INDEX
Explanations
the presence of formatting characters or special symbols at the beginning of text segments
New Auto-Interp
Negative Logits
photolibrary
-1.29
Efq
-1.20
auffi
-1.20
$_"
-1.19
avoient
-1.17
doubtnut
-1.17
ſind
-1.15
verwijspagina
-1.15
་་
-1.14
dezelve
-1.14
POSITIVE LOGITS
<bos>
2.03
'
1.32
’
1.27
begin
1.23
1.09
2
0.80
0
0.80
of
0.78
-
0.74
to
0.72
Activations Density 0.325%