Token | bed |
Token ID | 3996 |
Feature activation | +1.20e+00 |
Pos logit contributions | |
idents | +0.026 |
batch | +0.022 |
package | +0.022 |
meal | +0.022 |
ksh | +0.021 |
Neg logit contributions | |
eight | -0.032 |
stretched | -0.031 |
ila | -0.031 |
Tw | -0.030 |
La | -0.029 |
Token | and |
Token ID | 290 |
Feature activation | +1.39e-01 |
Pos logit contributions | |
idents | +0.043 |
package | +0.043 |
meal | +0.041 |
tray | +0.041 |
batch | +0.041 |
Neg logit contributions | |
stretched | -0.026 |
ila | -0.026 |
Tw | -0.025 |
eight | -0.024 |
La | -0.023 |
Token | found |
Token ID | 1043 |
Feature activation | -3.85e-01 |
Pos logit contributions | |
idents | +0.005 |
package | +0.005 |
tray | +0.005 |
meal | +0.005 |
batch | +0.005 |
Neg logit contributions | |
ila | -0.003 |
stretched | -0.003 |
Tw | -0.003 |
eight | -0.003 |
Tweet | -0.003 |
Token | the |
Token ID | 262 |
Feature activation | -3.49e-01 |
Pos logit contributions | |
ila | +0.009 |
stretched | +0.009 |
Tw | +0.008 |
eight | +0.008 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.014 |
batch | -0.013 |
meal | -0.013 |
package | -0.013 |
tray | -0.013 |
Token | candy |
Token ID | 18550 |
Feature activation | +7.34e-01 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.010 |
Tw | +0.010 |
eight | +0.010 |
Tweet | +0.010 |
Neg logit contributions | |
idents | -0.010 |
package | -0.009 |
batch | -0.009 |
meal | -0.009 |
tray | -0.009 |
Token | wra |
Token ID | 7917 |
Feature activation | +6.01e+00 |
Pos logit contributions | |
idents | +0.031 |
batch | +0.028 |
stret | +0.028 |
aman | +0.028 |
isions | +0.027 |
Neg logit contributions | |
ila | -0.011 |
he | -0.011 |
stretched | -0.010 |
Tw | -0.010 |
fast | -0.010 |
Token | ppers |
Token ID | 11799 |
Feature activation | +1.30e+00 |
Pos logit contributions | |
batch | +0.171 |
idents | +0.169 |
meal | +0.168 |
dinners | +0.157 |
package | +0.154 |
Neg logit contributions | |
stretched | -0.178 |
eight | -0.172 |
ila | -0.164 |
Tweet | -0.159 |
A | -0.150 |
Token | . |
Token ID | 13 |
Feature activation | +5.15e-01 |
Pos logit contributions | |
idents | +0.050 |
package | +0.049 |
batch | +0.048 |
meal | +0.047 |
stret | +0.047 |
Neg logit contributions | |
stretched | -0.026 |
ila | -0.025 |
eight | -0.024 |
Tw | -0.023 |
Tweet | -0.020 |
Token |
|
Token ID | 198 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
idents | +0.018 |
package | +0.017 |
batch | +0.017 |
tray | +0.017 |
meal | +0.016 |
Neg logit contributions | |
stretched | -0.013 |
ila | -0.013 |
eight | -0.012 |
La | -0.012 |
Tw | -0.011 |
Token |
|
Token ID | 198 |
Feature activation | -2.15e-01 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.002 |
eight | +0.002 |
brav | +0.002 |
Tw | +0.002 |
Neg logit contributions | |
package | -0.004 |
batch | -0.004 |
tray | -0.004 |
stret | -0.004 |
meal | -0.004 |
Token | L |
Token ID | 43 |
Feature activation | -8.64e-01 |
Pos logit contributions | |
stretched | +0.005 |
ila | +0.005 |
Tw | +0.004 |
eight | +0.004 |
La | +0.004 |
Neg logit contributions | |
package | -0.008 |
tray | -0.008 |
batch | -0.008 |
idents | -0.008 |
stret | -0.007 |
Token | . |
Token ID | 13 |
Feature activation | +1.99e-01 |
Pos logit contributions | |
idents | +0.051 |
batch | +0.051 |
package | +0.050 |
stret | +0.049 |
tray | +0.049 |
Neg logit contributions | |
Tw | -0.013 |
stretched | -0.013 |
ila | -0.011 |
eight | -0.010 |
Ch | -0.009 |
Token | She |
Token ID | 1375 |
Feature activation | -1.61e+00 |
Pos logit contributions | |
package | +0.007 |
refrigerator | +0.007 |
idents | +0.007 |
tray | +0.007 |
appliance | +0.006 |
Neg logit contributions | |
stretched | -0.005 |
Tweet | -0.004 |
La | -0.004 |
eight | -0.004 |
ila | -0.004 |
Token | saw |
Token ID | 2497 |
Feature activation | -5.42e-01 |
Pos logit contributions | |
stretched | +0.051 |
Tweet | +0.047 |
ila | +0.047 |
La | +0.045 |
Tw | +0.045 |
Neg logit contributions | |
package | -0.044 |
tray | -0.040 |
batch | -0.040 |
mess | -0.039 |
meal | -0.038 |
Token | the |
Token ID | 262 |
Feature activation | -2.77e-01 |
Pos logit contributions | |
stretched | +0.014 |
ila | +0.013 |
Tweet | +0.013 |
La | +0.013 |
Tw | +0.013 |
Neg logit contributions | |
package | -0.018 |
idents | -0.017 |
tray | -0.017 |
meal | -0.017 |
batch | -0.017 |
Token | chocolate |
Token ID | 11311 |
Feature activation | +2.24e+00 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.009 |
eight | +0.009 |
Tw | +0.009 |
La | +0.009 |
Neg logit contributions | |
idents | -0.006 |
batch | -0.006 |
package | -0.006 |
meal | -0.006 |
dinners | -0.005 |
Token | wra |
Token ID | 7917 |
Feature activation | +5.90e+00 |
Pos logit contributions | |
idents | +0.086 |
isions | +0.084 |
rd | +0.081 |
dinners | +0.080 |
keeping | +0.079 |
Neg logit contributions | |
ila | -0.037 |
he | -0.035 |
stretched | -0.035 |
Tw | -0.033 |
eight | -0.033 |
Token | ppers |
Token ID | 11799 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
idents | +0.189 |
batch | +0.188 |
meal | +0.187 |
meals | +0.175 |
dinners | +0.175 |
Neg logit contributions | |
stretched | -0.167 |
eight | -0.163 |
ila | -0.156 |
Tweet | -0.145 |
A | -0.141 |
Token | on |
Token ID | 319 |
Feature activation | +3.37e-01 |
Pos logit contributions | |
idents | +0.039 |
package | +0.038 |
batch | +0.037 |
meal | +0.036 |
tray | +0.036 |
Neg logit contributions | |
stretched | -0.021 |
ila | -0.020 |
eight | -0.019 |
Tw | -0.018 |
Tweet | -0.017 |
Token | the |
Token ID | 262 |
Feature activation | +5.86e-01 |
Pos logit contributions | |
idents | +0.014 |
package | +0.014 |
meal | +0.013 |
batch | +0.013 |
tray | +0.013 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
Tw | -0.005 |
eight | -0.005 |
Tweet | -0.005 |
Token | floor |
Token ID | 4314 |
Feature activation | +1.85e+00 |
Pos logit contributions | |
idents | +0.011 |
batch | +0.010 |
package | +0.009 |
meal | +0.009 |
dinners | +0.009 |
Neg logit contributions | |
ila | -0.023 |
eight | -0.023 |
stretched | -0.023 |
Tw | -0.023 |
La | -0.022 |
Token | and |
Token ID | 290 |
Feature activation | -5.49e-01 |
Pos logit contributions | |
idents | +0.064 |
package | +0.063 |
meal | +0.062 |
tray | +0.061 |
batch | +0.061 |
Neg logit contributions | |
ila | -0.044 |
stretched | -0.044 |
Tw | -0.040 |
eight | -0.040 |
Tweet | -0.039 |
Token | box |
Token ID | 3091 |
Feature activation | +2.27e+00 |
Pos logit contributions | |
idents | +0.007 |
ics | +0.007 |
batch | +0.006 |
dinners | +0.006 |
stret | +0.006 |
Neg logit contributions | |
eight | -0.010 |
stretched | -0.010 |
Tw | -0.009 |
ila | -0.009 |
Fay | -0.009 |
Token | and |
Token ID | 290 |
Feature activation | +1.24e-01 |
Pos logit contributions | |
idents | +0.096 |
stret | +0.094 |
meal | +0.090 |
package | +0.088 |
batch | +0.088 |
Neg logit contributions | |
stretched | -0.043 |
ila | -0.037 |
eight | -0.037 |
Tw | -0.036 |
he | -0.028 |
Token | saw |
Token ID | 2497 |
Feature activation | -3.69e-01 |
Pos logit contributions | |
idents | +0.004 |
package | +0.004 |
meal | +0.004 |
tray | +0.004 |
batch | +0.004 |
Neg logit contributions | |
ila | -0.003 |
stretched | -0.003 |
Tw | -0.003 |
eight | -0.003 |
Tweet | -0.003 |
Token | a |
Token ID | 257 |
Feature activation | -6.30e-02 |
Pos logit contributions | |
stretched | +0.008 |
ila | +0.008 |
Tweet | +0.008 |
La | +0.008 |
Tw | +0.008 |
Neg logit contributions | |
package | -0.013 |
idents | -0.013 |
meal | -0.013 |
tray | -0.013 |
batch | -0.013 |
Token | j |
Token ID | 474 |
Feature activation | +8.21e-01 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.001 |
eight | +0.001 |
Tw | +0.001 |
Tweet | +0.001 |
Neg logit contributions | |
idents | -0.002 |
meal | -0.002 |
package | -0.002 |
batch | -0.002 |
tray | -0.002 |
Token | igsaw |
Token ID | 45636 |
Feature activation | +5.81e+00 |
Pos logit contributions | |
idents | +0.030 |
package | +0.030 |
meal | +0.029 |
tray | +0.029 |
batch | +0.029 |
Neg logit contributions | |
ila | -0.018 |
stretched | -0.018 |
Tw | -0.017 |
eight | -0.016 |
Tweet | -0.016 |
Token | puzzle |
Token ID | 15027 |
Feature activation | +1.49e+00 |
Pos logit contributions | |
idents | +0.194 |
stret | +0.183 |
lette | +0.182 |
aman | +0.182 |
urger | +0.179 |
Neg logit contributions | |
he | -0.134 |
Tw | -0.134 |
stretched | -0.133 |
fast | -0.130 |
blue | -0.127 |
Token | inside |
Token ID | 2641 |
Feature activation | +8.62e-01 |
Pos logit contributions | |
idents | +0.057 |
stret | +0.052 |
meal | +0.051 |
package | +0.051 |
aman | +0.051 |
Neg logit contributions | |
stretched | -0.032 |
ila | -0.031 |
eight | -0.029 |
Tw | -0.029 |
Tweet | -0.025 |
Token | . |
Token ID | 13 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
idents | +0.031 |
package | +0.030 |
tray | +0.030 |
meal | +0.030 |
batch | +0.029 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.019 |
Tw | -0.018 |
eight | -0.018 |
La | -0.017 |
Token | He |
Token ID | 679 |
Feature activation | -4.05e-01 |
Pos logit contributions | |
package | +0.005 |
refrigerator | +0.005 |
idents | +0.004 |
tray | +0.004 |
batch | +0.004 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
Tweet | -0.003 |
La | -0.003 |
Leave | -0.003 |
Token | paused |
Token ID | 24487 |
Feature activation | -1.60e+00 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.011 |
La | +0.010 |
Tweet | +0.010 |
Tw | +0.010 |
Neg logit contributions | |
package | -0.013 |
idents | -0.013 |
batch | -0.012 |
tray | -0.012 |
meal | -0.012 |
Token | wallet |
Token ID | 13008 |
Feature activation | +1.07e+00 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.010 |
Tw | +0.009 |
eight | +0.009 |
Tweet | +0.009 |
Neg logit contributions | |
idents | -0.015 |
package | -0.014 |
meal | -0.014 |
tray | -0.014 |
batch | -0.014 |
Token | had |
Token ID | 550 |
Feature activation | +2.48e-01 |
Pos logit contributions | |
package | +0.033 |
idents | +0.033 |
tray | +0.032 |
meal | +0.032 |
batch | +0.031 |
Neg logit contributions | |
ila | -0.029 |
stretched | -0.029 |
Tw | -0.027 |
La | -0.027 |
Tweet | -0.027 |
Token | coins |
Token ID | 10796 |
Feature activation | -3.62e-01 |
Pos logit contributions | |
package | +0.007 |
meal | +0.007 |
idents | +0.006 |
mess | +0.006 |
tray | +0.006 |
Neg logit contributions | |
ila | -0.008 |
Tweet | -0.007 |
stretched | -0.007 |
La | -0.007 |
Tw | -0.007 |
Token | , |
Token ID | 11 |
Feature activation | +4.21e-01 |
Pos logit contributions | |
stretched | +0.009 |
ila | +0.009 |
eight | +0.008 |
Tw | +0.008 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.012 |
package | -0.012 |
meal | -0.012 |
tray | -0.012 |
batch | -0.012 |
Token | gum |
Token ID | 27142 |
Feature activation | +1.03e+00 |
Pos logit contributions | |
package | +0.012 |
tray | +0.011 |
meal | +0.011 |
cake | +0.011 |
idents | +0.011 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.013 |
La | -0.013 |
Tweet | -0.013 |
Tw | -0.012 |
Token | wra |
Token ID | 7917 |
Feature activation | +5.76e+00 |
Pos logit contributions | |
idents | +0.038 |
package | +0.037 |
batch | +0.036 |
meal | +0.036 |
tray | +0.035 |
Neg logit contributions | |
ila | -0.022 |
stretched | -0.022 |
eight | -0.021 |
Tw | -0.020 |
Tweet | -0.019 |
Token | ppers |
Token ID | 11799 |
Feature activation | +1.53e+00 |
Pos logit contributions | |
meal | +0.167 |
batch | +0.167 |
idents | +0.164 |
dinners | +0.151 |
tray | +0.151 |
Neg logit contributions | |
eight | -0.175 |
stretched | -0.174 |
Tweet | -0.154 |
ila | -0.152 |
Tw | -0.147 |
Token | , |
Token ID | 11 |
Feature activation | +3.71e-01 |
Pos logit contributions | |
idents | +0.056 |
package | +0.055 |
batch | +0.054 |
tray | +0.054 |
meal | +0.054 |
Neg logit contributions | |
stretched | -0.033 |
ila | -0.032 |
eight | -0.031 |
Tw | -0.030 |
Tweet | -0.028 |
Token | buttons |
Token ID | 12163 |
Feature activation | +7.56e-01 |
Pos logit contributions | |
package | +0.010 |
meal | +0.010 |
tray | +0.010 |
idents | +0.010 |
sandwiches | +0.010 |
Neg logit contributions | |
ila | -0.012 |
stretched | -0.012 |
La | -0.011 |
Tw | -0.011 |
Tweet | -0.011 |
Token | , |
Token ID | 11 |
Feature activation | +2.14e-01 |
Pos logit contributions | |
idents | +0.025 |
package | +0.025 |
meal | +0.024 |
tray | +0.024 |
batch | +0.024 |
Neg logit contributions | |
ila | -0.019 |
stretched | -0.019 |
Tw | -0.017 |
eight | -0.017 |
Tweet | -0.017 |
Token | and |
Token ID | 290 |
Feature activation | -5.04e-01 |
Pos logit contributions | |
package | +0.006 |
idents | +0.006 |
tray | +0.006 |
meal | +0.006 |
cake | +0.006 |
Neg logit contributions | |
ila | -0.007 |
stretched | -0.006 |
La | -0.006 |
Tw | -0.006 |
Tweet | -0.006 |
Token | ksh |
Token ID | 50133 |
Feature activation | +6.32e+00 |
Pos logit contributions | |
package | +0.055 |
meal | +0.053 |
refrigerator | +0.051 |
tray | +0.050 |
idents | +0.049 |
Neg logit contributions | |
eight | -0.074 |
stretched | -0.072 |
Tweet | -0.070 |
ila | -0.069 |
Tw | -0.069 |
Token | akes |
Token ID | 1124 |
Feature activation | +1.77e+00 |
Pos logit contributions | |
package | +0.184 |
tray | +0.179 |
envelope | +0.176 |
mess | +0.175 |
stret | +0.172 |
Neg logit contributions | |
eight | -0.157 |
stretched | -0.150 |
ib | -0.142 |
Some | -0.141 |
ubby | -0.139 |
Token | . |
Token ID | 13 |
Feature activation | +7.23e-01 |
Pos logit contributions | |
idents | +0.060 |
package | +0.059 |
meal | +0.058 |
tray | +0.057 |
batch | +0.057 |
Neg logit contributions | |
ila | -0.042 |
stretched | -0.042 |
Tw | -0.039 |
eight | -0.039 |
Tweet | -0.038 |
Token | The |
Token ID | 383 |
Feature activation | -1.06e+00 |
Pos logit contributions | |
idents | +0.022 |
batch | +0.021 |
tray | +0.021 |
refrigerator | +0.021 |
package | +0.021 |
Neg logit contributions | |
stretched | -0.020 |
ila | -0.020 |
Tweet | -0.018 |
La | -0.018 |
brav | -0.017 |
Token | mil |
Token ID | 1465 |
Feature activation | +2.06e+00 |
Pos logit contributions | |
stretched | +0.035 |
ila | +0.035 |
La | +0.033 |
Tweet | +0.032 |
Tw | +0.032 |
Neg logit contributions | |
package | -0.029 |
tray | -0.029 |
meal | -0.028 |
batch | -0.028 |
cake | -0.026 |
Token | ksh |
Token ID | 50133 |
Feature activation | +5.72e+00 |
Pos logit contributions | |
package | +0.030 |
meal | +0.028 |
idents | +0.026 |
tray | +0.026 |
refrigerator | +0.024 |
Neg logit contributions | |
stretched | -0.090 |
eight | -0.089 |
ila | -0.088 |
Tw | -0.088 |
Tweet | -0.088 |
Token | akes |
Token ID | 1124 |
Feature activation | +1.25e+00 |
Pos logit contributions | |
package | +0.164 |
envelope | +0.162 |
tray | +0.156 |
stret | +0.153 |
ours | +0.150 |
Neg logit contributions | |
eight | -0.146 |
Bo | -0.142 |
stretched | -0.140 |
Tw | -0.136 |
Then | -0.132 |
Token | were |
Token ID | 547 |
Feature activation | +2.58e+00 |
Pos logit contributions | |
idents | +0.040 |
package | +0.039 |
meal | +0.038 |
tray | +0.038 |
batch | +0.038 |
Neg logit contributions | |
ila | -0.033 |
stretched | -0.033 |
Tw | -0.030 |
eight | -0.030 |
Tweet | -0.030 |
Token | blue |
Token ID | 4171 |
Feature activation | +1.79e-01 |
Pos logit contributions | |
package | +0.074 |
idents | +0.074 |
meal | +0.073 |
tray | +0.073 |
batch | +0.072 |
Neg logit contributions | |
ila | -0.075 |
stretched | -0.074 |
Tw | -0.070 |
Tweet | -0.070 |
eight | -0.069 |
Token | and |
Token ID | 290 |
Feature activation | +7.32e-01 |
Pos logit contributions | |
idents | +0.006 |
package | +0.006 |
meal | +0.006 |
tray | +0.006 |
batch | +0.006 |
Neg logit contributions | |
stretched | -0.004 |
ila | -0.004 |
Tw | -0.004 |
eight | -0.004 |
Tweet | -0.004 |
Token | pink |
Token ID | 11398 |
Feature activation | +5.86e-01 |
Pos logit contributions | |
package | +0.022 |
meal | +0.022 |
tray | +0.022 |
batch | +0.022 |
idents | +0.021 |
Neg logit contributions | |
ila | -0.020 |
stretched | -0.020 |
La | -0.019 |
Tweet | -0.019 |
Tw | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | -8.78e-02 |
Pos logit contributions | |
ila | +0.032 |
Tweet | +0.031 |
stretched | +0.030 |
eight | +0.030 |
La | +0.029 |
Neg logit contributions | |
package | -0.033 |
meal | -0.033 |
idents | -0.033 |
tray | -0.032 |
batch | -0.032 |
Token | She |
Token ID | 1375 |
Feature activation | -1.13e+00 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tweet | +0.002 |
La | +0.002 |
Tw | +0.001 |
Neg logit contributions | |
idents | -0.003 |
package | -0.003 |
refrigerator | -0.003 |
tray | -0.003 |
batch | -0.003 |
Token | grabbed |
Token ID | 13176 |
Feature activation | -1.11e-01 |
Pos logit contributions | |
stretched | +0.031 |
ila | +0.030 |
Tw | +0.028 |
La | +0.028 |
Tweet | +0.027 |
Neg logit contributions | |
package | -0.035 |
idents | -0.034 |
tray | -0.033 |
meal | -0.033 |
batch | -0.033 |
Token | the |
Token ID | 262 |
Feature activation | -4.34e-01 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
eight | +0.002 |
Tw | +0.002 |
Tweet | +0.002 |
Neg logit contributions | |
idents | -0.005 |
package | -0.004 |
meal | -0.004 |
tray | -0.004 |
batch | -0.004 |
Token | fire |
Token ID | 2046 |
Feature activation | +9.22e-01 |
Pos logit contributions | |
stretched | +0.012 |
eight | +0.012 |
ila | +0.012 |
Tw | +0.012 |
Tweet | +0.011 |
Neg logit contributions | |
idents | -0.013 |
batch | -0.012 |
dinners | -0.011 |
meal | -0.011 |
ics | -0.011 |
Token | extingu |
Token ID | 24995 |
Feature activation | +5.72e+00 |
Pos logit contributions | |
idents | +0.032 |
batch | +0.031 |
meal | +0.030 |
package | +0.029 |
sandwiches | +0.029 |
Neg logit contributions | |
stretched | -0.021 |
Tw | -0.020 |
eight | -0.020 |
ila | -0.018 |
Tweet | -0.016 |
Token | isher |
Token ID | 4828 |
Feature activation | -5.39e-01 |
Pos logit contributions | |
meal | +0.168 |
ics | +0.161 |
tray | +0.160 |
dining | +0.156 |
dinner | +0.154 |
Neg logit contributions | |
stretched | -0.169 |
ila | -0.158 |
eight | -0.132 |
Tw | -0.131 |
brav | -0.129 |
Token | and |
Token ID | 290 |
Feature activation | -1.76e-01 |
Pos logit contributions | |
stretched | +0.011 |
ila | +0.010 |
eight | +0.010 |
Tw | +0.010 |
La | +0.009 |
Neg logit contributions | |
idents | -0.021 |
meal | -0.020 |
batch | -0.020 |
package | -0.020 |
tray | -0.019 |
Token | put |
Token ID | 1234 |
Feature activation | -1.55e-01 |
Pos logit contributions | |
ila | +0.004 |
stretched | +0.004 |
eight | +0.003 |
Tw | +0.003 |
Tweet | +0.003 |
Neg logit contributions | |
idents | -0.007 |
package | -0.006 |
meal | -0.006 |
batch | -0.006 |
tray | -0.006 |
Token | out |
Token ID | 503 |
Feature activation | -9.22e-01 |
Pos logit contributions | |
ila | +0.004 |
stretched | +0.004 |
Tw | +0.004 |
Tweet | +0.004 |
eight | +0.004 |
Neg logit contributions | |
package | -0.005 |
idents | -0.005 |
tray | -0.005 |
meal | -0.005 |
batch | -0.005 |
Token | the |
Token ID | 262 |
Feature activation | -7.22e-01 |
Pos logit contributions | |
Tweet | +0.025 |
ila | +0.025 |
La | +0.023 |
Bo | +0.023 |
stretched | +0.023 |
Neg logit contributions | |
package | -0.029 |
tray | -0.028 |
meal | -0.027 |
dinner | -0.026 |
cake | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | -2.68e-02 |
Pos logit contributions | |
stretched | +0.008 |
ila | +0.008 |
Tw | +0.008 |
Tweet | +0.008 |
La | +0.008 |
Neg logit contributions | |
package | -0.011 |
idents | -0.011 |
tray | -0.011 |
meal | -0.011 |
batch | -0.011 |
Token | There |
Token ID | 1318 |
Feature activation | -6.84e-01 |
Pos logit contributions | |
stretched | +0.001 |
ila | +0.000 |
Tweet | +0.000 |
La | +0.000 |
Tw | +0.000 |
Neg logit contributions | |
idents | -0.001 |
package | -0.001 |
tray | -0.001 |
meal | -0.001 |
batch | -0.001 |
Token | were |
Token ID | 547 |
Feature activation | +1.79e+00 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.014 |
Tw | +0.014 |
eight | +0.013 |
Tweet | +0.012 |
Neg logit contributions | |
stret | -0.025 |
idents | -0.025 |
batch | -0.024 |
package | -0.024 |
meal | -0.024 |
Token | cans |
Token ID | 23916 |
Feature activation | +2.30e+00 |
Pos logit contributions | |
tray | +0.061 |
package | +0.060 |
sandwiches | +0.060 |
meal | +0.059 |
idents | +0.059 |
Neg logit contributions | |
stretched | -0.043 |
ila | -0.042 |
La | -0.041 |
Tweet | -0.040 |
Tw | -0.040 |
Token | , |
Token ID | 11 |
Feature activation | +2.33e-01 |
Pos logit contributions | |
idents | +0.079 |
package | +0.078 |
meal | +0.076 |
tray | +0.076 |
batch | +0.075 |
Neg logit contributions | |
ila | -0.055 |
stretched | -0.055 |
eight | -0.051 |
Tw | -0.051 |
Tweet | -0.050 |
Token | wra |
Token ID | 7917 |
Feature activation | +5.58e+00 |
Pos logit contributions | |
package | +0.005 |
idents | +0.005 |
tray | +0.005 |
meal | +0.005 |
sandwiches | +0.005 |
Neg logit contributions | |
stretched | -0.008 |
ila | -0.008 |
Tw | -0.008 |
La | -0.008 |
Tweet | -0.008 |
Token | ppers |
Token ID | 11799 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
batch | +0.175 |
idents | +0.172 |
meal | +0.170 |
dinners | +0.159 |
meals | +0.155 |
Neg logit contributions | |
eight | -0.162 |
stretched | -0.160 |
ila | -0.154 |
Tweet | -0.151 |
A | -0.137 |
Token | and |
Token ID | 290 |
Feature activation | -7.44e-01 |
Pos logit contributions | |
idents | +0.044 |
package | +0.043 |
meal | +0.042 |
tray | +0.042 |
batch | +0.042 |
Neg logit contributions | |
ila | -0.030 |
stretched | -0.030 |
eight | -0.028 |
Tw | -0.028 |
Tweet | -0.027 |
Token | bottles |
Token ID | 14096 |
Feature activation | +2.07e+00 |
Pos logit contributions | |
ila | +0.027 |
stretched | +0.027 |
Tw | +0.026 |
La | +0.026 |
Tweet | +0.025 |
Neg logit contributions | |
package | -0.017 |
meal | -0.017 |
tray | -0.016 |
idents | -0.016 |
sandwiches | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | -8.14e-02 |
Pos logit contributions | |
idents | +0.082 |
batch | +0.078 |
package | +0.077 |
meal | +0.077 |
stret | +0.076 |
Neg logit contributions | |
ila | -0.041 |
stretched | -0.041 |
eight | -0.037 |
Tw | -0.035 |
Tweet | -0.033 |
Token | Sara |
Token ID | 24799 |
Feature activation | -4.33e-01 |
Pos logit contributions | |
stretched | +0.002 |
Tweet | +0.002 |
ila | +0.002 |
La | +0.002 |
Tw | +0.001 |
Neg logit contributions | |
idents | -0.003 |
tray | -0.003 |
package | -0.003 |
meal | -0.003 |
batch | -0.003 |
Token | Alice |
Token ID | 14862 |
Feature activation | +5.95e-01 |
Pos logit contributions | |
ila | +0.019 |
Tw | +0.018 |
stretched | +0.018 |
eight | +0.018 |
La | +0.017 |
Neg logit contributions | |
idents | -0.036 |
stret | -0.035 |
meal | -0.034 |
batch | -0.034 |
isions | -0.034 |
Token | saw |
Token ID | 2497 |
Feature activation | -2.26e-01 |
Pos logit contributions | |
idents | +0.020 |
package | +0.019 |
tray | +0.019 |
meal | +0.019 |
batch | +0.019 |
Neg logit contributions | |
ila | -0.015 |
stretched | -0.015 |
Tw | -0.014 |
eight | -0.014 |
Tweet | -0.014 |
Token | Lisa |
Token ID | 15378 |
Feature activation | +7.12e-01 |
Pos logit contributions | |
ila | +0.004 |
stretched | +0.004 |
Tw | +0.004 |
eight | +0.004 |
La | +0.004 |
Neg logit contributions | |
idents | -0.009 |
package | -0.008 |
meal | -0.008 |
tray | -0.008 |
batch | -0.008 |
Token | with |
Token ID | 351 |
Feature activation | -7.82e-02 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
tray | +0.021 |
meal | +0.021 |
batch | +0.021 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.019 |
Tw | -0.018 |
eight | -0.018 |
Tweet | -0.017 |
Token | the |
Token ID | 262 |
Feature activation | +4.03e-01 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tweet | +0.002 |
Tw | +0.002 |
La | +0.002 |
Neg logit contributions | |
package | -0.003 |
idents | -0.003 |
tray | -0.003 |
meal | -0.003 |
batch | -0.003 |
Token | y |
Token ID | 331 |
Feature activation | +5.58e+00 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
tray | +0.007 |
meal | +0.007 |
batch | +0.007 |
