INDEX
Explanations
words related to textures and flavors
New Auto-Interp
Negative Logits
morire
-0.63
huriyet
-0.62
neque
-0.59
us
-0.52
es
-0.52
acja
-0.51
ın
-0.50
regola
-0.50
ének
-0.49
ation
-0.48
POSITIVE LOGITS
dusty
1.17
bumpy
1.16
windy
1.15
breezy
1.14
licky
1.12
greasy
1.10
uffy
1.09
smelly
1.08
sweaty
1.07
squishy
1.07
Activations Density 0.166%