closer | -0.025 |
attention | -0.024 |
ogging | -0.024 |
Token | These |
Token ID | 4711 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
sibling | +0.054 |
pap | +0.053 |
brothers | +0.053 |
sisters | +0.050 |
Mouse | +0.049 |
Neg logit contributions | |
away | -0.079 |
des | -0.076 |
contain | -0.071 |
of | -0.070 |
ogging | -0.069 |
Token | are |
Token ID | 389 |
Feature activation | +6.94e-02 |
Pos logit contributions | |
Ant | +0.165 |
hes | +0.157 |
Will | +0.156 |
Mouse | +0.155 |
Next | +0.154 |
Neg logit contributions | |
goodbye | -0.181 |
attention | -0.168 |
away | -0.160 |
ly | -0.156 |
back | -0.149 |
Token | not |
Token ID | 407 |
Feature activation | -5.14e-03 |
Pos logit contributions | |
� | +0.138 |
Ant | +0.132 |
Mouse | +0.132 |
Really | +0.128 |
Will | +0.127 |
Neg logit contributions | |
goodbye | -0.099 |
attention | -0.095 |
back | -0.083 |
away | -0.076 |
ly | -0.075 |
Token | toys |
Token ID | 14958 |
Feature activation | +6.27e-02 |
Pos logit contributions | |
goodbye | +0.006 |
attention | +0.006 |
away | +0.006 |
back | +0.005 |
ly | +0.005 |
Neg logit contributions | |
� | -0.012 |
Ant | -0.011 |
Mouse | -0.011 |
Will | -0.011 |
Really | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | +1.98e-02 |
Pos logit contributions | |
Ant | +0.122 |
Salt | +0.119 |
Fire | +0.119 |
hes | +0.117 |
� | +0.117 |
Neg logit contributions | |
goodbye | -0.088 |
farewell | -0.073 |
away | -0.073 |
ogging | -0.072 |
des | -0.071 |
Token | These |
Token ID | 2312 |
Feature activation | +9.64e-02 |
Pos logit contributions | |
Ant | +0.028 |
oyd | +0.028 |
kins | +0.027 |
iflower | +0.027 |
if | +0.026 |
Neg logit contributions | |
goodbye | -0.037 |
attention | -0.030 |
away | -0.029 |
ly | -0.029 |
farewell | -0.028 |
Token | Her |
Token ID | 9360 |
Feature activation | +1.79e-02 |
Pos logit contributions | |
goodbye | +0.030 |
away | +0.028 |
ogging | +0.027 |
farewell | +0.027 |
back | +0.026 |
Neg logit contributions | |
Ant | -0.021 |
hes | -0.021 |
poker | -0.021 |
chefs | -0.021 |
Salt | -0.020 |
Token | mother |
Token ID | 2802 |
Feature activation | +5.11e-03 |
Pos logit contributions | |
Ant | +0.016 |
Fire | +0.015 |
Next | +0.014 |
Salt | +0.014 |
� | +0.014 |
Neg logit contributions | |
goodbye | -0.045 |
away | -0.042 |
des | -0.041 |
ogging | -0.041 |
attention | -0.040 |
Token | replied |
Token ID | 8712 |
Feature activation | -7.61e-03 |
Pos logit contributions | |
Ant | +0.010 |
� | +0.010 |
hes | +0.009 |
Salt | +0.009 |
Next | +0.009 |
Neg logit contributions | |
goodbye | -0.008 |
ly | -0.006 |
ogging | -0.006 |
attention | -0.006 |
away | -0.006 |
Token | , |
Token ID | 11 |
Feature activation | -6.20e-03 |
Pos logit contributions | |
goodbye | +0.008 |
away | +0.007 |
attention | +0.007 |
des | +0.007 |
ogging | +0.007 |
Neg logit contributions | |
Ant | -0.017 |
� | -0.017 |
Salt | -0.016 |
Fire | -0.016 |
Next | -0.016 |
Token | " |
Token ID | 366 |
Feature activation | +5.89e-02 |
Pos logit contributions | |
goodbye | +0.012 |
back | +0.010 |
ly | +0.009 |
closer | +0.009 |
to | +0.009 |
Neg logit contributions | |
Ant | -0.010 |
Mouse | -0.009 |
poker | -0.009 |
Bees | -0.009 |
chefs | -0.009 |
Token | These |
Token ID | 4711 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
kins | +0.069 |
poker | +0.066 |
kin | +0.061 |
brothers | +0.060 |
ates | +0.060 |
Neg logit contributions | |
away | -0.117 |
of | -0.115 |
des | -0.112 |
goodbye | -0.110 |
ogging | -0.109 |
Token | are |
Token ID | 389 |
Feature activation | +7.90e-02 |
Pos logit contributions | |
Ant | +0.187 |
Salt | +0.173 |
Next | +0.173 |
Bees | +0.171 |
Mouse | +0.171 |
Neg logit contributions | |
goodbye | -0.165 |
attention | -0.153 |
away | -0.142 |
ly | -0.135 |
back | -0.132 |
Token | special |
Token ID | 2041 |
Feature activation | +7.10e-02 |
Pos logit contributions | |
Ant | +0.150 |
� | +0.142 |
Mouse | +0.141 |
Next | +0.141 |
Really | +0.138 |
Neg logit contributions | |
goodbye | -0.122 |
attention | -0.112 |
back | -0.098 |
away | -0.098 |
ly | -0.093 |
Token | keys |
Token ID | 8251 |
Feature activation | +3.98e-02 |
Pos logit contributions | |
Ant | +0.115 |
hes | +0.104 |
Salt | +0.103 |
Mouse | +0.103 |
Next | +0.103 |
Neg logit contributions | |
goodbye | -0.135 |
attention | -0.124 |
away | -0.120 |
ogging | -0.115 |
ly | -0.114 |
Token | . |
Token ID | 13 |
Feature activation | +6.26e-03 |
Pos logit contributions | |
Ant | +0.080 |
� | +0.077 |
Salt | +0.077 |
Fire | +0.077 |
Next | +0.077 |
Neg logit contributions | |
goodbye | -0.054 |
away | -0.045 |
ogging | -0.044 |
attention | -0.044 |
farewell | -0.043 |
Token | They |
Token ID | 1119 |
Feature activation | +1.78e-02 |
Pos logit contributions | |
kins | +0.011 |
Ant | +0.010 |
hes | +0.009 |
oyd | +0.009 |
kin | +0.009 |
Neg logit contributions | |
goodbye | -0.011 |
attention | -0.009 |
ly | -0.009 |
farewell | -0.009 |
ogging | -0.009 |
Token | a |
Token ID | 257 |
Feature activation | -3.87e-03 |
Pos logit contributions | |
Ant | +0.042 |
� | +0.040 |
Salt | +0.039 |
Next | +0.039 |
Fire | +0.039 |
Neg logit contributions | |
goodbye | -0.049 |
away | -0.042 |
attention | -0.042 |
ogging | -0.042 |
des | -0.040 |
Token | small |
Token ID | 1402 |
Feature activation | +1.84e-02 |
Pos logit contributions | |
goodbye | +0.006 |
away | +0.006 |
ogging | +0.005 |
back | +0.005 |
farewell | +0.005 |
Neg logit contributions | |
Next | -0.007 |
hes | -0.007 |
Ant | -0.006 |
� | -0.006 |
kins | -0.006 |
Token | bird |
Token ID | 6512 |
Feature activation | +2.47e-02 |
Pos logit contributions | |
hes | +0.026 |
Ant | +0.024 |
Next | +0.024 |
kins | +0.023 |
Sal | +0.023 |
Neg logit contributions | |
goodbye | -0.039 |
away | -0.035 |
ogging | -0.034 |
ly | -0.033 |
back | -0.033 |
Token | , |
Token ID | 11 |
Feature activation | -2.40e-02 |
Pos logit contributions | |
Ant | +0.034 |
Next | +0.031 |
Salt | +0.031 |
Fire | +0.031 |
� | +0.031 |
Neg logit contributions | |
goodbye | -0.051 |
attention | -0.045 |
away | -0.045 |
ogging | -0.044 |
farewell | -0.043 |
Token | a |
Token ID | 257 |
Feature activation | -1.60e-03 |
Pos logit contributions | |
goodbye | +0.036 |
closer | +0.029 |
back | +0.028 |
away | +0.027 |
farewell | +0.026 |
Neg logit contributions | |
� | -0.045 |
chefs | -0.044 |
Cere | -0.044 |
sibling | -0.043 |
ities | -0.043 |
Token | rob |
Token ID | 3857 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
goodbye | +0.003 |
away | +0.002 |
back | +0.002 |
closer | +0.002 |
farewell | +0.002 |
Neg logit contributions | |
Next | -0.003 |
hes | -0.003 |
� | -0.003 |
kins | -0.003 |
Ant | -0.003 |
Token | in |
Token ID | 259 |
Feature activation | +5.91e-02 |
Pos logit contributions | |
� | +0.172 |
Next | +0.169 |
Salt | +0.161 |
Cere | +0.161 |
Blocks | +0.160 |
Neg logit contributions | |
goodbye | -0.164 |
away | -0.163 |
ogging | -0.152 |
ly | -0.149 |
back | -0.141 |
Token | , |
Token ID | 11 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
Next | +0.125 |
Ant | +0.124 |
Salt | +0.124 |
� | +0.124 |
hes | +0.123 |
Neg logit contributions | |
goodbye | -0.071 |
away | -0.060 |
ly | -0.055 |
farewell | -0.051 |
ogging | -0.051 |
Token | who |
Token ID | 508 |
Feature activation | +6.89e-02 |
Pos logit contributions | |
goodbye | +0.033 |
back | +0.027 |
away | +0.027 |
ogging | +0.026 |
closer | +0.026 |
Neg logit contributions | |
� | -0.027 |
Cere | -0.027 |
kins | -0.027 |
chefs | -0.027 |
Salt | -0.027 |
Token | wanted |
Token ID | 2227 |
Feature activation | -9.62e-03 |
Pos logit contributions | |
chefs | +0.096 |
ters | +0.095 |
kins | +0.090 |
Really | +0.089 |
poker | +0.088 |
Neg logit contributions | |
goodbye | -0.129 |
away | -0.124 |
des | -0.118 |
farewell | -0.114 |
ly | -0.113 |
Token | to |
Token ID | 284 |
Feature activation | -6.51e-03 |
Pos logit contributions | |
goodbye | +0.010 |
ogging | +0.009 |
away | +0.008 |
des | +0.007 |
fy | +0.007 |
Neg logit contributions | |
Ant | -0.022 |
Fire | -0.022 |
Salt | -0.022 |
Next | -0.022 |
Will | -0.021 |
Token | not |
Token ID | 407 |
Feature activation | -1.58e-03 |
Pos logit contributions | |
Ant | +0.160 |
� | +0.157 |
Mouse | +0.153 |
Will | +0.152 |
Next | +0.150 |
Neg logit contributions | |
goodbye | -0.121 |
attention | -0.105 |
away | -0.093 |
back | -0.092 |
farewell | -0.090 |
Token | pe |
Token ID | 613 |
Feature activation | +3.86e-02 |
Pos logit contributions | |
goodbye | +0.002 |
attention | +0.002 |
away | +0.002 |
ogging | +0.002 |
back | +0.002 |
Neg logit contributions | |
Ant | -0.003 |
� | -0.003 |
Next | -0.003 |
Mouse | -0.003 |
Will | -0.003 |
Token | b |
Token ID | 65 |
Feature activation | +6.21e-02 |
Pos logit contributions | |
Ant | +0.058 |
kins | +0.054 |
hes | +0.054 |
Salt | +0.053 |
� | +0.052 |
Neg logit contributions | |
goodbye | -0.077 |
farewell | -0.065 |
away | -0.064 |
ogging | -0.064 |
des | -0.063 |
Token | bles |
Token ID | 7689 |
Feature activation | +2.74e-02 |
Pos logit contributions | |
Ant | +0.068 |
Next | +0.062 |
Salt | +0.061 |
Fire | +0.061 |
� | +0.061 |
Neg logit contributions | |
goodbye | -0.144 |
away | -0.130 |
attention | -0.129 |
ogging | -0.128 |
farewell | -0.125 |
Token | . |
Token ID | 13 |
Feature activation | +2.38e-02 |
Pos logit contributions | |
Ant | +0.059 |
Salt | +0.056 |
� | +0.056 |
Fire | +0.056 |
Next | +0.056 |
Neg logit contributions | |
goodbye | -0.034 |
away | -0.027 |
attention | -0.027 |
ogging | -0.026 |
farewell | -0.026 |
Token | These |
Token ID | 2312 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
Ant | +0.035 |
kins | +0.034 |
hes | +0.033 |
iflower | +0.032 |
oyd | +0.032 |
Neg logit contributions | |
goodbye | -0.044 |
attention | -0.037 |
away | -0.036 |
ly | -0.036 |
ogging | -0.035 |
Token | are |
Token ID | 389 |
Feature activation | +7.70e-02 |
Pos logit contributions | |
Ant | +0.183 |
� | +0.175 |
Next | +0.175 |
hes | +0.174 |
Salt | +0.173 |
Neg logit contributions | |
goodbye | -0.159 |
attention | -0.141 |
away | -0.139 |
ogging | -0.134 |
ly | -0.129 |
Token | surprise |
Token ID | 5975 |
Feature activation | +5.73e-02 |
Pos logit contributions | |
� | +0.149 |
Ant | +0.144 |
Mouse | +0.140 |
Next | +0.139 |
Will | +0.136 |
Neg logit contributions | |
goodbye | -0.114 |
attention | -0.107 |
back | -0.095 |
away | -0.093 |
ly | -0.087 |
Token | eggs |
Token ID | 9653 |
Feature activation | +5.15e-02 |
Pos logit contributions | |
Ant | +0.125 |
� | +0.119 |
Next | +0.119 |
Salt | +0.118 |
Fire | +0.118 |
Neg logit contributions | |
goodbye | -0.071 |
attention | -0.060 |
away | -0.058 |
ogging | -0.056 |
ly | -0.053 |
Token | !" |
Token ID | 2474 |
Feature activation | +6.50e-03 |
Pos logit contributions | |
Ant | +0.113 |
Next | +0.107 |
� | +0.107 |
Fire | +0.107 |
Salt | +0.107 |
Neg logit contributions | |
goodbye | -0.063 |
attention | -0.053 |
away | -0.052 |
ogging | -0.050 |
ly | -0.047 |
Token | she |
Token ID | 673 |
Feature activation | -2.26e-02 |
Pos logit contributions | |
Ant | +0.009 |
� | +0.009 |
Mouse | +0.008 |
Salt | +0.008 |
Fire | +0.008 |
Neg logit contributions | |
goodbye | -0.013 |
attention | -0.011 |
away | -0.011 |
ogging | -0.011 |
farewell | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | +1.36e-02 |
Pos logit contributions | |
Ant | +0.012 |
Mouse | +0.012 |
Next | +0.012 |
� | +0.011 |
hes | +0.011 |
Neg logit contributions | |
goodbye | -0.006 |
attention | -0.005 |
away | -0.004 |
ly | -0.004 |
ogging | -0.004 |
Token | They |
Token ID | 1119 |
Feature activation | +1.52e-03 |
Pos logit contributions | |
Ant | +0.019 |
kins | +0.018 |
Mouse | +0.016 |
iflower | +0.016 |
cut | +0.016 |
Neg logit contributions | |
goodbye | -0.029 |
attention | -0.025 |
away | -0.024 |
farewell | -0.024 |
back | -0.023 |
Token | say |
Token ID | 910 |
Feature activation | -4.57e-03 |
Pos logit contributions | |
Ant | +0.003 |
Next | +0.003 |
� | +0.003 |
Fire | +0.003 |
Salt | +0.003 |
Neg logit contributions | |
goodbye | -0.002 |
attention | -0.002 |
contain | -0.002 |
away | -0.002 |
farewell | -0.001 |
Token | , |
Token ID | 11 |
Feature activation | -4.95e-03 |
Pos logit contributions | |
goodbye | +0.005 |
away | +0.004 |
attention | +0.004 |
ogging | +0.004 |
des | +0.004 |
Neg logit contributions | |
Ant | -0.010 |
Fire | -0.010 |
Salt | -0.010 |
� | -0.010 |
Next | -0.010 |
Token | " |
Token ID | 366 |
Feature activation | +4.36e-02 |
Pos logit contributions | |
goodbye | +0.010 |
back | +0.008 |
farewell | +0.008 |
attention | +0.008 |
closer | +0.008 |
Neg logit contributions | |
Ant | -0.007 |
Mouse | -0.007 |
oyd | -0.007 |
kins | -0.006 |
chefs | -0.006 |
Token | These |
Token ID | 4711 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
kins | +0.055 |
brothers | +0.055 |
poker | +0.055 |
sisters | +0.051 |
Mouse | +0.050 |
Neg logit contributions | |
away | -0.086 |
des | -0.085 |
goodbye | -0.084 |
land | -0.081 |
of | -0.079 |
Token | are |
Token ID | 389 |
Feature activation | +6.81e-02 |
Pos logit contributions | |
Ant | +0.164 |
� | +0.154 |
Next | +0.154 |
Fire | +0.153 |
Salt | +0.153 |
Neg logit contributions | |
goodbye | -0.182 |
away | -0.160 |
ogging | -0.157 |
attention | -0.157 |
farewell | -0.151 |
Token | for |
Token ID | 329 |
Feature activation | +6.55e-02 |
Pos logit contributions | |
Ant | +0.146 |
� | +0.145 |
Mouse | +0.141 |
Next | +0.139 |
Really | +0.137 |
Neg logit contributions | |
goodbye | -0.090 |
attention | -0.081 |
back | -0.068 |
away | -0.066 |
farewell | -0.062 |
Token | Grand |
Token ID | 5675 |
Feature activation | +2.23e-03 |
Pos logit contributions | |
Ant | +0.132 |
Mouse | +0.129 |
Bees | +0.124 |
� | +0.123 |
hes | +0.122 |
Neg logit contributions | |
goodbye | -0.091 |
attention | -0.081 |
back | -0.071 |
away | -0.071 |
ogging | -0.070 |
Token | ma |
Token ID | 2611 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
Ant | +0.003 |
Next | +0.002 |
Salt | +0.002 |
� | +0.002 |
Fire | +0.002 |
Neg logit contributions | |
goodbye | -0.005 |
away | -0.004 |
attention | -0.004 |
ogging | -0.004 |
farewell | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +1.73e-02 |
Pos logit contributions | |
Ant | +0.032 |
Next | +0.031 |
� | +0.030 |
Fire | +0.030 |
Salt | +0.030 |
Neg logit contributions | |
goodbye | -0.018 |
attention | -0.015 |
away | -0.015 |
ogging | -0.014 |
ly | -0.013 |
Token | friends |
Token ID | 2460 |
Feature activation | +6.16e-03 |
Pos logit contributions | |
Ant | +0.007 |
Fire | +0.007 |
Next | +0.007 |
Salt | +0.007 |
� | +0.007 |
Neg logit contributions | |
goodbye | -0.012 |
away | -0.011 |
attention | -0.011 |
ogging | -0.011 |
des | -0.011 |
Token | laughed |
Token ID | 13818 |
Feature activation | -4.45e-02 |
Pos logit contributions | |
Ant | +0.009 |
Next | +0.008 |
Salt | +0.008 |
� | +0.008 |
Fire | +0.008 |
Neg logit contributions | |
goodbye | -0.012 |
away | -0.011 |
attention | -0.011 |
ogging | -0.010 |
farewell | -0.010 |
Token | at |
Token ID | 379 |
Feature activation | -1.06e-02 |
Pos logit contributions | |
goodbye | +0.066 |
des | +0.065 |
ogging | +0.064 |
away | +0.064 |
attention | +0.063 |
Neg logit contributions | |
Ant | -0.073 |
Next | -0.073 |
Fire | -0.072 |
� | -0.072 |
Salt | -0.071 |
Token | their |
Token ID | 511 |
Feature activation | -2.13e-02 |
Pos logit contributions | |
goodbye | +0.019 |
attention | +0.017 |
away | +0.017 |
farewell | +0.016 |
ogging | +0.015 |
Neg logit contributions | |
Ant | -0.017 |
hes | -0.017 |
Next | -0.016 |
� | -0.016 |
Salt | -0.016 |
Token | cle |
Token ID | 1190 |
Feature activation | -5.42e-02 |
Pos logit contributions | |
goodbye | +0.040 |
ogging | +0.038 |
away | +0.037 |
des | +0.037 |
ly | +0.035 |
Neg logit contributions | |
Ant | -0.030 |
Fire | -0.029 |
Next | -0.028 |
brothers | -0.028 |
sibling | -0.028 |
Token | vern |
Token ID | 933 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
goodbye | +0.085 |
away | +0.068 |
attention | +0.065 |
des | +0.065 |
ly | +0.063 |
Neg logit contributions | |
kins | -0.098 |
hes | -0.094 |
itt | -0.093 |
Ant | -0.092 |
ates | -0.090 |
Token | ess |
Token ID | 408 |
Feature activation | -8.19e-03 |
Pos logit contributions | |
goodbye | +0.216 |
ogging | +0.192 |
des | +0.190 |
away | +0.179 |
attention | +0.174 |
Neg logit contributions | |
Ant | -0.134 |
sibling | -0.127 |
Next | -0.122 |
abies | -0.120 |
ters | -0.119 |
Token | and |
Token ID | 290 |
Feature activation | +7.76e-03 |
Pos logit contributions | |
goodbye | +0.011 |
away | +0.010 |
des | +0.010 |
ogging | +0.010 |
attention | +0.009 |
Neg logit contributions | |
Ant | -0.015 |
Salt | -0.015 |
Fire | -0.015 |
� | -0.015 |
Next | -0.014 |
Token | went |
Token ID | 1816 |
Feature activation | -4.01e-02 |
Pos logit contributions | |
Ant | +0.013 |
Next | +0.013 |
� | +0.012 |
Salt | +0.012 |
Fire | +0.012 |
Neg logit contributions | |
goodbye | -0.014 |
attention | -0.012 |
farewell | -0.012 |
away | -0.012 |
back | -0.011 |
Token | off |
Token ID | 572 |
Feature activation | -2.65e-02 |
Pos logit contributions | |
des | +0.050 |
ly | +0.044 |
away | +0.043 |
goodbye | +0.042 |
ong | +0.042 |
Neg logit contributions | |
hes | -0.075 |
twins | -0.071 |
brothers | -0.071 |
Brush | -0.070 |
Mouse | -0.069 |
Token | to |
Token ID | 284 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
des | +0.033 |
away | +0.033 |
goodbye | +0.033 |
ly | +0.033 |
attention | +0.031 |
Neg logit contributions | |
Ant | -0.047 |
Fire | -0.047 |
Mouse | -0.047 |
Salt | -0.046 |
hes | -0.046 |
Token | or |
Token ID | 393 |
Feature activation | +2.92e-02 |
