EXPLANATIONS
attack (攻)
INFERENCE
IFRAME INTEGRATION
NEGATIVE LOGITS
aland-17.750
aver-17.125
IRR-16.500
avl-16.375
综-15.938
umped-15.250
razil-15.125
озв-15.000
Moment-14.875
吼-14.875
POSITIVE LOGITS
dating16.250
包围14.500
_primes14.438
dee14.313
tet13.875
_prim13.750
pragma13.750
girl13.563
[['13.250
times13.125
ACTIVATION TIMES 0.0083%
Activation histogram is not available.
Logits histogram is not available.
TOPACTIVATIONS
SUBSAMPLING80%
SUBSAMPLING60%
SUBSAMPLING40%
NONACTIVATING