INDEX
Explanations
online forum discussions
The neuron is keyed to isolated tokens that look like user handles or signature names (e.g. the commenter-ID lines you see after the “------” separators).
New Auto-Interp
Negative Logits
辅
-0.07
Orleans
-0.06
token
-0.06
contraception
-0.06
Style
-0.06
.coroutines
-0.06
VMware
-0.06
Afghanistan
-0.06
Guard
-0.06
lerine
-0.06
POSITIVE LOGITS
_Object
0.07
dib
0.06
νονται
0.06
федераль
0.06
キャ
0.06
VERBOSE
0.06
начале
0.06
rocking
0.06
xu
0.06
$_
0.06
Activations Density 0.008%