Pos logit contributions | |
goodbye | +0.040 |
away | +0.034 |
ogging | +0.034 |
attention | +0.034 |
des | +0.033 |
Neg logit contributions | |
Ant | -0.056 |
Fire | -0.055 |
Next | -0.055 |
Salt | -0.055 |
� | -0.054 |
Token | we |
Token ID | 356 |
Feature activation | -3.51e-02 |
Pos logit contributions | |
Ant | +0.051 |
� | +0.048 |
Next | +0.048 |
Salt | +0.048 |
Fire | +0.048 |
Neg logit contributions | |
goodbye | -0.049 |
attention | -0.042 |
away | -0.042 |
ogging | -0.040 |
farewell | -0.039 |
Token | won |
Token ID | 1839 |
Feature activation | -6.77e-02 |
Pos logit contributions | |
goodbye | +0.066 |
away | +0.058 |
attention | +0.058 |
ogging | +0.056 |
farewell | +0.055 |
Neg logit contributions | |
Ant | -0.054 |
Next | -0.051 |
Salt | -0.051 |
� | -0.050 |
Fire | -0.050 |
Token | â |
Token ID | 22940 |
Feature activation | -5.91e-02 |
Pos logit contributions | |
goodbye | +0.109 |
away | +0.094 |
attention | +0.094 |
ogging | +0.092 |
farewell | +0.088 |
Neg logit contributions | |
Ant | -0.120 |
Next | -0.114 |
� | -0.114 |
Salt | -0.112 |
hes | -0.112 |
Token | € |
Token ID | 26391 |
Feature activation | -6.41e-02 |
Pos logit contributions | |
goodbye | +0.097 |
attention | +0.084 |
away | +0.082 |
ogging | +0.080 |
farewell | +0.077 |
Neg logit contributions | |
Ant | -0.104 |
� | -0.101 |
Next | -0.099 |
hes | -0.098 |
Salt | -0.098 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.04e-01 |
Pos logit contributions | |
goodbye | +0.130 |
away | +0.118 |
attention | +0.117 |
ogging | +0.114 |
farewell | +0.111 |
Neg logit contributions | |
Ant | -0.087 |
Next | -0.082 |
� | -0.081 |
Fire | -0.081 |
Salt | -0.081 |
Token | t |
Token ID | 83 |
Feature activation | -7.96e-02 |
Pos logit contributions | |
goodbye | +0.171 |
away | +0.159 |
attention | +0.154 |
farewell | +0.148 |
des | +0.146 |
Neg logit contributions | |
Ant | -0.168 |
I | -0.164 |
hes | -0.162 |
Fire | -0.161 |
Salt | -0.161 |
Token | be |
Token ID | 307 |
Feature activation | -9.77e-03 |
Pos logit contributions | |
goodbye | +0.154 |
farewell | +0.126 |
des | +0.123 |
attention | +0.121 |
uits | +0.118 |
Neg logit contributions | |
Ant | -0.130 |
I | -0.119 |
Mouse | -0.117 |
Mouse | -0.115 |
Will | -0.112 |
Token | able |
Token ID | 1498 |
Feature activation | -2.02e-02 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.016 |
ogging | +0.016 |
attention | +0.016 |
des | +0.016 |
Neg logit contributions | |
Ant | -0.015 |
hes | -0.014 |
Fire | -0.014 |
Salt | -0.014 |
Next | -0.014 |
Token | to |
Token ID | 284 |
Feature activation | -1.73e-02 |
Pos logit contributions | |
goodbye | +0.025 |
away | +0.023 |
ogging | +0.023 |
farewell | +0.022 |
of | +0.022 |
Neg logit contributions | |
Fire | -0.038 |
twins | -0.037 |
chefs | -0.037 |
brothers | -0.036 |
Next | -0.036 |
Token | breathe |
Token ID | 18044 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
goodbye | +0.028 |
ogging | +0.025 |
ly | +0.023 |
attention | +0.023 |
des | +0.023 |
Neg logit contributions | |
Ant | -0.031 |
Fire | -0.029 |
Mouse | -0.028 |
Salt | -0.028 |
Mouse | -0.028 |
Token | if |
Token ID | 611 |
Feature activation | +3.99e-02 |
Pos logit contributions | |
Ant | +0.013 |
Salt | +0.012 |
Next | +0.012 |
� | +0.012 |
Fire | +0.012 |
Neg logit contributions | |
goodbye | -0.010 |
ogging | -0.009 |
away | -0.008 |
des | -0.008 |
attention | -0.008 |
Token | you |
Token ID | 345 |
Feature activation | -5.74e-02 |
Pos logit contributions | |
Ant | +0.075 |
Mouse | +0.072 |
hes | +0.072 |
Sal | +0.070 |
poker | +0.069 |
Neg logit contributions | |
goodbye | -0.063 |
attention | -0.056 |
ogging | -0.052 |
away | -0.050 |
ly | -0.047 |
Token | don |
Token ID | 836 |
Feature activation | -4.13e-02 |
Pos logit contributions | |
goodbye | +0.084 |
attention | +0.073 |
away | +0.072 |
ogging | +0.069 |
des | +0.067 |
Neg logit contributions | |
Ant | -0.110 |
Next | -0.105 |
� | -0.105 |
Fire | -0.104 |
Salt | -0.104 |
Token | â |
Token ID | 22940 |
Feature activation | -6.18e-02 |
Pos logit contributions | |
goodbye | +0.080 |
ogging | +0.072 |
away | +0.071 |
attention | +0.071 |
farewell | +0.068 |
Neg logit contributions | |
Ant | -0.060 |
Next | -0.056 |
Fire | -0.055 |
Salt | -0.055 |
� | -0.055 |
Token | € |
Token ID | 26391 |
Feature activation | -6.12e-02 |
Pos logit contributions | |
goodbye | +0.097 |
away | +0.084 |
ogging | +0.084 |
attention | +0.083 |
des | +0.080 |
Neg logit contributions | |
Ant | -0.114 |
Fire | -0.108 |
Salt | -0.107 |
Next | -0.106 |
Will | -0.105 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
goodbye | +0.112 |
ogging | +0.099 |
des | +0.091 |
attention | +0.089 |
farewell | +0.088 |
Neg logit contributions | |
Ant | -0.100 |
Mouse | -0.092 |
poker | -0.091 |
hes | -0.089 |
Bear | -0.087 |
Token | t |
Token ID | 83 |
Feature activation | -6.27e-02 |
Pos logit contributions | |
goodbye | +0.164 |
away | +0.144 |
attention | +0.144 |
farewell | +0.134 |
des | +0.131 |
Neg logit contributions | |
Ant | -0.185 |
� | -0.177 |
Salt | -0.176 |
Next | -0.175 |
I | -0.174 |
Token | know |
Token ID | 760 |
Feature activation | +5.94e-03 |
Pos logit contributions | |
goodbye | +0.091 |
away | +0.078 |
attention | +0.078 |
ogging | +0.076 |
farewell | +0.072 |
Neg logit contributions | |
Ant | -0.122 |
Next | -0.116 |
� | -0.116 |
Fire | -0.116 |
Salt | -0.115 |
Token | what |
Token ID | 644 |
Feature activation | +1.36e-02 |
Pos logit contributions | |
Ant | +0.011 |
Next | +0.010 |
� | +0.010 |
Fire | +0.010 |
hes | +0.010 |
Neg logit contributions | |
goodbye | -0.009 |
attention | -0.008 |
away | -0.008 |
ogging | -0.008 |
farewell | -0.007 |
Token | â |
Token ID | 22940 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
Ant | +0.028 |
Mouse | +0.027 |
hes | +0.026 |
Next | +0.025 |
Salt | +0.025 |
Neg logit contributions | |
goodbye | -0.019 |
attention | -0.016 |
ogging | -0.015 |
away | -0.015 |
back | -0.014 |
Token | € |
Token ID | 26391 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.017 |
attention | +0.017 |
ogging | +0.017 |
des | +0.016 |
Neg logit contributions | |
Ant | -0.025 |
Fire | -0.023 |
Next | -0.023 |
Salt | -0.023 |
� | -0.023 |
Token | that |
Token ID | 326 |
Feature activation | +2.32e-02 |
Pos logit contributions | |
goodbye | +0.061 |
away | +0.053 |
ogging | +0.052 |
attention | +0.050 |
des | +0.048 |
Neg logit contributions | |
Ant | -0.099 |
Salt | -0.095 |
Next | -0.095 |
� | -0.095 |
I | -0.094 |
Token | she |
Token ID | 673 |
Feature activation | -2.83e-02 |
Pos logit contributions | |
Ant | +0.045 |
� | +0.043 |
Mouse | +0.042 |
poker | +0.042 |
Next | +0.042 |
Neg logit contributions | |
goodbye | -0.038 |
attention | -0.032 |
ogging | -0.031 |
away | -0.029 |
des | -0.029 |
Token | wouldn |
Token ID | 3636 |
Feature activation | -6.10e-02 |
Pos logit contributions | |
goodbye | +0.054 |
attention | +0.046 |
away | +0.044 |
ly | +0.044 |
back | +0.044 |
Neg logit contributions | |
Ant | -0.046 |
Salt | -0.042 |
Fire | -0.042 |
� | -0.041 |
Next | -0.041 |
Token | â |
Token ID | 22940 |
Feature activation | -5.98e-02 |
Pos logit contributions | |
goodbye | +0.129 |
attention | +0.115 |
away | +0.115 |
ogging | +0.112 |
des | +0.110 |
Neg logit contributions | |
Ant | -0.080 |
� | -0.074 |
Salt | -0.073 |
Next | -0.073 |
Fire | -0.073 |
Token | € |
Token ID | 26391 |
Feature activation | -6.58e-02 |
Pos logit contributions | |
goodbye | +0.095 |
attention | +0.081 |
away | +0.081 |
ogging | +0.079 |
farewell | +0.076 |
Neg logit contributions | |
Ant | -0.109 |
� | -0.104 |
Next | -0.104 |
Salt | -0.103 |
hes | -0.102 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
goodbye | +0.133 |
ogging | +0.115 |
away | +0.114 |
attention | +0.114 |
des | +0.111 |
Neg logit contributions | |
Ant | -0.095 |
Fire | -0.087 |
Salt | -0.087 |
Next | -0.087 |
hes | -0.086 |
Token | t |
Token ID | 83 |
Feature activation | -7.28e-02 |
Pos logit contributions | |
goodbye | +0.157 |
away | +0.143 |
attention | +0.140 |
farewell | +0.132 |
des | +0.129 |
Neg logit contributions | |
Ant | -0.182 |
I | -0.178 |
Salt | -0.174 |
� | -0.174 |
Fire | -0.171 |
Token | disturb |
Token ID | 17037 |
Feature activation | -8.83e-03 |
Pos logit contributions | |
goodbye | +0.125 |
away | +0.106 |
attention | +0.104 |
ogging | +0.103 |
farewell | +0.102 |
Neg logit contributions | |
Ant | -0.124 |
Salt | -0.115 |
I | -0.114 |
Next | -0.114 |
� | -0.114 |
Token | the |
Token ID | 262 |
Feature activation | -2.69e-02 |
Pos logit contributions | |
goodbye | +0.014 |
attention | +0.012 |
away | +0.012 |
ogging | +0.011 |
ly | +0.011 |
Neg logit contributions | |
Ant | -0.017 |
� | -0.016 |
hes | -0.016 |
Next | -0.016 |
Fire | -0.016 |
Token | trophy |
Token ID | 16383 |
Feature activation | +1.07e-02 |
Pos logit contributions | |
goodbye | +0.048 |
away | +0.042 |
attention | +0.042 |
ogging | +0.041 |
farewell | +0.039 |
Neg logit contributions | |
Ant | -0.044 |
Next | -0.042 |
Salt | -0.041 |
� | -0.041 |
Fire | -0.041 |
Token | . |
Token ID | 13 |
Feature activation | -8.42e-04 |
Pos logit contributions | |
Ant | +0.023 |
Next | +0.022 |
Fire | +0.022 |
Salt | +0.022 |
� | +0.022 |
Neg logit contributions | |
goodbye | -0.013 |
away | -0.011 |
attention | -0.011 |
ogging | -0.010 |
farewell | -0.010 |
Token | time |
Token ID | 640 |
Feature activation | -1.10e-02 |
Pos logit contributions | |
goodbye | +0.083 |
away | +0.073 |
attention | +0.073 |
ogging | +0.072 |
farewell | +0.069 |
Neg logit contributions | |
Ant | -0.060 |
Next | -0.055 |
Salt | -0.055 |
hes | -0.055 |
Fire | -0.055 |
Token | she |
Token ID | 673 |
Feature activation | -1.74e-02 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.016 |
attention | +0.016 |
ogging | +0.016 |
farewell | +0.015 |
Neg logit contributions | |
Ant | -0.019 |
Next | -0.018 |
Salt | -0.018 |
� | -0.018 |
Fire | -0.018 |
Token | didn |
Token ID | 1422 |
Feature activation | -5.57e-02 |
Pos logit contributions | |
goodbye | +0.034 |
away | +0.029 |
ogging | +0.029 |
attention | +0.028 |
back | +0.028 |
Neg logit contributions | |
Ant | -0.027 |
Salt | -0.025 |
poker | -0.025 |
Next | -0.025 |
Fire | -0.025 |
Token | â |
Token ID | 22940 |
Feature activation | -5.60e-02 |
Pos logit contributions | |
goodbye | +0.121 |
attention | +0.109 |
away | +0.108 |
ogging | +0.105 |
des | +0.104 |
Neg logit contributions | |
Ant | -0.068 |
� | -0.065 |
Next | -0.064 |
Fire | -0.064 |
Salt | -0.064 |
Token | € |
Token ID | 26391 |
Feature activation | -6.39e-02 |
Pos logit contributions | |
goodbye | +0.086 |
away | +0.073 |
attention | +0.073 |
ogging | +0.071 |
farewell | +0.068 |
Neg logit contributions | |
Ant | -0.106 |
Next | -0.100 |
Salt | -0.100 |
Fire | -0.100 |
� | -0.100 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.03e-01 |
Pos logit contributions | |
goodbye | +0.122 |
ogging | +0.106 |
away | +0.104 |
attention | +0.103 |
des | +0.101 |
Neg logit contributions | |
Ant | -0.099 |
Salt | -0.090 |
Next | -0.090 |
Fire | -0.090 |
hes | -0.089 |
Token | t |
Token ID | 83 |
Feature activation | -7.29e-02 |
Pos logit contributions | |
goodbye | +0.159 |
away | +0.145 |
attention | +0.141 |
farewell | +0.134 |
des | +0.132 |
Neg logit contributions | |
Ant | -0.178 |
I | -0.173 |
� | -0.173 |
Salt | -0.171 |
hes | -0.170 |
Token | feel |
Token ID | 1254 |
Feature activation | -1.12e-02 |
Pos logit contributions | |
goodbye | +0.102 |
away | +0.086 |
attention | +0.085 |
ogging | +0.083 |
farewell | +0.079 |
Neg logit contributions | |
Ant | -0.147 |
Next | -0.141 |
Salt | -0.140 |
� | -0.140 |
Fire | -0.140 |
Token | so |
Token ID | 523 |
Feature activation | -9.35e-04 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.016 |
attention | +0.016 |
ogging | +0.015 |
farewell | +0.015 |
Neg logit contributions | |
Ant | -0.020 |
Next | -0.019 |
Salt | -0.019 |
Fire | -0.019 |
� | -0.019 |
Token | lonely |
Token ID | 21757 |
Feature activation | +1.03e-02 |
Pos logit contributions | |
goodbye | +0.002 |
away | +0.001 |
attention | +0.001 |
ogging | +0.001 |
des | +0.001 |
Neg logit contributions | |
Ant | -0.001 |
Next | -0.001 |
hes | -0.001 |
� | -0.001 |
Fire | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | -3.25e-03 |
Pos logit contributions | |
Ant | +0.020 |
� | +0.019 |
Salt | +0.019 |
Will | +0.019 |
Fire | +0.019 |
Neg logit contributions | |
goodbye | -0.014 |
attention | -0.012 |
away | -0.012 |
ogging | -0.011 |
des | -0.011 |
Token | the |
Token ID | 262 |
Feature activation | +7.28e-03 |
Pos logit contributions | |
goodbye | +0.002 |
away | +0.002 |
attention | +0.002 |
ogging | +0.002 |
farewell | +0.002 |
Neg logit contributions | |
Ant | -0.002 |
� | -0.002 |
Mouse | -0.002 |
Next | -0.002 |
Bees | -0.002 |
Token | cat |
Token ID | 3797 |
Feature activation | +2.09e-02 |
Pos logit contributions | |
Ant | +0.010 |
kins | +0.009 |
� | +0.009 |
hes | +0.008 |
poker | +0.008 |
Neg logit contributions | |
goodbye | -0.016 |
attention | -0.015 |
ogging | -0.014 |
back | -0.013 |
ly | -0.013 |
Token | wouldn |
Token ID | 3636 |
Feature activation | -6.00e-02 |
Pos logit contributions | |
Ant | +0.033 |
Next | +0.032 |
Salt | +0.032 |
Fire | +0.031 |
� | +0.031 |
Neg logit contributions | |
goodbye | -0.040 |
away | -0.033 |
back | -0.033 |
ly | -0.032 |
attention | -0.032 |
Token | â |
Token ID | 22940 |
Feature activation | -5.84e-02 |
Pos logit contributions | |
goodbye | +0.137 |
away | +0.126 |
attention | +0.125 |
des | +0.120 |
ogging | +0.119 |
Neg logit contributions | |
Ant | -0.068 |
I | -0.063 |
� | -0.062 |
Will | -0.061 |
Salt | -0.061 |
Token | € |
Token ID | 26391 |
Feature activation | -6.46e-02 |
Pos logit contributions | |
goodbye | +0.090 |
attention | +0.077 |
away | +0.076 |
ogging | +0.075 |
farewell | +0.071 |
Neg logit contributions | |
Ant | -0.109 |
� | -0.104 |
Next | -0.104 |
Salt | -0.103 |
hes | -0.102 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
goodbye | +0.134 |
away | +0.118 |
attention | +0.117 |
ogging | +0.117 |
des | +0.113 |
Neg logit contributions | |
Ant | -0.088 |
Next | -0.081 |
Fire | -0.080 |
Salt | -0.080 |
� | -0.079 |
Token | t |
Token ID | 83 |
Feature activation | -7.30e-02 |
Pos logit contributions | |
goodbye | +0.156 |
away | +0.142 |
attention | +0.139 |
farewell | +0.131 |
des | +0.127 |
Neg logit contributions | |
Ant | -0.177 |
I | -0.175 |
hes | -0.169 |
Salt | -0.168 |
� | -0.168 |
Token | catch |
Token ID | 4929 |
Feature activation | -3.45e-02 |
Pos logit contributions | |
goodbye | +0.129 |
farewell | +0.107 |
des | +0.107 |
ogging | +0.106 |
uits | +0.104 |
Neg logit contributions | |
Ant | -0.118 |
I | -0.115 |
Will | -0.106 |
You | -0.104 |
Mouse | -0.103 |
Token | him |
Token ID | 683 |
Feature activation | -2.69e-02 |
Pos logit contributions | |
goodbye | +0.053 |
ogging | +0.046 |
attention | +0.045 |
away | +0.045 |
des | +0.043 |
Neg logit contributions | |
Ant | -0.063 |
Salt | -0.060 |
Next | -0.060 |
Fire | -0.060 |
I | -0.059 |
Token | trying |
Token ID | 2111 |
Feature activation | -3.30e-02 |
Pos logit contributions | |
goodbye | +0.036 |
attention | +0.030 |
away | +0.030 |
ogging | +0.029 |
des | +0.028 |
Neg logit contributions | |
Ant | -0.055 |
Salt | -0.052 |
� | -0.052 |
Next | -0.051 |
Fire | -0.051 |
Token | to |
Token ID | 284 |
Feature activation | -1.85e-02 |
Pos logit contributions | |
goodbye | +0.045 |
des | +0.042 |
fy | +0.041 |
ogging | +0.041 |
ly | +0.040 |
Neg logit contributions | |
Ant | -0.057 |
Salt | -0.056 |
Will | -0.054 |
I | -0.054 |
chefs | -0.054 |
Token | and |
Token ID | 290 |
Feature activation | -1.17e-02 |
Pos logit contributions | |
� | +0.012 |
Ant | +0.011 |
Next | +0.011 |
Really | +0.011 |
Bees | +0.011 |
Neg logit contributions | |
goodbye | -0.009 |
ly | -0.007 |
back | -0.007 |
ogging | -0.007 |
farewell | -0.007 |
Token | Jack |
Token ID | 3619 |
Feature activation | -3.50e-02 |
Pos logit contributions | |
goodbye | +0.021 |
attention | +0.017 |
ogging | +0.017 |
farewell | +0.016 |
away | +0.016 |
Neg logit contributions | |
Ant | -0.020 |
hes | -0.020 |
Mouse | -0.020 |
poker | -0.020 |
� | -0.019 |
Token | didn |
Token ID | 1422 |
Feature activation | -4.48e-02 |
Pos logit contributions | |
goodbye | +0.063 |
farewell | +0.051 |
back | +0.051 |
ly | +0.051 |
away | +0.050 |
Neg logit contributions | |
Ant | -0.058 |
Next | -0.057 |
� | -0.056 |
Salt | -0.056 |
poker | -0.056 |
Token | â |
Token ID | 22940 |
Feature activation | -6.03e-02 |
Pos logit contributions | |
goodbye | +0.098 |
away | +0.089 |
attention | +0.088 |
ogging | +0.086 |
des | +0.085 |
Neg logit contributions | |
Ant | -0.055 |
Salt | -0.051 |
Next | -0.051 |
Fire | -0.051 |
� | -0.051 |
Token | € |
Token ID | 26391 |
Feature activation | -6.18e-02 |
Pos logit contributions | |
goodbye | +0.095 |
attention | +0.081 |
away | +0.080 |
ogging | +0.080 |
farewell | +0.075 |
Neg logit contributions | |
Ant | -0.110 |
Next | -0.106 |
� | -0.106 |
Salt | -0.104 |
Fire | -0.103 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.01e-01 |
Pos logit contributions | |
goodbye | +0.115 |
ogging | +0.100 |
attention | +0.098 |
away | +0.097 |
des | +0.095 |
Neg logit contributions | |
Ant | -0.097 |
Next | -0.090 |
Mouse | -0.089 |
poker | -0.088 |
Fire | -0.087 |
Token | t |
Token ID | 83 |
Feature activation | -7.12e-02 |
Pos logit contributions | |
goodbye | +0.162 |
away | +0.160 |
attention | +0.149 |
des | +0.146 |
farewell | +0.138 |
Neg logit contributions | |
Ant | -0.166 |
hes | -0.161 |
Salt | -0.158 |
Fire | -0.155 |
itt | -0.155 |
