INDEX
Negative Logits
ه
1.11
ти
1.06
when
1.01
for
1.00
if
0.91
ці
0.91
𝙠
0.91
𝙧
0.90
ط
0.90
³
0.88
POSITIVE LOGITS
in
1.54
at
1.36
as
1.18
ou
1.13
ance
1.10
ur
1.04
1.04
on
1.02
or
1.00
اب
0.96
Activations Density 0.005%
ه
ти
when
for
if
ці
𝙠
𝙧
ط
³
in
at
as
ou
ance
ur
on
or
اب