INDEX
Explanations
The neuron responds to thematic section-header vocabulary—labels or descriptors (e.g. “national,” “global,” “security,” “terrorism,” “intelligence”) that mark a news section or category.
New Auto-Interp
Negative Logits
宝
-0.07
فنی
-0.07
(@"%@",
-0.07
่วน
-0.06
Showing
-0.06
寶
-0.06
bs
-0.06
'http
-0.06
olas
-0.06
//----------------------------------------------------------------
-0.06
POSITIVE LOGITS
Chair
0.07
.Builder
0.07
Cult
0.06
Сам
0.06
-preview
0.06
,Th
0.06
Strike
0.06
planes
0.06
Gibraltar
0.06
intéress
0.06
Activations Density 0.028%