INDEX
Explanations
The neuron lights up on multi-syllable, domain-specific jargon words—especially medical or software/technical terms.
New Auto-Interp
Negative Logits
copyrighted
-0.07
کارگرد
-0.06
password
-0.06
intens
-0.06
ComVisible
-0.06
BigInt
-0.06
炮
-0.06
photo
-0.06
-wow
-0.06
collectionView
-0.06
POSITIVE LOGITS
held
0.06
.sul
0.06
acios
0.06
154
0.06
neust
0.06
glimpse
0.06
Depths
0.06
getLocale
0.06
uale
0.06
enable
0.06
Activations Density 0.331%