INDEX
Explanations
references to Christmas and related holiday themes
New Auto-Interp
Negative Logits
ker
-0.18
ched
-0.17
age
-0.15
ilon
-0.15
ster
-0.15
alo
-0.15
led
-0.15
lement
-0.14
usc
-0.14
chap
-0.14
POSITIVE LOGITS
like
0.17
bane
0.17
gow
0.17
-time
0.17
asaki
0.16
rp
0.15
MAS
0.15
-themed
0.15
ophe
0.15
ÑĢÑĮ
0.14
Activations Density 0.014%