INDEX
Explanations
words followed by punctuation
New Auto-Interp
Negative Logits
-0.83
another
-0.80
where
-0.79
genü
-0.75
what
-0.75
actually
-0.74
crafted
-0.73
things
-0.73
everybody
-0.72
えて
-0.71
POSITIVE LOGITS
}{@0.92
\\\\
0.92
enfermos
0.90
backslash
0.88
especialmente
0.85
externe
0.84
."/
0.84
termilk
0.84
ownic
0.82
</caption>
0.81
Activations Density 0.422%