Token | give |
Token ID | 1577 |
Feature activation | -9.50e-03 |
Pos logit contributions | |
goodbye | +0.098 |
away | +0.080 |
ogging | +0.079 |
attention | +0.078 |
farewell | +0.076 |
Neg logit contributions | |
Ant | -0.145 |
Salt | -0.138 |
hes | -0.137 |
Fire | -0.137 |
Next | -0.136 |
Token | up |
Token ID | 510 |
Feature activation | +8.96e-03 |
Pos logit contributions | |
goodbye | +0.012 |
away | +0.010 |
ogging | +0.010 |
des | +0.010 |
attention | +0.010 |
Neg logit contributions | |
Ant | -0.020 |
Salt | -0.019 |
Fire | -0.019 |
Next | -0.019 |
hes | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | -2.28e-03 |
Pos logit contributions | |
Ant | +0.020 |
Next | +0.019 |
� | +0.019 |
Fire | +0.019 |
Salt | +0.019 |
Neg logit contributions | |
goodbye | -0.011 |
away | -0.009 |
attention | -0.009 |
ogging | -0.008 |
farewell | -0.008 |
Token | They |
Token ID | 1119 |
Feature activation | -3.74e-03 |
Pos logit contributions | |
goodbye | +0.005 |
ogging | +0.004 |
farewell | +0.004 |
back | +0.004 |
attention | +0.004 |
Neg logit contributions | |
Ant | -0.003 |
kins | -0.003 |
hes | -0.003 |
poker | -0.003 |
Mouse | -0.003 |
Token | things |
Token ID | 1243 |
Feature activation | -1.57e-02 |
Pos logit contributions | |
goodbye | +0.016 |
away | +0.014 |
attention | +0.014 |
ogging | +0.014 |
des | +0.013 |
Neg logit contributions | |
Ant | -0.016 |
Salt | -0.015 |
Fire | -0.015 |
Next | -0.015 |
� | -0.015 |
Token | he |
Token ID | 339 |
Feature activation | -4.04e-02 |
Pos logit contributions | |
goodbye | +0.023 |
away | +0.020 |
attention | +0.019 |
ogging | +0.019 |
farewell | +0.018 |
Neg logit contributions | |
Ant | -0.030 |
Next | -0.029 |
Salt | -0.029 |
� | -0.029 |
Fire | -0.029 |
Token | couldn |
Token ID | 3521 |
Feature activation | -4.98e-02 |
Pos logit contributions | |
goodbye | +0.078 |
attention | +0.070 |
away | +0.069 |
ly | +0.068 |
des | +0.067 |
Neg logit contributions | |
Ant | -0.058 |
Salt | -0.055 |
Fire | -0.054 |
poker | -0.054 |
� | -0.053 |
Token | â |
Token ID | 22940 |
Feature activation | -6.27e-02 |
Pos logit contributions | |
goodbye | +0.101 |
away | +0.090 |
attention | +0.089 |
ogging | +0.088 |
farewell | +0.085 |
Neg logit contributions | |
Ant | -0.069 |
Next | -0.064 |
Salt | -0.064 |
� | -0.064 |
Fire | -0.064 |
Token | € |
Token ID | 26391 |
Feature activation | -6.49e-02 |
Pos logit contributions | |
goodbye | +0.098 |
away | +0.084 |
attention | +0.083 |
ogging | +0.082 |
farewell | +0.078 |
Neg logit contributions | |
Ant | -0.117 |
Next | -0.110 |
Fire | -0.110 |
Salt | -0.110 |
� | -0.109 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
goodbye | +0.126 |
ogging | +0.111 |
away | +0.107 |
attention | +0.106 |
des | +0.105 |
Neg logit contributions | |
Ant | -0.098 |
Fire | -0.089 |
Next | -0.089 |
Salt | -0.088 |
Mouse | -0.088 |
Token | t |
Token ID | 83 |
Feature activation | -7.27e-02 |
Pos logit contributions | |
goodbye | +0.157 |
away | +0.136 |
attention | +0.136 |
farewell | +0.130 |
des | +0.124 |
Neg logit contributions | |
Ant | -0.180 |
Salt | -0.173 |
I | -0.172 |
hes | -0.172 |
� | -0.170 |
Token | do |
Token ID | 466 |
Feature activation | -7.51e-03 |
Pos logit contributions | |
goodbye | +0.104 |
ogging | +0.085 |
attention | +0.085 |
away | +0.083 |
des | +0.081 |
Neg logit contributions | |
Ant | -0.146 |
� | -0.136 |
Salt | -0.135 |
Next | -0.134 |
Mouse | -0.134 |
Token | without |
Token ID | 1231 |
Feature activation | +3.19e-02 |
Pos logit contributions | |
goodbye | +0.010 |
away | +0.009 |
attention | +0.008 |
ogging | +0.008 |
farewell | +0.008 |
Neg logit contributions | |
Ant | -0.015 |
Next | -0.015 |
Salt | -0.015 |
� | -0.015 |
Fire | -0.014 |
Token | it |
Token ID | 340 |
Feature activation | -3.93e-02 |
Pos logit contributions | |
Ant | +0.057 |
� | +0.054 |
Next | +0.054 |
Fire | +0.053 |
Salt | +0.053 |
Neg logit contributions | |
goodbye | -0.053 |
attention | -0.046 |
away | -0.045 |
ogging | -0.044 |
farewell | -0.042 |
Token | . |
Token ID | 13 |
Feature activation | -3.33e-03 |
Pos logit contributions | |
goodbye | +0.052 |
away | +0.044 |
attention | +0.042 |
ogging | +0.042 |
des | +0.041 |
Neg logit contributions | |
Ant | -0.080 |
Salt | -0.077 |
Next | -0.077 |
Fire | -0.077 |
� | -0.076 |
Token | that |
Token ID | 326 |
Feature activation | +1.97e-02 |
Pos logit contributions | |
Ant | +0.017 |
� | +0.017 |
Salt | +0.017 |
Fire | +0.017 |
Next | +0.017 |
Neg logit contributions | |
goodbye | -0.011 |
away | -0.010 |
des | -0.009 |
attention | -0.009 |
ogging | -0.009 |
Token | she |
Token ID | 673 |
Feature activation | -1.52e-02 |
Pos logit contributions | |
Ant | +0.037 |
Mouse | +0.034 |
poker | +0.033 |
Sal | +0.033 |
Next | +0.033 |
Neg logit contributions | |
goodbye | -0.036 |
attention | -0.030 |
ogging | -0.028 |
away | -0.028 |
back | -0.027 |
Token | couldn |
Token ID | 3521 |
Feature activation | -3.38e-02 |
Pos logit contributions | |
goodbye | +0.031 |
away | +0.027 |
ogging | +0.026 |
attention | +0.026 |
back | +0.026 |
Neg logit contributions | |
Ant | -0.022 |
Fire | -0.021 |
Salt | -0.021 |
poker | -0.021 |
Next | -0.021 |
Token | â |
Token ID | 22940 |
Feature activation | -5.59e-02 |
Pos logit contributions | |
goodbye | +0.068 |
away | +0.060 |
attention | +0.060 |
ogging | +0.060 |
farewell | +0.057 |
Neg logit contributions | |
Ant | -0.048 |
Next | -0.044 |
Salt | -0.044 |
Fire | -0.044 |
� | -0.043 |
Token | € |
Token ID | 26391 |
Feature activation | -6.33e-02 |
Pos logit contributions | |
goodbye | +0.092 |
attention | +0.079 |
away | +0.078 |
ogging | +0.077 |
farewell | +0.073 |
Neg logit contributions | |
Ant | -0.097 |
� | -0.096 |
Next | -0.094 |
Salt | -0.092 |
hes | -0.092 |
Token | ™ |
Token ID | 8151 |
Feature activation | -1.00e-01 |
Pos logit contributions | |
goodbye | +0.123 |
ogging | +0.108 |
attention | +0.105 |
away | +0.105 |
des | +0.102 |
Neg logit contributions | |
Ant | -0.096 |
Fire | -0.088 |
Salt | -0.087 |
Next | -0.087 |
poker | -0.086 |
Token | t |
Token ID | 83 |
Feature activation | -6.20e-02 |
Pos logit contributions | |
goodbye | +0.158 |
away | +0.145 |
attention | +0.142 |
farewell | +0.135 |
des | +0.132 |
Neg logit contributions | |
Ant | -0.167 |
I | -0.167 |
hes | -0.164 |
Salt | -0.161 |
� | -0.160 |
Token | help |
Token ID | 1037 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
goodbye | +0.074 |
ogging | +0.060 |
attention | +0.060 |
away | +0.060 |
des | +0.056 |
Neg logit contributions | |
Ant | -0.134 |
� | -0.129 |
I | -0.128 |
Salt | -0.127 |
Will | -0.127 |
Token | but |
Token ID | 475 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
goodbye | +0.043 |
away | +0.037 |
attention | +0.036 |
ogging | +0.036 |
des | +0.035 |
Neg logit contributions | |
Ant | -0.049 |
Salt | -0.046 |
Next | -0.046 |
� | -0.046 |
Fire | -0.046 |
Token | feel |
Token ID | 1254 |
Feature activation | +1.67e-02 |
Pos logit contributions | |
goodbye | +0.025 |
away | +0.021 |
attention | +0.021 |
ogging | +0.020 |
farewell | +0.019 |
Neg logit contributions | |
Ant | -0.037 |
Next | -0.035 |
� | -0.035 |
Salt | -0.035 |
Fire | -0.035 |
Token | a |
Token ID | 257 |
Feature activation | -6.29e-03 |
Pos logit contributions | |
Ant | +0.032 |
� | +0.030 |
Next | +0.030 |
Fire | +0.030 |
Salt | +0.029 |
Neg logit contributions | |
goodbye | -0.026 |
attention | -0.022 |
away | -0.022 |
ogging | -0.021 |
farewell | -0.021 |
Token | , |
Token ID | 11 |
Feature activation | +8.45e-03 |
Pos logit contributions | |
goodbye | +0.008 |
attention | +0.007 |
away | +0.007 |
ogging | +0.006 |
ly | +0.006 |
Neg logit contributions | |
Ant | -0.015 |
Mouse | -0.014 |
Next | -0.014 |
hes | -0.014 |
Fire | -0.014 |
Token | she |
Token ID | 673 |
Feature activation | -4.56e-03 |
Pos logit contributions | |
Ant | +0.016 |
Mouse | +0.015 |
hes | +0.015 |
poker | +0.015 |
Sal | +0.015 |
Neg logit contributions | |
goodbye | -0.014 |
attention | -0.012 |
ogging | -0.011 |
back | -0.011 |
away | -0.011 |
Token | won |
Token ID | 1839 |
Feature activation | -5.90e-02 |
Pos logit contributions | |
goodbye | +0.008 |
back | +0.007 |
attention | +0.007 |
ly | +0.007 |
away | +0.007 |
Neg logit contributions | |
Ant | -0.008 |
Salt | -0.007 |
� | -0.007 |
Really | -0.007 |
Mouse | -0.007 |
Token | â |
Token ID | 22940 |
Feature activation | -5.36e-02 |
Pos logit contributions | |
goodbye | +0.083 |
attention | +0.071 |
ogging | +0.071 |
away | +0.068 |
back | +0.065 |
Neg logit contributions | |
Ant | -0.117 |
Next | -0.110 |
Really | -0.109 |
Bees | -0.108 |
� | -0.108 |
Token | € |
Token ID | 26391 |
Feature activation | -5.96e-02 |
Pos logit contributions | |
goodbye | +0.088 |
away | +0.076 |
attention | +0.076 |
ogging | +0.074 |
farewell | +0.071 |
Neg logit contributions | |
Ant | -0.095 |
Next | -0.090 |
Salt | -0.089 |
Fire | -0.089 |
� | -0.089 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.95e-02 |
Pos logit contributions | |
goodbye | +0.113 |
ogging | +0.099 |
des | +0.094 |
attention | +0.093 |
away | +0.093 |
Neg logit contributions | |
Ant | -0.095 |
poker | -0.086 |
Mouse | -0.086 |
Fire | -0.085 |
Next | -0.085 |
Token | t |
Token ID | 83 |
Feature activation | -7.55e-02 |
Pos logit contributions | |
goodbye | +0.154 |
away | +0.140 |
attention | +0.136 |
farewell | +0.129 |
des | +0.127 |
Neg logit contributions | |
Ant | -0.176 |
I | -0.170 |
Salt | -0.169 |
hes | -0.169 |
Fire | -0.168 |
Token | be |
Token ID | 307 |
Feature activation | -7.51e-03 |
Pos logit contributions | |
goodbye | +0.116 |
away | +0.099 |
attention | +0.098 |
ogging | +0.096 |
farewell | +0.092 |
Neg logit contributions | |
Ant | -0.142 |
Next | -0.134 |
Salt | -0.134 |
� | -0.134 |
Fire | -0.134 |
Token | able |
Token ID | 1498 |
Feature activation | -1.87e-02 |
Pos logit contributions | |
goodbye | +0.012 |
attention | +0.010 |
away | +0.010 |
ogging | +0.010 |
farewell | +0.009 |
Neg logit contributions | |
Ant | -0.014 |
� | -0.014 |
Next | -0.013 |
Fire | -0.013 |
Salt | -0.013 |
Token | to |
Token ID | 284 |
Feature activation | -1.55e-02 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.017 |
farewell | +0.016 |
des | +0.016 |
ogging | +0.015 |
Neg logit contributions | |
Ant | -0.040 |
Fire | -0.040 |
Next | -0.040 |
Salt | -0.040 |
twins | -0.039 |
Token | teach |
Token ID | 4545 |
Feature activation | -1.31e-02 |
Pos logit contributions | |
goodbye | +0.024 |
away | +0.020 |
attention | +0.020 |
ogging | +0.020 |
farewell | +0.019 |
Neg logit contributions | |
Ant | -0.029 |
Salt | -0.027 |
Fire | -0.027 |
Next | -0.027 |
� | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | -9.44e-03 |
Pos logit contributions | |
goodbye | +0.010 |
des | +0.009 |
ogging | +0.009 |
away | +0.009 |
farewell | +0.009 |
Neg logit contributions | |
hes | -0.012 |
kins | -0.012 |
Ant | -0.012 |
ters | -0.011 |
Salt | -0.011 |
Token | Jack |
Token ID | 3619 |
Feature activation | -3.13e-02 |
Pos logit contributions | |
goodbye | +0.020 |
ogging | +0.016 |
ly | +0.016 |
farewell | +0.016 |
attention | +0.016 |
Neg logit contributions | |
Ant | -0.015 |
Mouse | -0.013 |
kins | -0.013 |
poker | -0.012 |
hes | -0.012 |
Token | didn |
Token ID | 1422 |
Feature activation | -4.91e-02 |
Pos logit contributions | |
goodbye | +0.050 |
ogging | +0.039 |
back | +0.038 |
farewell | +0.038 |
ly | +0.038 |
Neg logit contributions | |
Ant | -0.058 |
� | -0.056 |
Next | -0.056 |
Really | -0.056 |
Fire | -0.055 |
Token | â |
Token ID | 22940 |
Feature activation | -6.16e-02 |
Pos logit contributions | |
goodbye | +0.116 |
away | +0.108 |
attention | +0.107 |
des | +0.102 |
ly | +0.102 |
Neg logit contributions | |
Ant | -0.049 |
hes | -0.048 |
Salt | -0.048 |
� | -0.046 |
Fire | -0.046 |
Token | € |
Token ID | 26391 |
Feature activation | -5.50e-02 |
Pos logit contributions | |
goodbye | +0.103 |
attention | +0.089 |
away | +0.087 |
ogging | +0.086 |
farewell | +0.082 |
Neg logit contributions | |
Ant | -0.106 |
� | -0.103 |
Next | -0.103 |
Salt | -0.100 |
hes | -0.100 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.94e-02 |
Pos logit contributions | |
goodbye | +0.112 |
away | +0.098 |
ogging | +0.098 |
attention | +0.098 |
des | +0.094 |
Neg logit contributions | |
Ant | -0.077 |
Next | -0.071 |
Fire | -0.070 |
Salt | -0.070 |
� | -0.070 |
Token | t |
Token ID | 83 |
Feature activation | -7.09e-02 |
Pos logit contributions | |
goodbye | +0.163 |
away | +0.153 |
attention | +0.148 |
farewell | +0.140 |
des | +0.137 |
Neg logit contributions | |
Ant | -0.163 |
hes | -0.158 |
I | -0.158 |
Salt | -0.156 |
itt | -0.152 |
Token | want |
Token ID | 765 |
Feature activation | -3.25e-02 |
Pos logit contributions | |
goodbye | +0.104 |
away | +0.084 |
attention | +0.084 |
farewell | +0.083 |
ogging | +0.083 |
Neg logit contributions | |
Ant | -0.136 |
Salt | -0.128 |
I | -0.127 |
hes | -0.127 |
You | -0.127 |
Token | it |
Token ID | 340 |
Feature activation | -3.91e-02 |
Pos logit contributions | |
goodbye | +0.032 |
away | +0.027 |
ogging | +0.025 |
des | +0.024 |
attention | +0.024 |
Neg logit contributions | |
Ant | -0.076 |
Salt | -0.075 |
Next | -0.074 |
Fire | -0.074 |
I | -0.074 |
Token | because |
Token ID | 780 |
Feature activation | -1.09e-03 |
Pos logit contributions | |
goodbye | +0.044 |
away | +0.040 |
des | +0.037 |
farewell | +0.037 |
ogging | +0.036 |
Neg logit contributions | |
I | -0.079 |
Ant | -0.079 |
Salt | -0.078 |
Next | -0.077 |
Fire | -0.077 |
Token | it |
Token ID | 340 |
Feature activation | +1.91e-02 |
Pos logit contributions | |
goodbye | +0.001 |
ogging | +0.001 |
attention | +0.001 |
away | +0.001 |
farewell | +0.001 |
Neg logit contributions | |
Ant | -0.002 |
Mouse | -0.002 |
� | -0.002 |
Sal | -0.002 |
Really | -0.002 |
Token | his |
Token ID | 465 |
Feature activation | +2.61e-02 |
Pos logit contributions | |
Ant | +0.010 |
� | +0.009 |
Next | +0.009 |
Fire | +0.009 |
Mouse | +0.009 |
Neg logit contributions | |
goodbye | -0.009 |
attention | -0.008 |
away | -0.007 |
ogging | -0.007 |
ly | -0.007 |
Token | parents |
Token ID | 3397 |
Feature activation | -1.26e-02 |
Pos logit contributions | |
Ant | +0.038 |
hes | +0.036 |
� | +0.035 |
Salt | +0.035 |
Next | +0.035 |
Neg logit contributions | |
goodbye | -0.053 |
attention | -0.047 |
away | -0.045 |
ogging | -0.045 |
farewell | -0.044 |
Token | couldn |
Token ID | 3521 |
Feature activation | -3.05e-02 |
Pos logit contributions | |
goodbye | +0.023 |
ogging | +0.019 |
away | +0.019 |
ly | +0.019 |
contain | +0.019 |
Neg logit contributions | |
Ant | -0.021 |
Next | -0.020 |
Salt | -0.019 |
poker | -0.019 |
Fire | -0.019 |
Token | â |
Token ID | 22940 |
Feature activation | -5.56e-02 |
Pos logit contributions | |
goodbye | +0.057 |
ogging | +0.050 |
away | +0.050 |
attention | +0.049 |
farewell | +0.048 |
Neg logit contributions | |
Ant | -0.047 |
Salt | -0.043 |
Next | -0.043 |
kins | -0.042 |
� | -0.042 |
Token | € |
Token ID | 26391 |
Feature activation | -6.64e-02 |
Pos logit contributions | |
goodbye | +0.087 |
away | +0.075 |
attention | +0.074 |
ogging | +0.073 |
farewell | +0.070 |
Neg logit contributions | |
Ant | -0.103 |
Next | -0.097 |
� | -0.097 |
Salt | -0.097 |
Fire | -0.097 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.94e-02 |
Pos logit contributions | |
goodbye | +0.129 |
away | +0.114 |
attention | +0.113 |
ogging | +0.112 |
farewell | +0.108 |
Neg logit contributions | |
Ant | -0.098 |
Next | -0.091 |
Salt | -0.090 |
Fire | -0.090 |
� | -0.090 |
Token | t |
Token ID | 83 |
Feature activation | -6.31e-02 |
Pos logit contributions | |
goodbye | +0.155 |
away | +0.139 |
attention | +0.136 |
farewell | +0.131 |
des | +0.128 |
Neg logit contributions | |
Ant | -0.172 |
Salt | -0.170 |
I | -0.168 |
hes | -0.167 |
Fire | -0.167 |
Token | take |
Token ID | 1011 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
goodbye | +0.088 |
ogging | +0.073 |
attention | +0.073 |
away | +0.071 |
farewell | +0.068 |
Neg logit contributions | |
Ant | -0.121 |
I | -0.119 |
Will | -0.118 |
� | -0.117 |
Salt | -0.115 |
Token | him |
Token ID | 683 |
Feature activation | -4.14e-02 |
Pos logit contributions | |
goodbye | +0.021 |
des | +0.019 |
away | +0.019 |
ogging | +0.019 |
contain | +0.018 |
Neg logit contributions | |
Salt | -0.020 |
Ant | -0.019 |
Fire | -0.019 |
Next | -0.019 |
Brush | -0.019 |
Token | away |
Token ID | 1497 |
Feature activation | -7.11e-03 |
Pos logit contributions | |
goodbye | +0.049 |
des | +0.043 |
ogging | +0.041 |
away | +0.041 |
farewell | +0.041 |
Neg logit contributions | |
Salt | -0.078 |
Fire | -0.078 |
brothers | -0.077 |
Ant | -0.076 |
Next | -0.075 |
Token | from |
Token ID | 422 |
Feature activation | +1.91e-02 |
Pos logit contributions | |
goodbye | +0.010 |
away | +0.008 |
attention | +0.008 |
ogging | +0.008 |
des | +0.008 |
Neg logit contributions | |
Ant | -0.014 |
Salt | -0.014 |
Fire | -0.014 |
Next | -0.014 |
� | -0.014 |
Token | , |
Token ID | 11 |
Feature activation | +2.26e-03 |
Pos logit contributions | |
Ant | +0.057 |
Next | +0.055 |
Fire | +0.054 |
Salt | +0.053 |
hes | +0.053 |
Neg logit contributions | |
goodbye | -0.036 |
away | -0.030 |
attention | -0.029 |
ogging | -0.029 |
ly | -0.027 |
Token | you |
Token ID | 345 |
Feature activation | -3.29e-02 |
Pos logit contributions | |
Ant | +0.004 |
� | +0.004 |
Next | +0.004 |
Mouse | +0.004 |
Fire | +0.004 |
Neg logit contributions | |
goodbye | -0.004 |
attention | -0.003 |
away | -0.003 |
ogging | -0.003 |
farewell | -0.003 |
Token | won |
Token ID | 1839 |
Feature activation | -7.26e-02 |