Neg logit contributions | |
ila | -0.016 |
stretched | -0.015 |
eight | -0.015 |
Tw | -0.015 |
Tweet | -0.015 |
Token | ummy |
Token ID | 13513 |
Feature activation | +3.47e+00 |
Pos logit contributions | |
stret | +0.189 |
idents | +0.179 |
tray | +0.171 |
package | +0.170 |
isions | +0.168 |
Neg logit contributions | |
stretched | -0.145 |
ila | -0.127 |
eight | -0.123 |
ubby | -0.114 |
ib | -0.113 |
Token | cand |
Token ID | 2658 |
Feature activation | +1.52e+00 |
Pos logit contributions | |
idents | +0.079 |
isions | +0.075 |
stret | +0.071 |
aman | +0.071 |
aghetti | +0.070 |
Neg logit contributions | |
ila | -0.119 |
stretched | -0.114 |
eight | -0.113 |
Tw | -0.112 |
Tweet | -0.111 |
Token | ies |
Token ID | 444 |
Feature activation | +2.20e-01 |
Pos logit contributions | |
idents | +0.049 |
stret | +0.048 |
package | +0.047 |
tray | +0.045 |
doorstep | +0.045 |
Neg logit contributions | |
ila | -0.041 |
stretched | -0.040 |
eight | -0.036 |
Tw | -0.034 |
Tweet | -0.033 |
Token | , |
Token ID | 11 |
Feature activation | +8.42e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
batch | +0.008 |
tray | +0.008 |
meal | +0.008 |
Neg logit contributions | |
stretched | -0.005 |
ila | -0.005 |
Tw | -0.004 |
eight | -0.004 |
Tweet | -0.004 |
Token | she |
Token ID | 673 |
Feature activation | -5.64e-02 |
Pos logit contributions | |
idents | +0.036 |
package | +0.035 |
meal | +0.035 |
tray | +0.035 |
batch | +0.035 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.013 |
Tw | -0.012 |
eight | -0.011 |
La | -0.011 |
Token | filled |
Token ID | 5901 |
Feature activation | +1.83e+00 |
Pos logit contributions | |
package | +0.040 |
tray | +0.039 |
meal | +0.039 |
batch | +0.039 |
idents | +0.037 |
Neg logit contributions | |
ila | -0.039 |
stretched | -0.038 |
Tweet | -0.038 |
La | -0.037 |
Tw | -0.036 |
Token | with |
Token ID | 351 |
Feature activation | +5.40e-01 |
Pos logit contributions | |
idents | +0.070 |
package | +0.067 |
stret | +0.066 |
meal | +0.065 |
tray | +0.064 |
Neg logit contributions | |
ila | -0.036 |
Tw | -0.036 |
stretched | -0.035 |
eight | -0.033 |
La | -0.033 |
Token | all |
Token ID | 477 |
Feature activation | +1.48e+00 |
Pos logit contributions | |
package | +0.017 |
idents | +0.016 |
meal | +0.016 |
tray | +0.016 |
batch | +0.016 |
Neg logit contributions | |
ila | -0.015 |
stretched | -0.015 |
Tweet | -0.014 |
Tw | -0.014 |
La | -0.014 |
Token | sorts |
Token ID | 10524 |
Feature activation | +2.31e+00 |
Pos logit contributions | |
idents | +0.048 |
package | +0.046 |
meal | +0.045 |
tray | +0.045 |
batch | +0.044 |
Neg logit contributions | |
stretched | -0.039 |
ila | -0.039 |
Tw | -0.036 |
eight | -0.036 |
La | -0.035 |
Token | of |
Token ID | 286 |
Feature activation | +1.15e+00 |
Pos logit contributions | |
package | +0.095 |
idents | +0.093 |
meal | +0.093 |
tray | +0.092 |
batch | +0.092 |
Neg logit contributions | |
ila | -0.037 |
La | -0.036 |
stretched | -0.035 |
Tw | -0.035 |
Tweet | -0.034 |
Token | y |
Token ID | 331 |
Feature activation | +5.56e+00 |
Pos logit contributions | |
package | +0.032 |
idents | +0.031 |
meal | +0.031 |
tray | +0.030 |
batch | +0.030 |
Neg logit contributions | |
ila | -0.035 |
stretched | -0.035 |
Tw | -0.034 |
eight | -0.033 |
Tweet | -0.033 |
Token | ummy |
Token ID | 13513 |
Feature activation | +3.16e+00 |
Pos logit contributions | |
stret | +0.173 |
tray | +0.167 |
package | +0.167 |
idents | +0.161 |
doorstep | +0.149 |
Neg logit contributions | |
stretched | -0.163 |
ila | -0.145 |
eight | -0.144 |
Tw | -0.132 |
Tweet | -0.127 |
Token | colours |
Token ID | 18915 |
Feature activation | +2.26e-01 |
Pos logit contributions | |
idents | +0.064 |
doorstep | +0.056 |
package | +0.054 |
tray | +0.054 |
stret | +0.053 |
Neg logit contributions | |
ila | -0.118 |
stretched | -0.116 |
eight | -0.114 |
Tw | -0.113 |
La | -0.109 |
Token | . |
Token ID | 13 |
Feature activation | +4.26e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
meal | +0.008 |
tray | +0.008 |
batch | +0.008 |
Neg logit contributions | |
stretched | -0.005 |
ila | -0.005 |
Tw | -0.005 |
eight | -0.005 |
Tweet | -0.005 |
Token | She |
Token ID | 1375 |
Feature activation | -7.48e-01 |
Pos logit contributions | |
batch | +0.014 |
refrigerator | +0.013 |
idents | +0.013 |
meal | +0.013 |
package | +0.013 |
Neg logit contributions | |
stretched | -0.011 |
ila | -0.011 |
La | -0.010 |
Tweet | -0.010 |
eight | -0.009 |
Token | wanted |
Token ID | 2227 |
Feature activation | +5.81e-01 |
Pos logit contributions | |
stretched | +0.022 |
ila | +0.022 |
Tweet | +0.021 |
La | +0.020 |
Tw | +0.020 |
Neg logit contributions | |
batch | -0.022 |
tray | -0.021 |
package | -0.021 |
meal | -0.020 |
idents | -0.020 |
Token | quickly |
Token ID | 2952 |
Feature activation | -8.46e-01 |
Pos logit contributions | |
idents | +0.016 |
package | +0.016 |
meal | +0.016 |
tray | +0.015 |
batch | +0.015 |
Neg logit contributions | |
ila | -0.010 |
stretched | -0.010 |
eight | -0.009 |
Tw | -0.009 |
Tweet | -0.009 |
Token | checked |
Token ID | 10667 |
Feature activation | -9.79e-01 |
Pos logit contributions | |
stretched | +0.018 |
ila | +0.017 |
eight | +0.017 |
Tweet | +0.016 |
Tw | +0.016 |
Neg logit contributions | |
idents | -0.033 |
package | -0.031 |
batch | -0.031 |
tray | -0.031 |
meal | -0.030 |
Token | for |
Token ID | 329 |
Feature activation | +1.96e+00 |
Pos logit contributions | |
Tweet | +0.028 |
La | +0.028 |
stretched | +0.027 |
Tw | +0.027 |
Tw | +0.025 |
Neg logit contributions | |
package | -0.030 |
tray | -0.028 |
refrigerator | -0.027 |
idents | -0.027 |
dinner | -0.027 |
Token | a |
Token ID | 257 |
Feature activation | +1.77e-01 |
Pos logit contributions | |
package | +0.074 |
meal | +0.074 |
idents | +0.072 |
tray | +0.072 |
dinner | +0.070 |
Neg logit contributions | |
ila | -0.042 |
stretched | -0.041 |
Tweet | -0.038 |
La | -0.037 |
Tw | -0.037 |
Token | fire |
Token ID | 2046 |
Feature activation | -1.12e-01 |
Pos logit contributions | |
idents | +0.007 |
package | +0.006 |
batch | +0.006 |
meal | +0.006 |
tray | +0.006 |
Neg logit contributions | |
stretched | -0.004 |
ila | -0.004 |
eight | -0.004 |
Tw | -0.003 |
Tweet | -0.003 |
Token | extingu |
Token ID | 24995 |
Feature activation | +5.54e+00 |
Pos logit contributions | |
stretched | +0.003 |
eight | +0.003 |
Tw | +0.003 |
ila | +0.003 |
Tweet | +0.002 |
Neg logit contributions | |
idents | -0.004 |
batch | -0.004 |
desserts | -0.003 |
goodies | -0.003 |
meal | -0.003 |
Token | isher |
Token ID | 4828 |
Feature activation | -1.63e+00 |
Pos logit contributions | |
dining | +0.155 |
meal | +0.154 |
ics | +0.152 |
idents | +0.152 |
tray | +0.149 |
Neg logit contributions | |
stretched | -0.164 |
ila | -0.157 |
brav | -0.133 |
eight | -0.133 |
racer | -0.129 |
Token | , |
Token ID | 11 |
Feature activation | +6.94e-01 |
Pos logit contributions | |
stretched | +0.033 |
ila | +0.032 |
Tw | +0.030 |
eight | +0.030 |
Tweet | +0.028 |
Neg logit contributions | |
idents | -0.062 |
package | -0.060 |
batch | -0.060 |
meal | -0.059 |
tray | -0.059 |
Token | but |
Token ID | 475 |
Feature activation | +4.10e-01 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
tray | +0.021 |
meal | +0.021 |
batch | +0.021 |
Neg logit contributions | |
ila | -0.019 |
stretched | -0.019 |
La | -0.017 |
Tw | -0.017 |
Tweet | -0.017 |
Token | there |
Token ID | 612 |
Feature activation | -6.94e-01 |
Pos logit contributions | |
package | +0.016 |
idents | +0.016 |
meal | +0.015 |
tray | +0.015 |
dinner | +0.015 |
Neg logit contributions | |
stretched | -0.008 |
ila | -0.008 |
La | -0.008 |
Tweet | -0.008 |
Tw | -0.007 |
Token | wasn |
Token ID | 2492 |
Feature activation | -4.75e-01 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.015 |
Tw | +0.014 |
eight | +0.014 |
Tweet | +0.014 |
Neg logit contributions | |
idents | -0.025 |
batch | -0.024 |
package | -0.024 |
tray | -0.024 |
meal | -0.024 |
Token | empty |
Token ID | 6565 |
Feature activation | +2.66e+00 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
eight | +0.002 |
Tweet | +0.002 |
Leave | +0.002 |
Neg logit contributions | |
idents | -0.002 |
batch | -0.002 |
aman | -0.002 |
package | -0.002 |
tray | -0.002 |
Token | bag |
Token ID | 6131 |
Feature activation | +1.73e+00 |
Pos logit contributions | |
idents | +0.072 |
ksh | +0.066 |
isions | +0.064 |
batch | +0.063 |
aman | +0.063 |
Neg logit contributions | |
ila | -0.079 |
eight | -0.078 |
stretched | -0.077 |
Tweet | -0.073 |
Tw | -0.071 |
Token | and |
Token ID | 290 |
Feature activation | -2.97e-01 |
Pos logit contributions | |
idents | +0.072 |
package | +0.067 |
meal | +0.067 |
batch | +0.067 |
dinners | +0.065 |
Neg logit contributions | |
stretched | -0.033 |
ila | -0.032 |
eight | -0.028 |
Tweet | -0.027 |
Tw | -0.026 |
Token | the |
Token ID | 262 |
Feature activation | -5.33e-01 |
Pos logit contributions | |
stretched | +0.007 |
ila | +0.007 |
La | +0.007 |
Tw | +0.007 |
Tweet | +0.006 |
Neg logit contributions | |
package | -0.010 |
idents | -0.010 |
batch | -0.010 |
tray | -0.010 |
meal | -0.010 |
Token | candy |
Token ID | 18550 |
Feature activation | +5.17e-01 |
Pos logit contributions | |
ila | +0.018 |
stretched | +0.018 |
eight | +0.017 |
Tw | +0.017 |
Tweet | +0.017 |
Neg logit contributions | |
idents | -0.013 |
package | -0.013 |
meal | -0.012 |
tray | -0.012 |
batch | -0.012 |
Token | wra |
Token ID | 7917 |
Feature activation | +5.53e+00 |
Pos logit contributions | |
idents | +0.020 |
aman | +0.019 |
ksh | +0.019 |
batch | +0.019 |
isions | +0.018 |
Neg logit contributions | |
ila | -0.010 |
stretched | -0.010 |
he | -0.009 |
eight | -0.008 |
Tw | -0.008 |
Token | ppers |
Token ID | 11799 |
Feature activation | +9.67e-01 |
Pos logit contributions | |
meal | +0.175 |
idents | +0.175 |
batch | +0.173 |
dinners | +0.165 |
meals | +0.163 |
Neg logit contributions | |
stretched | -0.158 |
ila | -0.150 |
eight | -0.145 |
Tweet | -0.141 |
He | -0.133 |
Token | on |
Token ID | 319 |
Feature activation | +1.12e-01 |
Pos logit contributions | |
idents | +0.038 |
package | +0.037 |
batch | +0.036 |
meal | +0.036 |
tray | +0.035 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.019 |
eight | -0.017 |
Tw | -0.016 |
Tweet | -0.015 |
Token | the |
Token ID | 262 |
Feature activation | +7.46e-01 |
Pos logit contributions | |
idents | +0.005 |
package | +0.005 |
meal | +0.005 |
batch | +0.005 |
tray | +0.004 |
Neg logit contributions | |
ila | -0.002 |
stretched | -0.002 |
eight | -0.002 |
Tw | -0.002 |
Tweet | -0.002 |
Token | floor |
Token ID | 4314 |
Feature activation | +1.88e+00 |
Pos logit contributions | |
idents | +0.016 |
batch | +0.015 |
package | +0.015 |
meal | +0.014 |
tray | +0.014 |
Neg logit contributions | |
eight | -0.027 |
ila | -0.027 |
stretched | -0.026 |
Tw | -0.026 |
Tweet | -0.025 |
Token | . |
Token ID | 13 |
Feature activation | +2.40e-01 |
Pos logit contributions | |
idents | +0.071 |
package | +0.070 |
meal | +0.068 |
tray | +0.067 |
batch | +0.067 |
Neg logit contributions | |
stretched | -0.040 |
ila | -0.040 |
eight | -0.036 |
Tw | -0.035 |
Tweet | -0.034 |
Token | would |
Token ID | 561 |
Feature activation | -1.69e+00 |
Pos logit contributions | |
stretched | +0.005 |
ila | +0.005 |
Tw | +0.005 |
La | +0.005 |
Tweet | +0.005 |
Neg logit contributions | |
package | -0.006 |
meal | -0.006 |
tray | -0.006 |
batch | -0.006 |
idents | -0.006 |
Token | take |
Token ID | 1011 |
Feature activation | -7.28e-01 |
Pos logit contributions | |
ila | +0.045 |
stretched | +0.045 |
Tw | +0.042 |
eight | +0.042 |
Tweet | +0.041 |
Neg logit contributions | |
idents | -0.053 |
package | -0.053 |
meal | -0.052 |
tray | -0.051 |
batch | -0.051 |
Token | out |
Token ID | 503 |
Feature activation | -6.37e-01 |
Pos logit contributions | |
ila | +0.017 |
stretched | +0.017 |
Tw | +0.016 |
eight | +0.016 |
Tweet | +0.015 |
Neg logit contributions | |
idents | -0.025 |
package | -0.025 |
meal | -0.025 |
tray | -0.025 |
batch | -0.024 |
Token | a |
Token ID | 257 |
Feature activation | -1.13e+00 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.015 |
Tweet | +0.015 |
La | +0.015 |
Tw | +0.014 |
Neg logit contributions | |
package | -0.021 |
meal | -0.021 |
tray | -0.021 |
batch | -0.021 |
idents | -0.020 |
Token | j |
Token ID | 474 |
Feature activation | +6.15e-01 |
Pos logit contributions | |
stretched | +0.029 |
eight | +0.028 |
ila | +0.028 |
Tw | +0.028 |
Tweet | +0.027 |
Neg logit contributions | |
idents | -0.037 |
package | -0.034 |
meal | -0.033 |
batch | -0.033 |
stret | -0.033 |
Token | igsaw |
Token ID | 45636 |
Feature activation | +5.50e+00 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
meal | +0.021 |
tray | +0.021 |
batch | +0.021 |
Neg logit contributions | |
stretched | -0.014 |
ila | -0.014 |
Tw | -0.013 |
eight | -0.013 |
Tweet | -0.013 |
Token | puzzle |
Token ID | 15027 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
idents | +0.160 |
stret | +0.155 |
aman | +0.151 |
lette | +0.145 |
urger | +0.143 |
Neg logit contributions | |
stretched | -0.153 |
Tw | -0.152 |
eight | -0.144 |
br | -0.142 |
fast | -0.141 |
Token | and |
Token ID | 290 |
Feature activation | -1.54e-01 |
Pos logit contributions | |
idents | +0.046 |
package | +0.042 |
meal | +0.041 |
batch | +0.041 |
stret | +0.041 |
Neg logit contributions | |
stretched | -0.024 |
ila | -0.023 |
Tw | -0.022 |
eight | -0.022 |
Tweet | -0.021 |
Token | try |
Token ID | 1949 |
Feature activation | -1.57e+00 |
Pos logit contributions | |
ila | +0.004 |
stretched | +0.004 |
Tweet | +0.004 |
Tw | +0.004 |
La | +0.004 |
Neg logit contributions | |
idents | -0.005 |
package | -0.005 |
meal | -0.005 |
tray | -0.005 |
batch | -0.005 |
Token | to |
Token ID | 284 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
ila | +0.041 |
stretched | +0.040 |
Tweet | +0.038 |
eight | +0.037 |
Tw | +0.036 |
Neg logit contributions | |
package | -0.050 |
idents | -0.050 |
meal | -0.050 |
tray | -0.049 |
batch | -0.049 |
Token | solve |
Token ID | 8494 |
Feature activation | -2.29e+00 |
Pos logit contributions | |
idents | +0.046 |
package | +0.046 |
meal | +0.044 |
batch | +0.044 |
tray | +0.044 |
Neg logit contributions | |
stretched | -0.034 |
ila | -0.033 |
eight | -0.032 |
Tw | -0.032 |
Tweet | -0.031 |
Token | They |
Token ID | 1119 |
Feature activation | +1.14e+00 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.002 |
Tweet | +0.002 |
La | +0.002 |
eight | +0.002 |
Neg logit contributions | |
idents | -0.002 |
tray | -0.002 |
package | -0.002 |
meal | -0.002 |
batch | -0.002 |
Token | said |
Token ID | 531 |
Feature activation | +5.00e-01 |
Pos logit contributions | |
idents | +0.038 |
package | +0.037 |
meal | +0.036 |
tray | +0.035 |
batch | +0.035 |
Neg logit contributions | |
ila | -0.029 |
stretched | -0.028 |
Tw | -0.027 |
eight | -0.027 |
Tweet | -0.026 |
Token | , |
Token ID | 11 |
Feature activation | +1.57e+00 |
Pos logit contributions | |
idents | +0.016 |
meal | +0.015 |
package | +0.015 |
tray | +0.015 |
batch | +0.015 |
Neg logit contributions | |
ila | -0.014 |
stretched | -0.013 |
Tweet | -0.013 |
La | -0.013 |
Tw | -0.012 |
Token | " |
Token ID | 366 |
Feature activation | -2.78e-01 |
Pos logit contributions | |
tray | +0.051 |
idents | +0.050 |
meal | +0.049 |
package | +0.049 |
batch | +0.048 |
Neg logit contributions | |
ila | -0.041 |
stretched | -0.040 |
Tweet | -0.037 |
La | -0.036 |
eight | -0.034 |
Token | Bon |
Token ID | 20682 |
Feature activation | +6.92e-01 |
Pos logit contributions | |
ila | +0.006 |
stretched | +0.006 |
Tw | +0.005 |
eight | +0.005 |
Tweet | +0.005 |
Neg logit contributions | |
idents | -0.010 |
package | -0.010 |
meal | -0.010 |
tray | -0.010 |
batch | -0.010 |
Token | appet |
Token ID | 16422 |
Feature activation | +5.49e+00 |
Pos logit contributions | |
idents | +0.011 |
stret | +0.011 |
package | +0.011 |
meal | +0.011 |
tray | +0.011 |
Neg logit contributions | |
ila | -0.028 |
Tw | -0.027 |
eight | -0.027 |
stretched | -0.026 |
Tweet | -0.026 |
Token | it |
Token ID | 270 |
Feature activation | -1.83e+00 |
Pos logit contributions | |
meal | +0.182 |
stret | +0.176 |
package | +0.169 |
tray | +0.166 |
doorstep | +0.166 |
Neg logit contributions | |
ila | -0.155 |
A | -0.143 |
I | -0.139 |
Tw | -0.137 |
Tweet | -0.137 |
Token | !" |
Token ID | 2474 |
Feature activation | +3.75e-01 |
Pos logit contributions | |
Tw | +0.041 |
ila | +0.041 |
eight | +0.040 |
stretched | +0.039 |
La | +0.035 |
Neg logit contributions | |
idents | -0.064 |
batch | -0.062 |
stret | -0.061 |
package | -0.061 |
meal | -0.059 |
Token | They |
Token ID | 1119 |
Feature activation | +5.90e-01 |
Pos logit contributions | |
idents | +0.011 |
meal | +0.011 |
tray | +0.011 |
batch | +0.011 |
package | +0.011 |
Neg logit contributions | |
stretched | -0.011 |
ila | -0.010 |
Tweet | -0.010 |
La | -0.010 |
brav | -0.009 |
Token | watched |
Token ID | 7342 |
Feature activation | -9.72e-01 |
Pos logit contributions | |
idents | +0.021 |
package | +0.020 |
meal | +0.019 |
batch | +0.019 |
tray | +0.019 |
Neg logit contributions | |
ila | -0.014 |
stretched | -0.014 |
eight | -0.013 |
Tw | -0.013 |
Tweet | -0.013 |
Token | their |
Token ID | 511 |
Feature activation | +1.52e+00 |
Pos logit contributions | |
stretched | +0.021 |
ila | +0.021 |
Tw | +0.020 |
eight | +0.019 |
Tweet | +0.018 |
Neg logit contributions | |
idents | -0.036 |
package | -0.036 |
tray | -0.034 |
meal | -0.034 |
batch | -0.034 |
Token | the |
Token ID | 262 |
Feature activation | +1.03e-01 |
Pos logit contributions | |
idents | +0.031 |
package | +0.031 |
batch | +0.030 |
tray | +0.030 |
meal | +0.030 |
Neg logit contributions | |
ila | -0.009 |
eight | -0.009 |
stretched | -0.009 |
Tw | -0.008 |
Tweet | -0.008 |
Token | kitchen |
Token ID | 9592 |
Feature activation | +1.84e+00 |
Pos logit contributions | |
idents | +0.003 |
urger | +0.003 |
batch | +0.002 |
stret | +0.002 |
lette | +0.002 |
Neg logit contributions | |
eight | -0.003 |
Tw | -0.003 |
stretched | -0.003 |
A | -0.003 |
ila | -0.003 |
Token | to |
Token ID | 284 |
Feature activation | +6.01e-01 |
Pos logit contributions | |
idents | +0.084 |
package | +0.082 |
stret | +0.082 |
advertisement | +0.079 |
isions | +0.079 |
Neg logit contributions | |
eight | -0.018 |
Tw | -0.015 |
ila | -0.014 |
stretched | -0.014 |
L | -0.012 |
Token | see |
Token ID | 766 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
idents | +0.022 |
package | +0.021 |
batch | +0.021 |
tray | +0.020 |
isions | +0.020 |
Neg logit contributions | |
ila | -0.013 |
eight | -0.013 |
stretched | -0.012 |
Tw | -0.012 |
Tweet | -0.012 |
Token | the |
Token ID | 262 |
Feature activation | +1.76e-01 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
Tw | +0.007 |
eight | +0.006 |
Tweet | +0.006 |
Neg logit contributions | |
idents | -0.010 |
package | -0.010 |
meal | -0.010 |
tray | -0.010 |
batch | -0.010 |
Token | y |
Token ID | 331 |
Feature activation | +5.48e+00 |
Pos logit contributions | |
idents | +0.006 |
stret | +0.005 |
urger | +0.005 |
keeping | +0.005 |
batch | +0.005 |
Neg logit contributions | |
Tw | -0.004 |
eight | -0.004 |
stretched | -0.004 |
ila | -0.004 |
I | -0.004 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.90e+00 |
Pos logit contributions | |
stret | +0.222 |
idents | +0.191 |
advertisement | +0.187 |
isions | +0.186 |
assed | +0.186 |
Neg logit contributions | |
stretched | -0.107 |
ubby | -0.092 |
ila | -0.082 |
Tw | -0.080 |
eight | -0.079 |
Token | food |
Token ID | 2057 |
Feature activation | +2.08e+00 |
Pos logit contributions | |
isions | +0.100 |
assed | +0.098 |
arre | +0.098 |
idents | +0.098 |
stret | +0.096 |
Neg logit contributions | |
blue | -0.064 |
of | -0.063 |
sun | -0.061 |
Tw | -0.060 |
she | -0.060 |
Token | . |
Token ID | 13 |
Feature activation | +4.92e-01 |
Pos logit contributions | |
idents | +0.095 |
stret | +0.091 |
package | +0.091 |
isions | +0.091 |
aman | +0.089 |
Neg logit contributions | |
ila | -0.026 |
stretched | -0.023 |
Tw | -0.021 |
he | -0.020 |
she | -0.018 |
Token | The |
Token ID | 383 |
Feature activation | -1.15e+00 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
meal | +0.022 |
tray | +0.022 |
batch | +0.022 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
Tw | -0.005 |
eight | -0.005 |
Tweet | -0.005 |
Token | baby |
Token ID | 5156 |
Feature activation | +1.50e+00 |
Pos logit contributions | |
eight | +0.028 |
stretched | +0.027 |
Tw | +0.027 |
ila | +0.025 |
A | +0.024 |
Neg logit contributions | |
idents | -0.039 |
stret | -0.036 |
batch | -0.035 |
isions | -0.034 |
aman | -0.034 |
Token | day |
Token ID | 1110 |
Feature activation | -1.02e+00 |
Pos logit contributions | |
ila | +0.005 |
stretched | +0.005 |
Tweet | +0.005 |
Tw | +0.005 |
La | +0.005 |
Neg logit contributions | |
idents | -0.007 |
package | -0.007 |
batch | -0.007 |
tray | -0.007 |
meal | -0.007 |
Token | drinking |
Token ID | 7722 |
Feature activation | -1.41e-01 |
Pos logit contributions | |
stretched | +0.027 |
ila | +0.025 |
Tw | +0.025 |
eight | +0.025 |
La | +0.023 |
Neg logit contributions | |
idents | -0.034 |
package | -0.034 |
tray | -0.032 |
refrigerator | -0.031 |
batch | -0.031 |
Token | y |
Token ID | 331 |
Feature activation | +4.15e+00 |
Pos logit contributions | |
ila | +0.003 |
stretched | +0.003 |
Tw | +0.003 |
eight | +0.003 |
Tweet | +0.003 |
Neg logit contributions | |
idents | -0.005 |
package | -0.005 |
isions | -0.005 |
doorstep | -0.005 |
batch | -0.005 |
Token | ummy |
Token ID | 13513 |
Feature activation | +1.97e+00 |
Pos logit contributions | |
stret | +0.138 |
package | +0.136 |
tray | +0.134 |
envelope | +0.127 |
idents | +0.125 |
Neg logit contributions | |
stretched | -0.109 |
ila | -0.094 |
eight | -0.087 |
Tw | -0.087 |
W | -0.082 |
Token | mil |
Token ID | 1465 |
Feature activation | +1.50e+00 |
Pos logit contributions | |
idents | +0.046 |
doorstep | +0.045 |
package | +0.042 |
box | +0.041 |
envelope | +0.041 |
Neg logit contributions | |
Tw | -0.062 |
stretched | -0.060 |
ila | -0.060 |
blue | -0.059 |
W | -0.058 |
Token | ksh |
Token ID | 50133 |
Feature activation | +5.45e+00 |
Pos logit contributions | |
package | +0.026 |
envelope | +0.023 |
tray | +0.023 |
meal | +0.022 |
refrigerator | +0.021 |
Neg logit contributions | |
Tweet | -0.061 |
Tw | -0.061 |
stretched | -0.060 |
eight | -0.059 |
La | -0.059 |
Token | akes |
Token ID | 1124 |
Feature activation | +4.05e-01 |
Pos logit contributions | |
envelope | +0.158 |
package | +0.156 |
stret | +0.153 |
mess | +0.147 |
tray | +0.147 |
Neg logit contributions | |
stretched | -0.140 |
Bo | -0.135 |
A | -0.131 |
Then | -0.131 |
eight | -0.131 |
Token | in |
Token ID | 287 |
Feature activation | -7.93e-01 |
Pos logit contributions | |
idents | +0.015 |
package | +0.015 |
batch | +0.014 |
meal | +0.014 |
tray | +0.014 |
Neg logit contributions | |
stretched | -0.008 |
Tw | -0.008 |
ila | -0.008 |
eight | -0.007 |
La | -0.007 |
Token | the |
Token ID | 262 |
Feature activation | +2.86e-01 |
Pos logit contributions | |
ila | +0.014 |
stretched | +0.014 |
Tw | +0.013 |
eight | +0.013 |
Tweet | +0.012 |
Neg logit contributions | |
idents | -0.032 |
package | -0.032 |
meal | -0.031 |
tray | -0.031 |
batch | -0.031 |
Token | sun |
Token ID | 4252 |
Feature activation | -7.25e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
tray | +0.008 |
meal | +0.008 |
batch | +0.008 |
Neg logit contributions | |
stretched | -0.008 |
ila | -0.008 |
Tw | -0.008 |
eight | -0.008 |
Tweet | -0.008 |
Token | . |
Token ID | 13 |
Feature activation | +3.92e-01 |
Pos logit contributions | |
stretched | +0.013 |
Tw | +0.013 |
ila | +0.012 |
eight | +0.011 |
Tweet | +0.011 |
Neg logit contributions | |
idents | -0.029 |
package | -0.028 |
tray | -0.027 |
batch | -0.027 |
stret | -0.027 |
Token | face |
