INDEX
Explanations
headings and structured sections in technical documents
New Auto-Interp
Negative Logits
olo
-0.16
-pad
-0.14
Leh
-0.13
Weinstein
-0.13
iar
-0.13
ým
-0.13
Pad
-0.13
ó
-0.13
ter
-0.13
ych
-0.13
POSITIVE LOGITS
abei
0.19
adro
0.15
aukee
0.15
елик
0.15
ogl
0.15
imdi
0.15
جستارÙĩاÛĮ
0.14
Ayrıca
0.14
άβ
0.14
abr
0.14
Activations Density 0.009%