Pos logit contributions | |
goodbye | +0.064 |
attention | +0.056 |
away | +0.056 |
ogging | +0.055 |
farewell | +0.053 |
Neg logit contributions | |
Ant | -0.049 |
Next | -0.046 |
Salt | -0.046 |
� | -0.046 |
Fire | -0.045 |
Token | â |
Token ID | 22940 |
Feature activation | -6.29e-02 |
Pos logit contributions | |
goodbye | +0.122 |
ogging | +0.108 |
attention | +0.107 |
away | +0.106 |
back | +0.102 |
Neg logit contributions | |
Ant | -0.120 |
Next | -0.113 |
Really | -0.112 |
kins | -0.112 |
� | -0.112 |
Token | € |
Token ID | 26391 |
Feature activation | -6.07e-02 |
Pos logit contributions | |
goodbye | +0.100 |
attention | +0.086 |
away | +0.086 |
ogging | +0.084 |
farewell | +0.080 |
Neg logit contributions | |
Ant | -0.115 |
Next | -0.109 |
� | -0.109 |
Salt | -0.108 |
Fire | -0.107 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.92e-02 |
Pos logit contributions | |
goodbye | +0.116 |
ogging | +0.103 |
des | +0.096 |
away | +0.095 |
farewell | +0.094 |
Neg logit contributions | |
Ant | -0.095 |
poker | -0.087 |
Fire | -0.086 |
Mouse | -0.086 |
Next | -0.085 |
Token | t |
Token ID | 83 |
Feature activation | -8.45e-02 |
Pos logit contributions | |
goodbye | +0.147 |
away | +0.132 |
attention | +0.130 |
farewell | +0.121 |
des | +0.119 |
Neg logit contributions | |
Ant | -0.185 |
Salt | -0.177 |
� | -0.176 |
I | -0.176 |
Fire | -0.175 |
Token | poke |
Token ID | 22620 |
Feature activation | -8.50e-03 |
Pos logit contributions | |
goodbye | +0.145 |
away | +0.123 |
attention | +0.122 |
ogging | +0.120 |
farewell | +0.118 |
Neg logit contributions | |
Ant | -0.146 |
Salt | -0.135 |
� | -0.135 |
Next | -0.134 |
Fire | -0.134 |
Token | it |
Token ID | 340 |
Feature activation | -3.55e-02 |
Pos logit contributions | |
goodbye | +0.014 |
away | +0.012 |
attention | +0.012 |
ogging | +0.012 |
farewell | +0.011 |
Neg logit contributions | |
Ant | -0.015 |
Next | -0.014 |
� | -0.014 |
Fire | -0.014 |
Salt | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -7.27e-05 |
Pos logit contributions | |
goodbye | +0.038 |
away | +0.030 |
attention | +0.030 |
ogging | +0.029 |
des | +0.027 |
Neg logit contributions | |
Ant | -0.082 |
Next | -0.079 |
Salt | -0.079 |
� | -0.079 |
Fire | -0.078 |
Token | Before |
Token ID | 7413 |
Feature activation | +4.28e-02 |
Pos logit contributions | |
goodbye | +0.000 |
ogging | +0.000 |
des | +0.000 |
owl | +0.000 |
farewell | +0.000 |
Neg logit contributions | |
kins | -0.000 |
Ant | -0.000 |
poker | -0.000 |
hes | -0.000 |
chefs | -0.000 |
Token | palm |
Token ID | 18057 |
Feature activation | +1.69e-02 |
Pos logit contributions | |
hes | +0.040 |
� | +0.039 |
Ant | +0.039 |
Salt | +0.039 |
kins | +0.038 |
Neg logit contributions | |
goodbye | -0.034 |
attention | -0.031 |
back | -0.028 |
farewell | -0.027 |
ogging | -0.027 |
Token | tree |
Token ID | 5509 |
Feature activation | +2.83e-02 |
Pos logit contributions | |
Ant | +0.029 |
Salt | +0.027 |
� | +0.027 |
Fire | +0.027 |
Next | +0.027 |
Neg logit contributions | |
goodbye | -0.030 |
ogging | -0.025 |
attention | -0.025 |
away | -0.025 |
ly | -0.024 |
Token | won |
Token ID | 1839 |
Feature activation | -6.28e-02 |
Pos logit contributions | |
Ant | +0.045 |
Salt | +0.042 |
Next | +0.042 |
Fire | +0.042 |
Mouse | +0.041 |
Neg logit contributions | |
goodbye | -0.053 |
attention | -0.046 |
away | -0.045 |
ogging | -0.044 |
ly | -0.043 |
Token | â |
Token ID | 22940 |
Feature activation | -4.98e-02 |
Pos logit contributions | |
goodbye | +0.110 |
attention | +0.096 |
away | +0.095 |
ogging | +0.095 |
farewell | +0.090 |
Neg logit contributions | |
Ant | -0.104 |
Next | -0.097 |
Salt | -0.096 |
� | -0.096 |
Fire | -0.096 |
Token | € |
Token ID | 26391 |
Feature activation | -5.42e-02 |
Pos logit contributions | |
goodbye | +0.081 |
attention | +0.069 |
away | +0.069 |
ogging | +0.067 |
farewell | +0.064 |
Neg logit contributions | |
Ant | -0.089 |
� | -0.085 |
Next | -0.085 |
Salt | -0.083 |
hes | -0.083 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.91e-02 |
Pos logit contributions | |
goodbye | +0.109 |
ogging | +0.095 |
away | +0.094 |
attention | +0.093 |
des | +0.090 |
Neg logit contributions | |
Ant | -0.078 |
Fire | -0.072 |
Next | -0.072 |
Salt | -0.072 |
� | -0.071 |
Token | t |
Token ID | 83 |
Feature activation | -7.63e-02 |
Pos logit contributions | |
goodbye | +0.147 |
attention | +0.131 |
away | +0.130 |
farewell | +0.121 |
des | +0.119 |
Neg logit contributions | |
Ant | -0.183 |
� | -0.176 |
I | -0.176 |
Salt | -0.175 |
hes | -0.174 |
Token | feel |
Token ID | 1254 |
Feature activation | -2.40e-03 |
Pos logit contributions | |
goodbye | +0.150 |
des | +0.128 |
away | +0.127 |
attention | +0.127 |
ogging | +0.126 |
Neg logit contributions | |
Ant | -0.110 |
I | -0.102 |
Will | -0.098 |
Mouse | -0.096 |
Salt | -0.096 |
Token | alone |
Token ID | 3436 |
Feature activation | +1.73e-02 |
Pos logit contributions | |
goodbye | +0.004 |
away | +0.003 |
ogging | +0.003 |
attention | +0.003 |
des | +0.003 |
Neg logit contributions | |
Ant | -0.004 |
Salt | -0.004 |
Next | -0.004 |
Fire | -0.004 |
hes | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +3.57e-03 |
Pos logit contributions | |
Ant | +0.033 |
Salt | +0.032 |
Next | +0.032 |
Fire | +0.032 |
Will | +0.032 |
Neg logit contributions | |
goodbye | -0.024 |
attention | -0.021 |
away | -0.020 |
ogging | -0.020 |
des | -0.020 |
Token | â |
Token ID | 22940 |
Feature activation | -2.76e-02 |
Pos logit contributions | |
kins | +0.007 |
poker | +0.006 |
hes | +0.006 |
Ant | +0.006 |
chefs | +0.006 |
Neg logit contributions | |
goodbye | -0.006 |
ogging | -0.005 |
farewell | -0.005 |
attention | -0.004 |
owl | -0.004 |
Token | Jack |
Token ID | 3619 |
Feature activation | -3.07e-02 |
Pos logit contributions | |
goodbye | +0.011 |
ogging | +0.009 |
farewell | +0.009 |
back | +0.008 |
attention | +0.008 |
Neg logit contributions | |
Ant | -0.012 |
poker | -0.011 |
hes | -0.011 |
kins | -0.011 |
Will | -0.011 |
Token | had |
Token ID | 550 |
Feature activation | +6.62e-03 |
Pos logit contributions | |
goodbye | +0.055 |
back | +0.046 |
ly | +0.045 |
farewell | +0.045 |
ogging | +0.044 |
Neg logit contributions | |
Ant | -0.051 |
poker | -0.050 |
Next | -0.049 |
Salt | -0.049 |
hes | -0.048 |
Token | used |
Token ID | 973 |
Feature activation | +2.98e-03 |
Pos logit contributions | |
Ant | +0.011 |
� | +0.010 |
Next | +0.010 |
Salt | +0.010 |
Fire | +0.010 |
Neg logit contributions | |
goodbye | -0.012 |
attention | -0.010 |
away | -0.010 |
ogging | -0.010 |
farewell | -0.010 |
Token | his |
Token ID | 465 |
Feature activation | +4.17e-03 |
Pos logit contributions | |
Ant | +0.006 |
� | +0.005 |
Next | +0.005 |
Mouse | +0.005 |
Bees | +0.005 |
Neg logit contributions | |
goodbye | -0.005 |
attention | -0.004 |
away | -0.004 |
ogging | -0.004 |
farewell | -0.004 |
Token | cle |
Token ID | 1190 |
Feature activation | -3.98e-02 |
Pos logit contributions | |
Ant | +0.007 |
Next | +0.007 |
hes | +0.007 |
� | +0.007 |
Salt | +0.007 |
Neg logit contributions | |
goodbye | -0.007 |
attention | -0.006 |
away | -0.006 |
ogging | -0.006 |
farewell | -0.006 |
Token | vern |
Token ID | 933 |
Feature activation | -9.91e-02 |
Pos logit contributions | |
goodbye | +0.055 |
away | +0.045 |
farewell | +0.042 |
des | +0.040 |
ogging | +0.039 |
Neg logit contributions | |
kins | -0.077 |
hes | -0.075 |
itt | -0.074 |
Ant | -0.073 |
brothers | -0.073 |
Token | ess |
Token ID | 408 |
Feature activation | -1.23e-02 |
Pos logit contributions | |
goodbye | +0.192 |
ogging | +0.164 |
des | +0.157 |
away | +0.157 |
farewell | +0.157 |
Neg logit contributions | |
Ant | -0.132 |
Next | -0.128 |
sibling | -0.128 |
Cere | -0.123 |
Salt | -0.122 |
Token | and |
Token ID | 290 |
Feature activation | +1.04e-02 |
Pos logit contributions | |
goodbye | +0.016 |
away | +0.014 |
attention | +0.013 |
ogging | +0.013 |
des | +0.012 |
Neg logit contributions | |
Ant | -0.025 |
Next | -0.024 |
Salt | -0.024 |
� | -0.024 |
Fire | -0.024 |
Token | creat |
Token ID | 1827 |
Feature activation | -4.02e-03 |
Pos logit contributions | |
Ant | +0.019 |
� | +0.017 |
Mouse | +0.017 |
Sal | +0.016 |
poker | +0.016 |
Neg logit contributions | |
goodbye | -0.019 |
attention | -0.016 |
farewell | -0.015 |
away | -0.015 |
ogging | -0.015 |
Token | iveness |
Token ID | 6517 |
Feature activation | +2.81e-02 |
Pos logit contributions | |
goodbye | +0.008 |
away | +0.007 |
attention | +0.007 |
ogging | +0.007 |
farewell | +0.007 |
Neg logit contributions | |
Ant | -0.006 |
Next | -0.005 |
� | -0.005 |
Salt | -0.005 |
Fire | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | +1.83e-02 |
Pos logit contributions | |
Ant | +0.049 |
� | +0.045 |
Mouse | +0.045 |
Really | +0.043 |
Will | +0.043 |
Neg logit contributions | |
goodbye | -0.050 |
attention | -0.045 |
back | -0.043 |
ly | -0.042 |
ogging | -0.041 |
Token | , |
Token ID | 11 |
Feature activation | -2.15e-03 |
Pos logit contributions | |
Ant | +0.041 |
Next | +0.040 |
Salt | +0.040 |
� | +0.039 |
I | +0.039 |
Neg logit contributions | |
goodbye | -0.031 |
away | -0.028 |
attention | -0.028 |
ogging | -0.028 |
des | -0.028 |
Token | she |
Token ID | 673 |
Feature activation | -4.18e-03 |
Pos logit contributions | |
goodbye | +0.004 |
away | +0.003 |
attention | +0.003 |
ogging | +0.003 |
farewell | +0.003 |
Neg logit contributions | |
Ant | -0.004 |
Next | -0.004 |
Salt | -0.004 |
� | -0.004 |
Fire | -0.004 |
Token | couldn |
Token ID | 3521 |
Feature activation | -2.60e-02 |
Pos logit contributions | |
goodbye | +0.008 |
away | +0.007 |
ogging | +0.007 |
attention | +0.007 |
farewell | +0.007 |
Neg logit contributions | |
Ant | -0.006 |
Fire | -0.006 |
Salt | -0.006 |
Next | -0.006 |
� | -0.006 |
Token | â |
Token ID | 22940 |
Feature activation | -5.33e-02 |
Pos logit contributions | |
goodbye | +0.053 |
away | +0.047 |
attention | +0.047 |
ogging | +0.046 |
farewell | +0.045 |
Neg logit contributions | |
Ant | -0.036 |
Next | -0.033 |
� | -0.033 |
Salt | -0.033 |
Fire | -0.033 |
Token | € |
Token ID | 26391 |
Feature activation | -5.82e-02 |
Pos logit contributions | |
goodbye | +0.091 |
attention | +0.080 |
away | +0.078 |
ogging | +0.076 |
closer | +0.073 |
Neg logit contributions | |
� | -0.088 |
Ant | -0.086 |
Next | -0.085 |
You | -0.085 |
I | -0.084 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.91e-02 |
Pos logit contributions | |
goodbye | +0.114 |
ogging | +0.100 |
away | +0.099 |
attention | +0.099 |
des | +0.095 |
Neg logit contributions | |
Ant | -0.086 |
Fire | -0.079 |
Next | -0.079 |
Salt | -0.079 |
� | -0.078 |
Token | t |
Token ID | 83 |
Feature activation | -6.08e-02 |
Pos logit contributions | |
goodbye | +0.161 |
away | +0.156 |
attention | +0.152 |
farewell | +0.139 |
des | +0.137 |
Neg logit contributions | |
I | -0.158 |
Ant | -0.157 |
hes | -0.152 |
Will | -0.147 |
You | -0.145 |
Token | wait |
Token ID | 4043 |
Feature activation | -4.88e-02 |
Pos logit contributions | |
goodbye | +0.078 |
ogging | +0.072 |
attention | +0.069 |
away | +0.068 |
des | +0.063 |
Neg logit contributions | |
Ant | -0.119 |
I | -0.116 |
Will | -0.111 |
� | -0.109 |
You | -0.109 |
Token | to |
Token ID | 284 |
Feature activation | -2.94e-02 |
Pos logit contributions | |
away | +0.065 |
oot | +0.064 |
des | +0.063 |
ly | +0.063 |
fy | +0.063 |
Neg logit contributions | |
Will | -0.073 |
hes | -0.071 |
I | -0.070 |
Let | -0.069 |
You | -0.068 |
Token | come |
Token ID | 1282 |
Feature activation | -8.65e-02 |
Pos logit contributions | |
goodbye | +0.043 |
away | +0.040 |
ly | +0.040 |
attention | +0.037 |
ogging | +0.037 |
Neg logit contributions | |
Ant | -0.053 |
Salt | -0.051 |
Fire | -0.051 |
Next | -0.051 |
brothers | -0.050 |
Token | back |
Token ID | 736 |
Feature activation | -1.32e-02 |
Pos logit contributions | |
goodbye | +0.084 |
ly | +0.080 |
fy | +0.076 |
ong | +0.075 |
contain | +0.074 |
Neg logit contributions | |
I | -0.189 |
Will | -0.187 |
� | -0.182 |
Next | -0.181 |
Ant | -0.181 |
Token | but |
Token ID | 475 |
Feature activation | +3.69e-03 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.017 |
attention | +0.017 |
ogging | +0.017 |
farewell | +0.016 |
Neg logit contributions | |
Ant | -0.028 |
Next | -0.026 |
Salt | -0.026 |
Fire | -0.026 |
� | -0.026 |
Token | he |
Token ID | 339 |
Feature activation | +8.75e-03 |
Pos logit contributions | |
Ant | +0.007 |
hes | +0.006 |
Bees | +0.006 |
Fire | +0.006 |
Mouse | +0.006 |
Neg logit contributions | |
goodbye | -0.007 |
attention | -0.005 |
away | -0.005 |
ly | -0.005 |
ogging | -0.005 |
Token | didn |
Token ID | 1422 |
Feature activation | -4.49e-02 |
Pos logit contributions | |
Ant | +0.014 |
� | +0.014 |
poker | +0.014 |
Blocks | +0.014 |
Fire | +0.013 |
Neg logit contributions | |
goodbye | -0.015 |
back | -0.013 |
ogging | -0.013 |
ly | -0.012 |
away | -0.012 |
Token | â |
Token ID | 22940 |
Feature activation | -5.72e-02 |
Pos logit contributions | |
goodbye | +0.092 |
away | +0.082 |
attention | +0.081 |
ogging | +0.080 |
farewell | +0.078 |
Neg logit contributions | |
Ant | -0.061 |
Next | -0.057 |
Salt | -0.057 |
� | -0.057 |
Fire | -0.057 |
Token | € |
Token ID | 26391 |
Feature activation | -5.96e-02 |
Pos logit contributions | |
goodbye | +0.092 |
attention | +0.079 |
away | +0.078 |
ogging | +0.076 |
farewell | +0.073 |
Neg logit contributions | |
Ant | -0.104 |
Next | -0.099 |
� | -0.098 |
Salt | -0.098 |
Fire | -0.097 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.90e-02 |
Pos logit contributions | |
goodbye | +0.116 |
ogging | +0.102 |
attention | +0.099 |
away | +0.099 |
des | +0.097 |
Neg logit contributions | |
Ant | -0.087 |
Fire | -0.080 |
Next | -0.079 |
Salt | -0.079 |
poker | -0.078 |
Token | t |
Token ID | 83 |
Feature activation | -6.38e-02 |
Pos logit contributions | |
goodbye | +0.156 |
away | +0.142 |
attention | +0.139 |
farewell | +0.129 |
des | +0.128 |
Neg logit contributions | |
Ant | -0.176 |
hes | -0.169 |
I | -0.167 |
Salt | -0.167 |
Will | -0.164 |
Token | know |
Token ID | 760 |
Feature activation | +6.35e-03 |
Pos logit contributions | |
goodbye | +0.102 |
attention | +0.082 |
farewell | +0.081 |
away | +0.081 |
ogging | +0.081 |
Neg logit contributions | |
Ant | -0.116 |
hes | -0.113 |
Salt | -0.111 |
I | -0.110 |
Next | -0.109 |
Token | which |
Token ID | 543 |
Feature activation | +2.99e-02 |
Pos logit contributions | |
Ant | +0.012 |
� | +0.011 |
Next | +0.011 |
Mouse | +0.011 |
Fire | +0.011 |
Neg logit contributions | |
goodbye | -0.010 |
attention | -0.009 |
away | -0.009 |
ogging | -0.009 |
ly | -0.008 |
Token | way |
Token ID | 835 |
Feature activation | +5.21e-03 |
Pos logit contributions | |
Ant | +0.053 |
� | +0.051 |
Next | +0.051 |
Really | +0.051 |
hes | +0.050 |
Neg logit contributions | |
goodbye | -0.050 |
attention | -0.045 |
away | -0.042 |
back | -0.041 |
ogging | -0.041 |
Token | to |
Token ID | 284 |
Feature activation | +7.23e-03 |
Pos logit contributions | |
Fire | +0.009 |
Salt | +0.008 |
I | +0.008 |
Ant | +0.008 |
brothers | +0.008 |
Neg logit contributions | |
goodbye | -0.008 |
des | -0.007 |
away | -0.007 |
farewell | -0.007 |
ogging | -0.007 |
Token | knew |
Token ID | 2993 |
Feature activation | +5.59e-03 |
Pos logit contributions | |
goodbye | +0.002 |
back | +0.002 |
farewell | +0.002 |
away | +0.002 |
des | +0.002 |
Neg logit contributions | |
poker | -0.002 |
ters | -0.002 |
Ant | -0.002 |
Blocks | -0.002 |
Salt | -0.002 |
Token | he |
Token ID | 339 |
Feature activation | -2.51e-02 |
Pos logit contributions | |
Ant | +0.011 |
� | +0.011 |
Mouse | +0.011 |
Next | +0.011 |
Bees | +0.010 |
Neg logit contributions | |
goodbye | -0.008 |
attention | -0.007 |
away | -0.007 |
ogging | -0.006 |
farewell | -0.006 |
Token | wouldn |
Token ID | 3636 |
Feature activation | -6.27e-02 |
Pos logit contributions | |
goodbye | +0.043 |
ly | +0.037 |
back | +0.037 |
away | +0.037 |
des | +0.036 |
Neg logit contributions | |
Ant | -0.042 |
poker | -0.041 |
Fire | -0.040 |
Blocks | -0.040 |
Salt | -0.040 |
Token | â |
Token ID | 22940 |
Feature activation | -5.71e-02 |
Pos logit contributions | |
goodbye | +0.122 |
ogging | +0.108 |
away | +0.108 |
attention | +0.107 |
farewell | +0.103 |
Neg logit contributions | |
Ant | -0.091 |
Next | -0.085 |
Fire | -0.084 |
Salt | -0.084 |
� | -0.083 |
Token | € |
Token ID | 26391 |
Feature activation | -5.96e-02 |
Pos logit contributions | |
goodbye | +0.093 |
away | +0.080 |
attention | +0.080 |
ogging | +0.078 |
farewell | +0.075 |
Neg logit contributions | |
Ant | -0.102 |
Next | -0.097 |
Salt | -0.096 |
Fire | -0.096 |
� | -0.096 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.89e-02 |
Pos logit contributions | |
goodbye | +0.115 |
ogging | +0.099 |
des | +0.097 |
attention | +0.095 |
away | +0.095 |
Neg logit contributions | |
Ant | -0.092 |
poker | -0.084 |
Mouse | -0.083 |
Fire | -0.082 |
Next | -0.082 |
Token | t |
Token ID | 83 |
Feature activation | -7.53e-02 |
Pos logit contributions | |
goodbye | +0.145 |
away | +0.129 |
attention | +0.128 |
farewell | +0.118 |
des | +0.116 |
Neg logit contributions | |
Ant | -0.186 |
� | -0.177 |
Salt | -0.177 |
I | -0.177 |
hes | -0.177 |
Token | forget |
Token ID | 6044 |
Feature activation | -3.45e-02 |
Pos logit contributions | |
goodbye | +0.119 |
away | +0.101 |
attention | +0.100 |
ogging | +0.098 |
farewell | +0.096 |
Neg logit contributions | |
Ant | -0.137 |
Salt | -0.129 |
Next | -0.129 |
� | -0.128 |
Fire | -0.128 |
Token | how |
Token ID | 703 |
Feature activation | +2.82e-03 |
Pos logit contributions | |