Token ID | 1986 |
Feature activation | +1.77e+00 |
Pos logit contributions | |
package | +0.060 |
meal | +0.059 |
idents | +0.058 |
batch | +0.058 |
tray | +0.057 |
Neg logit contributions | |
stretched | -0.058 |
ila | -0.058 |
La | -0.057 |
Tweet | -0.055 |
Tw | -0.055 |
Token | , |
Token ID | 11 |
Feature activation | +6.79e-01 |
Pos logit contributions | |
meal | +0.049 |
idents | +0.049 |
package | +0.048 |
tray | +0.047 |
batch | +0.047 |
Neg logit contributions | |
ila | -0.053 |
stretched | -0.052 |
Tweet | -0.051 |
La | -0.051 |
eight | -0.050 |
Token | the |
Token ID | 262 |
Feature activation | +2.01e-01 |
Pos logit contributions | |
meal | +0.020 |
idents | +0.020 |
package | +0.020 |
tray | +0.020 |
batch | +0.020 |
Neg logit contributions | |
ila | -0.019 |
stretched | -0.019 |
La | -0.018 |
Tweet | -0.018 |
Tw | -0.017 |
Token | world |
Token ID | 995 |
Feature activation | +9.33e-01 |
Pos logit contributions | |
package | +0.006 |
tray | +0.006 |
meal | +0.006 |
refrigerator | +0.006 |
idents | +0.006 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
La | -0.005 |
Tweet | -0.005 |
Tw | -0.005 |
Token | was |
Token ID | 373 |
Feature activation | +9.96e-01 |
Pos logit contributions | |
package | +0.028 |
idents | +0.028 |
tray | +0.028 |
meal | +0.027 |
batch | +0.027 |
Neg logit contributions | |
ila | -0.026 |
stretched | -0.025 |
Tweet | -0.024 |
eight | -0.024 |
Tw | -0.024 |
Token | ours |
Token ID | 16903 |
Feature activation | +5.43e+00 |
Pos logit contributions | |
package | +0.030 |
meal | +0.030 |
tray | +0.029 |
batch | +0.028 |
idents | +0.028 |
Neg logit contributions | |
ila | -0.029 |
stretched | -0.027 |
Tweet | -0.026 |
La | -0.026 |
Tw | -0.025 |
Token | to |
Token ID | 284 |
Feature activation | +6.76e-01 |
Pos logit contributions | |
idents | +0.193 |
package | +0.187 |
batch | +0.185 |
meal | +0.182 |
tray | +0.181 |
Neg logit contributions | |
ila | -0.122 |
stretched | -0.119 |
Tw | -0.112 |
eight | -0.111 |
Tweet | -0.106 |
Token | explore |
Token ID | 7301 |
Feature activation | +3.47e-01 |
Pos logit contributions | |
meal | +0.021 |
package | +0.021 |
idents | +0.021 |
tray | +0.021 |
batch | +0.020 |
Neg logit contributions | |
stretched | -0.018 |
ila | -0.018 |
Tw | -0.017 |
eight | -0.017 |
Tweet | -0.017 |
Token | . |
Token ID | 13 |
Feature activation | +3.17e-01 |
Pos logit contributions | |
package | +0.012 |
meal | +0.012 |
idents | +0.011 |
tray | +0.011 |
batch | +0.011 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.008 |
eight | -0.008 |
Tweet | -0.008 |
Tw | -0.008 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.72e+00 |
Pos logit contributions | |
package | +0.010 |
idents | +0.009 |
refrigerator | +0.009 |
batch | +0.009 |
tray | +0.009 |
Neg logit contributions | |
stretched | -0.009 |
ila | -0.008 |
La | -0.007 |
Tweet | -0.007 |
eight | -0.007 |
Token | J |
Token ID | 41 |
Feature activation | -1.29e-01 |
Pos logit contributions | |
package | +0.059 |
meal | +0.058 |
tray | +0.057 |
refrigerator | +0.056 |
dinner | +0.055 |
Neg logit contributions | |
stretched | -0.042 |
ila | -0.037 |
eight | -0.037 |
Tweet | -0.037 |
La | -0.034 |
Token | happy |
Token ID | 3772 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
La | +0.006 |
Tweet | +0.006 |
Tw | +0.006 |
Neg logit contributions | |
meal | -0.008 |
package | -0.008 |
tray | -0.008 |
batch | -0.007 |
idents | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | +3.65e-01 |
Pos logit contributions | |
package | +0.047 |
idents | +0.046 |
tray | +0.046 |
stret | +0.045 |
batch | +0.045 |
Neg logit contributions | |
stretched | -0.027 |
ila | -0.027 |
Tw | -0.026 |
eight | -0.025 |
Tweet | -0.024 |
Token | people |
Token ID | 661 |
Feature activation | -1.39e-02 |
Pos logit contributions | |
idents | +0.012 |
package | +0.012 |
tray | +0.012 |
meal | +0.011 |
batch | +0.011 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
Tw | -0.009 |
eight | -0.009 |
Tweet | -0.009 |
Token | saw |
Token ID | 2497 |
Feature activation | -5.85e-03 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
eight | +0.000 |
La | +0.000 |
Tw | +0.000 |
Neg logit contributions | |
package | -0.000 |
idents | -0.000 |
meal | -0.000 |
tray | -0.000 |
batch | -0.000 |
Token | the |
Token ID | 262 |
Feature activation | +3.74e-01 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
Tweet | +0.000 |
Tw | +0.000 |
eight | +0.000 |
Neg logit contributions | |
package | -0.000 |
idents | -0.000 |
meal | -0.000 |
tray | -0.000 |
batch | -0.000 |
Token | y |
Token ID | 331 |
Feature activation | +5.43e+00 |
Pos logit contributions | |
package | +0.012 |
meal | +0.012 |
tray | +0.012 |
idents | +0.012 |
mess | +0.011 |
Neg logit contributions | |
stretched | -0.011 |
ila | -0.010 |
eight | -0.009 |
La | -0.009 |
Tw | -0.009 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.94e+00 |
Pos logit contributions | |
stret | +0.176 |
tray | +0.165 |
package | +0.162 |
idents | +0.161 |
dining | +0.156 |
Neg logit contributions | |
stretched | -0.151 |
ila | -0.131 |
Tw | -0.123 |
Tweet | -0.123 |
eight | -0.122 |
Token | treat |
Token ID | 2190 |
Feature activation | +1.12e+00 |
Pos logit contributions | |
idents | +0.081 |
doorstep | +0.077 |
aman | +0.073 |
stret | +0.072 |
isions | +0.072 |
Neg logit contributions | |
ila | -0.092 |
Tw | -0.084 |
stretched | -0.082 |
Tweet | -0.081 |
mag | -0.079 |
Token | it |
Token ID | 340 |
Feature activation | +1.43e+00 |
Pos logit contributions | |
idents | +0.044 |
package | +0.042 |
stret | +0.041 |
batch | +0.040 |
meal | +0.040 |
Neg logit contributions | |
ila | -0.022 |
stretched | -0.021 |
Tw | -0.020 |
eight | -0.019 |
Tweet | -0.017 |
Token | had |
Token ID | 550 |
Feature activation | +2.68e-01 |
Pos logit contributions | |
idents | +0.047 |
package | +0.047 |
meal | +0.046 |
tray | +0.046 |
batch | +0.045 |
Neg logit contributions | |
ila | -0.036 |
stretched | -0.036 |
eight | -0.033 |
Tw | -0.033 |
Tweet | -0.033 |
Token | found |
Token ID | 1043 |
Feature activation | -1.64e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
tray | +0.008 |
batch | +0.008 |
meal | +0.008 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.008 |
Tw | -0.007 |
eight | -0.007 |
Tweet | -0.007 |
Token | he |
Token ID | 339 |
Feature activation | -6.02e-01 |
Pos logit contributions | |
Tweet | +0.025 |
ila | +0.024 |
stretched | +0.023 |
La | +0.022 |
Tw | +0.022 |
Neg logit contributions | |
dinner | -0.034 |
meal | -0.034 |
cake | -0.032 |
desserts | -0.032 |
dessert | -0.032 |
Token | saw |
Token ID | 2497 |
Feature activation | -6.54e-01 |
Pos logit contributions | |
stretched | +0.018 |
ila | +0.018 |
La | +0.017 |
Tw | +0.017 |
Tweet | +0.017 |
Neg logit contributions | |
meal | -0.017 |
batch | -0.017 |
package | -0.017 |
tray | -0.016 |
idents | -0.016 |
Token | the |
Token ID | 262 |
Feature activation | +1.13e-01 |
Pos logit contributions | |
stretched | +0.017 |
ila | +0.017 |
La | +0.016 |
Tw | +0.016 |
Tweet | +0.016 |
Neg logit contributions | |
package | -0.021 |
meal | -0.021 |
idents | -0.021 |
batch | -0.021 |
tray | -0.021 |
Token | bowl |
Token ID | 9396 |
Feature activation | +2.94e+00 |
Pos logit contributions | |
idents | +0.003 |
package | +0.003 |
tray | +0.003 |
batch | +0.003 |
stret | +0.003 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
eight | -0.003 |
Tweet | -0.003 |
Tw | -0.003 |
Token | of |
Token ID | 286 |
Feature activation | +5.78e-01 |
Pos logit contributions | |
idents | +0.117 |
package | +0.113 |
stret | +0.109 |
tray | +0.108 |
meal | +0.108 |
Neg logit contributions | |
stretched | -0.058 |
ila | -0.055 |
eight | -0.050 |
Tw | -0.048 |
Tweet | -0.047 |
Token | y |
Token ID | 331 |
Feature activation | +5.41e+00 |
Pos logit contributions | |
idents | +0.012 |
package | +0.012 |
meal | +0.012 |
tray | +0.012 |
batch | +0.011 |
Neg logit contributions | |
ila | -0.021 |
stretched | -0.021 |
Tw | -0.020 |
eight | -0.020 |
Tweet | -0.020 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.93e+00 |
Pos logit contributions | |
stret | +0.182 |
package | +0.174 |
tray | +0.171 |
idents | +0.162 |
dining | +0.157 |
Neg logit contributions | |
stretched | -0.147 |
ila | -0.128 |
eight | -0.124 |
ib | -0.118 |
Tweet | -0.117 |
Token | peanuts |
Token ID | 45612 |
Feature activation | +2.03e+00 |
Pos logit contributions | |
doorstep | +0.081 |
stret | +0.081 |
idents | +0.081 |
package | +0.080 |
tray | +0.079 |
Neg logit contributions | |
ila | -0.083 |
stretched | -0.080 |
Tweet | -0.077 |
eight | -0.075 |
he | -0.074 |
Token | ! |
Token ID | 0 |
Feature activation | +4.34e-01 |
Pos logit contributions | |
idents | +0.083 |
package | +0.081 |
stret | +0.079 |
batch | +0.076 |
isions | +0.075 |
Neg logit contributions | |
ila | -0.032 |
he | -0.032 |
stretched | -0.031 |
she | -0.028 |
Tw | -0.028 |
Token | They |
Token ID | 1119 |
Feature activation | +2.33e+00 |
Pos logit contributions | |
idents | +0.016 |
batch | +0.016 |
package | +0.016 |
meal | +0.015 |
refrigerator | +0.015 |
Neg logit contributions | |
stretched | -0.009 |
ila | -0.009 |
La | -0.009 |
Tweet | -0.008 |
eight | -0.008 |
Token | both |
Token ID | 1111 |
Feature activation | +3.56e-01 |
Pos logit contributions | |
idents | +0.081 |
package | +0.079 |
tray | +0.076 |
meal | +0.075 |
stret | +0.075 |
Neg logit contributions | |
stretched | -0.057 |
ila | -0.056 |
eight | -0.055 |
Tweet | -0.053 |
Tw | -0.053 |
Token | unp |
Token ID | 8593 |
Feature activation | +2.77e+00 |
Pos logit contributions | |
package | +0.028 |
meal | +0.027 |
idents | +0.027 |
tray | +0.027 |
batch | +0.026 |
Neg logit contributions | |
stretched | -0.026 |
ila | -0.025 |
Tweet | -0.024 |
Tw | -0.024 |
La | -0.023 |
Token | acked |
Token ID | 6021 |
Feature activation | +1.39e+00 |
Pos logit contributions | |
idents | +0.068 |
package | +0.067 |
batch | +0.064 |
tray | +0.063 |
meal | +0.063 |
Neg logit contributions | |
stretched | -0.094 |
ila | -0.094 |
Tweet | -0.090 |
eight | -0.090 |
Tw | -0.089 |
Token | all |
Token ID | 477 |
Feature activation | +1.85e+00 |
Pos logit contributions | |
idents | +0.041 |
package | +0.041 |
tray | +0.040 |
meal | +0.040 |
batch | +0.039 |
Neg logit contributions | |
ila | -0.040 |
stretched | -0.040 |
Tweet | -0.038 |
Tw | -0.037 |
eight | -0.037 |
Token | kinds |
Token ID | 6982 |
Feature activation | +2.05e+00 |
Pos logit contributions | |
package | +0.076 |
meal | +0.074 |
tray | +0.073 |
batch | +0.073 |
idents | +0.073 |
Neg logit contributions | |
ila | -0.032 |
stretched | -0.031 |
Tweet | -0.030 |
Tw | -0.029 |
eight | -0.029 |
Token | of |
Token ID | 286 |
Feature activation | +6.30e-01 |
Pos logit contributions | |
package | +0.083 |
meal | +0.082 |
idents | +0.081 |
tray | +0.081 |
batch | +0.081 |
Neg logit contributions | |
ila | -0.035 |
La | -0.034 |
stretched | -0.033 |
Tw | -0.032 |
Tweet | -0.032 |
Token | y |
Token ID | 331 |
Feature activation | +5.39e+00 |
Pos logit contributions | |
idents | +0.015 |
package | +0.015 |
meal | +0.015 |
tray | +0.015 |
batch | +0.014 |
Neg logit contributions | |
ila | -0.021 |
stretched | -0.021 |
Tw | -0.020 |
eight | -0.020 |
Tweet | -0.020 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.68e+00 |
Pos logit contributions | |
stret | +0.164 |
tray | +0.160 |
package | +0.159 |
idents | +0.155 |
envelope | +0.141 |
Neg logit contributions | |
stretched | -0.151 |
ila | -0.138 |
eight | -0.133 |
Tw | -0.123 |
Tweet | -0.120 |
Token | treats |
Token ID | 18432 |
Feature activation | +2.52e+00 |
Pos logit contributions | |
idents | +0.052 |
doorstep | +0.043 |
tray | +0.042 |
package | +0.042 |
aman | +0.042 |
Neg logit contributions | |
ila | -0.098 |
stretched | -0.095 |
eight | -0.093 |
Tw | -0.092 |
La | -0.089 |
Token | . |
Token ID | 13 |
Feature activation | +4.07e-01 |
Pos logit contributions | |
idents | +0.095 |
package | +0.092 |
batch | +0.090 |
tray | +0.089 |
stret | +0.089 |
Neg logit contributions | |
ila | -0.050 |
stretched | -0.048 |
Tw | -0.045 |
eight | -0.042 |
Tweet | -0.040 |
Token | They |
Token ID | 1119 |
Feature activation | +1.88e+00 |
Pos logit contributions | |
batch | +0.012 |
refrigerator | +0.012 |
meal | +0.012 |
idents | +0.012 |
package | +0.012 |
Neg logit contributions | |
stretched | -0.011 |
ila | -0.011 |
La | -0.010 |
Tweet | -0.010 |
eight | -0.009 |
Token | put |
Token ID | 1234 |
Feature activation | -3.84e-01 |
Pos logit contributions | |
package | +0.052 |
meal | +0.052 |
tray | +0.052 |
idents | +0.051 |
batch | +0.051 |
Neg logit contributions | |
ila | -0.056 |
stretched | -0.055 |
Tweet | -0.052 |
Tw | -0.052 |
La | -0.052 |
Token | Sandy |
Token ID | 17581 |
Feature activation | +6.08e-01 |
Pos logit contributions | |
idents | +0.042 |
meal | +0.040 |
tray | +0.040 |
batch | +0.040 |
package | +0.040 |
Neg logit contributions | |
stretched | -0.028 |
Tweet | -0.027 |
Tw | -0.026 |
La | -0.026 |
eight | -0.026 |
Token | saw |
Token ID | 2497 |
Feature activation | +2.53e-01 |
Pos logit contributions | |
idents | +0.020 |
package | +0.020 |
meal | +0.020 |
tray | +0.020 |
batch | +0.019 |
Neg logit contributions | |
ila | -0.015 |
stretched | -0.014 |
Tw | -0.014 |
eight | -0.014 |
La | -0.013 |
Token | a |
Token ID | 257 |
Feature activation | +2.93e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
meal | +0.008 |
batch | +0.008 |
tray | +0.008 |
Neg logit contributions | |
stretched | -0.006 |
ila | -0.006 |
Tw | -0.006 |
eight | -0.006 |
Tweet | -0.006 |
Token | bowl |
Token ID | 9396 |
Feature activation | +2.60e+00 |
Pos logit contributions | |
idents | +0.011 |
keeping | +0.010 |
dinners | +0.010 |
ics | +0.010 |
meal | +0.010 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
eight | -0.006 |
Tw | -0.005 |
Tweet | -0.005 |
Token | of |
Token ID | 286 |
Feature activation | +3.13e-01 |
Pos logit contributions | |
idents | +0.139 |
stret | +0.129 |
package | +0.126 |
isions | +0.125 |
meal | +0.124 |
Neg logit contributions | |
stretched | -0.023 |
ila | -0.022 |
eight | -0.013 |
Tw | -0.011 |
Tweet | -0.010 |
Token | y |
Token ID | 331 |
Feature activation | +5.38e+00 |
Pos logit contributions | |
idents | +0.008 |
doorstep | +0.008 |
stret | +0.007 |
package | +0.007 |
keeping | +0.007 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
eight | -0.009 |
Some | -0.009 |
Tweet | -0.009 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.94e+00 |
Pos logit contributions | |
stret | +0.213 |
package | +0.186 |
idents | +0.184 |
tray | +0.182 |
dining | +0.176 |
Neg logit contributions | |
stretched | -0.117 |
eight | -0.097 |
ila | -0.095 |
ib | -0.092 |
ubby | -0.091 |
Token | le |
Token ID | 443 |
Feature activation | +1.29e+00 |
Pos logit contributions | |
doorstep | +0.088 |
stret | +0.087 |
aman | +0.086 |
idents | +0.084 |
keeping | +0.080 |
Neg logit contributions | |
he | -0.079 |
blue | -0.079 |
of | -0.075 |
yellow | -0.073 |
green | -0.073 |
Token | mons |
Token ID | 11567 |
Feature activation | +2.09e+00 |
Pos logit contributions | |
package | +0.037 |
idents | +0.036 |
meal | +0.036 |
tray | +0.034 |
stret | +0.034 |
Neg logit contributions | |
ila | -0.039 |
stretched | -0.036 |
Tweet | -0.035 |
eight | -0.035 |
Tw | -0.034 |
Token | on |
Token ID | 319 |
Feature activation | +9.85e-01 |
Pos logit contributions | |
idents | +0.090 |
package | +0.086 |
stret | +0.083 |
meal | +0.082 |
doorstep | +0.081 |
Neg logit contributions | |
ila | -0.035 |
stretched | -0.033 |
eight | -0.027 |
Tweet | -0.026 |
Tw | -0.026 |
Token | the |
Token ID | 262 |
Feature activation | +1.41e+00 |
Pos logit contributions | |
package | +0.040 |
idents | +0.040 |
meal | +0.039 |
tray | +0.039 |
batch | +0.039 |
Neg logit contributions | |
stretched | -0.018 |
ila | -0.017 |
Tw | -0.016 |
Tweet | -0.015 |
eight | -0.015 |
Token | be |
Token ID | 307 |
Feature activation | -2.63e+00 |
Pos logit contributions | |
ila | +0.026 |
stretched | +0.026 |
Tw | +0.024 |
Tweet | +0.024 |
eight | +0.024 |
Neg logit contributions | |
package | -0.037 |
idents | -0.036 |
meal | -0.036 |
tray | -0.036 |
batch | -0.035 |
Token | ep |
Token ID | 538 |
Feature activation | -2.93e+00 |
Pos logit contributions | |
ila | +0.072 |
stretched | +0.072 |
Tw | +0.071 |
Tweet | +0.067 |
La | +0.067 |
Neg logit contributions | |
package | -0.076 |
meal | -0.075 |
tray | -0.075 |
batch | -0.072 |
stret | -0.071 |
Token | and |
Token ID | 290 |
Feature activation | -1.29e+00 |
Pos logit contributions | |
ila | +0.087 |
Tweet | +0.084 |
stretched | +0.084 |
La | +0.084 |
Tw | +0.080 |
Neg logit contributions | |
package | -0.082 |
idents | -0.078 |
meal | -0.078 |
tray | -0.077 |
stret | -0.075 |
Token | a |
Token ID | 257 |
Feature activation | -1.37e+00 |
Pos logit contributions | |
ila | +0.034 |
stretched | +0.033 |
Tweet | +0.033 |
La | +0.031 |
Tw | +0.031 |
Neg logit contributions | |
package | -0.042 |
meal | -0.041 |
tray | -0.040 |
idents | -0.040 |
dinner | -0.039 |
Token | hon |
Token ID | 3032 |
Feature activation | -2.15e+00 |
Pos logit contributions | |
ila | +0.039 |
stretched | +0.037 |
Tweet | +0.034 |
Tw | +0.034 |
La | +0.034 |
Neg logit contributions | |
package | -0.044 |
tray | -0.043 |
meal | -0.043 |
batch | -0.041 |
idents | -0.041 |
Token | k |
Token ID | 74 |
Feature activation | -6.21e+00 |
Pos logit contributions | |
ila | +0.072 |
stretched | +0.069 |
La | +0.067 |
Tweet | +0.067 |
Tw | +0.066 |
Neg logit contributions | |
idents | -0.053 |
package | -0.051 |
stret | -0.049 |
tray | -0.048 |
batch | -0.047 |
Token | , |
Token ID | 11 |
Feature activation | -5.17e-01 |
Pos logit contributions | |
ila | +0.182 |
Tweet | +0.180 |
La | +0.179 |
stretched | +0.170 |
Tw | +0.169 |
Neg logit contributions | |
package | -0.168 |
tray | -0.160 |
stret | -0.159 |
meal | -0.154 |
idents | -0.152 |
Token | the |
Token ID | 262 |
Feature activation | -1.03e+00 |
Pos logit contributions | |
ila | +0.014 |
Tweet | +0.014 |
stretched | +0.014 |
La | +0.013 |
Tw | +0.012 |
Neg logit contributions | |
package | -0.016 |
idents | -0.015 |
meal | -0.015 |
tray | -0.015 |
batch | -0.015 |
Token | gentle |
Token ID | 10296 |
Feature activation | -1.40e+00 |
Pos logit contributions | |
ila | +0.031 |
Tw | +0.030 |
La | +0.030 |
stretched | +0.029 |
Tweet | +0.028 |
Neg logit contributions | |
package | -0.033 |
meal | -0.031 |
envelope | -0.030 |
tray | -0.030 |
stret | -0.030 |
Token | van |
Token ID | 5719 |
Feature activation | +5.42e-01 |
Pos logit contributions | |
ila | +0.043 |
stretched | +0.041 |
Tw | +0.041 |
La | +0.040 |
eight | +0.039 |
Neg logit contributions | |
package | -0.041 |
idents | -0.039 |
meal | -0.039 |
tray | -0.038 |
batch | -0.037 |
Token | drove |
Token ID | 10357 |
Feature activation | -5.80e-01 |
Pos logit contributions | |
package | +0.016 |
idents | +0.015 |
meal | +0.015 |
tray | +0.015 |
doorstep | +0.014 |
Neg logit contributions | |
ila | -0.016 |
La | -0.015 |
stretched | -0.015 |
eight | -0.014 |
Tweet | -0.014 |
Token | would |
Token ID | 561 |
Feature activation | -1.44e+00 |
Pos logit contributions | |
stretched | +0.015 |
La | +0.015 |
ila | +0.015 |
eight | +0.014 |
Tw | +0.013 |
Neg logit contributions | |
package | -0.018 |
idents | -0.018 |
mess | -0.017 |
meal | -0.016 |
batch | -0.016 |
Token | have |
Token ID | 423 |
Feature activation | +4.29e-01 |
Pos logit contributions | |
ila | +0.044 |
stretched | +0.042 |
La | +0.040 |
eight | +0.038 |
Tw | +0.038 |
Neg logit contributions | |
idents | -0.041 |
package | -0.041 |
mess | -0.040 |
meal | -0.039 |
stret | -0.039 |
Token | to |
Token ID | 284 |
Feature activation | +5.35e-01 |
Pos logit contributions | |
package | +0.014 |
idents | +0.013 |
meal | +0.013 |
dinner | +0.013 |
batch | +0.013 |
Neg logit contributions | |
ila | -0.012 |
stretched | -0.012 |
La | -0.011 |
Tweet | -0.011 |
Tw | -0.011 |
Token | wear |
Token ID | 5806 |
Feature activation | -2.36e+00 |
Pos logit contributions | |
package | +0.016 |
idents | +0.016 |
meal | +0.015 |
tray | +0.015 |
stret | +0.015 |
Neg logit contributions | |
stretched | -0.016 |
ila | -0.015 |
Tw | -0.014 |
eight | -0.014 |
La | -0.014 |
Token | a |
Token ID | 257 |
Feature activation | -1.11e+00 |
Pos logit contributions | |
stretched | +0.073 |
ila | +0.072 |
Tweet | +0.068 |
eight | +0.068 |
Bo | +0.065 |
Neg logit contributions | |
package | -0.072 |
idents | -0.069 |
dinner | -0.066 |
tray | -0.064 |
stret | -0.064 |
Token | cast |
Token ID | 3350 |
Feature activation | -6.20e+00 |
Pos logit contributions | |
ila | +0.031 |
stretched | +0.031 |
Tw | +0.029 |
eight | +0.029 |
Tweet | +0.029 |
Neg logit contributions | |
package | -0.034 |
idents | -0.033 |
meal | -0.033 |
tray | -0.033 |
batch | -0.032 |
Token | for |
Token ID | 329 |
Feature activation | +1.25e+00 |
Pos logit contributions | |
stretched | +0.182 |
Tweet | +0.180 |
eight | +0.177 |
ila | +0.176 |
La | +0.173 |
Neg logit contributions | |
package | -0.175 |
tray | -0.169 |
stret | -0.167 |
meal | -0.161 |
envelope | -0.160 |
Token | a |
Token ID | 257 |
Feature activation | -4.44e-01 |
Pos logit contributions | |
dinner | +0.040 |
dessert | +0.037 |
meal | +0.037 |
cake | +0.037 |
package | +0.037 |
Neg logit contributions | |
stretched | -0.037 |
ila | -0.035 |
La | -0.033 |
Tweet | -0.033 |
Bo | -0.031 |
Token | long |
Token ID | 890 |
Feature activation | -6.71e-01 |
Pos logit contributions | |
ila | +0.012 |
stretched | +0.011 |
Tw | +0.010 |
La | +0.010 |
eight | +0.010 |
Neg logit contributions | |
package | -0.015 |
meal | -0.015 |
stret | -0.015 |
tray | -0.015 |
batch | -0.014 |
Token | time |
Token ID | 640 |
Feature activation | -1.43e+00 |
Pos logit contributions | |
ila | +0.020 |
stretched | +0.020 |
Tw | +0.018 |
La | +0.018 |
eight | +0.018 |
Neg logit contributions | |
package | -0.020 |
meal | -0.019 |
tray | -0.019 |
stret | -0.019 |
batch | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | +2.21e-01 |
Pos logit contributions | |
ila | +0.042 |
Tweet | +0.041 |
stretched | +0.040 |
La | +0.039 |
eight | +0.038 |
Neg logit contributions | |
package | -0.041 |
meal | -0.039 |
tray | -0.039 |
idents | -0.039 |
stret | -0.038 |
Token | . |
Token ID | 13 |
Feature activation | -1.49e-01 |
Pos logit contributions | |
ila | +0.021 |
stretched | +0.021 |
eight | +0.019 |
Tw | +0.019 |
Tweet | +0.019 |
Neg logit contributions | |
idents | -0.030 |
package | -0.030 |
meal | -0.030 |
tray | -0.030 |
batch | -0.029 |
Token | She |
Token ID | 1375 |
Feature activation | -1.54e+00 |
Pos logit contributions | |
stretched | +0.004 |
Tweet | +0.004 |
racer | +0.004 |
fastest | +0.004 |
brav | +0.004 |
Neg logit contributions | |
refrigerator | -0.004 |
package | -0.004 |
idents | -0.004 |
tray | -0.004 |
batch | -0.003 |
Token | has |
Token ID | 468 |
Feature activation | +3.14e-02 |
Pos logit contributions | |
stretched | +0.049 |
Tweet | +0.048 |
ila | +0.044 |
Tw | +0.042 |
eight | +0.042 |
Neg logit contributions | |
mess | -0.042 |
idents | -0.041 |
package | -0.040 |
refrigerator | -0.038 |
tray | -0.037 |
Token | a |
Token ID | 257 |
Feature activation | -1.15e+00 |
Pos logit contributions | |
package | +0.001 |
cake | +0.001 |
envelope | +0.001 |
dinner | +0.001 |
meal | +0.001 |
Neg logit contributions | |
stretched | -0.001 |
Tweet | -0.001 |
ila | -0.001 |
Bo | -0.001 |
Tw | -0.001 |
Token | big |
Token ID | 1263 |
Feature activation | +1.22e+00 |
Pos logit contributions | |
ila | +0.030 |
stretched | +0.030 |
Tw | +0.028 |
Tweet | +0.028 |
eight | +0.027 |
Neg logit contributions | |
package | -0.038 |
tray | -0.037 |
meal | -0.037 |
idents | -0.036 |
batch | -0.036 |
Token | cast |
Token ID | 3350 |
Feature activation | -6.06e+00 |
Pos logit contributions | |
idents | +0.041 |
package | +0.038 |
meal | +0.038 |
batch | +0.038 |
tray | +0.037 |
Neg logit contributions | |
