INDEX
Explanations
article snippets
This neuron activates on “read more” excerpts or similar inline prompts indicating continuation links in article previews.
New Auto-Interp
Negative Logits
bosses
-0.07
gender
-0.06
joins
-0.06
Boy
-0.06
ُر
-0.06
JSGlobal
-0.06
bites
-0.06
passes
-0.06
English
-0.06
hate
-0.06
POSITIVE LOGITS
Див
0.06
ickou
0.06
благодаря
0.06
CENT
0.06
CERT
0.06
(LL
0.06
Welfare
0.06
�
0.06
_PAD
0.06
ليف
0.06
Activations Density 0.000%