goodbye | +0.055 |
away | +0.048 |
attention | +0.047 |
ogging | +0.046 |
farewell | +0.044 |
Neg logit contributions | |
Ant | -0.063 |
Next | -0.059 |
Salt | -0.059 |
Fire | -0.059 |
� | -0.059 |
Token | delicious |
Token ID | 12625 |
Feature activation | +4.46e-02 |
Pos logit contributions | |
Ant | +0.005 |
Next | +0.005 |
Fire | +0.005 |
Salt | +0.005 |
� | +0.005 |
Neg logit contributions | |
goodbye | -0.005 |
away | -0.004 |
attention | -0.004 |
ogging | -0.004 |
farewell | -0.004 |
Token | it |
Token ID | 340 |
Feature activation | -1.41e-02 |
Pos logit contributions | |
Ant | +0.093 |
Mouse | +0.091 |
hes | +0.087 |
Ch | +0.087 |
� | +0.087 |
Neg logit contributions | |
goodbye | -0.062 |
attention | -0.054 |
away | -0.051 |
ly | -0.048 |
back | -0.048 |
Token | the |
Token ID | 262 |
Feature activation | +3.23e-03 |
Pos logit contributions | |
Ant | +0.046 |
Next | +0.044 |
� | +0.044 |
Fire | +0.043 |
Salt | +0.043 |
Neg logit contributions | |
goodbye | -0.021 |
attention | -0.017 |
away | -0.017 |
ogging | -0.016 |
farewell | -0.015 |
Token | girl |
Token ID | 2576 |
Feature activation | +2.93e-02 |
Pos logit contributions | |
Ant | +0.005 |
Salt | +0.004 |
Next | +0.004 |
� | +0.004 |
Fire | +0.004 |
Neg logit contributions | |
goodbye | -0.007 |
attention | -0.006 |
away | -0.006 |
ogging | -0.006 |
farewell | -0.006 |
Token | couldn |
Token ID | 3521 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
Next | +0.047 |
Ant | +0.047 |
� | +0.046 |
Fire | +0.045 |
Salt | +0.044 |
Neg logit contributions | |
goodbye | -0.054 |
back | -0.044 |
ly | -0.044 |
attention | -0.044 |
away | -0.044 |
Token | â |
Token ID | 22940 |
Feature activation | -5.36e-02 |
Pos logit contributions | |
goodbye | +0.041 |
ogging | +0.035 |
away | +0.035 |
attention | +0.035 |
farewell | +0.034 |
Neg logit contributions | |
Ant | -0.032 |
Next | -0.029 |
Fire | -0.029 |
Salt | -0.029 |
� | -0.029 |
Token | € |
Token ID | 26391 |
Feature activation | -5.95e-02 |
Pos logit contributions | |
goodbye | +0.087 |
attention | +0.075 |
away | +0.074 |
ogging | +0.073 |
farewell | +0.069 |
Neg logit contributions | |
Ant | -0.096 |
� | -0.092 |
Next | -0.091 |
Salt | -0.090 |
hes | -0.090 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.89e-02 |
Pos logit contributions | |
goodbye | +0.122 |
ogging | +0.106 |
away | +0.106 |
attention | +0.106 |
farewell | +0.102 |
Neg logit contributions | |
Ant | -0.083 |
Next | -0.076 |
Fire | -0.076 |
Salt | -0.075 |
� | -0.074 |
Token | t |
Token ID | 83 |
Feature activation | -6.49e-02 |
Pos logit contributions | |
goodbye | +0.167 |
away | +0.155 |
attention | +0.147 |
farewell | +0.144 |
des | +0.143 |
Neg logit contributions | |
I | -0.156 |
Ant | -0.155 |
hes | -0.150 |
Salt | -0.147 |
You | -0.143 |
Token | hold |
Token ID | 1745 |
Feature activation | +1.16e-03 |
Pos logit contributions | |
goodbye | +0.106 |
attention | +0.090 |
ogging | +0.087 |
away | +0.086 |
des | +0.086 |
Neg logit contributions | |
Ant | -0.119 |
Salt | -0.110 |
� | -0.109 |
Will | -0.109 |
Mouse | -0.109 |
Token | on |
Token ID | 319 |
Feature activation | +3.08e-03 |
Pos logit contributions | |
Ant | +0.002 |
Salt | +0.002 |
Next | +0.002 |
� | +0.002 |
Fire | +0.002 |
Neg logit contributions | |
goodbye | -0.002 |
away | -0.002 |
attention | -0.002 |
ogging | -0.002 |
des | -0.002 |
Token | any |
Token ID | 597 |
Feature activation | -6.25e-04 |
Pos logit contributions | |
Ant | +0.006 |
Salt | +0.006 |
Next | +0.006 |
Fire | +0.006 |
� | +0.006 |
Neg logit contributions | |
goodbye | -0.004 |
away | -0.004 |
ogging | -0.004 |
attention | -0.004 |
des | -0.004 |
Token | longer |
Token ID | 2392 |
Feature activation | +5.78e-03 |
Pos logit contributions | |
goodbye | +0.001 |
des | +0.001 |
away | +0.001 |
farewell | +0.001 |
ogging | +0.001 |
Neg logit contributions | |
Ant | -0.001 |
Salt | -0.001 |
Will | -0.001 |
Fire | -0.001 |
brothers | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | +1.71e-02 |
Pos logit contributions | |
goodbye | +0.041 |
attention | +0.035 |
away | +0.035 |
farewell | +0.034 |
des | +0.034 |
Neg logit contributions | |
I | -0.060 |
Ant | -0.060 |
Salt | -0.059 |
Will | -0.059 |
You | -0.058 |
Token | he |
Token ID | 339 |
Feature activation | -6.31e-03 |
Pos logit contributions | |
Ant | +0.029 |
� | +0.027 |
Next | +0.027 |
Fire | +0.026 |
Salt | +0.026 |
Neg logit contributions | |
goodbye | -0.031 |
attention | -0.026 |
away | -0.026 |
ogging | -0.026 |
farewell | -0.025 |
Token | didn |
Token ID | 1422 |
Feature activation | -5.33e-02 |
Pos logit contributions | |
goodbye | +0.013 |
farewell | +0.011 |
away | +0.010 |
back | +0.010 |
attention | +0.010 |
Neg logit contributions | |
Ant | -0.009 |
� | -0.009 |
Salt | -0.009 |
Next | -0.009 |
poker | -0.009 |
Token | â |
Token ID | 22940 |
Feature activation | -5.68e-02 |
Pos logit contributions | |
goodbye | +0.123 |
away | +0.113 |
attention | +0.112 |
des | +0.109 |
ly | +0.107 |
Neg logit contributions | |
Ant | -0.056 |
Next | -0.054 |
hes | -0.053 |
I | -0.053 |
Salt | -0.053 |
Token | € |
Token ID | 26391 |
Feature activation | -5.37e-02 |
Pos logit contributions | |
goodbye | +0.092 |
attention | +0.079 |
away | +0.078 |
ogging | +0.076 |
farewell | +0.073 |
Neg logit contributions | |
Ant | -0.101 |
� | -0.097 |
Next | -0.097 |
Salt | -0.095 |
hes | -0.095 |
Token | ™ |
Token ID | 8151 |
Feature activation | -9.87e-02 |
Pos logit contributions | |
goodbye | +0.105 |
ogging | +0.091 |
attention | +0.089 |
away | +0.089 |
des | +0.087 |
Neg logit contributions | |
Ant | -0.081 |
Fire | -0.074 |
Next | -0.074 |
Salt | -0.074 |
Mouse | -0.073 |
Token | t |
Token ID | 83 |
Feature activation | -7.07e-02 |
Pos logit contributions | |
goodbye | +0.160 |
away | +0.152 |
attention | +0.148 |
farewell | +0.138 |
des | +0.138 |
Neg logit contributions | |
I | -0.160 |
Ant | -0.159 |
hes | -0.154 |
Will | -0.151 |
Salt | -0.151 |
Token | have |
Token ID | 423 |
Feature activation | -6.72e-03 |
Pos logit contributions | |
goodbye | +0.106 |
away | +0.088 |
attention | +0.088 |
ogging | +0.086 |
farewell | +0.084 |
Neg logit contributions | |
Ant | -0.132 |
I | -0.125 |
Salt | -0.123 |
hes | -0.123 |
Will | -0.123 |
Token | any |
Token ID | 597 |
Feature activation | +1.39e-02 |
Pos logit contributions | |
goodbye | +0.008 |
away | +0.007 |
attention | +0.006 |
ogging | +0.006 |
des | +0.006 |
Neg logit contributions | |
Ant | -0.015 |
Next | -0.014 |
Salt | -0.014 |
Fire | -0.014 |
hes | -0.014 |
Token | friends |
Token ID | 2460 |
Feature activation | -2.07e-03 |
Pos logit contributions | |
Ant | +0.021 |
Next | +0.021 |
brothers | +0.021 |
chefs | +0.020 |
Fire | +0.020 |
Neg logit contributions | |
goodbye | -0.024 |
away | -0.022 |
farewell | -0.020 |
des | -0.020 |
ogging | -0.020 |
Token | either |
Token ID | 2035 |
Feature activation | -2.03e-02 |
Pos logit contributions | |
goodbye | +0.003 |
attention | +0.002 |
away | +0.002 |
farewell | +0.002 |
ogging | +0.002 |
Neg logit contributions | |
Ant | -0.004 |
Will | -0.004 |
I | -0.004 |
� | -0.004 |
Salt | -0.004 |
Token | did |
Token ID | 750 |
Feature activation | -7.44e-03 |
Pos logit contributions | |
� | +0.042 |
Ant | +0.040 |
Salt | +0.040 |
kins | +0.039 |
poker | +0.039 |
Neg logit contributions | |
goodbye | -0.046 |
ly | -0.039 |
away | -0.039 |
back | -0.038 |
attention | -0.038 |
Token | not |
Token ID | 407 |
Feature activation | -5.79e-02 |
Pos logit contributions | |
goodbye | +0.019 |
ogging | +0.019 |
away | +0.018 |
ly | +0.018 |
fy | +0.018 |
Neg logit contributions | |
hes | -0.005 |
Ant | -0.005 |
Will | -0.005 |
Next | -0.005 |
You | -0.004 |
Token | trust |
Token ID | 3774 |
Feature activation | -2.49e-02 |
Pos logit contributions | |
goodbye | +0.103 |
ogging | +0.086 |
farewell | +0.084 |
oured | +0.078 |
away | +0.077 |
Neg logit contributions | |
Ant | -0.098 |
You | -0.091 |
hes | -0.091 |
I | -0.089 |
Mouse | -0.088 |
Token | Tom |
Token ID | 4186 |
Feature activation | -4.12e-02 |
Pos logit contributions | |
goodbye | +0.046 |
des | +0.040 |
ogging | +0.040 |
away | +0.040 |
attention | +0.039 |
Neg logit contributions | |
Ant | -0.039 |
Salt | -0.037 |
Fire | -0.036 |
Next | -0.036 |
hes | -0.035 |
Token | because |
Token ID | 780 |
Feature activation | +1.11e-03 |
Pos logit contributions | |
goodbye | +0.064 |
away | +0.058 |
farewell | +0.055 |
des | +0.054 |
ogging | +0.054 |
Neg logit contributions | |
Ant | -0.071 |
Salt | -0.069 |
Fire | -0.068 |
hes | -0.067 |
Will | -0.067 |
Token | of |
Token ID | 286 |
Feature activation | +6.85e-02 |
Pos logit contributions | |
� | +0.003 |
Ant | +0.003 |
Mouse | +0.002 |
Sal | +0.002 |
Next | +0.002 |
Neg logit contributions | |
goodbye | -0.001 |
attention | -0.001 |
ogging | -0.001 |
away | -0.001 |
back | -0.001 |
Token | the |
Token ID | 262 |
Feature activation | -9.24e-03 |
Pos logit contributions | |
� | +0.140 |
Ant | +0.137 |
Mouse | +0.130 |
Really | +0.130 |
hes | +0.130 |
Neg logit contributions | |
goodbye | -0.093 |
attention | -0.085 |
away | -0.078 |
ogging | -0.075 |
back | -0.072 |
Token | whip |
Token ID | 22972 |
Feature activation | +3.76e-02 |
Pos logit contributions | |
goodbye | +0.015 |
attention | +0.015 |
away | +0.013 |
back | +0.013 |
ogging | +0.013 |
Neg logit contributions | |
Ant | -0.016 |
� | -0.015 |
kins | -0.015 |
Next | -0.015 |
Really | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | -6.75e-03 |
Pos logit contributions | |
Ant | +0.093 |
� | +0.090 |
Next | +0.089 |
Really | +0.087 |
Mouse | +0.087 |
Neg logit contributions | |
goodbye | -0.039 |
attention | -0.030 |
ogging | -0.028 |
back | -0.028 |
away | -0.027 |
Token | Tim |
Token ID | 5045 |
Feature activation | -1.16e-02 |
Pos logit contributions | |
goodbye | +0.012 |
attention | +0.010 |
ogging | +0.010 |
ly | +0.010 |
away | +0.009 |
Neg logit contributions | |
Ant | -0.011 |
kins | -0.011 |
� | -0.011 |
chick | -0.010 |
Will | -0.010 |
Token | would |
Token ID | 561 |
Feature activation | -1.38e-02 |
Pos logit contributions | |
goodbye | +0.020 |
ly | +0.018 |
back | +0.018 |
away | +0.017 |
ogging | +0.017 |
Neg logit contributions | |
Ant | -0.018 |
Salt | -0.018 |
� | -0.018 |
Fire | -0.018 |
Next | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -1.66e-02 |
Pos logit contributions | |
Ant | +0.062 |
Next | +0.059 |
� | +0.059 |
Fire | +0.059 |
Mouse | +0.059 |
Neg logit contributions | |
goodbye | -0.032 |
attention | -0.026 |
away | -0.025 |
ogging | -0.024 |
ly | -0.022 |
Token | |
Token ID | 220 |
Feature activation | -1.31e-02 |
Pos logit contributions | |
goodbye | +0.025 |
ogging | +0.021 |
back | +0.020 |
farewell | +0.019 |
toward | +0.018 |
Neg logit contributions | |
kins | -0.032 |
Ant | -0.030 |
hes | -0.029 |
web | -0.029 |
cut | -0.029 |
Token |
|
Token ID | 198 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
goodbye | +0.021 |
ogging | +0.019 |
attention | +0.018 |
back | +0.017 |
away | +0.017 |
Neg logit contributions | |
Ant | -0.022 |
� | -0.022 |
Really | -0.021 |
Mouse | -0.021 |
chefs | -0.021 |
Token | " |
Token ID | 1 |
Feature activation | +3.74e-02 |
Pos logit contributions | |
goodbye | +0.029 |
ogging | +0.027 |
of | +0.025 |
away | +0.025 |
des | +0.025 |
Neg logit contributions | |
poker | -0.029 |
kins | -0.029 |
ities | -0.028 |
� | -0.028 |
chefs | -0.028 |
Token | What |
Token ID | 2061 |
Feature activation | +5.88e-02 |
Pos logit contributions | |
kins | +0.049 |
poker | +0.047 |
brothers | +0.045 |
sibling | +0.045 |
bes | +0.044 |
Neg logit contributions | |
closer | -0.065 |
of | -0.064 |
ax | -0.060 |
des | -0.059 |
away | -0.059 |
Token | are |
Token ID | 389 |
Feature activation | +6.74e-02 |
Pos logit contributions | |
Ant | +0.109 |
Mouse | +0.108 |
� | +0.106 |
Really | +0.102 |
Next | +0.101 |
Neg logit contributions | |
goodbye | -0.090 |
attention | -0.076 |
ogging | -0.071 |
away | -0.070 |
of | -0.068 |
Token | you |
Token ID | 345 |
Feature activation | -2.11e-02 |
Pos logit contributions | |
Mouse | +0.139 |
Ant | +0.138 |
oyd | +0.131 |
Will | +0.131 |
� | +0.131 |
Neg logit contributions | |
goodbye | -0.087 |
attention | -0.078 |
ly | -0.065 |
ogging | -0.065 |
away | -0.065 |
Token | cooking |
Token ID | 10801 |
Feature activation | +4.11e-02 |
Pos logit contributions | |
goodbye | +0.035 |
attention | +0.033 |
ogging | +0.031 |
away | +0.030 |
back | +0.030 |
Neg logit contributions | |
Ant | -0.036 |
Mouse | -0.033 |
Really | -0.033 |
� | -0.032 |
Will | -0.032 |
Token | mummy |
Token ID | 38567 |
Feature activation | +7.74e-03 |
Pos logit contributions | |
Ant | +0.081 |
Mouse | +0.076 |
Next | +0.076 |
Fire | +0.075 |
� | +0.075 |
Neg logit contributions | |
goodbye | -0.060 |
attention | -0.051 |
away | -0.051 |
ogging | -0.049 |
back | -0.047 |
Token | ?" |
Token ID | 1701 |
Feature activation | -9.91e-03 |
Pos logit contributions | |
Ant | +0.016 |
Really | +0.015 |
Mouse | +0.014 |
Fire | +0.014 |
� | +0.014 |
Neg logit contributions | |
goodbye | -0.011 |
attention | -0.009 |
ogging | -0.009 |
ly | -0.009 |
away | -0.009 |
Token | asked |
Token ID | 1965 |
Feature activation | -1.35e-03 |
Pos logit contributions | |
goodbye | +0.017 |
ly | +0.014 |
ogging | +0.013 |
closer | +0.013 |
back | +0.013 |
Neg logit contributions | |
poker | -0.017 |
� | -0.017 |
ators | -0.017 |
Ant | -0.017 |
ates | -0.017 |
Token | " |
Token ID | 1 |
Feature activation | +4.13e-02 |
Pos logit contributions | |
Ant | +0.001 |
Salt | +0.001 |
Fire | +0.001 |
Next | +0.001 |
hes | +0.001 |
Neg logit contributions | |
goodbye | -0.002 |
away | -0.002 |
ogging | -0.002 |
attention | -0.002 |
farewell | -0.002 |
Token | No |
Token ID | 2949 |
Feature activation | +3.47e-02 |
Pos logit contributions | |
kins | +0.059 |
poker | +0.056 |
Fire | +0.056 |
Bees | +0.056 |
hes | +0.055 |
Neg logit contributions | |
goodbye | -0.077 |
away | -0.070 |
closer | -0.068 |
contain | -0.067 |
attention | -0.066 |
Token | , |
Token ID | 11 |
Feature activation | +2.62e-02 |
Pos logit contributions | |
Ant | +0.066 |
brothers | +0.066 |
Salt | +0.065 |
Fire | +0.065 |
twins | +0.064 |
Neg logit contributions | |
away | -0.047 |
goodbye | -0.046 |
des | -0.046 |
contain | -0.041 |
ogging | -0.040 |
Token | mine |
Token ID | 6164 |
Feature activation | +7.06e-02 |
Pos logit contributions | |
Ant | +0.045 |
Next | +0.042 |
Salt | +0.042 |
Fire | +0.042 |
� | +0.042 |
Neg logit contributions | |
goodbye | -0.045 |
away | -0.039 |
attention | -0.039 |
ogging | -0.038 |
farewell | -0.037 |
Token | , |
Token ID | 11 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
Ant | +0.138 |
� | +0.132 |
Next | +0.131 |
Salt | +0.131 |
Fire | +0.130 |
Neg logit contributions | |
goodbye | -0.102 |
des | -0.085 |
away | -0.085 |
ogging | -0.084 |
attention | -0.083 |
Token | mine |
Token ID | 6164 |
Feature activation | +7.21e-02 |
Pos logit contributions | |
Ant | +0.053 |
Salt | +0.051 |
Fire | +0.050 |
Next | +0.050 |
� | +0.050 |
Neg logit contributions | |
goodbye | -0.051 |
away | -0.046 |
attention | -0.045 |
ogging | -0.045 |
des | -0.043 |
Token | , |
Token ID | 11 |
Feature activation | +3.09e-02 |
Pos logit contributions | |
Ant | +0.160 |
� | +0.154 |
Next | +0.153 |
Salt | +0.152 |
Fire | +0.151 |
Neg logit contributions | |
goodbye | -0.085 |
des | -0.068 |
ogging | -0.067 |
away | -0.066 |
farewell | -0.065 |
Token | mine |
Token ID | 6164 |
Feature activation | +7.00e-02 |
Pos logit contributions | |
Ant | +0.060 |
Salt | +0.058 |
Fire | +0.058 |
Next | +0.058 |
� | +0.057 |
Neg logit contributions | |
goodbye | -0.043 |
away | -0.038 |
ogging | -0.037 |
attention | -0.037 |
des | -0.036 |
Token | !" |
Token ID | 2474 |
Feature activation | +3.56e-03 |
Pos logit contributions | |
Ant | +0.156 |
Next | +0.150 |
� | +0.149 |
Salt | +0.148 |
I | +0.147 |
Neg logit contributions | |
goodbye | -0.082 |
des | -0.066 |
ogging | -0.064 |
away | -0.063 |
farewell | -0.063 |
Token | Tom |
Token ID | 4186 |
Feature activation | -4.98e-02 |
Pos logit contributions | |
Ant | +0.005 |
Next | +0.005 |
Salt | +0.005 |
Fire | +0.005 |
� | +0.005 |
Neg logit contributions | |
goodbye | -0.007 |
away | -0.006 |
attention | -0.006 |
ogging | -0.006 |
des | -0.006 |
Token | says |
Token ID | 1139 |
Feature activation | -1.80e-02 |
Pos logit contributions | |
goodbye | +0.090 |
away | +0.082 |
des | +0.078 |
ogging | +0.077 |
attention | +0.076 |
Neg logit contributions | |
Ant | -0.078 |
Next | -0.076 |
Salt | -0.074 |
hes | -0.074 |
Fire | -0.073 |
Token | . |
Token ID | 13 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
Ant | +0.077 |
Salt | +0.074 |
Next | +0.074 |
� | +0.074 |
Fire | +0.073 |
Neg logit contributions | |
goodbye | -0.051 |
away | -0.043 |
attention | -0.043 |
ogging | -0.042 |
farewell | -0.040 |
Token | It |
Token ID | 632 |
Feature activation | +2.57e-02 |
Pos logit contributions | |
goodbye | +0.025 |
ogging | +0.021 |
attention | +0.021 |
farewell | +0.021 |
away | +0.021 |
Neg logit contributions | |
Ant | -0.017 |
Mouse | -0.015 |
kins | -0.015 |
hes | -0.015 |
� | -0.015 |
Token | was |
Token ID | 373 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
Ant | +0.037 |
chefs | +0.036 |
� | +0.035 |
Next | +0.035 |
Salt | +0.034 |
Neg logit contributions | |
goodbye | -0.051 |
back | -0.045 |
closer | -0.043 |
ly | -0.043 |
away | -0.043 |
Token | full |
Token ID | 1336 |
Feature activation | +1.74e-02 |
Pos logit contributions | |
Ant | +0.060 |
� | +0.059 |
Mouse | +0.056 |
Next | +0.056 |
kins | +0.055 |
Neg logit contributions | |