ila | -0.030 |
stretched | -0.030 |
eight | -0.028 |
Tw | -0.028 |
Tweet | -0.027 |
Token | on |
Token ID | 319 |
Feature activation | -5.39e-01 |
Pos logit contributions | |
Tweet | +0.178 |
La | +0.173 |
stretched | +0.172 |
Tw | +0.170 |
eight | +0.169 |
Neg logit contributions | |
package | -0.169 |
tray | -0.164 |
envelope | -0.161 |
stret | -0.156 |
meal | -0.153 |
Token | her |
Token ID | 607 |
Feature activation | -1.06e+00 |
Pos logit contributions | |
stretched | +0.016 |
ila | +0.016 |
Tweet | +0.016 |
Bo | +0.015 |
eight | +0.014 |
Neg logit contributions | |
tray | -0.015 |
package | -0.015 |
meal | -0.015 |
dinner | -0.015 |
idents | -0.014 |
Token | leg |
Token ID | 1232 |
Feature activation | -1.83e+00 |
Pos logit contributions | |
ila | +0.036 |
stretched | +0.036 |
Tweet | +0.034 |
Tw | +0.033 |
La | +0.032 |
Neg logit contributions | |
package | -0.027 |
tray | -0.027 |
meal | -0.027 |
idents | -0.026 |
stret | -0.026 |
Token | . |
Token ID | 13 |
Feature activation | -4.19e-01 |
Pos logit contributions | |
ila | +0.053 |
stretched | +0.053 |
eight | +0.051 |
Tweet | +0.050 |
La | +0.050 |
Neg logit contributions | |
package | -0.053 |
tray | -0.051 |
meal | -0.050 |
idents | -0.049 |
batch | -0.049 |
Token | She |
Token ID | 1375 |
Feature activation | -1.68e+00 |
Pos logit contributions | |
stretched | +0.013 |
Tweet | +0.012 |
fastest | +0.012 |
racer | +0.012 |
brav | +0.012 |
Neg logit contributions | |
refrigerator | -0.011 |
package | -0.010 |
idents | -0.009 |
tray | -0.009 |
batch | -0.009 |
Token | she |
Token ID | 673 |
Feature activation | -4.71e-01 |
Pos logit contributions | |
package | +0.020 |
mess | +0.019 |
idents | +0.019 |
meal | +0.018 |
dinner | +0.018 |
Neg logit contributions | |
La | -0.021 |
stretched | -0.019 |
Bo | -0.018 |
ila | -0.018 |
Tw | -0.018 |
Token | needed |
Token ID | 2622 |
Feature activation | -3.74e-01 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.012 |
La | +0.012 |
Tw | +0.011 |
eight | +0.011 |
Neg logit contributions | |
package | -0.016 |
idents | -0.016 |
mess | -0.015 |
meal | -0.014 |
batch | -0.014 |
Token | to |
Token ID | 284 |
Feature activation | +6.01e-01 |
Pos logit contributions | |
stretched | +0.012 |
ila | +0.012 |
La | +0.011 |
Tw | +0.011 |
Bo | +0.011 |
Neg logit contributions | |
package | -0.011 |
dinner | -0.011 |
cake | -0.010 |
meal | -0.010 |
dessert | -0.010 |
Token | wear |
Token ID | 5806 |
Feature activation | -2.45e+00 |
Pos logit contributions | |
package | +0.017 |
idents | +0.017 |
meal | +0.017 |
stret | +0.016 |
mess | +0.016 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.019 |
La | -0.017 |
eight | -0.017 |
Tw | -0.017 |
Token | a |
Token ID | 257 |
Feature activation | -9.58e-01 |
Pos logit contributions | |
stretched | +0.078 |
ila | +0.077 |
eight | +0.072 |
Tweet | +0.072 |
Bo | +0.070 |
Neg logit contributions | |
package | -0.073 |
idents | -0.070 |
dinner | -0.067 |
tray | -0.065 |
stret | -0.065 |
Token | cast |
Token ID | 3350 |
Feature activation | -6.04e+00 |
Pos logit contributions | |
ila | +0.027 |
stretched | +0.027 |
Tw | +0.026 |
eight | +0.025 |
Tweet | +0.025 |
Neg logit contributions | |
package | -0.029 |
idents | -0.028 |
meal | -0.028 |
tray | -0.028 |
batch | -0.027 |
Token | . |
Token ID | 13 |
Feature activation | +3.75e-01 |
Pos logit contributions | |
stretched | +0.181 |
Tweet | +0.177 |
eight | +0.175 |
ila | +0.173 |
La | +0.172 |
Neg logit contributions | |
package | -0.170 |
tray | -0.162 |
stret | -0.161 |
meal | -0.156 |
envelope | -0.156 |
Token | Lily |
Token ID | 20037 |
Feature activation | -4.36e-01 |
Pos logit contributions | |
idents | +0.015 |
batch | +0.014 |
package | +0.014 |
meal | +0.014 |
tray | +0.014 |
Neg logit contributions | |
stretched | -0.008 |
ila | -0.008 |
La | -0.007 |
eight | -0.007 |
Tweet | -0.007 |
Token | felt |
Token ID | 2936 |
Feature activation | -1.48e+00 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.009 |
Tw | +0.009 |
eight | +0.009 |
Tweet | +0.009 |
Neg logit contributions | |
idents | -0.016 |
package | -0.016 |
meal | -0.016 |
tray | -0.015 |
batch | -0.015 |
Token | sad |
Token ID | 6507 |
Feature activation | +3.10e-01 |
Pos logit contributions | |
ila | +0.044 |
stretched | +0.043 |
La | +0.042 |
eight | +0.041 |
Tw | +0.041 |
Neg logit contributions | |
package | -0.042 |
idents | -0.042 |
meal | -0.041 |
stret | -0.039 |
tray | -0.039 |
Token | and |
Token ID | 290 |
Feature activation | +5.78e-01 |
Pos logit contributions | |
idents | +0.009 |
package | +0.009 |
meal | +0.008 |
batch | +0.008 |
tray | +0.008 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
La | -0.009 |
Tweet | -0.009 |
eight | -0.009 |
Token | doctor |
Token ID | 6253 |
Feature activation | -1.51e+00 |
Pos logit contributions | |
stretched | +0.019 |
ila | +0.018 |
Tw | +0.017 |
La | +0.017 |
eight | +0.016 |
Neg logit contributions | |
package | -0.024 |
tray | -0.022 |
meal | -0.022 |
mess | -0.022 |
refrigerator | -0.022 |
Token | had |
Token ID | 550 |
Feature activation | +4.20e-01 |
Pos logit contributions | |
ila | +0.040 |
stretched | +0.040 |
Tw | +0.037 |
eight | +0.037 |
Tweet | +0.037 |
Neg logit contributions | |
idents | -0.048 |
package | -0.047 |
meal | -0.046 |
tray | -0.046 |
batch | -0.046 |
Token | to |
Token ID | 284 |
Feature activation | +5.27e-01 |
Pos logit contributions | |
idents | +0.016 |
package | +0.015 |
tray | +0.015 |
meal | +0.015 |
batch | +0.015 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
Tw | -0.008 |
eight | -0.008 |
Tweet | -0.008 |
Token | put |
Token ID | 1234 |
Feature activation | -3.53e-01 |
Pos logit contributions | |
idents | +0.017 |
package | +0.017 |
meal | +0.016 |
tray | +0.016 |
batch | +0.016 |
Neg logit contributions | |
ila | -0.014 |
stretched | -0.013 |
Tw | -0.013 |
eight | -0.013 |
Tweet | -0.012 |
Token | a |
Token ID | 257 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
ila | +0.008 |
stretched | +0.008 |
Tw | +0.008 |
eight | +0.008 |
Tweet | +0.007 |
Neg logit contributions | |
idents | -0.012 |
package | -0.012 |
meal | -0.012 |
tray | -0.012 |
batch | -0.012 |
Token | cast |
Token ID | 3350 |
Feature activation | -6.00e+00 |
Pos logit contributions | |
stretched | +0.027 |
ila | +0.026 |
eight | +0.025 |
Tweet | +0.025 |
Tw | +0.024 |
Neg logit contributions | |
idents | -0.041 |
meal | -0.037 |
batch | -0.037 |
dinners | -0.037 |
ics | -0.037 |
Token | on |
Token ID | 319 |
Feature activation | -4.96e-01 |
Pos logit contributions | |
stretched | +0.163 |
ila | +0.160 |
Tweet | +0.158 |
eight | +0.155 |
Tw | +0.155 |
Neg logit contributions | |
package | -0.186 |
tray | -0.179 |
stret | -0.176 |
meal | -0.174 |
idents | -0.171 |
Token | her |
Token ID | 607 |
Feature activation | -1.19e+00 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.013 |
Tweet | +0.012 |
La | +0.012 |
eight | +0.011 |
Neg logit contributions | |
package | -0.017 |
idents | -0.016 |
tray | -0.016 |
meal | -0.016 |
batch | -0.016 |
Token | foot |
Token ID | 2366 |
Feature activation | -1.70e+00 |
Pos logit contributions | |
ila | +0.038 |
stretched | +0.038 |
Tw | +0.036 |
La | +0.036 |
Tweet | +0.036 |
Neg logit contributions | |
package | -0.033 |
tray | -0.032 |
idents | -0.031 |
meal | -0.031 |
stret | -0.030 |
Token | . |
Token ID | 13 |
Feature activation | +1.98e-01 |
Pos logit contributions | |
ila | +0.046 |
stretched | +0.045 |
eight | +0.044 |
Tw | +0.043 |
La | +0.043 |
Neg logit contributions | |
package | -0.054 |
idents | -0.052 |
tray | -0.051 |
batch | -0.050 |
meal | -0.050 |
Token |
|
Token ID | 198 |
Feature activation | -6.89e-02 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
meal | +0.008 |
tray | +0.008 |
batch | +0.008 |
Neg logit contributions | |
ila | -0.003 |
stretched | -0.003 |
Tw | -0.003 |
eight | -0.003 |
Tweet | -0.003 |
Token | she |
Token ID | 673 |
Feature activation | -6.70e-01 |
Pos logit contributions | |
idents | +0.018 |
mess | +0.017 |
package | +0.017 |
dinner | +0.017 |
cake | +0.017 |
Neg logit contributions | |
stretched | -0.015 |
La | -0.015 |
ila | -0.014 |
eight | -0.014 |
Tweet | -0.014 |
Token | had |
Token ID | 550 |
Feature activation | +2.32e-01 |
Pos logit contributions | |
stretched | +0.017 |
ila | +0.016 |
La | +0.016 |
eight | +0.015 |
Tw | +0.015 |
Neg logit contributions | |
package | -0.023 |
idents | -0.022 |
meal | -0.022 |
tray | -0.021 |
batch | -0.021 |
Token | to |
Token ID | 284 |
Feature activation | +4.09e-01 |
Pos logit contributions | |
package | +0.007 |
idents | +0.007 |
meal | +0.006 |
mess | +0.006 |
dinner | +0.006 |
Neg logit contributions | |
ila | -0.007 |
stretched | -0.007 |
Tweet | -0.007 |
La | -0.006 |
Tw | -0.006 |
Token | wear |
Token ID | 5806 |
Feature activation | -2.35e+00 |
Pos logit contributions | |
package | +0.013 |
idents | +0.012 |
meal | +0.012 |
tray | +0.012 |
stret | +0.012 |
Neg logit contributions | |
stretched | -0.012 |
ila | -0.011 |
Tw | -0.011 |
eight | -0.011 |
La | -0.010 |
Token | a |
Token ID | 257 |
Feature activation | -7.31e-01 |
Pos logit contributions | |
ila | +0.075 |
stretched | +0.073 |
Tweet | +0.071 |
eight | +0.068 |
Bo | +0.066 |
Neg logit contributions | |
package | -0.069 |
idents | -0.066 |
dinner | -0.063 |
tray | -0.061 |
cake | -0.060 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.95e+00 |
Pos logit contributions | |
ila | +0.019 |
stretched | +0.019 |
Tw | +0.018 |
eight | +0.018 |
Tweet | +0.017 |
Neg logit contributions | |
package | -0.024 |
idents | -0.023 |
tray | -0.023 |
meal | -0.023 |
batch | -0.023 |
Token | so |
Token ID | 523 |
Feature activation | +3.73e-01 |
Pos logit contributions | |
stretched | +0.173 |
Tweet | +0.173 |
ila | +0.171 |
eight | +0.169 |
Bo | +0.163 |
Neg logit contributions | |
package | -0.174 |
tray | -0.167 |
stret | -0.160 |
meal | -0.159 |
envelope | -0.158 |
Token | it |
Token ID | 340 |
Feature activation | +1.25e+00 |
Pos logit contributions | |
dinner | +0.009 |
cake | +0.009 |
dessert | +0.008 |
package | +0.008 |
idents | +0.008 |
Neg logit contributions | |
La | -0.012 |
Ru | -0.012 |
female | -0.012 |
Rib | -0.012 |
Bo | -0.012 |
Token | would |
Token ID | 561 |
Feature activation | -7.48e-01 |
Pos logit contributions | |
idents | +0.034 |
package | +0.033 |
mess | +0.031 |
tray | +0.030 |
meal | +0.030 |
Neg logit contributions | |
ila | -0.039 |
stretched | -0.039 |
eight | -0.038 |
La | -0.038 |
Tweet | -0.036 |
Token | heal |
Token ID | 12035 |
Feature activation | -1.18e+00 |
Pos logit contributions | |
ila | +0.024 |
La | +0.023 |
stretched | +0.023 |
Nice | +0.021 |
eight | +0.021 |
Neg logit contributions | |
mess | -0.020 |
idents | -0.020 |
package | -0.019 |
meal | -0.019 |
tray | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | +1.68e-01 |
Pos logit contributions | |
stretched | +0.034 |
ila | +0.033 |
eight | +0.032 |
Tweet | +0.030 |
La | +0.030 |
Neg logit contributions | |
idents | -0.036 |
package | -0.035 |
meal | -0.034 |
tray | -0.033 |
dinner | -0.033 |
Token | and |
Token ID | 290 |
Feature activation | -6.77e-01 |
Pos logit contributions | |
package | +0.046 |
idents | +0.046 |
batch | +0.045 |
tray | +0.045 |
meal | +0.044 |
Neg logit contributions | |
stretched | -0.026 |
ila | -0.025 |
Tw | -0.024 |
eight | -0.023 |
Tweet | -0.022 |
Token | makes |
Token ID | 1838 |
Feature activation | -6.38e-01 |
Pos logit contributions | |
ila | +0.020 |
stretched | +0.020 |
Tweet | +0.018 |
La | +0.018 |
Tw | +0.017 |
Neg logit contributions | |
idents | -0.020 |
tray | -0.019 |
meal | -0.018 |
batch | -0.018 |
mess | -0.018 |
Token | a |
Token ID | 257 |
Feature activation | -1.13e+00 |
Pos logit contributions | |
ila | +0.014 |
stretched | +0.013 |
Tw | +0.013 |
eight | +0.013 |
Tweet | +0.012 |
Neg logit contributions | |
idents | -0.024 |
package | -0.024 |
tray | -0.023 |
meal | -0.023 |
batch | -0.023 |
Token | loud |
Token ID | 7812 |
Feature activation | -1.55e+00 |
Pos logit contributions | |
stretched | +0.023 |
eight | +0.022 |
ila | +0.022 |
Tweet | +0.022 |
Tw | +0.021 |
Neg logit contributions | |
idents | -0.042 |
package | -0.042 |
keeping | -0.041 |
dinners | -0.040 |
isions | -0.040 |
Token | hon |
Token ID | 3032 |
Feature activation | -1.83e+00 |
Pos logit contributions | |
stretched | +0.046 |
ila | +0.046 |
Tw | +0.045 |
Tweet | +0.045 |
eight | +0.044 |
Neg logit contributions | |
package | -0.046 |
idents | -0.042 |
tray | -0.042 |
doorstep | -0.042 |
dining | -0.042 |
Token | k |
Token ID | 74 |
Feature activation | -5.95e+00 |
Pos logit contributions | |
ila | +0.049 |
stretched | +0.048 |
Tw | +0.046 |
eight | +0.045 |
Tweet | +0.045 |
Neg logit contributions | |
package | -0.058 |
idents | -0.058 |
meal | -0.057 |
tray | -0.056 |
batch | -0.056 |
Token | . |
Token ID | 13 |
Feature activation | -1.74e-01 |
Pos logit contributions | |
ila | +0.154 |
stretched | +0.152 |
Tweet | +0.148 |
La | +0.144 |
Tw | +0.143 |
Neg logit contributions | |
idents | -0.177 |
tray | -0.176 |
package | -0.176 |
meal | -0.175 |
batch | -0.171 |
Token | Tom |
Token ID | 4186 |
Feature activation | +3.24e-01 |
Pos logit contributions | |
stretched | +0.005 |
Mart | +0.005 |
Chip | +0.005 |
La | +0.005 |
known | +0.005 |
Neg logit contributions | |
tray | -0.005 |
idents | -0.004 |
refrigerator | -0.004 |
batch | -0.004 |
grill | -0.004 |
Token | and |
Token ID | 290 |
Feature activation | -1.31e+00 |
Pos logit contributions | |
idents | +0.010 |
tray | +0.009 |
package | +0.009 |
meal | +0.009 |
batch | +0.009 |
Neg logit contributions | |
stretched | -0.009 |
ila | -0.009 |
Tweet | -0.009 |
La | -0.008 |
Tw | -0.008 |
Token | Mia |
Token ID | 32189 |
Feature activation | -3.36e-01 |
Pos logit contributions | |
stretched | +0.037 |
La | +0.033 |
ila | +0.031 |
Bo | +0.031 |
Tweet | +0.031 |
Neg logit contributions | |
tray | -0.042 |
batch | -0.040 |
cake | -0.040 |
meal | -0.039 |
package | -0.039 |
Token | stop |
Token ID | 2245 |
Feature activation | +1.05e+00 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.010 |
Tw | +0.010 |
Tweet | +0.010 |
eight | +0.010 |
Neg logit contributions | |
idents | -0.009 |
package | -0.009 |
tray | -0.009 |
meal | -0.009 |
batch | -0.009 |
Token | I |
Token ID | 40 |
Feature activation | -4.10e-01 |
Pos logit contributions | |
stretched | +0.017 |
ila | +0.017 |
eight | +0.017 |
Tw | +0.016 |
La | +0.015 |
Neg logit contributions | |
idents | -0.032 |
package | -0.031 |
batch | -0.030 |
meal | -0.029 |
advertisement | -0.029 |
Token | wish |
Token ID | 4601 |
Feature activation | -1.40e+00 |
Pos logit contributions | |
stretched | +0.011 |
ila | +0.011 |
eight | +0.011 |
Tw | +0.010 |
La | +0.010 |
Neg logit contributions | |
idents | -0.013 |
package | -0.013 |
meal | -0.012 |
batch | -0.012 |
tray | -0.012 |
Token | I |
Token ID | 314 |
Feature activation | -9.71e-01 |
Pos logit contributions | |
stretched | +0.035 |
La | +0.034 |
ila | +0.033 |
Tweet | +0.033 |
Bo | +0.031 |
Neg logit contributions | |
idents | -0.049 |
meal | -0.047 |
cake | -0.047 |
package | -0.047 |
dinner | -0.046 |
Token | had |
Token ID | 550 |
Feature activation | -8.82e-01 |
Pos logit contributions | |
stretched | +0.023 |
ila | +0.021 |
La | +0.020 |
eight | +0.019 |
Tw | +0.018 |
Neg logit contributions | |
package | -0.036 |
meal | -0.034 |
idents | -0.034 |
tray | -0.032 |
dinner | -0.031 |
Token | a |
Token ID | 257 |
Feature activation | -6.45e-01 |
Pos logit contributions | |
ila | +0.023 |
stretched | +0.023 |
Tweet | +0.021 |
Tw | +0.020 |
La | +0.020 |
Neg logit contributions | |
package | -0.029 |
meal | -0.028 |
idents | -0.028 |
tray | -0.027 |
dinner | -0.027 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.94e+00 |
Pos logit contributions | |
ila | +0.020 |
stretched | +0.020 |
Tweet | +0.018 |
Tw | +0.018 |
La | +0.017 |
Neg logit contributions | |
package | -0.020 |
meal | -0.018 |
tray | -0.018 |
idents | -0.018 |
stret | -0.018 |
Token | too |
Token ID | 1165 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
stretched | +0.178 |
ila | +0.177 |
Tweet | +0.171 |
La | +0.169 |
Tw | +0.167 |
Neg logit contributions | |
package | -0.169 |
idents | -0.162 |
tray | -0.161 |
meal | -0.161 |
stret | -0.155 |
Token | ," |
Token ID | 553 |
Feature activation | -1.04e-02 |
Pos logit contributions | |
ila | +0.041 |
stretched | +0.040 |
Tweet | +0.040 |
La | +0.039 |
Tw | +0.038 |
Neg logit contributions | |
package | -0.031 |
meal | -0.029 |
idents | -0.028 |
mess | -0.028 |
envelope | -0.028 |
Token | Sally |
Token ID | 25737 |
Feature activation | -5.01e-01 |
Pos logit contributions | |
stretched | +0.000 |
ila | +0.000 |
eight | +0.000 |
La | +0.000 |
Tw | +0.000 |
Neg logit contributions | |
idents | -0.000 |
package | -0.000 |
meal | -0.000 |
batch | -0.000 |
tray | -0.000 |
Token | said |
Token ID | 531 |
Feature activation | -6.78e-01 |
Pos logit contributions | |
ila | +0.014 |
stretched | +0.014 |
Tw | +0.013 |
eight | +0.013 |
Tweet | +0.013 |
Neg logit contributions | |
idents | -0.015 |
package | -0.015 |
meal | -0.015 |
tray | -0.015 |
batch | -0.015 |
Token | . |
Token ID | 13 |
Feature activation | +4.63e-01 |
Pos logit contributions | |
stretched | +0.013 |
Tw | +0.012 |
ila | +0.012 |
eight | +0.011 |
Tweet | +0.011 |
Neg logit contributions | |
package | -0.027 |
idents | -0.027 |
meal | -0.026 |
tray | -0.026 |
batch | -0.026 |
Token | he |
Token ID | 339 |
Feature activation | +5.82e-01 |
Pos logit contributions | |
idents | +0.024 |
package | +0.024 |
mess | +0.024 |
meal | +0.023 |
batch | +0.022 |
Neg logit contributions | |
La | -0.022 |
ila | -0.021 |
stretched | -0.021 |
Tweet | -0.020 |
Tw | -0.020 |
Token | had |
Token ID | 550 |
Feature activation | +9.11e-02 |
Pos logit contributions | |
idents | +0.017 |
package | +0.017 |
mess | +0.016 |
meal | +0.016 |
tray | +0.016 |
Neg logit contributions | |
ila | -0.017 |
stretched | -0.017 |
La | -0.016 |
Tweet | -0.016 |
Tw | -0.015 |
Token | to |
Token ID | 284 |
Feature activation | +4.25e-01 |
Pos logit contributions | |
package | +0.003 |
idents | +0.003 |
meal | +0.003 |
mess | +0.003 |
batch | +0.003 |
Neg logit contributions | |
ila | -0.003 |
La | -0.002 |
stretched | -0.002 |
Tweet | -0.002 |
Tw | -0.002 |
Token | wear |
Token ID | 5806 |
Feature activation | -2.42e+00 |
Pos logit contributions | |
idents | +0.014 |
package | +0.013 |
batch | +0.013 |
tray | +0.013 |
meal | +0.013 |
Neg logit contributions | |
ila | -0.011 |
stretched | -0.011 |
Tw | -0.010 |
eight | -0.010 |
Tweet | -0.010 |
Token | a |
Token ID | 257 |
Feature activation | -7.91e-01 |
Pos logit contributions | |
ila | +0.072 |
stretched | +0.071 |
Tweet | +0.068 |
eight | +0.066 |
La | +0.065 |
Neg logit contributions | |
package | -0.075 |
idents | -0.073 |
tray | -0.069 |
meal | -0.068 |
dinner | -0.067 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.93e+00 |
Pos logit contributions | |
stretched | +0.021 |
ila | +0.020 |
eight | +0.019 |
Tw | +0.019 |
Tweet | +0.019 |
Neg logit contributions | |
idents | -0.025 |
meal | -0.023 |
batch | -0.023 |
package | -0.023 |
tray | -0.023 |
Token | for |
Token ID | 329 |
Feature activation | +1.32e+00 |
Pos logit contributions | |
stretched | +0.172 |
Tweet | +0.170 |
ila | +0.165 |
eight | +0.165 |
La | +0.165 |
Neg logit contributions | |
package | -0.172 |
tray | -0.165 |
stret | -0.162 |
meal | -0.160 |
idents | -0.158 |
Token | a |
Token ID | 257 |
Feature activation | -2.99e-01 |
Pos logit contributions | |
dinner | +0.040 |
meal | +0.038 |
dessert | +0.037 |
package | +0.037 |
cake | +0.035 |
Neg logit contributions | |
stretched | -0.042 |
ila | -0.041 |
La | -0.039 |
Tweet | -0.038 |
Tw | -0.036 |
Token | long |
Token ID | 890 |
Feature activation | -6.15e-01 |
Pos logit contributions | |
ila | +0.008 |
stretched | +0.007 |
Tw | +0.007 |
La | +0.007 |
eight | +0.006 |
Neg logit contributions | |
package | -0.010 |
meal | -0.010 |
tray | -0.010 |
stret | -0.010 |
batch | -0.010 |
Token | time |
Token ID | 640 |
Feature activation | -1.30e+00 |
Pos logit contributions | |
ila | +0.017 |
stretched | +0.016 |
Tw | +0.015 |
eight | +0.015 |
La | +0.015 |
Neg logit contributions | |
package | -0.019 |
meal | -0.019 |
tray | -0.019 |
idents | -0.019 |
batch | -0.018 |
Token | . |
Token ID | 13 |
Feature activation | +1.83e-01 |
Pos logit contributions | |
ila | +0.035 |
stretched | +0.034 |
Tweet | +0.033 |
La | +0.032 |
Tw | +0.032 |
Neg logit contributions | |
package | -0.040 |
idents | -0.040 |
meal | -0.039 |
tray | -0.039 |
batch | -0.038 |
Token | boy |
Token ID | 2933 |
Feature activation | +5.07e-01 |
Pos logit contributions | |
stretched | +0.046 |
Tw | +0.041 |
La | +0.041 |
ila | +0.040 |
Bo | +0.040 |
Neg logit contributions | |
package | -0.045 |
tray | -0.044 |
mess | -0.043 |
cake | -0.042 |
dinner | -0.042 |
Token | had |
Token ID | 550 |
Feature activation | +2.61e-01 |
Pos logit contributions | |
idents | +0.016 |
package | +0.016 |
meal | +0.016 |
tray | +0.016 |
batch | +0.016 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.013 |
Tw | -0.012 |
eight | -0.012 |
Tweet | -0.012 |
Token | to |
Token ID | 284 |
Feature activation | +1.81e-01 |
Pos logit contributions | |
package | +0.008 |
meal | +0.007 |
idents | +0.007 |
dinner | +0.007 |
mess | +0.007 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.008 |
La | -0.007 |
Tw | -0.007 |
Tweet | -0.007 |
Token | wear |
Token ID | 5806 |
Feature activation | -2.37e+00 |
Pos logit contributions | |
package | +0.005 |
idents | +0.005 |
meal | +0.005 |
tray | +0.005 |
batch | +0.005 |
Neg logit contributions | |
stretched | -0.005 |
ila | -0.005 |
Tw | -0.005 |
eight | -0.005 |
Tweet | -0.005 |
Token | a |
Token ID | 257 |
Feature activation | -8.98e-01 |
Pos logit contributions | |
ila | +0.067 |
stretched | +0.066 |
Tweet | +0.064 |
eight | +0.061 |
Tw | +0.060 |
Neg logit contributions | |
package | -0.077 |
idents | -0.073 |
stret | -0.071 |
dinner | -0.070 |
tray | -0.070 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.89e+00 |
Pos logit contributions | |
ila | +0.024 |
stretched | +0.024 |
Tw | +0.022 |
eight | +0.022 |
Tweet | +0.022 |
Neg logit contributions | |
idents | -0.028 |
package | -0.028 |
meal | -0.027 |
tray | -0.027 |
batch | -0.027 |
Token | for |
Token ID | 329 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
Tweet | +0.178 |
stretched | +0.175 |
ila | +0.171 |
La | +0.170 |
eight | +0.170 |
Neg logit contributions | |
package | -0.161 |
tray | -0.155 |
stret | -0.155 |
envelope | -0.148 |
meal | -0.148 |
Token | a |
Token ID | 257 |
Feature activation | -3.82e-01 |
Pos logit contributions | |
dinner | +0.040 |
dessert | +0.037 |
meal | +0.037 |
cake | +0.036 |
sandwiches | +0.035 |
Neg logit contributions | |
stretched | -0.039 |
ila | -0.036 |
Tweet | -0.036 |
La | -0.035 |
Ru | -0.034 |
Token | long |
Token ID | 890 |
Feature activation | -5.58e-01 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.009 |
Tw | +0.009 |
La | +0.008 |
eight | +0.008 |
Neg logit contributions | |
package | -0.013 |
meal | -0.013 |
stret | -0.013 |
tray | -0.013 |
batch | -0.012 |
Token | time |
Token ID | 640 |
Feature activation | -1.34e+00 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.016 |
Tw | +0.015 |
La | +0.015 |
eight | +0.015 |
Neg logit contributions | |
package | -0.016 |
tray | -0.016 |
meal | -0.016 |
stret | -0.016 |
batch | -0.015 |
Token | . |
Token ID | 13 |
Feature activation | +1.89e-01 |
Pos logit contributions | |
Tweet | +0.042 |
ila | +0.041 |
La | +0.039 |
stretched | +0.039 |
Tw | +0.038 |
Neg logit contributions | |
package | -0.035 |
tray | -0.034 |
meal | -0.034 |
dinner | -0.034 |
mess | -0.033 |
Token | doctor |
Token ID | 6253 |
Feature activation | -1.51e+00 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.012 |
