INDEX
Explanations
code, variables, and punctuation
New Auto-Interp
Negative Logits
and
-1.62
in
-1.62
at
-1.45
or
-1.41
from
-1.38
one
-1.34
as
-1.19
on
-1.14
for
-1.09
out
-1.01
POSITIVE LOGITS
そうで
1.14
conséquences
1.07
רוב
1.05
Hilsen
1.05
mudou
1.02
quello
1.02
さり
1.01
そうです
1.01
véritable
1.00
revamped
0.99
Activations Density 0.003%