INDEX
Explanations
superlative forms indicating the highest quality or degree of comparison
New Auto-Interp
Negative Logits
nice
-0.65
poco
-0.63
InjectAttribute
-0.62
nice
-0.57
pouco
-0.56
weiterer
-0.56
InitVars
-0.56
NICE
-0.56
lüğü
-0.55
のでしょうか
-0.54
POSITIVE LOGITS
freshest
1.25
healthiest
1.22
lightest
1.21
clearest
1.18
widest
1.15
ariest
1.13
strongest
1.12
broadest
1.10
heaviest
1.08
iest
1.07
Activations Density 0.629%