goodbye | -0.047 |
attention | -0.040 |
back | -0.037 |
away | -0.036 |
farewell | -0.035 |
Token | of |
Token ID | 286 |
Feature activation | +7.89e-02 |
Pos logit contributions | |
Ant | +0.036 |
Next | +0.034 |
Salt | +0.034 |
Fire | +0.034 |
� | +0.034 |
Neg logit contributions | |
goodbye | -0.023 |
away | -0.020 |
attention | -0.019 |
ogging | -0.019 |
farewell | -0.018 |
Token | dolls |
Token ID | 36062 |
Feature activation | +7.06e-02 |
Pos logit contributions | |
Ant | +0.136 |
� | +0.132 |
Mouse | +0.123 |
Next | +0.121 |
Bees | +0.120 |
Neg logit contributions | |
goodbye | -0.142 |
attention | -0.131 |
away | -0.120 |
ogging | -0.118 |
ly | -0.113 |
Token | , |
Token ID | 11 |
Feature activation | -9.70e-03 |
Pos logit contributions | |
Ant | +0.146 |
Salt | +0.139 |
Next | +0.139 |
� | +0.139 |
Fire | +0.139 |
Neg logit contributions | |
goodbye | -0.094 |
away | -0.079 |
attention | -0.078 |
ogging | -0.076 |
farewell | -0.072 |
Token | puzzles |
Token ID | 24367 |
Feature activation | +6.69e-02 |
Pos logit contributions | |
goodbye | +0.019 |
attention | +0.016 |
ly | +0.015 |
away | +0.015 |
contain | +0.015 |
Neg logit contributions | |
Ant | -0.016 |
� | -0.015 |
Next | -0.014 |
Mouse | -0.014 |
Fire | -0.014 |
Token | , |
Token ID | 11 |
Feature activation | -2.73e-03 |
Pos logit contributions | |
Ant | +0.143 |
Next | +0.137 |
Salt | +0.137 |
� | +0.136 |
Fire | +0.136 |
Neg logit contributions | |
goodbye | -0.084 |
away | -0.069 |
attention | -0.069 |
ogging | -0.067 |
farewell | -0.063 |
Token | and |
Token ID | 290 |
Feature activation | +2.47e-02 |
Pos logit contributions | |
goodbye | +0.004 |
attention | +0.003 |
away | +0.003 |
ly | +0.003 |
ogging | +0.003 |
Neg logit contributions | |
Ant | -0.006 |
� | -0.005 |
Next | -0.005 |
Mouse | -0.005 |
Fire | -0.005 |
Token | books |
Token ID | 3835 |
Feature activation | +3.10e-02 |
Pos logit contributions | |
Ant | +0.044 |
� | +0.042 |
Mouse | +0.041 |
Next | +0.041 |
Salt | +0.040 |
Neg logit contributions | |
goodbye | -0.044 |
attention | -0.038 |
ly | -0.035 |
away | -0.034 |
back | -0.034 |
Token | Maybe |
Token ID | 6674 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
Ant | +0.027 |
hes | +0.027 |
kins | +0.026 |
� | +0.024 |
Salt | +0.024 |
Neg logit contributions | |
goodbye | -0.045 |
attention | -0.037 |
away | -0.037 |
farewell | -0.037 |
ly | -0.037 |
Token | . |
Token ID | 13 |
Feature activation | +1.94e-02 |
Pos logit contributions | |
goodbye | +0.030 |
away | +0.027 |
ogging | +0.027 |
attention | +0.027 |
farewell | +0.026 |
Neg logit contributions | |
Ant | -0.025 |
Next | -0.024 |
Fire | -0.023 |
Salt | -0.023 |
� | -0.023 |
Token | Maybe |
Token ID | 6674 |
Feature activation | -1.53e-02 |
Pos logit contributions | |
kins | +0.027 |
hes | +0.026 |
Ant | +0.026 |
poker | +0.023 |
Salt | +0.023 |
Neg logit contributions | |
goodbye | -0.042 |
attention | -0.034 |
farewell | -0.034 |
away | -0.034 |
back | -0.033 |
Token | not |
Token ID | 407 |
Feature activation | -1.54e-02 |
Pos logit contributions | |
goodbye | +0.028 |
away | +0.025 |
ogging | +0.025 |
attention | +0.025 |
farewell | +0.024 |
Neg logit contributions | |
Ant | -0.024 |
Next | -0.022 |
Fire | -0.022 |
Salt | -0.022 |
� | -0.022 |
Token | . |
Token ID | 13 |
Feature activation | +1.82e-02 |
Pos logit contributions | |
goodbye | +0.024 |
away | +0.021 |
attention | +0.021 |
ogging | +0.020 |
farewell | +0.019 |
Neg logit contributions | |
Ant | -0.028 |
Next | -0.027 |
Fire | -0.027 |
Salt | -0.027 |
� | -0.027 |
Token | Some |
Token ID | 2773 |
Feature activation | +7.69e-02 |
Pos logit contributions | |
hes | +0.023 |
Ant | +0.023 |
kins | +0.023 |
Salt | +0.021 |
� | +0.020 |
Neg logit contributions | |
goodbye | -0.040 |
attention | -0.034 |
away | -0.033 |
farewell | -0.033 |
des | -0.033 |
Token | strangers |
Token ID | 18981 |
Feature activation | -3.19e-03 |
Pos logit contributions | |
Ant | +0.117 |
Salt | +0.109 |
� | +0.108 |
Next | +0.107 |
Fire | +0.107 |
Neg logit contributions | |
goodbye | -0.151 |
attention | -0.132 |
away | -0.131 |
ogging | -0.128 |
ly | -0.125 |
Token | are |
Token ID | 389 |
Feature activation | +4.79e-02 |
Pos logit contributions | |
goodbye | +0.005 |
away | +0.005 |
attention | +0.005 |
ogging | +0.005 |
farewell | +0.004 |
Neg logit contributions | |
Ant | -0.005 |
Next | -0.005 |
Salt | -0.005 |
� | -0.005 |
Fire | -0.005 |
Token | nice |
Token ID | 3621 |
Feature activation | +2.95e-02 |
Pos logit contributions | |
Ant | +0.087 |
� | +0.082 |
Next | +0.082 |
Salt | +0.081 |
Fire | +0.081 |
Neg logit contributions | |
goodbye | -0.078 |
attention | -0.067 |
away | -0.066 |
ogging | -0.064 |
farewell | -0.062 |
Token | and |
Token ID | 290 |
Feature activation | +3.97e-02 |
Pos logit contributions | |
Ant | +0.053 |
Next | +0.053 |
chefs | +0.052 |
Fire | +0.052 |
Salt | +0.051 |
Neg logit contributions | |
goodbye | -0.044 |
farewell | -0.037 |
away | -0.037 |
ogging | -0.037 |
des | -0.036 |
Token | some |
Token ID | 617 |
Feature activation | +4.41e-02 |
Pos logit contributions | |
Ant | +0.073 |
Salt | +0.068 |
� | +0.068 |
Next | +0.068 |
hes | +0.067 |
Neg logit contributions | |
goodbye | -0.066 |
attention | -0.057 |
away | -0.055 |
farewell | -0.051 |
ogging | -0.051 |
Token | he |
Token ID | 339 |
Feature activation | +3.67e-04 |
Pos logit contributions | |
goodbye | +0.001 |
attention | +0.001 |
ogging | +0.001 |
away | +0.001 |
ly | +0.001 |
Neg logit contributions | |
Ant | -0.001 |
� | -0.001 |
Fire | -0.001 |
Salt | -0.001 |
Mouse | -0.001 |
Token | knew |
Token ID | 2993 |
Feature activation | +5.33e-03 |
Pos logit contributions | |
Blocks | +0.001 |
poker | +0.001 |
� | +0.001 |
Ant | +0.001 |
ters | +0.001 |
Neg logit contributions | |
goodbye | -0.001 |
back | -0.001 |
farewell | -0.001 |
ly | -0.001 |
away | -0.001 |
Token | his |
Token ID | 465 |
Feature activation | +2.05e-02 |
Pos logit contributions | |
Ant | +0.011 |
� | +0.011 |
Mouse | +0.010 |
Next | +0.010 |
Bees | +0.010 |
Neg logit contributions | |
goodbye | -0.008 |
attention | -0.007 |
ogging | -0.006 |
away | -0.006 |
back | -0.006 |
Token | mom |
Token ID | 1995 |
Feature activation | -6.29e-03 |
Pos logit contributions | |
Ant | +0.020 |
Next | +0.018 |
Fire | +0.018 |
Salt | +0.018 |
hes | +0.017 |
Neg logit contributions | |
goodbye | -0.049 |
away | -0.046 |
ogging | -0.045 |
attention | -0.045 |
des | -0.044 |
Token | was |
Token ID | 373 |
Feature activation | +1.06e-02 |
Pos logit contributions | |
goodbye | +0.012 |
ly | +0.010 |
back | +0.010 |
attention | +0.010 |
anymore | +0.009 |
Neg logit contributions | |
Ant | -0.011 |
� | -0.010 |
Really | -0.010 |
Blocks | -0.010 |
Mouse | -0.010 |
Token | right |
Token ID | 826 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
Ant | +0.022 |
� | +0.021 |
Next | +0.021 |
Salt | +0.021 |
Fire | +0.021 |
Neg logit contributions | |
goodbye | -0.015 |
attention | -0.012 |
away | -0.012 |
ogging | -0.012 |
farewell | -0.011 |
Token | . |
Token ID | 13 |
Feature activation | -1.29e-02 |
Pos logit contributions | |
Ant | +0.025 |
Salt | +0.025 |
Will | +0.025 |
kins | +0.025 |
chefs | +0.024 |
Neg logit contributions | |
goodbye | -0.027 |
ogging | -0.025 |
des | -0.024 |
farewell | -0.024 |
away | -0.023 |
Token | He |
Token ID | 679 |
Feature activation | +2.73e-02 |
Pos logit contributions | |
goodbye | +0.023 |
ly | +0.018 |
ogging | +0.018 |
farewell | +0.018 |
attention | +0.018 |
Neg logit contributions | |
kins | -0.024 |
Ant | -0.023 |
hes | -0.022 |
kin | -0.021 |
ters | -0.021 |
Token | got |
Token ID | 1392 |
Feature activation | -8.86e-04 |
Pos logit contributions | |
Ant | +0.046 |
� | +0.046 |
Blocks | +0.044 |
Salt | +0.044 |
Next | +0.043 |
Neg logit contributions | |
goodbye | -0.049 |
ly | -0.040 |
farewell | -0.040 |
back | -0.039 |
away | -0.039 |
Token | up |
Token ID | 510 |
Feature activation | -1.30e-02 |
Pos logit contributions | |
goodbye | +0.001 |
attention | +0.001 |
away | +0.001 |
ogging | +0.001 |
back | +0.001 |
Neg logit contributions | |
Ant | -0.002 |
� | -0.002 |
Next | -0.002 |
Salt | -0.002 |
Mouse | -0.002 |
Token | from |
Token ID | 422 |
Feature activation | +8.61e-03 |
Pos logit contributions | |
goodbye | +0.020 |
away | +0.018 |
des | +0.018 |
attention | +0.018 |
ogging | +0.018 |
Neg logit contributions | |
hes | -0.022 |
Salt | -0.022 |
Ant | -0.022 |
Fire | -0.021 |
Next | -0.021 |
Token | new |
Token ID | 649 |
Feature activation | +3.20e-02 |
Pos logit contributions | |
Ant | +0.033 |
Fire | +0.031 |
Next | +0.031 |
Salt | +0.031 |
� | +0.030 |
Neg logit contributions | |
goodbye | -0.047 |
away | -0.042 |
ogging | -0.041 |
attention | -0.041 |
des | -0.040 |
Token | friend |
Token ID | 1545 |
Feature activation | +3.15e-02 |
Pos logit contributions | |
Ant | +0.046 |
Next | +0.043 |
hes | +0.043 |
� | +0.043 |
Salt | +0.043 |
Neg logit contributions | |
goodbye | -0.064 |
attention | -0.057 |
away | -0.056 |
ogging | -0.055 |
farewell | -0.053 |
Token | . |
Token ID | 13 |
Feature activation | +1.96e-02 |
Pos logit contributions | |
Ant | +0.076 |
� | +0.072 |
hes | +0.072 |
Next | +0.072 |
Mouse | +0.072 |
Neg logit contributions | |
goodbye | -0.034 |
attention | -0.025 |
away | -0.024 |
back | -0.024 |
ogging | -0.023 |
Token | He |
Token ID | 679 |
Feature activation | +3.85e-02 |
Pos logit contributions | |
Ant | +0.028 |
kins | +0.027 |
hes | +0.026 |
� | +0.025 |
poker | +0.024 |
Neg logit contributions | |
goodbye | -0.041 |
attention | -0.035 |
des | -0.034 |
away | -0.034 |
farewell | -0.033 |
Token | was |
Token ID | 373 |
Feature activation | +3.47e-02 |
Pos logit contributions | |
Ant | +0.066 |
poker | +0.061 |
Really | +0.061 |
Salt | +0.061 |
� | +0.060 |
Neg logit contributions | |
goodbye | -0.067 |
attention | -0.057 |
back | -0.055 |
des | -0.054 |
away | -0.054 |
Token | in |
Token ID | 287 |
Feature activation | +2.96e-02 |
Pos logit contributions | |
Ant | +0.063 |
� | +0.062 |
Next | +0.060 |
Salt | +0.059 |
Fire | +0.058 |
Neg logit contributions | |
goodbye | -0.057 |
attention | -0.049 |
away | -0.048 |
ogging | -0.045 |
back | -0.044 |
Token | the |
Token ID | 262 |
Feature activation | -4.41e-03 |
Pos logit contributions | |
Ant | +0.055 |
� | +0.053 |
Next | +0.051 |
hes | +0.051 |
Bees | +0.051 |
Neg logit contributions | |
goodbye | -0.048 |
away | -0.041 |
attention | -0.041 |
ogging | -0.038 |
farewell | -0.037 |
Token | shelter |
Token ID | 11772 |
Feature activation | +3.27e-02 |
Pos logit contributions | |
goodbye | +0.010 |
attention | +0.009 |
away | +0.009 |
back | +0.009 |
farewell | +0.009 |
Neg logit contributions | |
Ant | -0.005 |
Next | -0.005 |
� | -0.005 |
kins | -0.005 |
hes | -0.005 |
Token | for |
Token ID | 329 |
Feature activation | +3.66e-02 |
Pos logit contributions | |
Ant | +0.070 |
Next | +0.066 |
� | +0.066 |
Salt | +0.066 |
Fire | +0.066 |
Neg logit contributions | |
goodbye | -0.043 |
away | -0.035 |
attention | -0.035 |
ogging | -0.034 |
farewell | -0.032 |
Token | a |
Token ID | 257 |
Feature activation | -4.69e-02 |
Pos logit contributions | |
Ant | +0.066 |
� | +0.066 |
Next | +0.063 |
Bees | +0.062 |
Mouse | +0.062 |
Neg logit contributions | |
goodbye | -0.059 |
attention | -0.051 |
away | -0.049 |
ogging | -0.048 |
farewell | -0.047 |
Token | long |
Token ID | 890 |
Feature activation | -5.79e-02 |
Pos logit contributions | |
goodbye | +0.074 |
ogging | +0.069 |
attention | +0.068 |
away | +0.065 |
ly | +0.063 |
Neg logit contributions | |
Ant | -0.082 |
Will | -0.077 |
Fire | -0.076 |
� | -0.074 |
Mouse | -0.072 |
Token | inside |
Token ID | 2641 |
Feature activation | +4.57e-02 |
Pos logit contributions | |
goodbye | +0.071 |
ogging | +0.063 |
away | +0.062 |
attention | +0.061 |
des | +0.059 |
Neg logit contributions | |
Ant | -0.066 |
Next | -0.064 |
� | -0.064 |
Will | -0.063 |
I | -0.062 |
Token | . |
Token ID | 13 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
Ant | +0.086 |
Salt | +0.082 |
Fire | +0.081 |
Next | +0.081 |
Will | +0.081 |
Neg logit contributions | |
goodbye | -0.069 |
attention | -0.058 |
ogging | -0.058 |
away | -0.057 |
farewell | -0.056 |
Token | But |
Token ID | 887 |
Feature activation | +2.40e-02 |
Pos logit contributions | |
kins | +0.029 |
hes | +0.027 |
Ant | +0.027 |
ins | +0.026 |
chefs | +0.025 |
Neg logit contributions | |
goodbye | -0.034 |
farewell | -0.028 |
attention | -0.027 |
des | -0.027 |
back | -0.026 |
Token | be |
Token ID | 307 |
Feature activation | +5.17e-02 |
Pos logit contributions | |
hes | +0.045 |
Ant | +0.044 |
Mouse | +0.042 |
kins | +0.042 |
Cere | +0.042 |
Neg logit contributions | |
goodbye | -0.037 |
attention | -0.033 |
away | -0.030 |
back | -0.030 |
closer | -0.029 |
Token | careful |
Token ID | 8161 |
Feature activation | +2.06e-02 |
Pos logit contributions | |
� | +0.092 |
Ant | +0.088 |
Fire | +0.083 |
poker | +0.083 |
Next | +0.083 |
Neg logit contributions | |
goodbye | -0.084 |
attention | -0.076 |
back | -0.073 |
away | -0.073 |
closer | -0.070 |
Token | , |
Token ID | 11 |
Feature activation | +1.82e-02 |
Pos logit contributions | |
Ant | +0.043 |
Next | +0.041 |
Salt | +0.041 |
� | +0.041 |
Fire | +0.040 |
Neg logit contributions | |
goodbye | -0.028 |
away | -0.023 |
attention | -0.023 |
ogging | -0.023 |
farewell | -0.021 |
Token | it |
Token ID | 340 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
Ant | +0.029 |
Next | +0.028 |
� | +0.028 |
Salt | +0.027 |
Fire | +0.027 |
Neg logit contributions | |
goodbye | -0.033 |
attention | -0.029 |
away | -0.029 |
ogging | -0.028 |
farewell | -0.027 |
Token | is |
Token ID | 318 |
Feature activation | +3.19e-02 |
Pos logit contributions | |
Really | +0.028 |
Ant | +0.028 |
� | +0.027 |
Next | +0.027 |
Daddy | +0.026 |
Neg logit contributions | |
goodbye | -0.025 |
attention | -0.022 |
back | -0.021 |
away | -0.021 |
ly | -0.021 |
Token | very |
Token ID | 845 |
Feature activation | +2.23e-02 |
Pos logit contributions | |
Ant | +0.057 |
� | +0.056 |
Next | +0.054 |
Mouse | +0.054 |
Fire | +0.053 |
Neg logit contributions | |
goodbye | -0.051 |
attention | -0.046 |
away | -0.044 |
back | -0.042 |
farewell | -0.040 |
Token | cozy |
Token ID | 37438 |
Feature activation | +3.29e-02 |
Pos logit contributions | |
Ant | +0.038 |
Next | +0.036 |
Fire | +0.036 |
hes | +0.036 |
Salt | +0.036 |
Neg logit contributions | |
goodbye | -0.039 |
ogging | -0.034 |
away | -0.033 |
farewell | -0.032 |
attention | -0.032 |
Token | and |
Token ID | 290 |
Feature activation | +2.64e-02 |
Pos logit contributions | |
Ant | +0.059 |
Next | +0.057 |
Fire | +0.056 |
Salt | +0.056 |
chefs | +0.056 |
Neg logit contributions | |
goodbye | -0.051 |
ogging | -0.044 |
attention | -0.044 |
farewell | -0.043 |
away | -0.043 |
Token | creative |
Token ID | 7325 |
Feature activation | +5.93e-03 |
Pos logit contributions | |
Ant | +0.012 |
� | +0.012 |
Salt | +0.011 |
Next | +0.011 |
Fire | +0.011 |
Neg logit contributions | |
goodbye | -0.009 |
attention | -0.007 |
away | -0.007 |
ogging | -0.007 |
farewell | -0.006 |
Token | with |
Token ID | 351 |
Feature activation | +4.86e-02 |
Pos logit contributions | |
Ant | +0.011 |
Next | +0.011 |
Salt | +0.011 |
� | +0.011 |
Fire | +0.011 |
Neg logit contributions | |
goodbye | -0.008 |
away | -0.007 |
ogging | -0.007 |
attention | -0.007 |
farewell | -0.007 |
Token | the |
Token ID | 262 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
Ant | +0.094 |
� | +0.090 |
Mouse | +0.087 |
Next | +0.086 |
Bees | +0.086 |
Neg logit contributions | |
goodbye | -0.079 |
attention | -0.069 |
away | -0.063 |
farewell | -0.061 |
back | -0.060 |
Token | jar |
Token ID | 17379 |
Feature activation | +1.37e-02 |
Pos logit contributions | |
goodbye | +0.028 |
attention | +0.026 |
away | +0.024 |
ogging | +0.023 |
ly | +0.023 |
Neg logit contributions | |
Ant | -0.023 |
� | -0.022 |
hes | -0.022 |
Next | -0.022 |
Salt | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -1.03e-02 |
Pos logit contributions | |
Ant | +0.026 |
hes | +0.025 |
Salt | +0.025 |
kins | +0.025 |
Next | +0.025 |
Neg logit contributions | |
goodbye | -0.020 |
away | -0.017 |
attention | -0.017 |
ogging | -0.016 |
farewell | -0.016 |
Token | She |
Token ID | 1375 |
Feature activation | +2.21e-02 |
Pos logit contributions | |
goodbye | +0.016 |
ogging | +0.013 |
farewell | +0.012 |
of | +0.011 |
back | +0.011 |
Neg logit contributions | |
kins | -0.021 |
poker | -0.019 |
chefs | -0.019 |
Ant | -0.019 |
hes | -0.019 |
Token | used |
Token ID | 973 |
Feature activation | +6.76e-03 |
Pos logit contributions | |
Ant | +0.039 |
poker | +0.037 |
Fire | +0.037 |
Salt | +0.037 |
Next | +0.037 |
Neg logit contributions | |
goodbye | -0.038 |
farewell | -0.030 |
back | -0.030 |
ly | -0.030 |
away | -0.029 |
Token | the |
Token ID | 262 |
Feature activation | -1.82e-02 |
Pos logit contributions | |
Ant | +0.014 |
� | +0.013 |
Mouse | +0.013 |
Bees | +0.013 |
Next | +0.013 |
Neg logit contributions | |
goodbye | -0.010 |
attention | -0.009 |
away | -0.008 |
farewell | -0.008 |
back | -0.008 |
Token | pieces |
Token ID | 5207 |
Feature activation | +8.79e-03 |
Pos logit contributions | |
goodbye | +0.036 |
attention | +0.033 |
away | +0.031 |
ogging | +0.030 |
farewell | +0.030 |
Neg logit contributions | |
Ant | -0.026 |
Next | -0.025 |
Fire | -0.025 |
hes | -0.025 |
� | -0.024 |
Token | of |
Token ID | 286 |
Feature activation | +4.46e-02 |
Pos logit contributions | |
Ant | +0.017 |
Salt | +0.016 |
hes | +0.016 |
kins | +0.016 |
Fire | +0.016 |
