INDEX
Explanations
software documentation
The neuron strongly activates on the very first word of sentences or section‐heading lines—i.e. uppercase tokens at the start of a sentence or heading.
New Auto-Interp
Negative Logits
ÜNİ
-0.07
HONE
-0.07
Lng
-0.06
�
-0.06
.dismiss
-0.06
Forget
-0.06
디어
-0.06
CAUSED
-0.06
.collection
-0.06
رح
-0.06
POSITIVE LOGITS
configur
0.06
biện
0.06
rebel
0.06
****************************************
0.06
(author
0.06
>j
0.06
port
0.06
війсь
0.06
op
0.06
Propagation
0.06
Activations Density 0.055%