INDEX
Explanations
descriptions of physical appearances, especially hair colors and styles
New Auto-Interp
Negative Logits
ilater
-0.92
route
-0.92
apego
-0.91
HER
-0.89
iseum
-0.88
Lost
-0.84
itionally
-0.82
failed
-0.81
undown
-0.81
Reviewer
-0.81
POSITIVE LOGITS
embro
1.08
accents
1.05
wrink
1.04
shaved
1.00
protr
0.99
hood
0.98
tint
0.98
robes
0.97
sleeves
0.97
translucent
0.96
Activations Density 9.115%