Neg logit contributions | |
goodbye | -0.012 |
ogging | -0.010 |
des | -0.010 |
away | -0.010 |
attention | -0.010 |
Token | the |
Token ID | 262 |
Feature activation | -2.27e-02 |
Pos logit contributions | |
� | +0.083 |
Ant | +0.083 |
Next | +0.082 |
Mouse | +0.080 |
hes | +0.080 |
Neg logit contributions | |
goodbye | -0.066 |
attention | -0.062 |
away | -0.057 |
ogging | -0.051 |
of | -0.051 |
Token | the |
Token ID | 262 |
Feature activation | +3.21e-03 |
Pos logit contributions | |
Ant | +0.052 |
hes | +0.049 |
kins | +0.048 |
� | +0.048 |
Salt | +0.048 |
Neg logit contributions | |
goodbye | -0.044 |
attention | -0.037 |
away | -0.036 |
ogging | -0.034 |
farewell | -0.034 |
Token | classroom |
Token ID | 15806 |
Feature activation | +3.23e-02 |
Pos logit contributions | |
Ant | +0.005 |
kins | +0.005 |
hes | +0.005 |
� | +0.005 |
ters | +0.005 |
Neg logit contributions | |
goodbye | -0.006 |
back | -0.005 |
away | -0.005 |
attention | -0.005 |
home | -0.005 |
Token | . |
Token ID | 13 |
Feature activation | -2.68e-03 |
Pos logit contributions | |
Ant | +0.076 |
Mouse | +0.072 |
Salt | +0.071 |
Fire | +0.071 |
� | +0.071 |
Neg logit contributions | |
goodbye | -0.037 |
away | -0.028 |
attention | -0.028 |
ogging | -0.028 |
ly | -0.026 |
Token | They |
Token ID | 1119 |
Feature activation | -1.12e-03 |
Pos logit contributions | |
goodbye | +0.005 |
back | +0.004 |
farewell | +0.004 |
anymore | +0.004 |
attention | +0.004 |
Neg logit contributions | |
if | -0.005 |
iflower | -0.005 |
we | -0.005 |
kins | -0.005 |
ins | -0.004 |
Token | liked |
Token ID | 8288 |
Feature activation | +1.70e-02 |
Pos logit contributions | |
goodbye | +0.002 |
away | +0.002 |
contain | +0.002 |
attention | +0.002 |
back | +0.002 |
Neg logit contributions | |
Ant | -0.001 |
Salt | -0.001 |
Mouse | -0.001 |
poker | -0.001 |
Fire | -0.001 |
Token | to |
Token ID | 284 |
Feature activation | +8.58e-03 |
Pos logit contributions | |
Ant | +0.047 |
� | +0.045 |
Next | +0.045 |
Salt | +0.045 |
Fire | +0.044 |
Neg logit contributions | |
goodbye | -0.012 |
attention | -0.008 |
away | -0.008 |
ogging | -0.007 |
farewell | -0.006 |
Token | play |
Token ID | 711 |
Feature activation | -5.50e-03 |
Pos logit contributions | |
Ant | +0.016 |
Next | +0.015 |
� | +0.015 |
Salt | +0.015 |
Fire | +0.015 |
Neg logit contributions | |
goodbye | -0.013 |
away | -0.011 |
attention | -0.011 |
ogging | -0.011 |
des | -0.011 |
Token | with |
Token ID | 351 |
Feature activation | +2.68e-02 |
Pos logit contributions | |
goodbye | +0.009 |
away | +0.008 |
attention | +0.008 |
ogging | +0.008 |
farewell | +0.007 |
Neg logit contributions | |
Ant | -0.010 |
Next | -0.009 |
� | -0.009 |
Fire | -0.009 |
Salt | -0.009 |
Token | the |
Token ID | 262 |
Feature activation | -1.89e-02 |
Pos logit contributions | |
Ant | +0.040 |
� | +0.040 |
Mouse | +0.038 |
Next | +0.038 |
Bees | +0.036 |
Neg logit contributions | |
goodbye | -0.051 |
attention | -0.047 |
away | -0.047 |
back | -0.043 |
ogging | -0.043 |
Token | toys |
Token ID | 14958 |
Feature activation | +4.74e-02 |
Pos logit contributions | |
goodbye | +0.037 |
attention | +0.033 |
away | +0.033 |
ogging | +0.032 |
farewell | +0.031 |
Neg logit contributions | |
Ant | -0.028 |
� | -0.026 |
Next | -0.026 |
Salt | -0.026 |
Fire | -0.025 |
Token | and |
Token ID | 290 |
Feature activation | +7.44e-03 |
Pos logit contributions | |
Ant | +0.105 |
Next | +0.101 |
� | +0.100 |
Fire | +0.100 |
Mouse | +0.099 |
Neg logit contributions | |
goodbye | -0.059 |
attention | -0.048 |
away | -0.047 |
ogging | -0.046 |
ly | -0.043 |
Token | not |
Token ID | 407 |
Feature activation | -4.41e-02 |
Pos logit contributions | |
goodbye | +0.004 |
away | +0.004 |
ogging | +0.004 |
attention | +0.004 |
des | +0.004 |
Neg logit contributions | |
Ant | -0.002 |
hes | -0.002 |
Salt | -0.002 |
Next | -0.002 |
Fire | -0.002 |
Token | breathe |
Token ID | 18044 |
Feature activation | -2.08e-04 |
Pos logit contributions | |
goodbye | +0.078 |
ogging | +0.066 |
attention | +0.065 |
des | +0.064 |
farewell | +0.063 |
Neg logit contributions | |
Ant | -0.081 |
I | -0.072 |
Salt | -0.070 |
Mouse | -0.070 |
Fire | -0.069 |
Token | . |
Token ID | 13 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
goodbye | +0.000 |
away | +0.000 |
attention | +0.000 |
ogging | +0.000 |
des | +0.000 |
Neg logit contributions | |
Ant | -0.000 |
Salt | -0.000 |
Fire | -0.000 |
Next | -0.000 |
hes | -0.000 |
Token | He |
Token ID | 679 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
� | +0.029 |
Ant | +0.028 |
kins | +0.028 |
sibling | +0.026 |
flower | +0.026 |
Neg logit contributions | |
goodbye | -0.030 |
ogging | -0.026 |
attention | -0.025 |
back | -0.024 |
away | -0.024 |
Token | was |
Token ID | 373 |
Feature activation | +2.76e-02 |
Pos logit contributions | |
� | +0.051 |
Next | +0.047 |
Ant | +0.046 |
Cere | +0.045 |
Salt | +0.045 |
Neg logit contributions | |
goodbye | -0.055 |
ly | -0.047 |
back | -0.046 |
away | -0.046 |
farewell | -0.045 |
Token | scared |
Token ID | 12008 |
Feature activation | -1.91e-02 |
Pos logit contributions | |
Ant | +0.049 |
� | +0.049 |
Next | +0.046 |
Mouse | +0.046 |
Salt | +0.046 |
Neg logit contributions | |
goodbye | -0.046 |
attention | -0.040 |
away | -0.038 |
back | -0.037 |
ogging | -0.037 |
Token | . |
Token ID | 13 |
Feature activation | +1.76e-02 |
Pos logit contributions | |
goodbye | +0.034 |
des | +0.032 |
farewell | +0.031 |
contain | +0.030 |
away | +0.029 |
Neg logit contributions | |
Fire | -0.029 |
Salt | -0.029 |
twins | -0.029 |
brothers | -0.029 |
Ant | -0.029 |
Token |
|
Token ID | 198 |
Feature activation | +5.85e-03 |
Pos logit contributions | |
� | +0.031 |
kins | +0.030 |
Ant | +0.029 |
sibling | +0.028 |
chick | +0.028 |
Neg logit contributions | |
goodbye | -0.028 |
ogging | -0.023 |
back | -0.023 |
attention | -0.023 |
away | -0.021 |
Token |
|
Token ID | 198 |
Feature activation | +4.57e-03 |
Pos logit contributions | |
Ant | +0.007 |
Next | +0.007 |
Salt | +0.006 |
Fire | +0.006 |
� | +0.006 |
Neg logit contributions | |
goodbye | -0.013 |
away | -0.012 |
attention | -0.012 |
ogging | -0.011 |
farewell | -0.011 |
Token | L |
Token ID | 43 |
Feature activation | +1.68e-02 |
Pos logit contributions | |
Ant | +0.006 |
� | +0.005 |
Next | +0.005 |
Fire | +0.005 |
Salt | +0.005 |
Neg logit contributions | |
goodbye | -0.009 |
away | -0.009 |
ogging | -0.008 |
attention | -0.008 |
des | -0.008 |
Token | ily |
Token ID | 813 |
Feature activation | -6.91e-04 |
Pos logit contributions | |
hes | +0.017 |
Ant | +0.017 |
Next | +0.016 |
I | +0.016 |
kins | +0.016 |
Neg logit contributions | |
goodbye | -0.040 |
attention | -0.036 |
away | -0.035 |
farewell | -0.035 |
des | -0.034 |
Token | " |
Token ID | 366 |
Feature activation | +5.28e-02 |
Pos logit contributions | |
goodbye | +0.018 |
attention | +0.016 |
ly | +0.016 |
farewell | +0.016 |
ogging | +0.015 |
Neg logit contributions | |
Ant | -0.009 |
hes | -0.008 |
Next | -0.008 |
Salt | -0.008 |
kins | -0.008 |
Token | Let |
Token ID | 5756 |
Feature activation | -1.51e-02 |
Pos logit contributions | |
poker | +0.043 |
hes | +0.042 |
kins | +0.041 |
brothers | +0.039 |
Ant | +0.038 |
Neg logit contributions | |
goodbye | -0.130 |
away | -0.129 |
des | -0.128 |
of | -0.122 |
ly | -0.121 |
Token | 's |
Token ID | 338 |
Feature activation | +1.98e-02 |
Pos logit contributions | |
goodbye | +0.033 |
away | +0.031 |
attention | +0.030 |
ogging | +0.030 |
contain | +0.030 |
Neg logit contributions | |
Next | -0.017 |
I | -0.017 |
Ant | -0.017 |
� | -0.017 |
Fire | -0.016 |
Token | go |
Token ID | 467 |
Feature activation | -5.02e-02 |
Pos logit contributions | |
Ant | +0.044 |
hes | +0.043 |
Next | +0.042 |
� | +0.042 |
Salt | +0.042 |
Neg logit contributions | |
goodbye | -0.023 |
attention | -0.019 |
away | -0.019 |
ogging | -0.017 |
farewell | -0.017 |
Token | to |
Token ID | 284 |
Feature activation | -3.27e-02 |
Pos logit contributions | |
goodbye | +0.091 |
farewell | +0.083 |
ly | +0.077 |
away | +0.075 |
des | +0.073 |
Neg logit contributions | |
chefs | -0.076 |
I | -0.073 |
brothers | -0.073 |
Mouse | -0.071 |
Will | -0.070 |
Token | the |
Token ID | 262 |
Feature activation | -3.97e-02 |
Pos logit contributions | |
goodbye | +0.067 |
ly | +0.061 |
ogging | +0.060 |
away | +0.060 |
attention | +0.060 |
Neg logit contributions | |
Fire | -0.043 |
Ant | -0.043 |
Salt | -0.041 |
Next | -0.041 |
Mouse | -0.039 |
Token | attic |
Token ID | 46113 |
Feature activation | +2.42e-02 |
Pos logit contributions | |
goodbye | +0.077 |
ogging | +0.070 |
away | +0.069 |
attention | +0.069 |
ly | +0.067 |
Neg logit contributions | |
Ant | -0.058 |
Fire | -0.056 |
Salt | -0.053 |
I | -0.053 |
Next | -0.053 |
Token | , |
Token ID | 11 |
Feature activation | +1.47e-02 |
Pos logit contributions | |
Ant | +0.044 |
� | +0.043 |
Fire | +0.043 |
Salt | +0.042 |
Next | +0.042 |
Neg logit contributions | |
goodbye | -0.035 |
attention | -0.030 |
farewell | -0.030 |
des | -0.030 |
ogging | -0.030 |
Token | Jack |
Token ID | 3619 |
Feature activation | -1.27e-02 |
Pos logit contributions | |
Ant | +0.024 |
Next | +0.022 |
Fire | +0.022 |
Salt | +0.022 |
� | +0.022 |
Neg logit contributions | |
goodbye | -0.026 |
away | -0.023 |
attention | -0.023 |
ogging | -0.023 |
des | -0.022 |
Token | ! |
Token ID | 0 |
Feature activation | -1.07e-02 |
Pos logit contributions | |
goodbye | +0.011 |
away | +0.009 |
attention | +0.009 |
ogging | +0.008 |
ly | +0.007 |
Neg logit contributions | |
Ant | -0.032 |
Next | -0.031 |
hes | -0.031 |
Fire | -0.030 |
� | -0.030 |
Token | I |
Token ID | 314 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
goodbye | +0.020 |
farewell | +0.016 |
back | +0.016 |
attention | +0.016 |
ogging | +0.016 |
Neg logit contributions | |
kins | -0.018 |
Ant | -0.017 |
hes | -0.017 |
iflower | -0.016 |
ins | -0.015 |
Token | something |
Token ID | 1223 |
Feature activation | -1.68e-02 |
Pos logit contributions | |
Ant | +0.037 |
� | +0.036 |
Next | +0.035 |
Salt | +0.034 |
Fire | +0.034 |
Neg logit contributions | |
goodbye | -0.038 |
attention | -0.033 |
away | -0.032 |
ogging | -0.031 |
farewell | -0.030 |
Token | new |
Token ID | 649 |
Feature activation | +2.81e-02 |
Pos logit contributions | |
goodbye | +0.027 |
away | +0.024 |
ogging | +0.023 |
des | +0.023 |
farewell | +0.023 |
Neg logit contributions | |
Ant | -0.029 |
Salt | -0.028 |
Will | -0.027 |
Fire | -0.027 |
hes | -0.027 |
Token | and |
Token ID | 290 |
Feature activation | +3.06e-02 |
Pos logit contributions | |
Ant | +0.054 |
Next | +0.051 |
Salt | +0.051 |
� | +0.051 |
Fire | +0.051 |
Neg logit contributions | |
goodbye | -0.041 |
away | -0.034 |
attention | -0.034 |
ogging | -0.034 |
des | -0.033 |
Token | exciting |
Token ID | 7895 |
Feature activation | +1.11e-02 |
Pos logit contributions | |
Ant | +0.060 |
Next | +0.057 |
Salt | +0.057 |
hes | +0.057 |
Fire | +0.057 |
Neg logit contributions | |
goodbye | -0.043 |
ogging | -0.037 |
away | -0.037 |
attention | -0.035 |
farewell | -0.034 |
Token | . |
Token ID | 13 |
Feature activation | -3.02e-03 |
Pos logit contributions | |
Ant | +0.018 |
Next | +0.018 |
kins | +0.018 |
Will | +0.018 |
� | +0.018 |
Neg logit contributions | |
goodbye | -0.017 |
ogging | -0.015 |
des | -0.015 |
away | -0.015 |
farewell | -0.015 |
Token | Lisa |
Token ID | 15378 |
Feature activation | -3.11e-02 |
Pos logit contributions | |
goodbye | +0.005 |
farewell | +0.004 |
ogging | +0.004 |
attention | +0.004 |
owl | +0.004 |
Neg logit contributions | |
kins | -0.006 |
Ant | -0.006 |
chefs | -0.005 |
ters | -0.005 |
poker | -0.005 |
Token | looked |
Token ID | 3114 |
Feature activation | -3.99e-02 |
Pos logit contributions | |
goodbye | +0.057 |
away | +0.048 |
ly | +0.047 |
ogging | +0.047 |
farewell | +0.047 |
Neg logit contributions | |
Ant | -0.050 |
Next | -0.048 |
Fire | -0.048 |
Salt | -0.048 |
hes | -0.047 |
Token | up |
Token ID | 510 |
Feature activation | +8.40e-03 |
Pos logit contributions | |
des | +0.070 |
ogging | +0.067 |
oot | +0.064 |
oor | +0.061 |
ong | +0.060 |
Neg logit contributions | |
I | -0.050 |
brothers | -0.049 |
Will | -0.049 |
Let | -0.049 |
Next | -0.049 |
Token | and |
Token ID | 290 |
Feature activation | -2.17e-03 |
Pos logit contributions | |
I | +0.011 |
Next | +0.010 |
� | +0.010 |
Will | +0.010 |
Let | +0.010 |
Neg logit contributions | |
des | -0.016 |
goodbye | -0.015 |
ooters | -0.015 |
oot | -0.015 |
ogging | -0.015 |
Token | down |
Token ID | 866 |
Feature activation | +2.42e-02 |
Pos logit contributions | |
goodbye | +0.004 |
away | +0.004 |
ogging | +0.004 |
attention | +0.004 |
des | +0.004 |
Neg logit contributions | |
Ant | -0.003 |
Next | -0.003 |
� | -0.003 |
Salt | -0.003 |
Fire | -0.003 |
Token | streets |
Token ID | 6483 |
Feature activation | +5.58e-03 |
Pos logit contributions | |
Ant | +0.042 |
� | +0.041 |
Next | +0.041 |
I | +0.041 |
Salt | +0.040 |
Neg logit contributions | |
goodbye | -0.036 |
des | -0.034 |
ogging | -0.033 |
attention | -0.032 |
away | -0.032 |
Token | ?" |
Token ID | 1701 |
Feature activation | +3.12e-04 |
Pos logit contributions | |
Ant | +0.061 |
Fire | +0.058 |
Next | +0.058 |
Salt | +0.057 |
� | +0.057 |
Neg logit contributions | |
goodbye | -0.072 |
attention | -0.062 |
away | -0.062 |
ogging | -0.061 |
des | -0.061 |
Token | Mom |
Token ID | 11254 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
Ant | +0.001 |
� | +0.001 |
Mouse | +0.000 |
poker | +0.000 |
kins | +0.000 |
Neg logit contributions | |
goodbye | -0.001 |
attention | -0.000 |
away | -0.000 |
ogging | -0.000 |
farewell | -0.000 |
Token | asks |
Token ID | 7893 |
Feature activation | +1.61e-02 |
Pos logit contributions | |
goodbye | +0.024 |
away | +0.021 |
attention | +0.021 |
ogging | +0.021 |
farewell | +0.020 |
Neg logit contributions | |
Ant | -0.018 |
Next | -0.017 |
� | -0.017 |
Salt | -0.017 |
Fire | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
Ant | +0.025 |
twins | +0.025 |
ters | +0.024 |
kins | +0.024 |
Salt | +0.024 |
Neg logit contributions | |
goodbye | -0.027 |
des | -0.026 |
away | -0.026 |
fy | -0.025 |
farewell | -0.024 |
Token | She |
Token ID | 1375 |
Feature activation | +2.86e-02 |
Pos logit contributions | |
Ant | +0.029 |
kins | +0.027 |
ates | +0.027 |
� | +0.026 |
poker | +0.026 |
Neg logit contributions | |
goodbye | -0.046 |
back | -0.039 |
attention | -0.038 |
ogging | -0.037 |
farewell | -0.037 |
Token | hears |
Token ID | 23180 |
Feature activation | -3.60e-02 |
Pos logit contributions | |
Ant | +0.048 |
Next | +0.046 |
Fire | +0.045 |
� | +0.045 |
Salt | +0.045 |
Neg logit contributions | |
goodbye | -0.050 |
away | -0.044 |
attention | -0.043 |
ogging | -0.042 |
des | -0.041 |
Token | them |
Token ID | 606 |
Feature activation | -5.31e-02 |
Pos logit contributions | |
goodbye | +0.064 |
away | +0.059 |
des | +0.059 |
ogging | +0.058 |
attention | +0.057 |
Neg logit contributions | |
Ant | -0.055 |
Fire | -0.055 |
Salt | -0.054 |
Next | -0.053 |
hes | -0.053 |
Token | crying |
Token ID | 13774 |
Feature activation | -1.15e-02 |
Pos logit contributions | |
goodbye | +0.083 |
away | +0.074 |
attention | +0.069 |
des | +0.069 |
farewell | +0.069 |
Neg logit contributions | |
Ant | -0.093 |
Fire | -0.091 |
hes | -0.091 |
Next | -0.090 |
Salt | -0.090 |
Token | and |
Token ID | 290 |
Feature activation | -2.35e-03 |
Pos logit contributions | |
goodbye | +0.018 |
des | +0.015 |
ogging | +0.015 |
farewell | +0.015 |
attention | +0.015 |
Neg logit contributions | |
Ant | -0.021 |
Salt | -0.020 |
Next | -0.020 |
Fire | -0.020 |
hes | -0.020 |
Token | comes |
Token ID | 2058 |
Feature activation | -5.27e-02 |
Pos logit contributions | |
goodbye | +0.004 |
away | +0.003 |
attention | +0.003 |
ogging | +0.003 |
farewell | +0.003 |
Neg logit contributions | |
Ant | -0.004 |
Next | -0.004 |
Salt | -0.004 |
� | -0.004 |
Fire | -0.004 |
Token | to |
Token ID | 284 |
Feature activation | -2.06e-02 |
Pos logit contributions | |
goodbye | +0.074 |
des | +0.068 |
fy | +0.064 |
ogging | +0.062 |
contain | +0.062 |
Neg logit contributions | |
Ant | -0.102 |
twins | -0.101 |
Fire | -0.101 |
brothers | -0.100 |
Will | -0.100 |
Token | to |
Token ID | 284 |
Feature activation | -1.25e-02 |
Pos logit contributions | |
goodbye | +0.007 |
away | +0.007 |
ogging | +0.006 |
des | +0.006 |
attention | +0.006 |
Neg logit contributions | |
Fire | -0.013 |
Ant | -0.013 |
Next | -0.013 |
I | -0.013 |
Will | -0.013 |
Token | play |
Token ID | 711 |
Feature activation | -1.05e-02 |
Pos logit contributions | |
goodbye | +0.023 |
ly | +0.021 |
ogging | +0.021 |
away | +0.021 |
attention | +0.020 |
Neg logit contributions | |
Ant | -0.019 |
Brush | -0.018 |
I | -0.018 |
Fire | -0.017 |
Will | -0.017 |
Token | with |
Token ID | 351 |
Feature activation | +3.02e-02 |
Pos logit contributions | |
goodbye | +0.022 |
attention | +0.020 |
ogging | +0.020 |
away | +0.020 |
farewell | +0.019 |
Neg logit contributions | |
Ant | -0.013 |
Will | -0.013 |
Next | -0.012 |
Salt | -0.012 |
Fire | -0.012 |
Token | me |
Token ID | 502 |
Feature activation | -1.05e-02 |
Pos logit contributions | |
Ant | +0.050 |
Fire | +0.047 |
Next | +0.047 |
Salt | +0.046 |
Will | +0.046 |
Neg logit contributions | |
goodbye | -0.054 |
attention | -0.048 |
ogging | -0.047 |
away | -0.046 |
des | -0.046 |
Token | ?" |
Token ID | 1701 |
Feature activation | +3.00e-03 |
Pos logit contributions | |
goodbye | +0.021 |
attention | +0.018 |
away | +0.018 |
des | +0.018 |
farewell | +0.018 |
Neg logit contributions | |
Ant | -0.015 |
Will | -0.014 |
