INDEX
Explanations
references to visual elements and descriptions in art
New Auto-Interp
Negative Logits
asu
-0.16
born
-0.15
outr
-0.15
force
-0.14
abstract
-0.14
hir
-0.14
bie
-0.14
usu
-0.13
triumph
-0.13
Tables
-0.13
POSITIVE LOGITS
folio
0.16
belong
0.16
belonged
0.15
unprotected
0.15
TAIL
0.15
oba
0.15
åĶ
0.14
aken
0.14
altung
0.14
رÙģØªÙĩ
0.14
Activations Density 0.374%