Tw | +0.012 |
La | +0.011 |
eight | +0.011 |
Neg logit contributions | |
package | -0.014 |
tray | -0.014 |
meal | -0.013 |
mess | -0.013 |
cake | -0.013 |
Token | said |
Token ID | 531 |
Feature activation | +1.00e+00 |
Pos logit contributions | |
stretched | +0.044 |
ila | +0.043 |
Tweet | +0.040 |
eight | +0.040 |
Tw | +0.040 |
Neg logit contributions | |
package | -0.045 |
idents | -0.044 |
batch | -0.043 |
tray | -0.043 |
meal | -0.043 |
Token | he |
Token ID | 339 |
Feature activation | +5.07e-01 |
Pos logit contributions | |
package | +0.033 |
idents | +0.033 |
meal | +0.033 |
dinner | +0.032 |
cake | +0.032 |
Neg logit contributions | |
stretched | -0.025 |
ila | -0.024 |
La | -0.024 |
Tweet | -0.023 |
eight | -0.022 |
Token | needed |
Token ID | 2622 |
Feature activation | -1.64e-01 |
Pos logit contributions | |
package | +0.016 |
idents | +0.016 |
meal | +0.014 |
envelope | +0.014 |
batch | +0.013 |
Neg logit contributions | |
stretched | -0.015 |
La | -0.014 |
ila | -0.014 |
Tw | -0.013 |
Tweet | -0.013 |
Token | a |
Token ID | 257 |
Feature activation | -2.46e-01 |
Pos logit contributions | |
stretched | +0.005 |
ila | +0.005 |
La | +0.005 |
Tweet | +0.005 |
Tw | +0.005 |
Neg logit contributions | |
dinner | -0.005 |
package | -0.005 |
meal | -0.005 |
dessert | -0.005 |
cake | -0.005 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.89e+00 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
Tw | +0.006 |
eight | +0.006 |
Tweet | +0.006 |
Neg logit contributions | |
package | -0.008 |
idents | -0.008 |
meal | -0.007 |
tray | -0.007 |
batch | -0.007 |
Token | on |
Token ID | 319 |
Feature activation | -4.14e-01 |
Pos logit contributions | |
stretched | +0.173 |
Tweet | +0.170 |
La | +0.168 |
ila | +0.167 |
Tw | +0.166 |
Neg logit contributions | |
package | -0.171 |
meal | -0.162 |
tray | -0.162 |
stret | -0.161 |
idents | -0.158 |
Token | his |
Token ID | 465 |
Feature activation | -4.89e-02 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.010 |
Tweet | +0.010 |
La | +0.010 |
Tw | +0.009 |
Neg logit contributions | |
package | -0.014 |
meal | -0.014 |
dinner | -0.014 |
tray | -0.013 |
batch | -0.013 |
Token | leg |
Token ID | 1232 |
Feature activation | -1.88e+00 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tweet | +0.002 |
Tw | +0.001 |
La | +0.001 |
Neg logit contributions | |
package | -0.001 |
tray | -0.001 |
meal | -0.001 |
stret | -0.001 |
idents | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | +2.42e-01 |
Pos logit contributions | |
ila | +0.052 |
stretched | +0.051 |
eight | +0.048 |
La | +0.048 |
Tw | +0.048 |
Neg logit contributions | |
package | -0.058 |
idents | -0.057 |
meal | -0.056 |
tray | -0.056 |
batch | -0.055 |
Token | Tim |
Token ID | 5045 |
Feature activation | +1.10e+00 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
batch | +0.007 |
meal | +0.007 |
tray | +0.007 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
Tw | -0.006 |
La | -0.006 |
eight | -0.006 |
Token | am |
Token ID | 321 |
Feature activation | +1.87e-01 |
Pos logit contributions | |
package | +0.020 |
meal | +0.020 |
batch | +0.020 |
tray | +0.020 |
idents | +0.019 |
Neg logit contributions | |
ila | -0.018 |
stretched | -0.017 |
eight | -0.016 |
Tweet | -0.016 |
Tw | -0.016 |
Token | away |
Token ID | 1497 |
Feature activation | +1.99e-01 |
Pos logit contributions | |
idents | +0.004 |
package | +0.003 |
tray | +0.003 |
mess | +0.003 |
meal | +0.003 |
Neg logit contributions | |
stretched | -0.007 |
ila | -0.007 |
La | -0.007 |
Tw | -0.007 |
Tweet | -0.007 |
Token | with |
Token ID | 351 |
Feature activation | -1.61e-01 |
Pos logit contributions | |
idents | +0.006 |
tray | +0.006 |
package | +0.005 |
meal | +0.005 |
batch | +0.005 |
Neg logit contributions | |
ila | -0.006 |
stretched | -0.006 |
Tweet | -0.006 |
La | -0.006 |
Tw | -0.005 |
Token | a |
Token ID | 257 |
Feature activation | -4.46e-01 |
Pos logit contributions | |
stretched | +0.003 |
ila | +0.003 |
Tw | +0.003 |
Tweet | +0.003 |
La | +0.003 |
Neg logit contributions | |
package | -0.006 |
idents | -0.006 |
tray | -0.006 |
meal | -0.006 |
batch | -0.006 |
Token | hon |
Token ID | 3032 |
Feature activation | -1.78e+00 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.011 |
Tw | +0.010 |
eight | +0.010 |
Tweet | +0.010 |
Neg logit contributions | |
idents | -0.015 |
package | -0.015 |
meal | -0.015 |
tray | -0.015 |
batch | -0.015 |
Token | k |
Token ID | 74 |
Feature activation | -5.85e+00 |
Pos logit contributions | |
stretched | +0.052 |
ila | +0.052 |
Tw | +0.049 |
Tweet | +0.048 |
eight | +0.048 |
Neg logit contributions | |
idents | -0.052 |
package | -0.051 |
tray | -0.049 |
meal | -0.049 |
batch | -0.048 |
Token | . |
Token ID | 13 |
Feature activation | -2.06e-01 |
Pos logit contributions | |
Tweet | +0.165 |
La | +0.163 |
ila | +0.160 |
stretched | +0.156 |
Tw | +0.156 |
Neg logit contributions | |
tray | -0.158 |
package | -0.157 |
stret | -0.151 |
envelope | -0.150 |
meal | -0.149 |
Token | It |
Token ID | 632 |
Feature activation | +3.10e-01 |
Pos logit contributions | |
stretched | +0.007 |
La | +0.006 |
Tweet | +0.006 |
Bo | +0.006 |
eight | +0.006 |
Neg logit contributions | |
idents | -0.006 |
package | -0.005 |
tray | -0.005 |
batch | -0.005 |
refrigerator | -0.005 |
Token | looked |
Token ID | 3114 |
Feature activation | +7.57e-01 |
Pos logit contributions | |
tray | +0.009 |
idents | +0.009 |
package | +0.009 |
meal | +0.009 |
mess | +0.009 |
Neg logit contributions | |
stretched | -0.009 |
ila | -0.009 |
La | -0.008 |
eight | -0.008 |
Tw | -0.008 |
Token | very |
Token ID | 845 |
Feature activation | +1.67e+00 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
meal | +0.021 |
tray | +0.021 |
batch | +0.021 |
Neg logit contributions | |
ila | -0.022 |
stretched | -0.022 |
Tw | -0.021 |
eight | -0.021 |
Tweet | -0.020 |
Token | pleased |
Token ID | 10607 |
Feature activation | +5.44e-01 |
Pos logit contributions | |
idents | +0.046 |
mess | +0.045 |
package | +0.042 |
meal | +0.042 |
tray | +0.041 |
Neg logit contributions | |
stretched | -0.050 |
La | -0.049 |
ila | -0.048 |
eight | -0.048 |
Tweet | -0.047 |
Token | leg |
Token ID | 1232 |
Feature activation | -1.09e+00 |
Pos logit contributions | |
idents | +0.019 |
package | +0.018 |
batch | +0.018 |
tray | +0.018 |
meal | +0.018 |
Neg logit contributions | |
ila | -0.016 |
stretched | -0.016 |
eight | -0.016 |
Tw | -0.015 |
Tweet | -0.015 |
Token | and |
Token ID | 290 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
ila | +0.031 |
stretched | +0.031 |
La | +0.030 |
Tw | +0.029 |
Tweet | +0.029 |
Neg logit contributions | |
package | -0.033 |
tray | -0.032 |
idents | -0.032 |
meal | -0.031 |
batch | -0.031 |
Token | he |
Token ID | 339 |
Feature activation | +2.31e-01 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.002 |
La | +0.002 |
Tw | +0.002 |
Tweet | +0.002 |
Neg logit contributions | |
package | -0.004 |
idents | -0.004 |
meal | -0.004 |
tray | -0.004 |
batch | -0.004 |
Token | needed |
Token ID | 2622 |
Feature activation | -2.61e-01 |
Pos logit contributions | |
idents | +0.007 |
package | +0.007 |
meal | +0.007 |
batch | +0.006 |
mess | +0.006 |
Neg logit contributions | |
ila | -0.007 |
stretched | -0.007 |
La | -0.007 |
Tw | -0.006 |
eight | -0.006 |
Token | a |
Token ID | 257 |
Feature activation | -3.47e-01 |
Pos logit contributions | |
ila | +0.008 |
stretched | +0.008 |
La | +0.008 |
Tw | +0.007 |
Tweet | +0.007 |
Neg logit contributions | |
package | -0.008 |
dinner | -0.008 |
meal | -0.008 |
dessert | -0.008 |
cake | -0.008 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.82e+00 |
Pos logit contributions | |
stretched | +0.009 |
ila | +0.009 |
eight | +0.009 |
Tw | +0.009 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.011 |
package | -0.011 |
batch | -0.010 |
meal | -0.010 |
tray | -0.010 |
Token | . |
Token ID | 13 |
Feature activation | +1.98e-01 |
Pos logit contributions | |
stretched | +0.176 |
La | +0.173 |
ila | +0.172 |
Tweet | +0.171 |
Tw | +0.170 |
Neg logit contributions | |
package | -0.166 |
meal | -0.157 |
tray | -0.156 |
stret | -0.154 |
dinner | -0.152 |
Token | Spot |
Token ID | 15899 |
Feature activation | +1.66e+00 |
Pos logit contributions | |
idents | +0.007 |
package | +0.007 |
meal | +0.007 |
tray | +0.007 |
batch | +0.007 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
Tw | -0.004 |
eight | -0.004 |
Tweet | -0.004 |
Token | didn |
Token ID | 1422 |
Feature activation | -1.12e+00 |
Pos logit contributions | |
idents | +0.057 |
package | +0.056 |
tray | +0.055 |
meal | +0.055 |
batch | +0.055 |
Neg logit contributions | |
ila | -0.039 |
stretched | -0.039 |
Tw | -0.037 |
eight | -0.037 |
Tweet | -0.037 |
Token | 't |
Token ID | 470 |
Feature activation | -8.14e-02 |
Pos logit contributions | |
stretched | +0.020 |
ila | +0.019 |
La | +0.017 |
eight | +0.017 |
Tw | +0.017 |
Neg logit contributions | |
idents | -0.045 |
package | -0.045 |
tray | -0.044 |
meal | -0.044 |
batch | -0.043 |
Token | like |
Token ID | 588 |
Feature activation | -1.38e-01 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tw | +0.002 |
eight | +0.002 |
Tweet | +0.002 |
Neg logit contributions | |
idents | -0.003 |
package | -0.003 |
meal | -0.003 |
tray | -0.003 |
batch | -0.003 |
Token | . |
Token ID | 13 |
Feature activation | +1.82e-02 |
Pos logit contributions | |
idents | +0.014 |
package | +0.014 |
batch | +0.013 |
meal | +0.013 |
tray | +0.013 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.007 |
Tw | -0.007 |
La | -0.007 |
eight | -0.007 |
Token | They |
Token ID | 1119 |
Feature activation | +9.60e-01 |
Pos logit contributions | |
package | +0.001 |
idents | +0.001 |
refrigerator | +0.001 |
tray | +0.000 |
meal | +0.000 |
Neg logit contributions | |
stretched | -0.001 |
ila | -0.000 |
Tweet | -0.000 |
eight | -0.000 |
brav | -0.000 |
Token | hear |
Token ID | 3285 |
Feature activation | -1.06e+00 |
Pos logit contributions | |
idents | +0.028 |
package | +0.028 |
tray | +0.027 |
meal | +0.027 |
stret | +0.026 |
Neg logit contributions | |
stretched | -0.028 |
ila | -0.027 |
Tw | -0.026 |
Tweet | -0.026 |
La | -0.025 |
Token | a |
Token ID | 257 |
Feature activation | -9.46e-01 |
Pos logit contributions | |
stretched | +0.024 |
ila | +0.024 |
eight | +0.022 |
Tw | +0.022 |
Tweet | +0.022 |
Neg logit contributions | |
idents | -0.037 |
package | -0.037 |
meal | -0.036 |
tray | -0.036 |
batch | -0.036 |
Token | hon |
Token ID | 3032 |
Feature activation | -1.57e+00 |
Pos logit contributions | |
stretched | +0.022 |
ila | +0.022 |
eight | +0.021 |
Tw | +0.021 |
Tweet | +0.020 |
Neg logit contributions | |
idents | -0.033 |
package | -0.032 |
meal | -0.032 |
tray | -0.032 |
batch | -0.032 |
Token | k |
Token ID | 74 |
Feature activation | -5.79e+00 |
Pos logit contributions | |
ila | +0.038 |
stretched | +0.038 |
eight | +0.036 |
Tw | +0.035 |
La | +0.035 |
Neg logit contributions | |
package | -0.055 |
meal | -0.055 |
idents | -0.055 |
batch | -0.054 |
tray | -0.053 |
Token | . |
Token ID | 13 |
Feature activation | -4.36e-02 |
Pos logit contributions | |
ila | +0.148 |
stretched | +0.147 |
Tw | +0.137 |
Tweet | +0.137 |
eight | +0.136 |
Neg logit contributions | |
idents | -0.188 |
package | -0.187 |
tray | -0.182 |
meal | -0.182 |
batch | -0.179 |
Token | They |
Token ID | 1119 |
Feature activation | +8.33e-01 |
Pos logit contributions | |
stretched | +0.001 |
ila | +0.001 |
Tweet | +0.001 |
eight | +0.001 |
La | +0.001 |
Neg logit contributions | |
package | -0.001 |
idents | -0.001 |
tray | -0.001 |
meal | -0.001 |
refrigerator | -0.001 |
Token | hear |
Token ID | 3285 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
package | +0.026 |
idents | +0.026 |
tray | +0.025 |
meal | +0.025 |
stret | +0.024 |
Neg logit contributions | |
stretched | -0.022 |
ila | -0.022 |
Tweet | -0.021 |
Tw | -0.021 |
La | -0.020 |
Token | a |
Token ID | 257 |
Feature activation | -9.65e-01 |
Pos logit contributions | |
stretched | +0.026 |
ila | +0.026 |
Tweet | +0.024 |
eight | +0.024 |
Tw | +0.024 |
Neg logit contributions | |
idents | -0.041 |
package | -0.041 |
meal | -0.040 |
tray | -0.040 |
batch | -0.040 |
Token | w |
Token ID | 266 |
Feature activation | -9.26e-01 |
Pos logit contributions | |
stretched | +0.022 |
ila | +0.022 |
eight | +0.020 |
Tw | +0.020 |
Tweet | +0.020 |
Neg logit contributions | |
idents | -0.034 |
package | -0.034 |
meal | -0.033 |
tray | -0.033 |
batch | -0.033 |
Token | to |
Token ID | 284 |
Feature activation | +7.13e-01 |
Pos logit contributions | |
package | +0.010 |
meal | +0.009 |
dinner | +0.009 |
cake | +0.009 |
tray | +0.009 |
Neg logit contributions | |
stretched | -0.009 |
Tweet | -0.009 |
ila | -0.009 |
La | -0.008 |
Tw | -0.008 |
Token | put |
Token ID | 1234 |
Feature activation | -4.13e-01 |
Pos logit contributions | |
package | +0.022 |
idents | +0.022 |
meal | +0.022 |
tray | +0.021 |
batch | +0.021 |
Neg logit contributions | |
ila | -0.019 |
stretched | -0.019 |
Tw | -0.018 |
eight | -0.018 |
Tweet | -0.018 |
Token | her |
Token ID | 607 |
Feature activation | -1.53e+00 |
Pos logit contributions | |
stretched | +0.011 |
ila | +0.011 |
Tw | +0.010 |
eight | +0.010 |
Tweet | +0.010 |
Neg logit contributions | |
idents | -0.013 |
package | -0.013 |
meal | -0.013 |
tray | -0.013 |
batch | -0.013 |
Token | in |
Token ID | 287 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
ila | +0.050 |
stretched | +0.049 |
Tweet | +0.047 |
La | +0.046 |
Tw | +0.046 |
Neg logit contributions | |
package | -0.042 |
idents | -0.040 |
tray | -0.039 |
meal | -0.039 |
batch | -0.038 |
Token | a |
Token ID | 257 |
Feature activation | -1.46e+00 |
Pos logit contributions | |
stretched | +0.029 |
ila | +0.029 |
Tweet | +0.028 |
La | +0.026 |
Tw | +0.026 |
Neg logit contributions | |
package | -0.044 |
idents | -0.042 |
tray | -0.042 |
meal | -0.042 |
batch | -0.042 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.77e+00 |
Pos logit contributions | |
stretched | +0.041 |
ila | +0.041 |
Tw | +0.038 |
eight | +0.038 |
Tweet | +0.038 |
Neg logit contributions | |
idents | -0.044 |
package | -0.042 |
meal | -0.042 |
tray | -0.041 |
batch | -0.041 |
Token | so |
Token ID | 523 |
Feature activation | +1.02e-01 |
Pos logit contributions | |
stretched | +0.166 |
Tweet | +0.161 |
ila | +0.161 |
La | +0.153 |
eight | +0.151 |
Neg logit contributions | |
package | -0.173 |
tray | -0.171 |
meal | -0.164 |
idents | -0.162 |
stret | -0.161 |
Token | it |
Token ID | 340 |
Feature activation | +1.05e+00 |
Pos logit contributions | |
dinner | +0.003 |
idents | +0.003 |
package | +0.003 |
cake | +0.003 |
tray | +0.002 |
Neg logit contributions | |
Tweet | -0.003 |
La | -0.003 |
Ru | -0.003 |
stretched | -0.003 |
Nice | -0.003 |
Token | can |
Token ID | 460 |
Feature activation | -1.05e+00 |
Pos logit contributions | |
idents | +0.032 |
package | +0.031 |
tray | +0.030 |
mess | +0.030 |
meal | +0.029 |
Neg logit contributions | |
ila | -0.031 |
stretched | -0.030 |
La | -0.029 |
Tweet | -0.028 |
eight | -0.027 |
Token | heal |
Token ID | 12035 |
Feature activation | -1.49e+00 |
Pos logit contributions | |
stretched | +0.032 |
ila | +0.032 |
La | +0.031 |
eight | +0.029 |
Nice | +0.029 |
Neg logit contributions | |
package | -0.027 |
idents | -0.027 |
mess | -0.027 |
meal | -0.027 |
doorstep | -0.027 |
Token | ." |
Token ID | 526 |
Feature activation | +2.03e-01 |
Pos logit contributions | |
stretched | +0.039 |
ila | +0.038 |
eight | +0.036 |
Tweet | +0.035 |
La | +0.034 |
Neg logit contributions | |
idents | -0.049 |
package | -0.048 |
meal | -0.047 |
tray | -0.047 |
batch | -0.046 |
Token | arm |
Token ID | 3211 |
Feature activation | -3.38e-01 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.011 |
Tw | +0.011 |
eight | +0.011 |
Tweet | +0.011 |
Neg logit contributions | |
package | -0.009 |
idents | -0.009 |
tray | -0.009 |
meal | -0.009 |
batch | -0.009 |
Token | . |
Token ID | 13 |
Feature activation | +2.82e-01 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.010 |
eight | +0.009 |
La | +0.009 |
Tw | +0.009 |
Neg logit contributions | |
package | -0.010 |
idents | -0.010 |
tray | -0.010 |
meal | -0.010 |
batch | -0.009 |
Token | He |
Token ID | 679 |
Feature activation | -7.69e-01 |
Pos logit contributions | |
idents | +0.011 |
package | +0.011 |
batch | +0.011 |
tray | +0.011 |
refrigerator | +0.011 |
Neg logit contributions | |
stretched | -0.005 |
Tweet | -0.005 |
ila | -0.005 |
Tw | -0.005 |
La | -0.005 |
Token | put |
Token ID | 1234 |
Feature activation | -6.15e-01 |
Pos logit contributions | |
stretched | +0.027 |
La | +0.026 |
ila | +0.025 |
eight | +0.025 |
Tw | +0.024 |
Neg logit contributions | |
package | -0.020 |
batch | -0.018 |
meal | -0.017 |
envelope | -0.017 |
mess | -0.017 |
Token | a |
Token ID | 257 |
Feature activation | -1.11e+00 |
Pos logit contributions | |
stretched | +0.016 |
ila | +0.015 |
Tweet | +0.015 |
Tw | +0.015 |
La | +0.015 |
Neg logit contributions | |
package | -0.020 |
idents | -0.020 |
tray | -0.020 |
meal | -0.019 |
batch | -0.019 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.75e+00 |
Pos logit contributions | |
ila | +0.031 |
stretched | +0.031 |
Tw | +0.029 |
eight | +0.029 |
Tweet | +0.028 |
Neg logit contributions | |
idents | -0.034 |
package | -0.034 |
meal | -0.033 |
tray | -0.033 |
batch | -0.032 |
Token | on |
Token ID | 319 |
Feature activation | -3.76e-01 |
Pos logit contributions | |
Tweet | +0.163 |
stretched | +0.161 |
La | +0.157 |
Tw | +0.155 |
ila | +0.154 |
Neg logit contributions | |
package | -0.168 |
tray | -0.162 |
stret | -0.160 |
envelope | -0.158 |
meal | -0.150 |
Token | his |
Token ID | 465 |
Feature activation | -1.71e-01 |
Pos logit contributions | |
stretched | +0.009 |
ila | +0.008 |
Tw | +0.008 |
Tweet | +0.008 |
eight | +0.008 |
Neg logit contributions | |
package | -0.013 |
idents | -0.013 |
tray | -0.013 |
meal | -0.013 |
batch | -0.013 |
Token | arm |
Token ID | 3211 |
Feature activation | -3.34e-01 |
Pos logit contributions | |
stretched | +0.006 |
ila | +0.006 |
Tweet | +0.006 |
Tw | +0.006 |
La | +0.005 |
Neg logit contributions | |
package | -0.004 |
tray | -0.004 |
stret | -0.004 |
meal | -0.004 |
idents | -0.004 |
Token | and |
Token ID | 290 |
Feature activation | -2.69e-01 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.010 |
eight | +0.010 |
Bo | +0.010 |
La | +0.010 |
Neg logit contributions | |
package | -0.009 |
stret | -0.008 |
tray | -0.008 |
batch | -0.008 |
idents | -0.008 |
Token | told |
Token ID | 1297 |
Feature activation | +2.67e-01 |
Pos logit contributions | |
stretched | +0.008 |
ila | +0.007 |
Tweet | +0.007 |
Tw | +0.007 |
La | +0.007 |
Neg logit contributions | |
package | -0.009 |
idents | -0.008 |
meal | -0.008 |
tray | -0.008 |
batch | -0.008 |
Token | arm |
Token ID | 3211 |
Feature activation | -6.02e-01 |
Pos logit contributions | |
ila | +0.026 |
stretched | +0.025 |
La | +0.024 |
Tweet | +0.024 |
eight | +0.023 |
Neg logit contributions | |
package | -0.020 |
meal | -0.020 |
tray | -0.019 |
idents | -0.019 |
batch | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | -6.29e-01 |
Pos logit contributions | |
stretched | +0.018 |
ila | +0.018 |
La | +0.018 |
Tweet | +0.017 |
Bo | +0.017 |
Neg logit contributions | |
package | -0.018 |
tray | -0.017 |
meal | -0.017 |
refrigerator | -0.016 |
cake | -0.016 |
Token | She |
Token ID | 1375 |
Feature activation | -2.21e+00 |
Pos logit contributions | |
stretched | +0.017 |
ila | +0.015 |
Tweet | +0.015 |
known | +0.014 |
La | +0.014 |
Neg logit contributions | |
package | -0.020 |
idents | -0.020 |
refrigerator | -0.020 |
meal | -0.019 |
tray | -0.018 |
Token | needs |
Token ID | 2476 |
Feature activation | -3.85e-01 |
Pos logit contributions | |
stretched | +0.064 |
ila | +0.059 |
Tweet | +0.057 |
La | +0.057 |
Tw | +0.054 |
Neg logit contributions | |
package | -0.069 |
idents | -0.069 |
meal | -0.067 |
tray | -0.065 |
batch | -0.064 |
Token | a |
Token ID | 257 |
Feature activation | -8.29e-01 |
Pos logit contributions | |
ila | +0.012 |
stretched | +0.012 |
Tweet | +0.012 |
La | +0.012 |
Bo | +0.011 |
Neg logit contributions | |
package | -0.011 |
dinner | -0.011 |
meal | -0.011 |
cake | -0.010 |
dessert | -0.010 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.74e+00 |
Pos logit contributions | |
ila | +0.026 |
stretched | +0.026 |
Tweet | +0.024 |
Tw | +0.024 |
La | +0.024 |
Neg logit contributions | |
package | -0.024 |
meal | -0.023 |
tray | -0.023 |
idents | -0.022 |
batch | -0.022 |
Token | and |
Token ID | 290 |
Feature activation | -7.89e-01 |
Pos logit contributions | |
stretched | +0.182 |
Tweet | +0.176 |
La | +0.172 |
ila | +0.168 |
Bo | +0.165 |
Neg logit contributions | |
package | -0.159 |
tray | -0.153 |
meal | -0.149 |
envelope | -0.147 |
stret | -0.144 |
Token | a |
Token ID | 257 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
stretched | +0.026 |
Tweet | +0.024 |
ila | +0.024 |
La | +0.023 |
Bo | +0.023 |
Neg logit contributions | |
dinner | -0.023 |
package | -0.022 |
meal | -0.022 |
cake | -0.021 |
idents | -0.021 |
Token | sticker |
Token ID | 27735 |
Feature activation | -3.15e+00 |
Pos logit contributions | |
ila | +0.039 |
stretched | +0.037 |
La | +0.035 |
Tweet | +0.035 |
Tw | +0.035 |
Neg logit contributions | |
package | -0.036 |
meal | -0.035 |
tray | -0.035 |
stret | -0.033 |
batch | -0.033 |
Token | ," |
Token ID | 553 |
Feature activation | -4.73e-01 |
Pos logit contributions | |
stretched | +0.091 |
ila | +0.084 |
Tweet | +0.084 |
La | +0.082 |
eight | +0.082 |
Neg logit contributions | |
package | -0.098 |
tray | -0.095 |
meal | -0.092 |
idents | -0.092 |
envelope | -0.089 |
Token | Sam |
Token ID | 3409 |
Feature activation | -1.46e+00 |
Pos logit contributions | |
stretched | +0.015 |
reetings | +0.013 |
La | +0.013 |
Tweet | +0.012 |
Bo | +0.012 |
Neg logit contributions | |
cake | -0.014 |
dinner | -0.014 |
idents | -0.013 |
package | -0.013 |
dessert | -0.013 |
Token | they |
Token ID | 484 |
Feature activation | +1.90e+00 |
Pos logit contributions | |
package | +0.015 |
meal | +0.015 |
tray | +0.015 |
idents | +0.015 |
dinner | +0.014 |
Neg logit contributions | |
stretched | -0.010 |
ila | -0.009 |
Tweet | -0.009 |
La | -0.009 |
eight | -0.008 |
Token | heard |
Token ID | 2982 |
Feature activation | -1.71e+00 |
Pos logit contributions | |
tray | +0.072 |
package | +0.072 |
idents | +0.072 |
meal | +0.071 |
refrigerator | +0.069 |
Neg logit contributions | |
ila | -0.036 |
stretched | -0.036 |
Tweet | -0.034 |
Tw | -0.032 |
La | -0.030 |
Token | a |
Token ID | 257 |
Feature activation | -8.93e-01 |
Pos logit contributions | |
stretched | +0.041 |
ila | +0.041 |
Tweet | +0.037 |
eight | +0.036 |
La | +0.035 |
Neg logit contributions | |
package | -0.059 |
meal | -0.058 |
idents | -0.057 |
tray | -0.057 |
batch | -0.056 |
Token | loud |
Token ID | 7812 |
Feature activation | -1.16e+00 |
Pos logit contributions | |
stretched | +0.020 |
ila | +0.020 |
Tw | +0.018 |
eight | +0.018 |
Tweet | +0.018 |
Neg logit contributions | |
idents | -0.032 |
package | -0.032 |
meal | -0.031 |
tray | -0.031 |
batch | -0.031 |
Token | hon |
Token ID | 3032 |
Feature activation | -1.69e+00 |
Pos logit contributions | |
Tw | +0.032 |
Tweet | +0.030 |
eight | +0.029 |
stretched | +0.029 |
ila | +0.029 |
Neg logit contributions | |
idents | -0.037 |
dinners | -0.036 |
package | -0.035 |
goodies | -0.035 |
dining | -0.035 |
Token | k |
Token ID | 74 |
Feature activation | -5.73e+00 |
Pos logit contributions | |
ila | +0.046 |
stretched | +0.046 |
Tw | +0.044 |
eight | +0.043 |
Tweet | +0.043 |
Neg logit contributions | |
idents | -0.052 |
package | -0.052 |
meal | -0.051 |
tray | -0.051 |
batch | -0.050 |
Token | . |
Token ID | 13 |
Feature activation | -5.57e-01 |
Pos logit contributions | |
