INDEX
Explanations
country names
The neuron activates strongly on country or nationality descriptors (demonyms).
New Auto-Interp
Negative Logits
.onerror
-0.06
wget
-0.06
OnTriggerEnter
-0.06
<!--[
-0.06
belongsTo
-0.06
Cin
-0.06
uni
-0.06
laması
-0.06
repercussions
-0.06
فوت
-0.06
POSITIVE LOGITS
robes
0.08
가지
0.07
Sync
0.07
бороть
0.07
obble
0.07
并
0.07
condolences
0.07
并
0.06
gth
0.06
productive
0.06
Activations Density 0.008%