INDEX
Explanations
references to divine authority and traditional religious concepts
New Auto-Interp
Negative Logits
lette
-0.17
zin
-0.17
ozÃŃ
-0.17
_DAC
-0.17
zbek
-0.16
LETTE
-0.16
indeb
-0.16
rador
-0.16
jom
-0.15
539
-0.15
POSITIVE LOGITS
Cay
0.18
ery
0.17
ona
0.16
eight
0.16
uns
0.16
Ŀ¼
0.16
ONA
0.16
iet
0.15
ayne
0.15
8
0.15
Activations Density 0.105%