INDEX
Explanations
separating clauses with punctuation
New Auto-Interp
Negative Logits
βιβ
-1.66
takich
-1.57
Such
-1.56
vandens
-1.55
bular
-1.54
όπως
-1.50
Method
-1.50
ceea
-1.49
tercih
-1.47
estos
-1.46
POSITIVE LOGITS
was
1.83
became
1.41
！
1.41
has
1.34
！(
1.32
1.28
helped
1.21
but
1.20
magazin
1.20
!（
1.19
Activations Density 0.162%