Salt | -0.014 |
Next | -0.014 |
hes | -0.014 |
Token | Leo |
Token ID | 19632 |
Feature activation | -3.95e-02 |
Pos logit contributions | |
kins | +0.005 |
� | +0.005 |
Ant | +0.005 |
poker | +0.005 |
hes | +0.005 |
Neg logit contributions | |
goodbye | -0.005 |
ogging | -0.004 |
farewell | -0.004 |
back | -0.004 |
attention | -0.004 |
Token | said |
Token ID | 531 |
Feature activation | -1.99e-02 |
Pos logit contributions | |
goodbye | +0.064 |
ly | +0.054 |
back | +0.052 |
farewell | +0.051 |
away | +0.051 |
Neg logit contributions | |
hes | -0.070 |
poker | -0.070 |
Fire | -0.069 |
Salt | -0.068 |
Blocks | -0.068 |
Token | , |
Token ID | 11 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
away | +0.028 |
des | +0.028 |
ogging | +0.027 |
attention | +0.026 |
land | +0.026 |
Neg logit contributions | |
Ant | -0.035 |
Will | -0.035 |
Salt | -0.034 |
kins | -0.033 |
Fire | -0.033 |
Token | " |
Token ID | 366 |
Feature activation | +4.97e-02 |
Pos logit contributions | |
goodbye | +0.034 |
back | +0.028 |
farewell | +0.027 |
ogging | +0.027 |
ly | +0.026 |
Neg logit contributions | |
Ant | -0.019 |
kins | -0.018 |
Next | -0.018 |
� | -0.018 |
Mouse | -0.018 |
Token | I |
Token ID | 40 |
Feature activation | -4.85e-03 |
Pos logit contributions | |
kins | +0.070 |
hes | +0.068 |
Fire | +0.067 |
Salt | +0.067 |
poker | +0.066 |
Neg logit contributions | |
goodbye | -0.100 |
away | -0.091 |
ogging | -0.087 |
des | -0.087 |
contain | -0.085 |
Token | am |
Token ID | 716 |
Feature activation | +2.34e-02 |
Pos logit contributions | |
goodbye | +0.009 |
ogging | +0.009 |
attention | +0.008 |
away | +0.008 |
ly | +0.008 |
Neg logit contributions | |
Ant | -0.007 |
I | -0.007 |
� | -0.006 |
Fire | -0.006 |
Will | -0.006 |
Token | catch |
Token ID | 4929 |
Feature activation | -3.07e-02 |
Pos logit contributions | |
goodbye | +0.055 |
away | +0.047 |
attention | +0.047 |
ogging | +0.046 |
farewell | +0.044 |
Neg logit contributions | |
Ant | -0.062 |
Next | -0.059 |
Salt | -0.058 |
Fire | -0.058 |
� | -0.058 |
Token | up |
Token ID | 510 |
Feature activation | -8.71e-03 |
Pos logit contributions | |
goodbye | +0.050 |
away | +0.043 |
attention | +0.043 |
ogging | +0.042 |
farewell | +0.041 |
Neg logit contributions | |
Ant | -0.055 |
Next | -0.052 |
Salt | -0.051 |
Fire | -0.051 |
� | -0.051 |
Token | and |
Token ID | 290 |
Feature activation | -1.41e-02 |
Pos logit contributions | |
goodbye | +0.013 |
attention | +0.011 |
away | +0.011 |
ogging | +0.010 |
des | +0.010 |
Neg logit contributions | |
Ant | -0.017 |
Next | -0.016 |
Salt | -0.016 |
� | -0.016 |
Fire | -0.016 |
Token | just |
Token ID | 655 |
Feature activation | -1.47e-02 |
Pos logit contributions | |
goodbye | +0.023 |
attention | +0.019 |
away | +0.019 |
ogging | +0.019 |
farewell | +0.018 |
Neg logit contributions | |
Ant | -0.026 |
Next | -0.024 |
� | -0.024 |
Salt | -0.024 |
Fire | -0.024 |
Token | watched |
Token ID | 7342 |
Feature activation | -2.96e-02 |
Pos logit contributions | |
goodbye | +0.028 |
away | +0.025 |
attention | +0.025 |
ogging | +0.024 |
farewell | +0.023 |
Neg logit contributions | |
Ant | -0.022 |
Next | -0.021 |
Salt | -0.021 |
� | -0.021 |
Fire | -0.021 |
Token | Jack |
Token ID | 3619 |
Feature activation | -6.88e-02 |
Pos logit contributions | |
goodbye | +0.047 |
away | +0.041 |
attention | +0.040 |
ogging | +0.040 |
des | +0.038 |
Neg logit contributions | |
Ant | -0.054 |
Next | -0.051 |
Salt | -0.051 |
Fire | -0.051 |
� | -0.051 |
Token | reach |
Token ID | 3151 |
Feature activation | -4.26e-02 |
Pos logit contributions | |
goodbye | +0.102 |
away | +0.089 |
attention | +0.087 |
ogging | +0.085 |
des | +0.084 |
Neg logit contributions | |
Ant | -0.128 |
Salt | -0.121 |
Next | -0.120 |
� | -0.120 |
hes | -0.119 |
Token | the |
Token ID | 262 |
Feature activation | -4.82e-02 |
Pos logit contributions | |
goodbye | +0.067 |
ogging | +0.060 |
des | +0.060 |
away | +0.060 |
farewell | +0.056 |
Neg logit contributions | |
Ant | -0.072 |
Salt | -0.071 |
Fire | -0.070 |
Next | -0.069 |
hes | -0.068 |
Token | end |
Token ID | 886 |
Feature activation | -2.12e-02 |
Pos logit contributions | |
goodbye | +0.089 |
away | +0.079 |
attention | +0.078 |
ogging | +0.078 |
des | +0.075 |
Neg logit contributions | |
Ant | -0.075 |
Fire | -0.070 |
Salt | -0.070 |
Next | -0.069 |
� | -0.069 |
Token | of |
Token ID | 286 |
Feature activation | +2.30e-02 |
Pos logit contributions | |
goodbye | +0.030 |
away | +0.025 |
attention | +0.025 |
ogging | +0.024 |
des | +0.024 |
Neg logit contributions | |
Ant | -0.042 |
Salt | -0.040 |
Fire | -0.040 |
Next | -0.040 |
hes | -0.039 |
Token | the |
Token ID | 262 |
Feature activation | -2.35e-02 |
Pos logit contributions | |
Ant | +0.044 |
� | +0.044 |
hes | +0.042 |
Bees | +0.042 |
Next | +0.042 |
Neg logit contributions | |
goodbye | -0.035 |
attention | -0.030 |
away | -0.028 |
ogging | -0.027 |
farewell | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
Ant | +0.115 |
Next | +0.110 |
� | +0.109 |
Salt | +0.109 |
Fire | +0.109 |
Neg logit contributions | |
goodbye | -0.077 |
away | -0.065 |
attention | -0.064 |
ogging | -0.063 |
farewell | -0.059 |
Token |
|
Token ID | 198 |
Feature activation | +9.84e-03 |
Pos logit contributions | |
kins | +0.037 |
Ant | +0.036 |
poker | +0.034 |
sibling | +0.034 |
� | +0.034 |
Neg logit contributions | |
goodbye | -0.040 |
ogging | -0.032 |
farewell | -0.031 |
away | -0.029 |
back | -0.029 |
Token |
|
Token ID | 198 |
Feature activation | +8.13e-03 |
Pos logit contributions | |
Ant | +0.011 |
Next | +0.010 |
Fire | +0.010 |
Salt | +0.010 |
hes | +0.010 |
Neg logit contributions | |
goodbye | -0.023 |
away | -0.021 |
attention | -0.021 |
ogging | -0.020 |
des | -0.020 |
Token | Anna |
Token ID | 31160 |
Feature activation | -1.05e-02 |
Pos logit contributions | |
Fire | +0.011 |
poker | +0.011 |
Ant | +0.011 |
Salt | +0.011 |
sibling | +0.011 |
Neg logit contributions | |
goodbye | -0.015 |
away | -0.014 |
ogging | -0.014 |
of | -0.013 |
farewell | -0.013 |
Token | takes |
Token ID | 2753 |
Feature activation | -1.46e-02 |
Pos logit contributions | |
goodbye | +0.017 |
attention | +0.014 |
ogging | +0.014 |
away | +0.014 |
farewell | +0.014 |
Neg logit contributions | |
Ant | -0.019 |
Salt | -0.018 |
� | -0.018 |
Fire | -0.018 |
Next | -0.018 |
Token | a |
Token ID | 257 |
Feature activation | -6.21e-02 |
Pos logit contributions | |
goodbye | +0.027 |
ogging | +0.025 |
away | +0.025 |
des | +0.024 |
attention | +0.023 |
Neg logit contributions | |
Ant | -0.022 |
Salt | -0.021 |
hes | -0.021 |
Next | -0.021 |
Fire | -0.021 |
Token | deep |
Token ID | 2769 |
Feature activation | -7.37e-02 |
Pos logit contributions | |
goodbye | +0.103 |
des | +0.094 |
ogging | +0.094 |
attention | +0.090 |
away | +0.087 |
Neg logit contributions | |
Ant | -0.105 |
Fire | -0.097 |
poker | -0.096 |
Mouse | -0.096 |
Will | -0.096 |
Token | breath |
Token ID | 8033 |
Feature activation | -7.46e-02 |
Pos logit contributions | |
goodbye | +0.097 |
ogging | +0.085 |
land | +0.084 |
attention | +0.083 |
des | +0.082 |
Neg logit contributions | |
Ant | -0.149 |
poker | -0.139 |
Salt | -0.138 |
Mouse | -0.137 |
kins | -0.136 |
Token | . |
Token ID | 13 |
Feature activation | +8.78e-03 |
Pos logit contributions | |
goodbye | +0.147 |
des | +0.142 |
icol | +0.141 |
tour | +0.140 |
attention | +0.139 |
Neg logit contributions | |
hes | -0.085 |
You | -0.081 |
I | -0.080 |
Salt | -0.078 |
Will | -0.075 |
Token | She |
Token ID | 1375 |
Feature activation | +3.08e-02 |
Pos logit contributions | |
kins | +0.014 |
Ant | +0.013 |
Mouse | +0.012 |
brothers | +0.012 |
sibling | +0.012 |
Neg logit contributions | |
goodbye | -0.018 |
ogging | -0.014 |
farewell | -0.014 |
attention | -0.014 |
away | -0.014 |
Token | reaches |
Token ID | 12229 |
Feature activation | -5.65e-02 |
Pos logit contributions | |
Ant | +0.054 |
Next | +0.051 |
Salt | +0.050 |
� | +0.050 |
Fire | +0.050 |
Neg logit contributions | |
goodbye | -0.051 |
away | -0.044 |
attention | -0.044 |
ogging | -0.043 |
farewell | -0.042 |
Token | much |
Token ID | 881 |
Feature activation | +2.21e-02 |
Pos logit contributions | |
goodbye | +0.004 |
attention | +0.003 |
away | +0.003 |
contain | +0.003 |
ly | +0.003 |
Neg logit contributions | |
Ant | -0.004 |
Next | -0.003 |
hes | -0.003 |
Salt | -0.003 |
I | -0.003 |
Token | fun |
Token ID | 1257 |
Feature activation | -1.19e-02 |
Pos logit contributions | |
Next | +0.033 |
hes | +0.031 |
Salt | +0.031 |
You | +0.031 |
I | +0.030 |
Neg logit contributions | |
goodbye | -0.042 |
away | -0.039 |
contain | -0.035 |
farewell | -0.035 |
ly | -0.034 |
Token | that |
Token ID | 326 |
Feature activation | +2.19e-02 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.016 |
attention | +0.015 |
farewell | +0.015 |
des | +0.015 |
Neg logit contributions | |
Ant | -0.022 |
Salt | -0.021 |
Will | -0.021 |
Next | -0.021 |
� | -0.021 |
Token | she |
Token ID | 673 |
Feature activation | -1.33e-02 |
Pos logit contributions | |
Ant | +0.032 |
Next | +0.029 |
Fire | +0.029 |
� | +0.029 |
Salt | +0.029 |
Neg logit contributions | |
goodbye | -0.044 |
away | -0.039 |
attention | -0.039 |
ogging | -0.038 |
farewell | -0.037 |
Token | didn |
Token ID | 1422 |
Feature activation | -5.45e-02 |
Pos logit contributions | |
goodbye | +0.026 |
attention | +0.024 |
away | +0.024 |
ogging | +0.023 |
des | +0.022 |
Neg logit contributions | |
Ant | -0.019 |
Next | -0.017 |
� | -0.017 |
Salt | -0.017 |
hes | -0.017 |
Token | 't |
Token ID | 470 |
Feature activation | -6.24e-02 |
Pos logit contributions | |
goodbye | +0.126 |
away | +0.115 |
attention | +0.114 |
des | +0.109 |
ly | +0.108 |
Neg logit contributions | |
Ant | -0.061 |
Salt | -0.057 |
hes | -0.057 |
Next | -0.056 |
I | -0.056 |
Token | notice |
Token ID | 4003 |
Feature activation | -6.67e-03 |
Pos logit contributions | |
goodbye | +0.092 |
attention | +0.074 |
ogging | +0.074 |
away | +0.073 |
des | +0.073 |
Neg logit contributions | |
Ant | -0.121 |
I | -0.114 |
Salt | -0.112 |
You | -0.112 |
Will | -0.110 |
Token | when |
Token ID | 618 |
Feature activation | +1.34e-02 |
Pos logit contributions | |
goodbye | +0.012 |
away | +0.011 |
attention | +0.010 |
des | +0.010 |
ogging | +0.010 |
Neg logit contributions | |
Ant | -0.010 |
Salt | -0.010 |
Next | -0.010 |
Fire | -0.010 |
Will | -0.010 |
Token | her |
Token ID | 607 |
Feature activation | +1.98e-02 |
Pos logit contributions | |
Ant | +0.021 |
Salt | +0.020 |
Next | +0.020 |
Fire | +0.020 |
� | +0.020 |
Neg logit contributions | |
goodbye | -0.025 |
attention | -0.022 |
away | -0.022 |
ogging | -0.022 |
farewell | -0.021 |
Token | mom |
Token ID | 1995 |
Feature activation | -2.42e-02 |
Pos logit contributions | |
Ant | +0.026 |
Next | +0.024 |
Salt | +0.024 |
� | +0.024 |
Fire | +0.024 |
Neg logit contributions | |
goodbye | -0.041 |
away | -0.037 |
attention | -0.036 |
ogging | -0.036 |
des | -0.035 |
Token | called |
Token ID | 1444 |
Feature activation | -4.29e-02 |
Pos logit contributions | |
goodbye | +0.046 |
away | +0.041 |
attention | +0.041 |
ogging | +0.040 |
farewell | +0.039 |
Neg logit contributions | |
Ant | -0.036 |
Next | -0.034 |
� | -0.033 |
Salt | -0.033 |
Fire | -0.033 |
Token | . |
Token ID | 13 |
Feature activation | -8.05e-03 |
Pos logit contributions | |
Ant | +0.016 |
Next | +0.015 |
Salt | +0.015 |
� | +0.015 |
Fire | +0.015 |
Neg logit contributions | |
goodbye | -0.012 |
away | -0.011 |
attention | -0.010 |
ogging | -0.010 |
farewell | -0.010 |
Token | Suddenly |
Token ID | 24975 |
Feature activation | +2.82e-02 |
Pos logit contributions | |
goodbye | +0.015 |
ogging | +0.012 |
farewell | +0.012 |
back | +0.010 |
des | +0.010 |
Neg logit contributions | |
kins | -0.014 |
Ant | -0.013 |
hes | -0.013 |
poker | -0.013 |
chefs | -0.013 |
Token | , |
Token ID | 11 |
Feature activation | +2.37e-03 |
Pos logit contributions | |
Ant | +0.065 |
Next | +0.062 |
� | +0.062 |
Salt | +0.062 |
Fire | +0.062 |
Neg logit contributions | |
goodbye | -0.031 |
away | -0.025 |
attention | -0.025 |
ogging | -0.024 |
farewell | -0.022 |
Token | they |
Token ID | 484 |
Feature activation | -3.47e-02 |
Pos logit contributions | |
Ant | +0.004 |
� | +0.004 |
Next | +0.004 |
hes | +0.004 |
Mouse | +0.004 |
Neg logit contributions | |
goodbye | -0.004 |
attention | -0.003 |
away | -0.003 |
farewell | -0.003 |
back | -0.003 |
Token | heard |
Token ID | 2982 |
Feature activation | -3.83e-02 |
Pos logit contributions | |
goodbye | +0.064 |
des | +0.060 |
ogging | +0.060 |
attention | +0.060 |
away | +0.059 |
Neg logit contributions | |
Ant | -0.052 |
Fire | -0.049 |
I | -0.048 |
hes | -0.047 |
brothers | -0.047 |
Token | a |
Token ID | 257 |
Feature activation | -5.77e-02 |
Pos logit contributions | |
goodbye | +0.076 |
ogging | +0.071 |
away | +0.071 |
attention | +0.069 |
des | +0.069 |
Neg logit contributions | |
Fire | -0.051 |
Ant | -0.049 |
Salt | -0.048 |
Next | -0.047 |
hes | -0.047 |
Token | loud |
Token ID | 7812 |
Feature activation | -6.69e-02 |
Pos logit contributions | |
goodbye | +0.096 |
attention | +0.084 |
ogging | +0.083 |
away | +0.082 |
des | +0.080 |
Neg logit contributions | |
Ant | -0.102 |
Fire | -0.096 |
Will | -0.094 |
Salt | -0.094 |
Mouse | -0.093 |
Token | noise |
Token ID | 7838 |
Feature activation | -9.44e-03 |
Pos logit contributions | |
goodbye | +0.125 |
attention | +0.113 |
des | +0.110 |
ogging | +0.110 |
away | +0.109 |
Neg logit contributions | |
Ant | -0.100 |
Fire | -0.097 |
Salt | -0.094 |
Next | -0.092 |
Mouse | -0.091 |
Token | . |
Token ID | 13 |
Feature activation | -6.88e-03 |
Pos logit contributions | |
goodbye | +0.017 |
des | +0.015 |
farewell | +0.015 |
attention | +0.015 |
ogging | +0.014 |
Neg logit contributions | |
Fire | -0.014 |
Salt | -0.014 |
Will | -0.014 |
Ant | -0.014 |
hes | -0.014 |
Token | Jack |
Token ID | 3619 |
Feature activation | -3.58e-02 |
Pos logit contributions | |
goodbye | +0.013 |
farewell | +0.010 |
ogging | +0.010 |
back | +0.009 |
owl | +0.009 |
Neg logit contributions | |
kins | -0.011 |
Ant | -0.011 |
poker | -0.011 |
� | -0.011 |
hes | -0.010 |
Token | said |
Token ID | 531 |
Feature activation | -1.97e-02 |
Pos logit contributions | |
goodbye | +0.066 |
away | +0.058 |
ogging | +0.057 |
attention | +0.057 |
ly | +0.056 |
Neg logit contributions | |
Ant | -0.055 |
Next | -0.053 |
� | -0.052 |
Salt | -0.052 |
hes | -0.052 |
Token | he |
Token ID | 339 |
Feature activation | +3.00e-04 |
Pos logit contributions | |
Ant | +0.036 |
� | +0.034 |
Next | +0.033 |
Salt | +0.033 |
Fire | +0.033 |
Neg logit contributions | |
goodbye | -0.041 |
attention | -0.036 |
away | -0.036 |
ogging | -0.035 |
farewell | -0.034 |
Token | stayed |
Token ID | 9658 |
Feature activation | -9.41e-03 |
Pos logit contributions | |
� | +0.000 |
Ant | +0.000 |
Next | +0.000 |
Blocks | +0.000 |
Salt | +0.000 |
Neg logit contributions | |
goodbye | -0.001 |
back | -0.000 |
farewell | -0.000 |
away | -0.000 |
ly | -0.000 |
Token | there |
Token ID | 612 |
Feature activation | -1.63e-02 |
Pos logit contributions | |
goodbye | +0.018 |
ogging | +0.017 |
attention | +0.016 |
ly | +0.016 |
away | +0.016 |
Neg logit contributions | |
Ant | -0.012 |
kins | -0.012 |
hes | -0.012 |
Will | -0.012 |
Next | -0.012 |
Token | for |
Token ID | 329 |
Feature activation | +2.62e-02 |
Pos logit contributions | |
goodbye | +0.019 |
ogging | +0.017 |
away | +0.017 |
des | +0.016 |
attention | +0.016 |
Neg logit contributions | |
Ant | -0.034 |
Salt | -0.033 |
Next | -0.033 |
Fire | -0.033 |
� | -0.032 |
Token | a |
Token ID | 257 |
Feature activation | -4.88e-02 |
Pos logit contributions | |
Ant | +0.039 |
Fire | +0.037 |
Salt | +0.037 |
Next | +0.036 |
Will | +0.036 |
Neg logit contributions | |
goodbye | -0.049 |
ogging | -0.045 |
away | -0.044 |
attention | -0.044 |
des | -0.042 |
Token | long |
Token ID | 890 |
Feature activation | -6.15e-02 |
Pos logit contributions | |
goodbye | +0.078 |
ogging | +0.073 |
attention | +0.071 |
away | +0.069 |
ly | +0.066 |
Neg logit contributions | |
Ant | -0.086 |
Will | -0.080 |
� | -0.078 |
Fire | -0.077 |
Salt | -0.076 |
Token | time |
Token ID | 640 |
Feature activation | -1.24e-02 |
Pos logit contributions | |
goodbye | +0.103 |
attention | +0.092 |
ogging | +0.089 |
away | +0.089 |
des | +0.086 |
Neg logit contributions | |
Ant | -0.104 |
� | -0.098 |
Fire | -0.096 |
Next | -0.096 |
Salt | -0.096 |
Token | . |
Token ID | 13 |
Feature activation | -7.16e-03 |
Pos logit contributions | |
goodbye | +0.018 |
away | +0.015 |
attention | +0.015 |
ogging | +0.015 |
farewell | +0.014 |
Neg logit contributions | |
Ant | -0.024 |
Salt | -0.023 |
Fire | -0.023 |
Next | -0.023 |
� | -0.023 |
Token | He |
Token ID | 679 |
Feature activation | +2.70e-02 |
Pos logit contributions | |
goodbye | +0.011 |
farewell | +0.009 |
ogging | +0.008 |
back | +0.008 |
des | +0.008 |
Neg logit contributions | |
kins | -0.015 |
chefs | -0.013 |
Ant | -0.013 |
hes | -0.013 |
ters | -0.013 |
Token | was |
Token ID | 373 |
Feature activation | +1.75e-02 |
Pos logit contributions | |
Ant | +0.041 |
� | +0.041 |
Next | +0.041 |
Salt | +0.040 |
Blocks | +0.040 |
Neg logit contributions | |
goodbye | -0.052 |
farewell | -0.043 |
away | -0.043 |
back | -0.042 |
des | -0.041 |
Token | very |
Token ID | 845 |
Feature activation | +3.78e-03 |
Pos logit contributions | |
Ant | +0.038 |
� | +0.037 |
Next | +0.035 |
Mouse | +0.035 |
Salt | +0.035 |
Neg logit contributions | |
goodbye | -0.025 |
attention | -0.019 |
back | -0.018 |
farewell | -0.018 |
away | -0.018 |