ila | +0.148 |
stretched | +0.144 |
Tweet | +0.141 |
eight | +0.136 |
Tw | +0.136 |
Neg logit contributions | |
package | -0.180 |
idents | -0.178 |
tray | -0.175 |
meal | -0.174 |
batch | -0.169 |
Token | The |
Token ID | 383 |
Feature activation | -1.74e+00 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.012 |
Tweet | +0.011 |
La | +0.011 |
eight | +0.011 |
Neg logit contributions | |
package | -0.020 |
tray | -0.020 |
idents | -0.019 |
meal | -0.019 |
refrigerator | -0.019 |
Token | bus |
Token ID | 1323 |
Feature activation | -8.61e-01 |
Pos logit contributions | |
stretched | +0.053 |
ila | +0.052 |
Tw | +0.049 |
Tweet | +0.049 |
La | +0.048 |
Neg logit contributions | |
package | -0.054 |
tray | -0.050 |
meal | -0.049 |
stret | -0.047 |
mess | -0.047 |
Token | had |
Token ID | 550 |
Feature activation | -2.07e-01 |
Pos logit contributions | |
ila | +0.022 |
stretched | +0.022 |
Tw | +0.021 |
La | +0.021 |
Tweet | +0.020 |
Neg logit contributions | |
package | -0.028 |
idents | -0.027 |
tray | -0.027 |
meal | -0.026 |
batch | -0.026 |
Token | arrived |
Token ID | 5284 |
Feature activation | +1.04e-01 |
Pos logit contributions | |
ila | +0.005 |
stretched | +0.005 |
Tweet | +0.005 |
Tw | +0.005 |
La | +0.004 |
Neg logit contributions | |
package | -0.007 |
tray | -0.007 |
meal | -0.007 |
idents | -0.006 |
sandwiches | -0.006 |
Token | and |
Token ID | 290 |
Feature activation | -2.25e-01 |
Pos logit contributions | |
idents | +0.022 |
package | +0.022 |
meal | +0.021 |
tray | +0.021 |
batch | +0.021 |
Neg logit contributions | |
stretched | -0.015 |
ila | -0.015 |
Tw | -0.014 |
eight | -0.014 |
Tweet | -0.014 |
Token | makes |
Token ID | 1838 |
Feature activation | -6.01e-01 |
Pos logit contributions | |
stretched | +0.006 |
ila | +0.006 |
Tweet | +0.006 |
eight | +0.006 |
La | +0.005 |
Neg logit contributions | |
idents | -0.007 |
meal | -0.007 |
tray | -0.007 |
package | -0.007 |
batch | -0.007 |
Token | a |
Token ID | 257 |
Feature activation | -1.21e+00 |
Pos logit contributions | |
ila | +0.013 |
stretched | +0.013 |
Tw | +0.012 |
eight | +0.012 |
Tweet | +0.012 |
Neg logit contributions | |
idents | -0.022 |
package | -0.022 |
meal | -0.021 |
tray | -0.021 |
batch | -0.021 |
Token | loud |
Token ID | 7812 |
Feature activation | -1.48e+00 |
Pos logit contributions | |
stretched | +0.025 |
eight | +0.025 |
ila | +0.024 |
Tweet | +0.023 |
Tw | +0.023 |
Neg logit contributions | |
package | -0.045 |
keeping | -0.044 |
idents | -0.044 |
dinners | -0.043 |
isions | -0.042 |
Token | hon |
Token ID | 3032 |
Feature activation | -1.62e+00 |
Pos logit contributions | |
ila | +0.045 |
stretched | +0.045 |
Tw | +0.045 |
Tweet | +0.044 |
eight | +0.044 |
Neg logit contributions | |
package | -0.042 |
idents | -0.039 |
tray | -0.038 |
doorstep | -0.038 |
aman | -0.037 |
Token | k |
Token ID | 74 |
Feature activation | -5.73e+00 |
Pos logit contributions | |
ila | +0.043 |
stretched | +0.042 |
Tw | +0.040 |
eight | +0.040 |
Tweet | +0.039 |
Neg logit contributions | |
package | -0.053 |
idents | -0.052 |
meal | -0.051 |
tray | -0.051 |
batch | -0.051 |
Token | . |
Token ID | 13 |
Feature activation | -2.29e-01 |
Pos logit contributions | |
ila | +0.152 |
stretched | +0.150 |
Tweet | +0.148 |
La | +0.142 |
Tw | +0.139 |
Neg logit contributions | |
idents | -0.168 |
tray | -0.166 |
meal | -0.164 |
package | -0.163 |
batch | -0.160 |
Token |
|
Token ID | 198 |
Feature activation | -1.58e-01 |
Pos logit contributions | |
stretched | +0.007 |
La | +0.006 |
Tweet | +0.006 |
brav | +0.006 |
Mart | +0.006 |
Neg logit contributions | |
idents | -0.006 |
pizza | -0.006 |
dinner | -0.006 |
tray | -0.006 |
sandwich | -0.006 |
Token |
|
Token ID | 198 |
Feature activation | -3.11e-01 |
Pos logit contributions | |
stretched | +0.005 |
brav | +0.004 |
Leave | +0.004 |
La | +0.004 |
eight | +0.004 |
Neg logit contributions | |
batch | -0.004 |
meal | -0.004 |
mess | -0.004 |
tray | -0.004 |
dinner | -0.004 |
Token | " |
Token ID | 1 |
Feature activation | +1.13e-02 |
Pos logit contributions | |
stretched | +0.010 |
Fay | +0.008 |
Tw | +0.008 |
known | +0.008 |
brav | +0.008 |
Neg logit contributions | |
tray | -0.008 |
burgers | -0.008 |
batch | -0.008 |
mess | -0.008 |
grill | -0.008 |
Token | More |
Token ID | 5167 |
Feature activation | -4.70e-01 |
Pos logit contributions | |
idents | +0.000 |
tray | +0.000 |
batch | +0.000 |
package | +0.000 |
meal | +0.000 |
Neg logit contributions | |
stretched | -0.000 |
ila | -0.000 |
eight | -0.000 |
Tw | -0.000 |
Bubble | -0.000 |
Token | . |
Token ID | 13 |
Feature activation | +3.90e-01 |
Pos logit contributions | |
ila | +0.021 |
stretched | +0.021 |
Tw | +0.019 |
La | +0.019 |
Tweet | +0.019 |
Neg logit contributions | |
package | -0.025 |
idents | -0.025 |
tray | -0.025 |
meal | -0.024 |
batch | -0.024 |
Token | Despite |
Token ID | 7945 |
Feature activation | -2.40e+00 |
Pos logit contributions | |
batch | +0.010 |
idents | +0.010 |
aman | +0.010 |
refrigerator | +0.010 |
meal | +0.009 |
Neg logit contributions | |
La | -0.012 |
stretched | -0.012 |
ila | -0.012 |
Tw | -0.011 |
Tw | -0.011 |
Token | wearing |
Token ID | 5762 |
Feature activation | -1.06e+00 |
Pos logit contributions | |
stretched | +0.066 |
Tweet | +0.065 |
ila | +0.064 |
La | +0.062 |
Tw | +0.061 |
Neg logit contributions | |
package | -0.071 |
meal | -0.071 |
batch | -0.069 |
dinner | -0.069 |
idents | -0.068 |
Token | a |
Token ID | 257 |
Feature activation | -5.97e-01 |
Pos logit contributions | |
ila | +0.029 |
stretched | +0.028 |
Tweet | +0.028 |
eight | +0.026 |
Tw | +0.026 |
Neg logit contributions | |
package | -0.035 |
tray | -0.033 |
idents | -0.033 |
meal | -0.032 |
dinner | -0.032 |
Token | pink |
Token ID | 11398 |
Feature activation | +4.33e-01 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.015 |
Tw | +0.015 |
Tweet | +0.014 |
La | +0.014 |
Neg logit contributions | |
package | -0.019 |
tray | -0.019 |
idents | -0.019 |
meal | -0.019 |
batch | -0.018 |
Token | cast |
Token ID | 3350 |
Feature activation | -5.73e+00 |
Pos logit contributions | |
tray | +0.012 |
package | +0.012 |
meal | +0.011 |
batch | +0.011 |
idents | +0.011 |
Neg logit contributions | |
ila | -0.014 |
stretched | -0.014 |
La | -0.013 |
Tw | -0.013 |
Tweet | -0.013 |
Token | , |
Token ID | 11 |
Feature activation | +6.48e-01 |
Pos logit contributions | |
Tweet | +0.172 |
La | +0.171 |
ila | +0.168 |
stretched | +0.165 |
Bo | +0.163 |
Neg logit contributions | |
package | -0.153 |
tray | -0.153 |
meal | -0.144 |
stret | -0.142 |
dinner | -0.139 |
Token | Sally |
Token ID | 25737 |
Feature activation | +1.10e+00 |
Pos logit contributions | |
dinner | +0.021 |
package | +0.021 |
meal | +0.021 |
batch | +0.021 |
idents | +0.021 |
Neg logit contributions | |
stretched | -0.016 |
La | -0.016 |
Tweet | -0.015 |
Tw | -0.015 |
ila | -0.015 |
Token | never |
Token ID | 1239 |
Feature activation | -1.56e+00 |
Pos logit contributions | |
package | +0.033 |
idents | +0.032 |
meal | +0.031 |
tray | +0.031 |
batch | +0.031 |
Neg logit contributions | |
ila | -0.031 |
stretched | -0.030 |
Tweet | -0.028 |
La | -0.028 |
Tw | -0.027 |
Token | went |
Token ID | 1816 |
Feature activation | -2.71e-01 |
Pos logit contributions | |
ila | +0.047 |
stretched | +0.047 |
Tweet | +0.044 |
Tw | +0.044 |
La | +0.043 |
Neg logit contributions | |
mess | -0.042 |
meal | -0.042 |
package | -0.042 |
idents | -0.041 |
dinner | -0.039 |
Token | back |
Token ID | 736 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
ila | +0.010 |
Bo | +0.010 |
Tweet | +0.010 |
stretched | +0.010 |
Tw | +0.010 |
Neg logit contributions | |
package | -0.006 |
dinner | -0.005 |
stret | -0.005 |
mess | -0.005 |
meal | -0.005 |
Token | happy |
Token ID | 3772 |
Feature activation | +1.27e+00 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
La | +0.006 |
Tweet | +0.006 |
Tw | +0.006 |
Neg logit contributions | |
meal | -0.008 |
package | -0.008 |
tray | -0.008 |
batch | -0.007 |
idents | -0.007 |
Token | and |
Token ID | 290 |
Feature activation | +3.65e-01 |
Pos logit contributions | |
package | +0.047 |
idents | +0.046 |
tray | +0.046 |
stret | +0.045 |
batch | +0.045 |
Neg logit contributions | |
stretched | -0.027 |
ila | -0.027 |
Tw | -0.026 |
eight | -0.025 |
Tweet | -0.024 |
Token | people |
Token ID | 661 |
Feature activation | -1.39e-02 |
Pos logit contributions | |
idents | +0.012 |
package | +0.012 |
tray | +0.012 |
meal | +0.011 |
batch | +0.011 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
Tw | -0.009 |
eight | -0.009 |
Tweet | -0.009 |
Token | saw |
Token ID | 2497 |
Feature activation | -5.85e-03 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
eight | +0.000 |
La | +0.000 |
Tw | +0.000 |
Neg logit contributions | |
package | -0.000 |
idents | -0.000 |
meal | -0.000 |
tray | -0.000 |
batch | -0.000 |
Token | the |
Token ID | 262 |
Feature activation | +3.74e-01 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
Tweet | +0.000 |
Tw | +0.000 |
eight | +0.000 |
Neg logit contributions | |
package | -0.000 |
idents | -0.000 |
meal | -0.000 |
tray | -0.000 |
batch | -0.000 |
Token | y |
Token ID | 331 |
Feature activation | +5.43e+00 |
Pos logit contributions | |
package | +0.012 |
meal | +0.012 |
tray | +0.012 |
idents | +0.012 |
mess | +0.011 |
Neg logit contributions | |
stretched | -0.011 |
ila | -0.010 |
eight | -0.009 |
La | -0.009 |
Tw | -0.009 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.94e+00 |
Pos logit contributions | |
stret | +0.176 |
tray | +0.165 |
package | +0.162 |
idents | +0.161 |
dining | +0.156 |
Neg logit contributions | |
stretched | -0.151 |
ila | -0.131 |
Tw | -0.123 |
Tweet | -0.123 |
eight | -0.122 |
Token | treat |
Token ID | 2190 |
Feature activation | +1.12e+00 |
Pos logit contributions | |
idents | +0.081 |
doorstep | +0.077 |
aman | +0.073 |
stret | +0.072 |
isions | +0.072 |
Neg logit contributions | |
ila | -0.092 |
Tw | -0.084 |
stretched | -0.082 |
Tweet | -0.081 |
mag | -0.079 |
Token | it |
Token ID | 340 |
Feature activation | +1.43e+00 |
Pos logit contributions | |
idents | +0.044 |
package | +0.042 |
stret | +0.041 |
batch | +0.040 |
meal | +0.040 |
Neg logit contributions | |
ila | -0.022 |
stretched | -0.021 |
Tw | -0.020 |
eight | -0.019 |
Tweet | -0.017 |
Token | had |
Token ID | 550 |
Feature activation | +2.68e-01 |
Pos logit contributions | |
idents | +0.047 |
package | +0.047 |
meal | +0.046 |
tray | +0.046 |
batch | +0.045 |
Neg logit contributions | |
ila | -0.036 |
stretched | -0.036 |
eight | -0.033 |
Tw | -0.033 |
Tweet | -0.033 |
Token | found |
Token ID | 1043 |
Feature activation | -1.64e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
tray | +0.008 |
batch | +0.008 |
meal | +0.008 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.008 |
Tw | -0.007 |
eight | -0.007 |
Tweet | -0.007 |
Token | big |
Token ID | 1263 |
Feature activation | +2.47e+00 |
Pos logit contributions | |
idents | +0.001 |
package | +0.001 |
meal | +0.001 |
tray | +0.001 |
batch | +0.001 |
Neg logit contributions | |
ila | -0.000 |
stretched | -0.000 |
Tw | -0.000 |
eight | -0.000 |
Tweet | -0.000 |
Token | tree |
Token ID | 5509 |
Feature activation | -5.59e-01 |
Pos logit contributions | |
idents | +0.098 |
keeping | +0.094 |
isions | +0.093 |
dinners | +0.092 |
ksh | +0.090 |
Neg logit contributions | |
eight | -0.045 |
Tw | -0.045 |
stretched | -0.043 |
ila | -0.042 |
Tweet | -0.040 |
Token | with |
Token ID | 351 |
Feature activation | -1.40e-01 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.012 |
Tw | +0.011 |
eight | +0.011 |
Tweet | +0.011 |
Neg logit contributions | |
idents | -0.021 |
package | -0.020 |
meal | -0.020 |
stret | -0.019 |
dinners | -0.019 |
Token | lots |
Token ID | 6041 |
Feature activation | +1.72e+00 |
Pos logit contributions | |
ila | +0.004 |
stretched | +0.003 |
La | +0.003 |
Tw | +0.003 |
Tweet | +0.003 |
Neg logit contributions | |
package | -0.005 |
meal | -0.004 |
tray | -0.004 |
batch | -0.004 |
sandwiches | -0.004 |
Token | of |
Token ID | 286 |
Feature activation | -2.74e-01 |
Pos logit contributions | |
package | +0.066 |
meal | +0.064 |
batch | +0.064 |
tray | +0.063 |
cake | +0.063 |
Neg logit contributions | |
La | -0.034 |
ila | -0.033 |
Tw | -0.032 |
Chip | -0.031 |
eight | -0.030 |
Token | y |
Token ID | 331 |
Feature activation | +4.96e+00 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
Tw | +0.007 |
La | +0.006 |
eight | +0.006 |
Neg logit contributions | |
package | -0.009 |
idents | -0.009 |
meal | -0.009 |
tray | -0.009 |
batch | -0.009 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.57e+00 |
Pos logit contributions | |
stret | +0.160 |
package | +0.154 |
idents | +0.154 |
tray | +0.153 |
dining | +0.148 |
Neg logit contributions | |
stretched | -0.145 |
eight | -0.129 |
ila | -0.124 |
Tweet | -0.117 |
Tw | -0.114 |
Token | berries |
Token ID | 36322 |
Feature activation | +2.26e+00 |
Pos logit contributions | |
idents | +0.084 |
doorstep | +0.078 |
dining | +0.074 |
aman | +0.072 |
tray | +0.072 |
Neg logit contributions | |
ila | -0.069 |
Tweet | -0.069 |
eight | -0.068 |
stretched | -0.067 |
Tw | -0.064 |
Token | . |
Token ID | 13 |
Feature activation | +3.45e-01 |
Pos logit contributions | |
idents | +0.101 |
package | +0.096 |
batch | +0.094 |
stret | +0.094 |
isions | +0.092 |
Neg logit contributions | |
ila | -0.034 |
stretched | -0.034 |
Tweet | -0.027 |
eight | -0.026 |
Tw | -0.026 |
Token | Blue |
Token ID | 4518 |
Feature activation | +3.92e-01 |
Pos logit contributions | |
package | +0.015 |
idents | +0.015 |
tray | +0.014 |
meal | +0.014 |
batch | +0.014 |
Neg logit contributions | |
stretched | -0.005 |
Tweet | -0.005 |
Tw | -0.005 |
ila | -0.005 |
eight | -0.005 |
Token | rushed |
Token ID | 13999 |
Feature activation | +9.51e-01 |
Pos logit contributions | |
idents | +0.013 |
package | +0.013 |
meal | +0.013 |
batch | +0.013 |
tray | +0.013 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
La | -0.009 |
Tw | -0.008 |
eight | -0.008 |
Token | ches |
Token ID | 2052 |
Feature activation | -2.50e+00 |
Pos logit contributions | |
stretched | +0.051 |
Does | +0.049 |
Some | +0.049 |
La | +0.048 |
People | +0.048 |
Neg logit contributions | |
package | -0.021 |
batch | -0.020 |
tray | -0.020 |
meal | -0.019 |
idents | -0.019 |
Token | , |
Token ID | 11 |
Feature activation | +3.72e-01 |
Pos logit contributions | |
stretched | +0.073 |
ila | +0.072 |
La | +0.068 |
Tw | +0.068 |
Tweet | +0.068 |
Neg logit contributions | |
idents | -0.072 |
package | -0.071 |
meal | -0.070 |
tray | -0.070 |
batch | -0.068 |
Token | I |
Token ID | 314 |
Feature activation | -2.88e-01 |
Pos logit contributions | |
package | +0.012 |
idents | +0.012 |
tray | +0.012 |
meal | +0.012 |
batch | +0.012 |
Neg logit contributions | |
stretched | -0.009 |
ila | -0.009 |
La | -0.009 |
Tweet | -0.009 |
Tw | -0.008 |
Token | found |
Token ID | 1043 |
Feature activation | -5.95e-01 |
Pos logit contributions | |
stretched | +0.008 |
ila | +0.007 |
Tw | +0.007 |
eight | +0.007 |
Tweet | +0.007 |
Neg logit contributions | |
idents | -0.009 |
package | -0.009 |
meal | -0.009 |
tray | -0.009 |
batch | -0.009 |
Token | the |
Token ID | 262 |
Feature activation | -3.38e-01 |
Pos logit contributions | |
stretched | +0.014 |
ila | +0.014 |
Tweet | +0.013 |
La | +0.013 |
Tw | +0.013 |
Neg logit contributions | |
package | -0.021 |
idents | -0.020 |
meal | -0.020 |
tray | -0.020 |
batch | -0.020 |
Token | y |
Token ID | 331 |
Feature activation | +5.13e+00 |
Pos logit contributions | |
ila | +0.009 |
stretched | +0.009 |
eight | +0.009 |
Tw | +0.009 |
La | +0.008 |
Neg logit contributions | |
idents | -0.011 |
meal | -0.010 |
package | -0.010 |
batch | -0.010 |
tray | -0.010 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.67e+00 |
Pos logit contributions | |
stret | +0.205 |
ics | +0.169 |
idents | +0.164 |
dining | +0.162 |
tray | +0.159 |
Neg logit contributions | |
stretched | -0.113 |
ila | -0.103 |
ubby | -0.102 |
eight | -0.101 |
ib | -0.090 |
Token | paper |
Token ID | 3348 |
Feature activation | -4.06e-01 |
Pos logit contributions | |
stret | +0.092 |
idents | +0.091 |
aman | +0.088 |
ics | +0.087 |
doorstep | +0.085 |
Neg logit contributions | |
ila | -0.063 |
eight | -0.059 |
Tw | -0.057 |
blue | -0.052 |
I | -0.052 |
Token | for |
Token ID | 329 |
Feature activation | +1.97e+00 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.011 |
Tw | +0.010 |
eight | +0.010 |
Tweet | +0.010 |
Neg logit contributions | |
idents | -0.012 |
package | -0.012 |
meal | -0.012 |
batch | -0.012 |
tray | -0.012 |
Token | us |
Token ID | 514 |
Feature activation | +1.17e+00 |
Pos logit contributions | |
idents | +0.066 |
package | +0.065 |
meal | +0.065 |
tray | +0.064 |
batch | +0.063 |
Neg logit contributions | |
stretched | -0.049 |
ila | -0.049 |
Tw | -0.045 |
Tweet | -0.045 |
eight | -0.044 |
Token | ! |
Token ID | 0 |
Feature activation | +2.77e-01 |
Pos logit contributions | |
idents | +0.038 |
package | +0.038 |
meal | +0.037 |
tray | +0.037 |
batch | +0.036 |
Neg logit contributions | |
ila | -0.030 |
stretched | -0.030 |
Tw | -0.028 |
Tweet | -0.027 |
eight | -0.027 |
Token | opened |
Token ID | 4721 |
Feature activation | -2.45e+00 |
Pos logit contributions | |
stretched | +0.051 |
Tweet | +0.049 |
La | +0.046 |
Leave | +0.046 |
ila | +0.046 |
Neg logit contributions | |
package | -0.041 |
meal | -0.039 |
batch | -0.038 |
tray | -0.038 |
envelope | -0.038 |
Token | it |
Token ID | 340 |
Feature activation | +1.21e+00 |
Pos logit contributions | |
ila | +0.044 |
stretched | +0.044 |
Tw | +0.041 |
eight | +0.040 |
Tweet | +0.038 |
Neg logit contributions | |
idents | -0.099 |
package | -0.097 |
meal | -0.096 |
tray | -0.095 |
batch | -0.094 |
Token | and |
Token ID | 290 |
Feature activation | -6.55e-02 |
Pos logit contributions | |
idents | +0.042 |
package | +0.041 |
meal | +0.041 |
tray | +0.040 |
batch | +0.040 |
Neg logit contributions | |
ila | -0.028 |
stretched | -0.028 |
Tw | -0.026 |
eight | -0.026 |
Tweet | -0.025 |
Token | all |
Token ID | 477 |
Feature activation | +4.68e-01 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tw | +0.002 |
eight | +0.002 |
Tweet | +0.001 |
Neg logit contributions | |
idents | -0.002 |
package | -0.002 |
meal | -0.002 |
tray | -0.002 |
batch | -0.002 |
Token | the |
Token ID | 262 |
Feature activation | -2.79e-01 |
Pos logit contributions | |
package | +0.016 |
idents | +0.016 |
meal | +0.016 |
tray | +0.015 |
batch | +0.015 |
Neg logit contributions | |
ila | -0.011 |
stretched | -0.011 |
Tw | -0.010 |
Tweet | -0.010 |
eight | -0.010 |
Token | y |
Token ID | 331 |
Feature activation | +4.86e+00 |
Pos logit contributions | |
stretched | +0.009 |
ila | +0.009 |
eight | +0.009 |
Tw | +0.009 |
Tweet | +0.009 |
Neg logit contributions | |
idents | -0.007 |
batch | -0.006 |
package | -0.006 |
tray | -0.006 |
meal | -0.006 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.42e+00 |
Pos logit contributions | |
stret | +0.176 |
tray | +0.161 |
package | +0.156 |
idents | +0.151 |
advertisement | +0.146 |
Neg logit contributions | |
stretched | -0.122 |
ila | -0.103 |
ubby | -0.094 |
eight | -0.093 |
ib | -0.092 |
Token | treats |
Token ID | 18432 |
Feature activation | +1.75e+00 |
Pos logit contributions | |
idents | +0.052 |
aman | +0.045 |
stret | +0.044 |
doorstep | +0.043 |
isions | +0.043 |
Neg logit contributions | |
ila | -0.084 |
stretched | -0.082 |
eight | -0.082 |
he | -0.079 |
Tw | -0.079 |
Token | came |
Token ID | 1625 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
idents | +0.071 |
stret | +0.068 |
batch | +0.067 |
isions | +0.066 |
package | +0.066 |
Neg logit contributions | |
ila | -0.031 |
stretched | -0.028 |
Tw | -0.024 |
eight | -0.024 |
she | -0.023 |
Token | t |
Token ID | 256 |
Feature activation | +1.60e+00 |
Pos logit contributions | |
stretched | +0.001 |
ila | +0.001 |
eight | +0.001 |
Tw | +0.001 |
Tweet | +0.001 |
Neg logit contributions | |
idents | -0.001 |
meal | -0.001 |
batch | -0.001 |
package | -0.001 |
tray | -0.001 |
Token | umbling |
Token ID | 14739 |
Feature activation | +1.98e+00 |
Pos logit contributions | |
meal | +0.052 |
tray | +0.051 |
idents | +0.050 |
package | +0.049 |
stret | +0.048 |
Neg logit contributions | |
ila | -0.044 |
stretched | -0.043 |
Tweet | -0.038 |
A | -0.038 |
Tw | -0.037 |
Token | " |
Token ID | 1 |
Feature activation | +5.87e-01 |
Pos logit contributions | |
stretched | +0.004 |
ila | +0.004 |
Tweet | +0.003 |
Tw | +0.003 |
eight | +0.003 |
Neg logit contributions | |
package | -0.006 |
tray | -0.006 |
idents | -0.006 |
batch | -0.006 |
stret | -0.006 |
Token | This |
Token ID | 1212 |
Feature activation | +2.79e-01 |
Pos logit contributions | |
package | +0.020 |
meal | +0.020 |
batch | +0.020 |
tray | +0.020 |
idents | +0.019 |
Neg logit contributions | |
stretched | -0.014 |
ila | -0.014 |
eight | -0.012 |
Tw | -0.012 |
Tweet | -0.012 |
Token | cake |
Token ID | 12187 |
Feature activation | +2.40e+00 |
Pos logit contributions | |
idents | +0.009 |
stret | +0.009 |
keeping | +0.009 |
isions | +0.009 |
ksh | +0.008 |
Neg logit contributions | |
ila | -0.007 |
eight | -0.006 |
Tw | -0.006 |
I | -0.006 |
One | -0.006 |
Token | is |
Token ID | 318 |
Feature activation | +1.96e+00 |
Pos logit contributions | |
stret | +0.085 |
idents | +0.079 |
package | +0.077 |
meal | +0.076 |
doorstep | +0.076 |
Neg logit contributions | |
ila | -0.056 |
Tw | -0.050 |
I | -0.049 |
eight | -0.048 |
stretched | -0.047 |
Token | so |
Token ID | 523 |
Feature activation | +1.64e+00 |
Pos logit contributions | |
idents | +0.059 |
stret | +0.056 |
package | +0.056 |
tray | +0.055 |
batch | +0.054 |
Neg logit contributions | |
ila | -0.053 |
stretched | -0.052 |
Tw | -0.052 |
eight | -0.052 |
Tweet | -0.048 |
Token | y |
Token ID | 331 |
Feature activation | +4.87e+00 |
Pos logit contributions | |
idents | +0.039 |
package | +0.038 |
meal | +0.037 |
tray | +0.037 |
batch | +0.037 |
Neg logit contributions | |
ila | -0.057 |
stretched | -0.057 |
Tw | -0.054 |
eight | -0.053 |
Tweet | -0.053 |
Token | ummy |
Token ID | 13513 |
Feature activation | +2.12e+00 |
Pos logit contributions | |
stret | +0.158 |
package | +0.145 |
idents | +0.144 |
tray | +0.142 |
dining | +0.138 |
Neg logit contributions | |
stretched | -0.135 |
ila | -0.133 |
Tw | -0.122 |
Tweet | -0.121 |
eight | -0.119 |
Token | , |
Token ID | 11 |
Feature activation | -2.74e-02 |
Pos logit contributions | |
idents | +0.072 |
package | +0.068 |
tray | +0.068 |
stret | +0.068 |
doorstep | +0.066 |
Neg logit contributions | |
ila | -0.050 |
Tw | -0.046 |
eight | -0.045 |
stretched | -0.044 |
I | -0.043 |
Token | Isabel |
Token ID | 41838 |
Feature activation | -2.36e+00 |
Pos logit contributions | |
ila | +0.001 |
stretched | +0.001 |
Tw | +0.001 |
eight | +0.001 |
Tweet | +0.001 |
Neg logit contributions | |
idents | -0.001 |
package | -0.001 |
meal | -0.001 |
tray | -0.001 |
batch | -0.001 |
Token | le |
Token ID | 293 |
Feature activation | -1.17e-01 |
Pos logit contributions | |
stretched | +0.087 |
eight | +0.081 |
Tweet | +0.081 |
ila | +0.080 |
La | +0.080 |
Neg logit contributions | |
package | -0.050 |
meal | -0.050 |
batch | -0.049 |
tray | -0.048 |
idents | -0.047 |
Token | . |
Token ID | 13 |
Feature activation | +1.29e-01 |
Pos logit contributions | |
ila | +0.003 |
stretched | +0.003 |
Tw | +0.003 |
eight | +0.003 |
Tweet | +0.003 |
Neg logit contributions | |
idents | -0.003 |
package | -0.003 |
meal | -0.003 |
tray | -0.003 |
batch | -0.003 |
Token | enjoyed |
Token ID | 8359 |
Feature activation | +1.70e-01 |
Pos logit contributions | |
idents | +0.001 |
package | +0.001 |
batch | +0.001 |
meal | +0.001 |
tray | +0.001 |
Neg logit contributions | |
stretched | -0.001 |
ila | -0.001 |
Tweet | -0.001 |
eight | -0.001 |
