INDEX
Explanations
meet, wanted, favor, do, aware, selective
New Auto-Interp
Negative Logits
ׇ
-1.86
֩
-1.45
踔
-1.39
τὴν
-1.38
cucchiai
-1.37
triom
-1.35
applau
-1.34
vítimas
-1.32
frambo
-1.31
vanil
-1.31
POSITIVE LOGITS
ַּ
2.03
ּוֹ
1.79
"
1.77
ִּ
1.64
it
1.55
“
1.45
"[
1.41
</h3>
1.37
וֹ
1.37
était
1.31
Activations Density 0.006%