INDEX
Explanations
references to visual elements, particularly in photographs and portraits
New Auto-Interp
Negative Logits
plots
-0.18
presentations
-0.16
mails
-0.15
captions
-0.15
speeches
-0.15
videos
-0.15
scripts
-0.14
_plots
-0.14
ems
-0.14
Append
-0.14
POSITIVE LOGITS
image
0.55
picture
0.50
photo
0.47
photograph
0.43
image
0.42
drawing
0.37
photo
0.37
'image
0.36
pic
0.36
picture
0.35
Activations Density 0.344%