Tw | -0.001 |
Token | her |
Token ID | 607 |
Feature activation | -1.36e+00 |
Pos logit contributions | |
meal | +0.005 |
batch | +0.005 |
dinner | +0.005 |
idents | +0.005 |
package | +0.005 |
Neg logit contributions | |
ila | -0.005 |
stretched | -0.004 |
Tweet | -0.004 |
Tw | -0.004 |
eight | -0.004 |
Token | special |
Token ID | 2041 |
Feature activation | +4.88e-01 |
Pos logit contributions | |
ila | +0.046 |
stretched | +0.046 |
Tw | +0.044 |
eight | +0.044 |
Tweet | +0.043 |
Neg logit contributions | |
idents | -0.033 |
package | -0.033 |
meal | -0.032 |
tray | -0.031 |
batch | -0.031 |
Token | juice |
Token ID | 13135 |
Feature activation | +1.68e+00 |
Pos logit contributions | |
idents | +0.010 |
package | +0.010 |
stret | +0.009 |
tray | +0.009 |
isions | +0.009 |
Neg logit contributions | |
stretched | -0.017 |
ila | -0.017 |
eight | -0.017 |
Tw | -0.017 |
Tweet | -0.016 |
Token | ! |
Token ID | 0 |
Feature activation | +2.37e-01 |
Pos logit contributions | |
idents | +0.059 |
package | +0.058 |
meal | +0.056 |
tray | +0.056 |
batch | +0.056 |
Neg logit contributions | |
stretched | -0.038 |
ila | -0.037 |
Tw | -0.036 |
eight | -0.034 |
Tweet | -0.034 |
Token | <|endoftext|> |
Token ID | 50256 |
Feature activation | +1.62e+00 |
Pos logit contributions | |
package | +0.007 |
idents | +0.007 |
batch | +0.006 |
doorstep | +0.006 |
amas | +0.006 |
Neg logit contributions | |
Tweet | -0.007 |
ila | -0.006 |
stretched | -0.006 |
Leave | -0.006 |
Rib | -0.006 |
Token | Once |
Token ID | 7454 |
Feature activation | +1.82e-01 |
Pos logit contributions | |
package | +0.057 |
meal | +0.055 |
tray | +0.055 |
refrigerator | +0.054 |
idents | +0.053 |
Neg logit contributions | |
stretched | -0.038 |
Tweet | -0.034 |
eight | -0.034 |
ila | -0.034 |
La | -0.031 |
Token | upon |
Token ID | 2402 |
Feature activation | -7.47e-01 |
Pos logit contributions | |
package | +0.006 |
idents | +0.006 |
tray | +0.006 |
batch | +0.006 |
meal | +0.006 |
Neg logit contributions | |
Tweet | -0.004 |
stretched | -0.004 |
La | -0.004 |
Tw | -0.004 |
Leave | -0.004 |
Token | a |
Token ID | 257 |
Feature activation | +2.95e-01 |
Pos logit contributions | |
ila | +0.015 |
stretched | +0.015 |
Tw | +0.014 |
eight | +0.014 |
Tweet | +0.014 |
Neg logit contributions | |
idents | -0.028 |
package | -0.028 |
meal | -0.027 |
tray | -0.027 |
batch | -0.027 |
Token | time |
Token ID | 640 |
Feature activation | -4.31e-01 |
Pos logit contributions | |
idents | +0.013 |
package | +0.012 |
keeping | +0.012 |
goodies | +0.011 |
meal | +0.011 |
Neg logit contributions | |
ila | -0.005 |
stretched | -0.005 |
La | -0.005 |
One | -0.005 |
eight | -0.005 |
Token | there |
Token ID | 612 |
Feature activation | -8.98e-01 |
Pos logit contributions | |
ila | +0.010 |
stretched | +0.009 |
Tw | +0.009 |
eight | +0.009 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.016 |
package | -0.016 |
meal | -0.015 |
batch | -0.015 |
tray | -0.015 |
Token | The |
Token ID | 383 |
Feature activation | -1.77e+00 |
Pos logit contributions | |
stretched | +0.004 |
ila | +0.004 |
Tweet | +0.003 |
La | +0.003 |
known | +0.003 |
Neg logit contributions | |
refrigerator | -0.005 |
idents | -0.005 |
package | -0.005 |
meal | -0.004 |
dining | -0.004 |
Token | hair |
Token ID | 4190 |
Feature activation | -3.10e+00 |
Pos logit contributions | |
ila | +0.049 |
stretched | +0.048 |
Tw | +0.045 |
La | +0.045 |
eight | +0.044 |
Neg logit contributions | |
package | -0.055 |
idents | -0.055 |
meal | -0.054 |
tray | -0.054 |
batch | -0.053 |
Token | helps |
Token ID | 5419 |
Feature activation | -4.50e-01 |
Pos logit contributions | |
stretched | +0.091 |
ila | +0.089 |
Tw | +0.085 |
eight | +0.085 |
Tweet | +0.085 |
Neg logit contributions | |
idents | -0.090 |
package | -0.090 |
tray | -0.087 |
meal | -0.087 |
batch | -0.085 |
Token | him |
Token ID | 683 |
Feature activation | +7.87e-02 |
Pos logit contributions | |
stretched | +0.014 |
ila | +0.014 |
Tweet | +0.013 |
La | +0.013 |
Tw | +0.013 |
Neg logit contributions | |
package | -0.012 |
meal | -0.012 |
idents | -0.012 |
tray | -0.012 |
batch | -0.011 |
Token | stay |
Token ID | 2652 |
Feature activation | +5.60e-02 |
Pos logit contributions | |
idents | +0.002 |
meal | +0.002 |
package | +0.002 |
tray | +0.002 |
stret | +0.002 |
Neg logit contributions | |
stretched | -0.002 |
ila | -0.002 |
Tweet | -0.002 |
La | -0.002 |
eight | -0.002 |
Token | warm |
Token ID | 5814 |
Feature activation | +1.58e+00 |
Pos logit contributions | |
idents | +0.002 |
package | +0.002 |
batch | +0.002 |
tray | +0.002 |
meal | +0.002 |
Neg logit contributions | |
ila | -0.002 |
stretched | -0.002 |
Tw | -0.002 |
eight | -0.001 |
Tweet | -0.001 |
Token | . |
Token ID | 13 |
Feature activation | +9.42e-02 |
Pos logit contributions | |
idents | +0.055 |
package | +0.054 |
tray | +0.053 |
batch | +0.053 |
meal | +0.052 |
Neg logit contributions | |
ila | -0.036 |
stretched | -0.036 |
Tw | -0.034 |
eight | -0.033 |
Tweet | -0.033 |
Token | But |
Token ID | 887 |
Feature activation | -1.28e+00 |
Pos logit contributions | |
refrigerator | +0.003 |
idents | +0.003 |
package | +0.003 |
meal | +0.003 |
ksh | +0.003 |
Neg logit contributions | |
stretched | -0.002 |
ila | -0.002 |
Tweet | -0.002 |
La | -0.002 |
known | -0.002 |
Token | not |
Token ID | 407 |
Feature activation | +5.64e-01 |
Pos logit contributions | |
stretched | +0.031 |
ila | +0.028 |
La | +0.028 |
Tweet | +0.028 |
eight | +0.025 |
Neg logit contributions | |
package | -0.046 |
idents | -0.046 |
meal | -0.045 |
dinner | -0.044 |
tray | -0.044 |
Token | all |
Token ID | 477 |
Feature activation | -1.08e+00 |
Pos logit contributions | |
package | +0.018 |
idents | +0.018 |
meal | +0.018 |
tray | +0.017 |
batch | +0.017 |
Neg logit contributions | |
ila | -0.015 |
stretched | -0.015 |
Tweet | -0.014 |
La | -0.014 |
Tw | -0.014 |
Token | elephants |
Token ID | 25368 |
Feature activation | -1.55e+00 |
Pos logit contributions | |
stretched | +0.025 |
ila | +0.025 |
Tweet | +0.023 |
Tw | +0.023 |
La | +0.023 |
Neg logit contributions | |
package | -0.038 |
idents | -0.038 |
meal | -0.037 |
tray | -0.037 |
batch | -0.036 |
Token | and |
Token ID | 290 |
Feature activation | -6.67e-01 |
Pos logit contributions | |
package | +0.019 |
idents | +0.019 |
meal | +0.019 |
tray | +0.019 |
batch | +0.018 |
Neg logit contributions | |
ila | -0.010 |
stretched | -0.009 |
Tw | -0.009 |
eight | -0.009 |
Tweet | -0.009 |
Token | slid |
Token ID | 27803 |
Feature activation | -5.98e-01 |
Pos logit contributions | |
stretched | +0.015 |
ila | +0.015 |
Tw | +0.014 |
eight | +0.014 |
La | +0.014 |
Neg logit contributions | |
idents | -0.024 |
package | -0.023 |
meal | -0.023 |
tray | -0.023 |
batch | -0.023 |
Token | down |
Token ID | 866 |
Feature activation | -5.94e-01 |
Pos logit contributions | |
stretched | +0.021 |
ila | +0.021 |
Tweet | +0.020 |
Tw | +0.019 |
eight | +0.019 |
Neg logit contributions | |
idents | -0.014 |
package | -0.014 |
tray | -0.014 |
batch | -0.013 |
meal | -0.013 |
Token | together |
Token ID | 1978 |
Feature activation | -4.92e-01 |
Pos logit contributions | |
ila | +0.013 |
stretched | +0.013 |
Tweet | +0.013 |
La | +0.013 |
Tw | +0.012 |
Neg logit contributions | |
tray | -0.020 |
package | -0.020 |
idents | -0.020 |
batch | -0.020 |
meal | -0.019 |
Token | . |
Token ID | 13 |
Feature activation | +5.96e-02 |
Pos logit contributions | |
ila | +0.012 |
stretched | +0.012 |
Tw | +0.011 |
eight | +0.011 |
Tweet | +0.011 |
Neg logit contributions | |
idents | -0.017 |
package | -0.017 |
meal | -0.017 |
tray | -0.016 |
batch | -0.016 |
Token | They |
Token ID | 1119 |
Feature activation | +1.86e+00 |
Pos logit contributions | |
idents | +0.002 |
batch | +0.002 |
tray | +0.002 |
package | +0.002 |
meal | +0.002 |
Neg logit contributions | |
stretched | -0.002 |
ila | -0.002 |
La | -0.001 |
eight | -0.001 |
Tweet | -0.001 |
Token | laughed |
Token ID | 13818 |
Feature activation | +5.71e-01 |
Pos logit contributions | |
tray | +0.051 |
idents | +0.051 |
package | +0.051 |
meal | +0.050 |
batch | +0.050 |
Neg logit contributions | |
ila | -0.053 |
stretched | -0.053 |
Tw | -0.050 |
Tweet | -0.049 |
La | -0.049 |
Token | and |
Token ID | 290 |
Feature activation | +4.94e-01 |
Pos logit contributions | |
package | +0.020 |
idents | +0.020 |
meal | +0.019 |
tray | +0.019 |
batch | +0.019 |
Neg logit contributions | |
stretched | -0.014 |
ila | -0.013 |
Tw | -0.013 |
eight | -0.013 |
Tweet | -0.012 |
Token | had |
Token ID | 550 |
Feature activation | +2.25e-01 |
Pos logit contributions | |
idents | +0.016 |
package | +0.015 |
meal | +0.015 |
tray | +0.015 |
batch | +0.015 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.013 |
Tw | -0.012 |
eight | -0.012 |
Tweet | -0.012 |
Token | so |
Token ID | 523 |
Feature activation | +3.51e-01 |
Pos logit contributions | |
idents | +0.007 |
package | +0.007 |
tray | +0.007 |
meal | +0.007 |
batch | +0.007 |
Neg logit contributions | |
stretched | -0.006 |
ila | -0.006 |
eight | -0.006 |
Tw | -0.005 |
La | -0.005 |
Token | much |
Token ID | 881 |
Feature activation | +8.20e-01 |
Pos logit contributions | |
idents | +0.013 |
package | +0.013 |
meal | +0.013 |
tray | +0.013 |
batch | +0.013 |
Neg logit contributions | |
ila | -0.007 |
stretched | -0.007 |
Tw | -0.007 |
eight | -0.007 |
Tweet | -0.007 |
Token |
|
Token ID | 198 |
Feature activation | -1.06e-01 |
Pos logit contributions | |
stretched | +0.001 |
brav | +0.001 |
eight | +0.001 |
Leave | +0.001 |
racer | +0.001 |
Neg logit contributions | |
batch | -0.001 |
package | -0.001 |
meal | -0.001 |
tray | -0.001 |
ics | -0.001 |
Token | She |
Token ID | 3347 |
Feature activation | -9.05e-01 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.002 |
eight | +0.002 |
Tw | +0.002 |
Chip | +0.002 |
Neg logit contributions | |
package | -0.004 |
batch | -0.004 |
tray | -0.004 |
idents | -0.003 |
aman | -0.003 |
Token | lly |
Token ID | 12810 |
Feature activation | -1.30e+00 |
Pos logit contributions | |
stretched | +0.026 |
ila | +0.025 |
Tweet | +0.023 |
Tw | +0.023 |
La | +0.023 |
Neg logit contributions | |
package | -0.028 |
batch | -0.027 |
tray | -0.027 |
idents | -0.026 |
meal | -0.026 |
Token | took |
Token ID | 1718 |
Feature activation | -5.81e-01 |
Pos logit contributions | |
stretched | +0.035 |
ila | +0.034 |
Tw | +0.032 |
eight | +0.032 |
Tweet | +0.032 |
Neg logit contributions | |
package | -0.041 |
idents | -0.041 |
tray | -0.040 |
meal | -0.039 |
batch | -0.039 |
Token | out |
Token ID | 503 |
Feature activation | -4.91e-01 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.013 |
Tweet | +0.012 |
Tw | +0.012 |
eight | +0.012 |
Neg logit contributions | |
package | -0.021 |
meal | -0.020 |
idents | -0.020 |
tray | -0.020 |
batch | -0.020 |
Token | some |
Token ID | 617 |
Feature activation | +1.94e+00 |
Pos logit contributions | |
ila | +0.012 |
stretched | +0.012 |
Tweet | +0.012 |
Tw | +0.011 |
La | +0.011 |
Neg logit contributions | |
package | -0.017 |
tray | -0.016 |
meal | -0.016 |
batch | -0.016 |
sandwiches | -0.016 |
Token | paints |
Token ID | 32281 |
Feature activation | -1.42e+00 |
Pos logit contributions | |
package | +0.060 |
tray | +0.058 |
meal | +0.058 |
batch | +0.057 |
idents | +0.057 |
Neg logit contributions | |
ila | -0.055 |
stretched | -0.054 |
Tw | -0.052 |
Tweet | -0.051 |
eight | -0.050 |
Token | and |
Token ID | 290 |
Feature activation | -1.57e-01 |
Pos logit contributions | |
ila | +0.035 |
stretched | +0.034 |
Tw | +0.032 |
eight | +0.032 |
Tweet | +0.031 |
Neg logit contributions | |
idents | -0.048 |
package | -0.047 |
meal | -0.047 |
tray | -0.046 |
batch | -0.046 |
Token | started |
Token ID | 2067 |
Feature activation | -2.22e+00 |
Pos logit contributions | |
stretched | +0.005 |
ila | +0.005 |
Tw | +0.005 |
Tweet | +0.005 |
Bo | +0.005 |
Neg logit contributions | |
package | -0.004 |
cake | -0.004 |
tray | -0.004 |
batch | -0.004 |
cakes | -0.004 |
Token | painting |
Token ID | 12036 |
Feature activation | -9.53e-01 |
Pos logit contributions | |
ila | +0.064 |
stretched | +0.063 |
Tweet | +0.060 |
Tw | +0.059 |
eight | +0.057 |
Neg logit contributions | |
package | -0.065 |
idents | -0.064 |
batch | -0.064 |
tray | -0.063 |
meal | -0.063 |
Token | . |
Token ID | 13 |
Feature activation | +2.75e-01 |
Pos logit contributions | |
ila | +0.023 |
stretched | +0.023 |
Tweet | +0.022 |
eight | +0.022 |
Tw | +0.021 |
Neg logit contributions | |
package | -0.032 |
idents | -0.032 |
tray | -0.031 |
batch | -0.031 |
meal | -0.031 |
Token | cleaning |
Token ID | 12724 |
Feature activation | +2.23e+00 |
Pos logit contributions | |
ila | +0.002 |
stretched | +0.002 |
Tw | +0.002 |
eight | +0.002 |
La | +0.002 |
Neg logit contributions | |
idents | -0.003 |
package | -0.003 |
batch | -0.003 |
tray | -0.003 |
meal | -0.003 |
Token | the |
Token ID | 262 |
Feature activation | +7.97e-01 |
Pos logit contributions | |
idents | +0.088 |
meal | +0.085 |
stret | +0.084 |
package | +0.084 |
batch | +0.083 |
Neg logit contributions | |
ila | -0.041 |
Tw | -0.037 |
stretched | -0.035 |
eight | -0.032 |
La | -0.032 |
Token | furniture |
Token ID | 13091 |
Feature activation | +9.49e-01 |
Pos logit contributions | |
idents | +0.021 |
batch | +0.019 |
aman | +0.019 |
urger | +0.019 |
meal | +0.018 |
Neg logit contributions | |
eight | -0.024 |
Tw | -0.023 |
ila | -0.023 |
stretched | -0.023 |
A | -0.023 |
Token | . |
Token ID | 13 |
Feature activation | +7.33e-01 |
Pos logit contributions | |
idents | +0.038 |
package | +0.037 |
meal | +0.036 |
batch | +0.036 |
stret | +0.035 |
Neg logit contributions | |
ila | -0.017 |
stretched | -0.016 |
Tw | -0.015 |
eight | -0.015 |
La | -0.013 |
Token | " |
Token ID | 366 |
Feature activation | -8.10e-01 |
Pos logit contributions | |
idents | +0.031 |
package | +0.030 |
tray | +0.030 |
meal | +0.030 |
batch | +0.029 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.012 |
Tweet | -0.011 |
Tw | -0.011 |
eight | -0.011 |
Token | Mom |
Token ID | 29252 |
Feature activation | +1.62e+00 |
Pos logit contributions | |
ila | +0.011 |
stretched | +0.010 |
Tweet | +0.009 |
La | +0.009 |
Tw | +0.009 |
Neg logit contributions | |
idents | -0.037 |
meal | -0.036 |
package | -0.036 |
tray | -0.036 |
batch | -0.036 |
Token | my |
Token ID | 1820 |
Feature activation | -4.82e-01 |
Pos logit contributions | |
idents | +0.049 |
package | +0.048 |
tray | +0.047 |
meal | +0.047 |
batch | +0.046 |
Neg logit contributions | |
ila | -0.045 |
stretched | -0.045 |
Tw | -0.043 |
eight | -0.043 |
Tweet | -0.042 |
Token | , |
Token ID | 11 |
Feature activation | +7.86e-01 |
Pos logit contributions | |
ila | +0.009 |
Tw | +0.009 |
eight | +0.009 |
stretched | +0.009 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.018 |
tray | -0.018 |
meal | -0.018 |
batch | -0.018 |
stret | -0.018 |
Token | can |
Token ID | 460 |
Feature activation | -8.91e-01 |
Pos logit contributions | |
idents | +0.033 |
batch | +0.032 |
stret | +0.032 |
meal | +0.031 |
tray | +0.031 |
Neg logit contributions | |
ila | -0.013 |
eight | -0.012 |
Tw | -0.012 |
stretched | -0.011 |
I | -0.010 |
Token | I |
Token ID | 314 |
Feature activation | -1.23e+00 |
Pos logit contributions | |
stretched | +0.020 |
ila | +0.019 |
La | +0.017 |
Tweet | +0.017 |
eight | +0.016 |
Neg logit contributions | |
idents | -0.034 |
package | -0.033 |
tray | -0.032 |
meal | -0.032 |
batch | -0.031 |
Token | help |
Token ID | 1037 |
Feature activation | -2.10e-02 |
Pos logit contributions | |
ila | +0.030 |
stretched | +0.030 |
Tw | +0.027 |
eight | +0.027 |
Tweet | +0.027 |
Neg logit contributions | |
package | -0.042 |
idents | -0.042 |
meal | -0.041 |
tray | -0.041 |
batch | -0.040 |
Token | house |
Token ID | 2156 |
Feature activation | +9.15e-01 |
Pos logit contributions | |
package | +0.028 |
tray | +0.027 |
meal | +0.027 |
idents | +0.026 |
refrigerator | +0.026 |
Neg logit contributions | |
ila | -0.027 |
stretched | -0.027 |
Tweet | -0.025 |
La | -0.025 |
Tw | -0.024 |
Token | . |
Token ID | 13 |
Feature activation | -3.13e-02 |
Pos logit contributions | |
idents | +0.030 |
package | +0.030 |
meal | +0.030 |
tray | +0.030 |
batch | +0.029 |
Neg logit contributions | |
ila | -0.023 |
stretched | -0.023 |
Tw | -0.021 |
eight | -0.021 |
Tweet | -0.021 |
Token | He |
Token ID | 679 |
Feature activation | -8.99e-01 |
Pos logit contributions | |
stretched | +0.001 |
La | +0.001 |
Tweet | +0.001 |
brav | +0.001 |
fastest | +0.001 |
Neg logit contributions | |
refrigerator | -0.001 |
package | -0.001 |
tray | -0.001 |
stret | -0.001 |
meal | -0.001 |
Token | was |
Token ID | 373 |
Feature activation | +2.39e-01 |
Pos logit contributions | |
stretched | +0.030 |
ila | +0.026 |
Tweet | +0.026 |
eight | +0.026 |
La | +0.025 |
Neg logit contributions | |
package | -0.022 |
tray | -0.021 |
meal | -0.021 |
mess | -0.021 |
batch | -0.020 |
Token | still |
Token ID | 991 |
Feature activation | -1.76e+00 |
Pos logit contributions | |
tray | +0.006 |
package | +0.006 |
meal | +0.006 |
mess | +0.006 |
dining | +0.006 |
Neg logit contributions | |
stretched | -0.007 |
ila | -0.007 |
Tweet | -0.007 |
Tw | -0.006 |
La | -0.006 |
Token | crying |
Token ID | 13774 |
Feature activation | -1.89e-01 |
Pos logit contributions | |
stretched | +0.049 |
ila | +0.049 |
Tweet | +0.047 |
Tw | +0.045 |
eight | +0.045 |
Neg logit contributions | |
package | -0.052 |
tray | -0.051 |
idents | -0.051 |
meal | -0.051 |
batch | -0.050 |
Token | . |
Token ID | 13 |
Feature activation | +8.21e-02 |
Pos logit contributions | |
ila | +0.006 |
Tweet | +0.006 |
stretched | +0.006 |
La | +0.006 |
eight | +0.006 |
Neg logit contributions | |
idents | -0.005 |
package | -0.005 |
tray | -0.005 |
refrigerator | -0.005 |
mess | -0.005 |
Token | Sara |
Token ID | 24799 |
Feature activation | -1.26e-01 |
Pos logit contributions | |
refrigerator | +0.002 |
package | +0.002 |
dining | +0.002 |
dessert | +0.002 |
meal | +0.002 |
Neg logit contributions | |
stretched | -0.002 |
Tweet | -0.002 |
La | -0.002 |
spun | -0.002 |
eight | -0.002 |
Token | called |
Token ID | 1444 |
Feature activation | -5.69e-01 |
Pos logit contributions | |
stretched | +0.004 |
ila | +0.003 |
Tweet | +0.003 |
Tw | +0.003 |
eight | +0.003 |
Neg logit contributions | |
package | -0.004 |
idents | -0.004 |
tray | -0.004 |
meal | -0.004 |
batch | -0.004 |
Token | out |
Token ID | 503 |
Feature activation | +1.51e-01 |
Pos logit contributions | |
ila | +0.016 |
stretched | +0.016 |
Tw | +0.015 |
eight | +0.015 |
Tweet | +0.015 |
Neg logit contributions | |
idents | -0.017 |
package | -0.017 |
batch | -0.016 |
meal | -0.016 |
tray | -0.016 |
Token | , |
Token ID | 11 |
Feature activation | +9.53e-01 |
Pos logit contributions | |
idents | +0.005 |
package | +0.005 |
meal | +0.005 |
tray | +0.005 |
batch | +0.005 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
Tw | -0.004 |
Tweet | -0.004 |
eight | -0.003 |
Token | collect |
Token ID | 2824 |
Feature activation | -1.67e+00 |
Pos logit contributions | |
stretched | +0.024 |
ila | +0.023 |
La | +0.023 |
Tw | +0.022 |
Tweet | +0.021 |
Neg logit contributions | |
package | -0.026 |
idents | -0.026 |
meal | -0.026 |
tray | -0.025 |
batch | -0.025 |
Token | it |
Token ID | 340 |
Feature activation | +6.15e-01 |
Pos logit contributions | |
stretched | +0.054 |
ila | +0.048 |
La | +0.048 |
Leave | +0.046 |
Tweet | +0.046 |
Neg logit contributions | |
package | -0.047 |
sandwiches | -0.046 |
cake | -0.046 |
meal | -0.046 |
batch | -0.044 |
Token | . |
Token ID | 13 |
Feature activation | +6.58e-02 |
Pos logit contributions | |
tray | +0.017 |
meal | +0.017 |
idents | +0.017 |
package | +0.017 |
batch | +0.016 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.018 |
La | -0.018 |
Tweet | -0.017 |
Leave | -0.017 |
Token | But |
Token ID | 887 |
Feature activation | -9.79e-01 |
Pos logit contributions | |
idents | +0.002 |
package | +0.002 |
tray | +0.002 |
meal | +0.002 |
refrigerator | +0.002 |
Neg logit contributions | |
ila | -0.002 |
stretched | -0.002 |
La | -0.002 |
eight | -0.001 |
Tweet | -0.001 |
Token | be |
Token ID | 307 |
Feature activation | -7.58e-01 |
Pos logit contributions | |
stretched | +0.027 |
ila | +0.025 |
La | +0.025 |
Tweet | +0.023 |
eight | +0.022 |
Neg logit contributions | |
idents | -0.033 |
package | -0.032 |
meal | -0.031 |
tray | -0.031 |
batch | -0.030 |
Token | careful |
Token ID | 8161 |
Feature activation | -4.21e-01 |
Pos logit contributions | |
stretched | +0.019 |
ila | +0.018 |
Tw | +0.017 |
Tweet | +0.017 |
eight | +0.017 |
Neg logit contributions | |
package | -0.026 |
idents | -0.025 |
tray | -0.025 |
meal | -0.025 |
batch | -0.025 |
Token | , |
Token ID | 11 |
Feature activation | +1.31e-01 |
Pos logit contributions | |
stretched | +0.012 |
ila | +0.012 |
La | +0.011 |
Tweet | +0.011 |
Tw | +0.011 |
Neg logit contributions | |
idents | -0.013 |
package | -0.012 |
tray | -0.012 |
meal | -0.012 |
batch | -0.012 |
Token | it |
Token ID | 340 |
Feature activation | +2.92e-01 |
Pos logit contributions | |
idents | +0.004 |
package | +0.004 |
meal | +0.004 |
tray | +0.004 |
envelope | +0.004 |
Neg logit contributions | |
stretched | -0.004 |
ila | -0.003 |
La | -0.003 |
Tweet | -0.003 |
Tw | -0.003 |
Token | 's |
Token ID | 338 |
Feature activation | +2.08e+00 |
Pos logit contributions | |
idents | +0.009 |
package | +0.009 |
meal | +0.009 |
tray | +0.009 |
batch | +0.009 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.008 |
Tw | -0.008 |
eight | -0.008 |
Tweet | -0.007 |
Token | in |
Token ID | 287 |
Feature activation | -8.34e-01 |
Pos logit contributions | |
idents | +0.066 |
package | +0.065 |
meal | +0.064 |
tray | +0.064 |
batch | +0.063 |
Neg logit contributions | |
ila | -0.055 |
stretched | -0.055 |
Tw | -0.051 |
eight | -0.051 |
Tweet | -0.051 |
Token | the |
Token ID | 262 |
Feature activation | -1.50e-02 |
Pos logit contributions | |
stretched | +0.017 |
ila | +0.017 |
Tw | +0.016 |
Tweet | +0.016 |
eight | +0.015 |
Neg logit contributions | |
package | -0.031 |
idents | -0.031 |
meal | -0.031 |
tray | -0.031 |
batch | -0.030 |
Token | y |
Token ID | 331 |
Feature activation | +4.18e+00 |
Pos logit contributions | |
meal | +0.009 |
package | +0.009 |
batch | +0.009 |
idents | +0.009 |
tray | +0.008 |
Neg logit contributions | |
ila | -0.014 |
stretched | -0.014 |
La | -0.013 |
Tw | -0.013 |
eight | -0.013 |
Token | ummy |
Token ID | 13513 |
Feature activation | +1.60e+00 |
Pos logit contributions | |
stret | +0.116 |
package | +0.115 |
tray | +0.114 |
idents | +0.112 |
meal | +0.108 |
Neg logit contributions | |
stretched | -0.133 |
ila | -0.127 |
eight | -0.119 |
Tweet | -0.116 |
Tw | -0.115 |
Token | looking |
Token ID | 2045 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
idents | +0.051 |
package | +0.049 |
tray | +0.048 |
batch | +0.048 |
meal | +0.047 |
Neg logit contributions | |
ila | -0.042 |
stretched | -0.041 |
Tw | -0.039 |
eight | -0.039 |
Tweet | -0.037 |
Token | ! |
Token ID | 0 |
Feature activation | +2.18e-01 |
Pos logit contributions | |
idents | +0.041 |
package | +0.041 |
meal | +0.040 |
tray | +0.040 |
batch | +0.039 |
Neg logit contributions | |
ila | -0.039 |
stretched | -0.038 |
Tw | -0.036 |
Tweet | -0.036 |
La | -0.036 |
Token | He |
Token ID | 679 |
Feature activation | -4.95e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
meal | +0.008 |
tray | +0.008 |
batch | +0.008 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
Tw | -0.004 |
eight | -0.004 |
Tweet | -0.004 |
Token | couldn |
Token ID | 3521 |
Feature activation | -2.46e+00 |
Pos logit contributions | |
ila | +0.014 |
stretched | +0.014 |
La | +0.013 |
Tw | +0.012 |
Tweet | +0.012 |
Neg logit contributions | |
batch | -0.015 |
meal | -0.015 |
package | -0.014 |
tray | -0.014 |
idents | -0.014 |
