INDEX
Explanations
document followed by descriptor
New Auto-Interp
Negative Logits
to
1.28
.
1.25
’
1.23
OL
0.99
as
0.93
to
0.93
AR
0.91
ão
0.89
şi
0.88
0.88
POSITIVE LOGITS
ي
1.63
<0x80>
1.28
ק
1.20
n
1.18
л
1.18
ن
1.16
יות
1.14
nul
1.09
י
1.06
０
1.04
Activations Density 0.035%