INDEX
Negative Logits
atable
-1.52
scientists
-1.50
scientist
-1.46
Scientists
-1.41
scienti
-1.38
InputBorder
-1.31
Efq
-1.29
itſelf
-1.29
doubtnut
-1.25
שוליים
-1.23
POSITIVE LOGITS
’
0.70
'
0.64
0.63
,
0.62
in
0.60
(
0.56
to
0.55
W
0.54
<eos>
0.54
on
0.53
Activations Density 0.060%