Token | 't |
Token ID | 470 |
Feature activation | -7.85e-01 |
Pos logit contributions | |
stretched | +0.050 |
ila | +0.047 |
La | +0.044 |
Tw | +0.042 |
eight | +0.042 |
Neg logit contributions | |
idents | -0.094 |
package | -0.092 |
tray | -0.092 |
meal | -0.091 |
batch | -0.091 |
Token | resist |
Token ID | 4180 |
Feature activation | -9.24e-01 |
Pos logit contributions | |
ila | +0.032 |
stretched | +0.031 |
La | +0.030 |
Tw | +0.030 |
eight | +0.029 |
Neg logit contributions | |
idents | -0.014 |
meal | -0.014 |
package | -0.013 |
tray | -0.013 |
doorstep | -0.012 |
Token | picking |
Token ID | 10868 |
Feature activation | +2.32e-01 |
Pos logit contributions | |
ila | +0.029 |
stretched | +0.028 |
Tweet | +0.028 |
La | +0.027 |
Tw | +0.027 |
Neg logit contributions | |
idents | -0.025 |
meal | -0.024 |
package | -0.024 |
tray | -0.024 |
batch | -0.023 |
Token | it |
Token ID | 340 |
Feature activation | +7.58e-01 |
Pos logit contributions | |
idents | +0.008 |
package | +0.008 |
meal | +0.008 |
tray | +0.008 |
batch | +0.008 |
Neg logit contributions | |
stretched | -0.005 |
ila | -0.005 |
Tw | -0.005 |
eight | -0.005 |
Tweet | -0.005 |
Token | up |
Token ID | 510 |
Feature activation | -6.99e-01 |
Pos logit contributions | |
stret | +0.026 |
package | +0.026 |
idents | +0.025 |
dining | +0.024 |
tray | +0.024 |
Neg logit contributions | |
stretched | -0.018 |
ila | -0.017 |
Tw | -0.017 |
eight | -0.016 |
Tweet | -0.016 |
Token | . |
Token ID | 13 |
Feature activation | -5.79e-01 |
Pos logit contributions | |
ila | +0.018 |
stretched | +0.018 |
Tw | +0.017 |
eight | +0.017 |
Tweet | +0.016 |
Neg logit contributions | |
idents | -0.024 |
package | -0.023 |
meal | -0.023 |
tray | -0.023 |
batch | -0.023 |
Token | I |
Token ID | 314 |
Feature activation | -8.13e-01 |
Pos logit contributions | |
stretched | +0.021 |
ila | +0.019 |
La | +0.018 |
Tweet | +0.018 |
known | +0.018 |
Neg logit contributions | |
idents | -0.015 |
package | -0.013 |
isions | -0.013 |
ksh | -0.013 |
tray | -0.011 |
Token | am |
Token ID | 716 |
Feature activation | -5.14e-01 |
Pos logit contributions | |
stretched | +0.022 |
ila | +0.020 |
eight | +0.020 |
La | +0.020 |
Tw | +0.019 |
Neg logit contributions | |
idents | -0.027 |
package | -0.026 |
meal | -0.025 |
tray | -0.025 |
batch | -0.025 |
Token | wise |
Token ID | 10787 |
Feature activation | -1.28e+00 |
Pos logit contributions | |
stretched | +0.016 |
ila | +0.015 |
Tweet | +0.015 |
La | +0.014 |
eight | +0.014 |
Neg logit contributions | |
package | -0.015 |
idents | -0.014 |
meal | -0.014 |
tray | -0.014 |
stret | -0.014 |
Token | and |
Token ID | 290 |
Feature activation | -6.63e-01 |
Pos logit contributions | |
ila | +0.031 |
stretched | +0.031 |
Tweet | +0.028 |
Tw | +0.028 |
eight | +0.028 |
Neg logit contributions | |
idents | -0.044 |
package | -0.044 |
meal | -0.043 |
tray | -0.043 |
batch | -0.042 |
Token | know |
Token ID | 760 |
Feature activation | -1.03e+00 |
Pos logit contributions | |
stretched | +0.020 |
La | +0.018 |
ila | +0.018 |
Tweet | +0.018 |
Tw | +0.017 |
Neg logit contributions | |
package | -0.022 |
idents | -0.021 |
envelope | -0.019 |
ours | -0.019 |
meal | -0.018 |
Token | the |
Token ID | 262 |
Feature activation | -1.29e+00 |
Pos logit contributions | |
stretched | +0.033 |
ila | +0.032 |
Tweet | +0.031 |
La | +0.030 |
Bo | +0.029 |
Neg logit contributions | |
package | -0.028 |
idents | -0.027 |
meal | -0.026 |
cake | -0.026 |
dinner | -0.026 |
Token | way |
Token ID | 835 |
Feature activation | -1.16e-01 |
Pos logit contributions | |
ila | +0.041 |
stretched | +0.041 |
Tweet | +0.038 |
Tw | +0.038 |
La | +0.038 |
Neg logit contributions | |
package | -0.038 |
meal | -0.035 |
tray | -0.034 |
idents | -0.034 |
mess | -0.033 |
Token | to |
Token ID | 284 |
Feature activation | +5.73e-01 |
Pos logit contributions | |
stretched | +0.004 |
ila | +0.004 |
Tweet | +0.004 |
La | +0.004 |
nine | +0.004 |
Neg logit contributions | |
package | -0.003 |
idents | -0.003 |
tray | -0.003 |
doorstep | -0.003 |
meal | -0.003 |
Token | grandma |
Token ID | 49890 |
Feature activation | -7.94e-01 |
Pos logit contributions | |
package | +0.016 |
meal | +0.015 |
dinner | +0.015 |
stret | +0.014 |
doorstep | +0.014 |
Neg logit contributions | |
stretched | -0.019 |
ila | -0.018 |
eight | -0.017 |
La | -0.017 |
Tweet | -0.016 |
Token | 's |
Token ID | 338 |
Feature activation | +1.24e+00 |
Pos logit contributions | |
stretched | +0.023 |
ila | +0.023 |
eight | +0.022 |
Tweet | +0.021 |
La | +0.021 |
Neg logit contributions | |
package | -0.024 |
idents | -0.024 |
meal | -0.023 |
tray | -0.023 |
batch | -0.022 |
Token | pretty |
Token ID | 2495 |
Feature activation | +3.39e-01 |
Pos logit contributions | |
idents | +0.061 |
package | +0.060 |
meal | +0.059 |
tray | +0.058 |
batch | +0.058 |
Neg logit contributions | |
ila | -0.058 |
stretched | -0.058 |
Tw | -0.055 |
eight | -0.055 |
Tweet | -0.054 |
Token | . |
Token ID | 13 |
Feature activation | +5.01e-01 |
Pos logit contributions | |
idents | +0.013 |
package | +0.013 |
batch | +0.012 |
tray | +0.012 |
meal | +0.012 |
Neg logit contributions | |
stretched | -0.007 |
ila | -0.007 |
Tw | -0.006 |
eight | -0.006 |
Tweet | -0.006 |
Token | Lily |
Token ID | 20037 |
Feature activation | -2.71e-01 |
Pos logit contributions | |
idents | +0.019 |
batch | +0.018 |
package | +0.018 |
tray | +0.018 |
meal | +0.018 |
Neg logit contributions | |
stretched | -0.012 |
ila | -0.011 |
eight | -0.010 |
La | -0.010 |
Tweet | -0.010 |
Token | was |
Token ID | 373 |
Feature activation | -1.72e-02 |
Pos logit contributions | |
ila | +0.007 |
stretched | +0.007 |
Tw | +0.006 |
eight | +0.006 |
Tweet | +0.006 |
Neg logit contributions | |
idents | -0.009 |
package | -0.009 |
meal | -0.009 |
tray | -0.009 |
batch | -0.009 |
Token | happy |
Token ID | 3772 |
Feature activation | +1.65e+00 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
eight | +0.000 |
Tw | +0.000 |
La | +0.000 |
Neg logit contributions | |
idents | -0.001 |
package | -0.001 |
meal | -0.001 |
tray | -0.001 |
batch | -0.001 |
Token | and |
Token ID | 290 |
Feature activation | +3.95e-01 |
Pos logit contributions | |
package | +0.063 |
idents | +0.063 |
tray | +0.062 |
stret | +0.061 |
meal | +0.061 |
Neg logit contributions | |
ila | -0.033 |
stretched | -0.032 |
Tw | -0.031 |
eight | -0.029 |
Tweet | -0.028 |
Token | proud |
Token ID | 6613 |
Feature activation | +7.67e-01 |
Pos logit contributions | |
package | +0.014 |
idents | +0.014 |
tray | +0.014 |
batch | +0.014 |
meal | +0.014 |
Neg logit contributions | |
ila | -0.009 |
stretched | -0.009 |
Tw | -0.008 |
eight | -0.008 |
Tweet | -0.008 |
Token | of |
Token ID | 286 |
Feature activation | -5.62e-01 |
Pos logit contributions | |
meal | +0.028 |
batch | +0.028 |
package | +0.027 |
tray | +0.027 |
idents | +0.027 |
Neg logit contributions | |
Tweet | -0.017 |
ila | -0.017 |
La | -0.016 |
stretched | -0.016 |
eight | -0.016 |
Token | herself |
Token ID | 5223 |
Feature activation | -1.58e+00 |
Pos logit contributions | |
stretched | +0.014 |
ila | +0.014 |
eight | +0.013 |
Tweet | +0.013 |
Tw | +0.013 |
Neg logit contributions | |
idents | -0.019 |
package | -0.019 |
meal | -0.018 |
tray | -0.018 |
batch | -0.018 |
Token | for |
Token ID | 329 |
Feature activation | +1.70e+00 |
Pos logit contributions | |
Tw | +0.028 |
stretched | +0.028 |
eight | +0.026 |
ila | +0.024 |
W | +0.023 |
Neg logit contributions | |
idents | -0.064 |
aman | -0.060 |
stret | -0.060 |
package | -0.058 |
goodies | -0.057 |
Token | helping |
Token ID | 5742 |
Feature activation | +5.45e-01 |
Pos logit contributions | |
idents | +0.056 |
package | +0.053 |
tray | +0.052 |
meal | +0.051 |
batch | +0.050 |
Neg logit contributions | |
ila | -0.046 |
stretched | -0.045 |
eight | -0.044 |
Tw | -0.043 |
Tweet | -0.041 |
Token | and |
Token ID | 290 |
Feature activation | +8.06e-01 |
Pos logit contributions | |
idents | +0.002 |
package | +0.002 |
meal | +0.002 |
tray | +0.002 |
batch | +0.002 |
Neg logit contributions | |
ila | -0.002 |
stretched | -0.002 |
Tweet | -0.002 |
Tw | -0.002 |
La | -0.002 |
Token | removed |
Token ID | 4615 |
Feature activation | -1.32e+00 |
Pos logit contributions | |
package | +0.026 |
idents | +0.026 |
batch | +0.026 |
meal | +0.026 |
tray | +0.025 |
Neg logit contributions | |
stretched | -0.021 |
ila | -0.020 |
Tweet | -0.019 |
La | -0.019 |
Tw | -0.019 |
Token | all |
Token ID | 477 |
Feature activation | +6.06e-01 |
Pos logit contributions | |
ila | +0.022 |
stretched | +0.022 |
Tw | +0.020 |
eight | +0.019 |
Tweet | +0.019 |
Neg logit contributions | |
idents | -0.055 |
package | -0.055 |
meal | -0.054 |
tray | -0.054 |
batch | -0.053 |
Token | the |
Token ID | 262 |
Feature activation | -3.96e-02 |
Pos logit contributions | |
idents | +0.028 |
meal | +0.027 |
package | +0.027 |
stret | +0.027 |
tray | +0.026 |
Neg logit contributions | |
ila | -0.008 |
stretched | -0.008 |
eight | -0.007 |
Tw | -0.006 |
Tweet | -0.006 |
Token | packaging |
Token ID | 16846 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
other | +0.001 |
blue | +0.001 |
two | +0.001 |
small | +0.001 |
Some | +0.001 |
Neg logit contributions | |
ics | -0.001 |
idents | -0.001 |
Birthday | -0.001 |
aman | -0.001 |
keeping | -0.001 |
Token | carefully |
Token ID | 7773 |
Feature activation | -2.59e+00 |
Pos logit contributions | |
idents | +0.051 |
meal | +0.050 |
package | +0.050 |
batch | +0.049 |
stret | +0.048 |
Neg logit contributions | |
ila | -0.029 |
stretched | -0.027 |
eight | -0.025 |
Tw | -0.024 |
fast | -0.021 |
Token | . |
Token ID | 13 |
Feature activation | -3.50e-02 |
Pos logit contributions | |
stretched | +0.054 |
ila | +0.053 |
eight | +0.049 |
Tw | +0.049 |
Tweet | +0.045 |
Neg logit contributions | |
idents | -0.097 |
package | -0.094 |
meal | -0.092 |
tray | -0.092 |
batch | -0.091 |
Token | They |
Token ID | 1119 |
Feature activation | +2.24e+00 |
Pos logit contributions | |
stretched | +0.001 |
ila | +0.001 |
La | +0.001 |
Tweet | +0.001 |
Tw | +0.001 |
Neg logit contributions | |
idents | -0.001 |
batch | -0.001 |
package | -0.001 |
refrigerator | -0.001 |
aman | -0.001 |
Token | both |
Token ID | 1111 |
Feature activation | -7.55e-02 |
Pos logit contributions | |
idents | +0.083 |
package | +0.076 |
meal | +0.075 |
tray | +0.074 |
batch | +0.073 |
Neg logit contributions | |
stretched | -0.051 |
ila | -0.050 |
eight | -0.049 |
Tw | -0.045 |
Tweet | -0.045 |
Token | worked |
Token ID | 3111 |
Feature activation | -2.10e-01 |
Pos logit contributions | |
stretched | +0.002 |
ila | +0.002 |
eight | +0.002 |
Tw | +0.002 |
Tweet | +0.002 |
Neg logit contributions | |
idents | -0.003 |
package | -0.003 |
meal | -0.002 |
tray | -0.002 |
batch | -0.002 |
Token | together |
Token ID | 1978 |
Feature activation | -2.95e-01 |
Pos logit contributions | |
stretched | +0.006 |
ila | +0.006 |
eight | +0.006 |
Tw | +0.006 |
Tweet | +0.005 |
Neg logit contributions | |
idents | -0.006 |
package | -0.006 |
meal | -0.006 |
tray | -0.006 |
dining | -0.006 |
Token | Can |
Token ID | 6090 |
Feature activation | +4.51e-01 |
Pos logit contributions | |
ila | +0.001 |
stretched | +0.001 |
Tw | +0.001 |
eight | +0.001 |
Tweet | +0.001 |
Neg logit contributions | |
package | -0.003 |
idents | -0.003 |
tray | -0.003 |
meal | -0.003 |
batch | -0.003 |
Token | I |
Token ID | 314 |
Feature activation | -1.58e+00 |
Pos logit contributions | |
idents | +0.017 |
package | +0.016 |
cake | +0.016 |
tray | +0.015 |
meal | +0.015 |
Neg logit contributions | |
stretched | -0.012 |
La | -0.010 |
ila | -0.010 |
Bo | -0.010 |
Tweet | -0.009 |
Token | have |
Token ID | 423 |
Feature activation | -3.68e-01 |
Pos logit contributions | |
stretched | +0.040 |
ila | +0.039 |
La | +0.038 |
Tw | +0.038 |
Tweet | +0.036 |
Neg logit contributions | |
package | -0.054 |
meal | -0.054 |
tray | -0.052 |
idents | -0.051 |
cake | -0.050 |
Token | some |
Token ID | 617 |
Feature activation | +1.37e+00 |
Pos logit contributions | |
ila | +0.008 |
stretched | +0.008 |
eight | +0.008 |
Tw | +0.008 |
Tweet | +0.008 |
Neg logit contributions | |
idents | -0.013 |
package | -0.013 |
tray | -0.013 |
meal | -0.012 |
batch | -0.012 |
Token | ?" |
Token ID | 1701 |
Feature activation | -5.31e-01 |
Pos logit contributions | |
idents | +0.030 |
package | +0.029 |
meal | +0.028 |
tray | +0.028 |
batch | +0.028 |
Neg logit contributions | |
ila | -0.050 |
stretched | -0.049 |
Tw | -0.047 |
eight | -0.047 |
Tweet | -0.047 |
Token | Her |
Token ID | 2332 |
Feature activation | -3.13e+00 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.013 |
Tweet | +0.012 |
eight | +0.012 |
La | +0.012 |
Neg logit contributions | |
batch | -0.017 |
idents | -0.017 |
package | -0.017 |
meal | -0.017 |
dinner | -0.017 |
Token | mom |
Token ID | 1995 |
Feature activation | +3.40e-03 |
Pos logit contributions | |
ila | +0.057 |
stretched | +0.056 |
eight | +0.052 |
Tw | +0.051 |
Tweet | +0.047 |
Neg logit contributions | |
idents | -0.126 |
meal | -0.123 |
package | -0.123 |
batch | -0.121 |
tray | -0.120 |
Token | said |
Token ID | 531 |
Feature activation | -5.12e-02 |
Pos logit contributions | |
idents | +0.000 |
package | +0.000 |
meal | +0.000 |
tray | +0.000 |
stret | +0.000 |
Neg logit contributions | |
ila | -0.000 |
stretched | -0.000 |
Tw | -0.000 |
eight | -0.000 |
Tweet | -0.000 |
Token | , |
Token ID | 11 |
Feature activation | +1.03e+00 |
Pos logit contributions | |
ila | +0.001 |
stretched | +0.001 |
Tweet | +0.001 |
La | +0.001 |
Tw | +0.001 |
Neg logit contributions | |
idents | -0.002 |
package | -0.002 |
tray | -0.002 |
meal | -0.002 |
batch | -0.002 |
Token | " |
Token ID | 366 |
Feature activation | -3.41e-01 |
Pos logit contributions | |
tray | +0.034 |
idents | +0.034 |
batch | +0.033 |
package | +0.033 |
meal | +0.033 |
Neg logit contributions | |
ila | -0.026 |
stretched | -0.025 |
Tweet | -0.024 |
La | -0.024 |
Tw | -0.022 |
Token | Yes |
Token ID | 5297 |
Feature activation | -2.06e-03 |
Pos logit contributions | |
stretched | +0.008 |
ila | +0.008 |
Tw | +0.008 |
eight | +0.008 |
La | +0.007 |
Neg logit contributions | |
batch | -0.011 |
package | -0.011 |
idents | -0.011 |
tray | -0.010 |
meal | -0.010 |
Token | that |
Token ID | 326 |
Feature activation | +8.39e-01 |
Pos logit contributions | |
meal | +0.016 |
idents | +0.016 |
dinner | +0.016 |
package | +0.016 |
cake | +0.015 |
Neg logit contributions | |
stretched | -0.017 |
ila | -0.017 |
Tweet | -0.016 |
La | -0.016 |
Tw | -0.015 |
Token | it |
Token ID | 340 |
Feature activation | +7.43e-01 |
Pos logit contributions | |
meal | +0.030 |
package | +0.030 |
idents | +0.030 |
tray | +0.029 |
batch | +0.029 |
Neg logit contributions | |
ila | -0.019 |
stretched | -0.019 |
Tweet | -0.017 |
eight | -0.017 |
La | -0.017 |
Token | was |
Token ID | 373 |
Feature activation | +9.96e-01 |
Pos logit contributions | |
idents | +0.019 |
package | +0.019 |
meal | +0.018 |
tray | +0.018 |
batch | +0.018 |
Neg logit contributions | |
ila | -0.024 |
stretched | -0.024 |
eight | -0.023 |
Tw | -0.023 |
Tweet | -0.023 |
Token | only |
Token ID | 691 |
Feature activation | -1.55e+00 |
Pos logit contributions | |
package | +0.028 |
meal | +0.028 |
dinner | +0.026 |
idents | +0.026 |
tray | +0.026 |
Neg logit contributions | |
ila | -0.031 |
stretched | -0.030 |
La | -0.029 |
Tweet | -0.029 |
eight | -0.028 |
Token | a |
Token ID | 257 |
Feature activation | -5.30e-01 |
Pos logit contributions | |
ila | +0.041 |
stretched | +0.041 |
Tweet | +0.039 |
La | +0.038 |
Tw | +0.037 |
Neg logit contributions | |
package | -0.050 |
meal | -0.049 |
idents | -0.048 |
tray | -0.047 |
batch | -0.046 |
Token | dream |
Token ID | 4320 |
Feature activation | -3.67e+00 |
Pos logit contributions | |
ila | +0.017 |
stretched | +0.016 |
Tweet | +0.015 |
La | +0.015 |
Tw | +0.015 |
Neg logit contributions | |
package | -0.016 |
meal | -0.016 |
tray | -0.015 |
idents | -0.015 |
mess | -0.014 |
Token | . |
Token ID | 13 |
Feature activation | +7.44e-02 |
Pos logit contributions | |
ila | +0.097 |
stretched | +0.095 |
Tweet | +0.093 |
eight | +0.090 |
La | +0.089 |
Neg logit contributions | |
package | -0.118 |
meal | -0.115 |
idents | -0.114 |
tray | -0.112 |
batch | -0.108 |
Token | She |
Token ID | 1375 |
Feature activation | -1.37e+00 |
Pos logit contributions | |
idents | +0.003 |
package | +0.003 |
meal | +0.003 |
tray | +0.003 |
batch | +0.003 |
Neg logit contributions | |
ila | -0.002 |
stretched | -0.002 |
Tweet | -0.002 |
Tw | -0.001 |
La | -0.001 |
Token | explained |
Token ID | 4893 |
Feature activation | +4.84e-01 |
Pos logit contributions | |
stretched | +0.041 |
ila | +0.040 |
Tweet | +0.038 |
eight | +0.037 |
La | +0.037 |
Neg logit contributions | |
package | -0.039 |
meal | -0.038 |
batch | -0.037 |
idents | -0.037 |
envelope | -0.036 |
Token | that |
Token ID | 326 |
Feature activation | +9.36e-01 |
Pos logit contributions | |
package | +0.015 |
idents | +0.015 |
meal | +0.015 |
batch | +0.014 |
tray | +0.014 |
Neg logit contributions | |
ila | -0.013 |
stretched | -0.013 |
Tweet | -0.012 |
La | -0.012 |
Tw | -0.012 |
Token | even |
Token ID | 772 |
Feature activation | -1.77e-01 |
Pos logit contributions | |
idents | +0.034 |
package | +0.033 |
meal | +0.033 |
tray | +0.033 |
batch | +0.032 |
Neg logit contributions | |
stretched | -0.021 |
ila | -0.021 |
Tweet | -0.019 |
eight | -0.019 |
Tw | -0.019 |
Token | permission |
Token ID | 7170 |
Feature activation | -6.93e-01 |
Pos logit contributions | |
dinner | +0.074 |
cake | +0.072 |
meal | +0.070 |
mess | +0.069 |
package | +0.069 |
Neg logit contributions | |
stretched | -0.066 |
Tweet | -0.065 |
ila | -0.063 |
Tw | -0.061 |
La | -0.060 |
Token | before |
Token ID | 878 |
Feature activation | -8.07e-01 |
Pos logit contributions | |
Tweet | +0.022 |
stretched | +0.022 |
ila | +0.021 |
La | +0.020 |
Leave | +0.020 |
Neg logit contributions | |
idents | -0.019 |
package | -0.019 |
dinner | -0.018 |
envelope | -0.018 |
tray | -0.018 |
Token | doing |
Token ID | 1804 |
Feature activation | -4.47e-01 |
Pos logit contributions | |
ila | +0.026 |
stretched | +0.026 |
Tweet | +0.025 |
La | +0.024 |
Tw | +0.024 |
Neg logit contributions | |
idents | -0.022 |
meal | -0.021 |
package | -0.021 |
dinner | -0.020 |
batch | -0.020 |
Token | something |
Token ID | 1223 |
Feature activation | +1.12e+00 |
Pos logit contributions | |
stretched | +0.013 |
ila | +0.013 |
Tweet | +0.013 |
Tw | +0.012 |
eight | +0.012 |
Neg logit contributions | |
idents | -0.013 |
package | -0.013 |
meal | -0.013 |
tray | -0.012 |
batch | -0.012 |
Token | that |
Token ID | 326 |
Feature activation | -7.43e-04 |
Pos logit contributions | |
idents | +0.033 |
package | +0.032 |
meal | +0.031 |
mess | +0.031 |
tray | +0.030 |
Neg logit contributions | |
stretched | -0.034 |
Tweet | -0.033 |
ila | -0.032 |
La | -0.032 |
eight | -0.031 |
Token | might |
Token ID | 1244 |
Feature activation | -2.74e+00 |
Pos logit contributions | |
ila | +0.000 |
stretched | +0.000 |
Tw | +0.000 |
eight | +0.000 |
Tweet | +0.000 |
Neg logit contributions | |
idents | -0.000 |
package | -0.000 |
tray | -0.000 |
meal | -0.000 |
batch | -0.000 |
Token | affect |
Token ID | 2689 |
Feature activation | +2.36e-01 |
Pos logit contributions | |
stretched | +0.082 |
ila | +0.081 |
La | +0.078 |
Tw | +0.075 |
Tweet | +0.074 |
Neg logit contributions | |
idents | -0.078 |
meal | -0.077 |
package | -0.076 |
mess | -0.075 |
tray | -0.074 |
Token | others |
Token ID | 1854 |
Feature activation | -2.64e+00 |
Pos logit contributions | |
package | +0.009 |
idents | +0.009 |
tray | +0.009 |
batch | +0.009 |
meal | +0.009 |
Neg logit contributions | |
ila | -0.004 |
stretched | -0.004 |
Tw | -0.004 |
eight | -0.004 |
Tweet | -0.004 |
Token | . |
Token ID | 13 |
Feature activation | +4.95e-01 |
Pos logit contributions | |
ila | +0.056 |
stretched | +0.055 |
Tw | +0.052 |
eight | +0.051 |
La | +0.049 |
Neg logit contributions | |
package | -0.096 |
idents | -0.096 |
batch | -0.095 |
tray | -0.094 |
meal | -0.094 |
Token |
|
Token ID | 198 |
Feature activation | +1.36e-01 |
Pos logit contributions | |
idents | +0.018 |
batch | +0.018 |
package | +0.017 |
meal | +0.017 |
tray | +0.017 |
Neg logit contributions | |
stretched | -0.012 |
ila | -0.011 |
La | -0.010 |
eight | -0.010 |
Tweet | -0.010 |
Token |
|
Token ID | 198 |
Feature activation | +7.19e-02 |
Pos logit contributions | |
idents | +0.005 |
package | +0.005 |
meal | +0.005 |
tray | +0.005 |
batch | +0.005 |
Neg logit contributions | |
stretched | -0.003 |
ila | -0.003 |
Tw | -0.003 |
eight | -0.003 |
Tweet | -0.003 |
Token | not |
Token ID | 407 |
Feature activation | +7.28e-01 |
Pos logit contributions | |
mess | +0.030 |
package | +0.030 |
treats | +0.029 |
dinner | +0.028 |
idents | +0.028 |
Neg logit contributions | |
stretched | -0.054 |
La | -0.052 |
ila | -0.051 |
Tw | -0.048 |
eight | -0.048 |
Token | fun |
Token ID | 1257 |
Feature activation | -5.31e-01 |
Pos logit contributions | |
package | +0.021 |
idents | +0.021 |
meal | +0.021 |
tray | +0.020 |
batch | +0.020 |
Neg logit contributions | |
ila | -0.022 |
stretched | -0.021 |
La | -0.020 |
Tw | -0.020 |
Tweet | -0.019 |
Token | and |
Token ID | 290 |
Feature activation | -9.72e-01 |
Pos logit contributions | |
ila | +0.014 |
stretched | +0.014 |
Tweet | +0.013 |
La | +0.013 |
eight | +0.013 |
Neg logit contributions | |
idents | -0.017 |
package | -0.017 |
meal | -0.016 |
tray | -0.016 |
batch | -0.016 |
Token | can |
Token ID | 460 |
Feature activation | -1.85e+00 |
Pos logit contributions | |
stretched | +0.028 |
ila | +0.028 |
La | +0.027 |
Tw | +0.026 |
Tweet | +0.025 |
Neg logit contributions | |
idents | -0.030 |
package | -0.028 |
meal | -0.027 |
mess | -0.027 |
tray | -0.026 |
Token | hurt |
Token ID | 5938 |
Feature activation | -2.54e-01 |
Pos logit contributions | |
stretched | +0.056 |
ila | +0.056 |
eight | +0.053 |
La | +0.053 |
Tw | +0.052 |
Neg logit contributions | |
idents | -0.052 |
mess | -0.051 |
meal | -0.048 |
package | -0.046 |
cake | -0.045 |
Token | others |
Token ID | 1854 |
Feature activation | -2.99e+00 |
Pos logit contributions | |
stretched | +0.007 |
La | +0.007 |
ila | +0.007 |
Tw | +0.007 |
Does | +0.006 |
Neg logit contributions | |
idents | -0.007 |
mess | -0.007 |
meal | -0.007 |
package | -0.007 |
dinner | -0.007 |
Token | . |
Token ID | 13 |
Feature activation | +3.95e-01 |
Pos logit contributions | |
ila | +0.072 |
stretched | +0.071 |
Tw | +0.066 |
eight | +0.066 |
Tweet | +0.065 |
Neg logit contributions | |
idents | -0.102 |
package | -0.101 |
meal | -0.099 |
tray | -0.099 |
batch | -0.098 |
Token | They |
Token ID | 1119 |
Feature activation | +1.40e+00 |
Pos logit contributions | |
idents | +0.010 |
package | +0.008 |
meal | +0.008 |
batch | +0.007 |
refrigerator | +0.007 |
Neg logit contributions | |
stretched | -0.014 |
ila | -0.013 |
brav | -0.013 |
eight | -0.013 |
Tweet | -0.012 |
Token | learned |
Token ID | 4499 |
Feature activation | -5.48e-01 |
Pos logit contributions | |
idents | +0.048 |
package | +0.047 |
meal | +0.046 |
tray | +0.045 |
batch | +0.045 |
Neg logit contributions | |
stretched | -0.034 |
ila | -0.033 |
eight | -0.031 |
Tw | -0.031 |
La | -0.030 |
Token | that |
Token ID | 326 |
Feature activation | +8.62e-01 |
Pos logit contributions | |
stretched | +0.016 |
ila | +0.015 |
eight | +0.014 |
Tweet | +0.014 |
La | +0.014 |
Neg logit contributions | |
idents | -0.017 |
package | -0.017 |
meal | -0.016 |
batch | -0.016 |
tray | -0.016 |
Token | being |
Token ID | 852 |
Feature activation | -1.69e-02 |
Pos logit contributions | |
idents | +0.025 |
package | +0.025 |
meal | +0.025 |
tray | +0.024 |
batch | +0.024 |
Neg logit contributions | |
ila | -0.025 |
stretched | -0.025 |
eight | -0.023 |
Tw | -0.023 |
Tweet | -0.023 |