INDEX
Explanations
Obstruction
The neuron fires on words (especially “way”) in obstructive contexts—i.e. it detects phrases about something being “in the way.”
New Auto-Interp
Negative Logits
bson
-0.06
bisc
-0.06
nants
-0.06
remorse
-0.06
่อไป
-0.06
eru
-0.06
Conj
-0.06
소
-0.06
".";↵
-0.05
Retorna
-0.05
POSITIVE LOGITS
kapı
0.07
فن
0.07
painfully
0.07
continual
0.06
((-
0.06
TED
0.06
orada
0.06
країн
0.06
_INTERRUPT
0.06
Comcast
0.06
Activations Density 0.018%