INDEX
Explanations
references to vision-related terms
New Auto-Interp
Negative Logits
cla
-0.17
endir
-0.17
çļĦå¿ĥ
-0.16
icos
-0.15
ÃŃda
-0.15
ment
-0.15
andr
-0.15
olist
-0.14
ico
-0.14
hr
-0.14
POSITIVE LOGITS
ight
0.27
-catching
0.21
-eye
0.21
IGHT
0.20
prints
0.17
eye
0.17
eye
0.17
ING
0.16
-eyed
0.16
sert
0.16
Activations Density 0.046%