INDEX
Explanations
references to Islam and Islamic identity
New Auto-Interp
Negative Logits
acman
-0.15
rowned
-0.15
iless
-0.15
datable
-0.15
jes
-0.14
509
-0.14
Dix
-0.14
éĤ
-0.14
Catholic
-0.14
904
-0.13
POSITIVE LOGITS
abad
0.18
ized
0.17
å¾Ĵ
0.15
ห
0.15
udd
0.15
arus
0.15
_generic
0.15
/non
0.15
>window
0.15
-Christian
0.15
Activations Density 0.016%