INDEX
Explanations
Information/Disclaimer
The neuron fires on boilerplate statements that say the text exists “to provide/present/assist” information or “to educate/advance awareness and understanding” of a topic.
New Auto-Interp
Negative Logits
Kiş
-0.06
Supports
-0.06
Message
-0.06
UpperCase
-0.06
pur
-0.06
تشخیص
-0.06
idal
-0.06
اع
-0.06
verify
-0.06
süz
-0.06
POSITIVE LOGITS
Fran
0.07
Proposition
0.06
Redirect
0.06
Nisan
0.06
lya
0.06
Zoom
0.06
!↵↵↵↵↵↵
0.06
Funds
0.06
ingen
0.06
时代
0.06
Activations Density 0.021%