INDEX
Explanations
references to the photo-sharing website Flickr
references to the photo-sharing platform Flickr
New Auto-Interp
Negative Logits
gran
-0.81
accur
-0.72
loud
-0.67
stals
-0.66
neys
-0.65
xon
-0.63
lass
-0.63
convincing
-0.63
bred
-0.62
ordial
-0.62
POSITIVE LOGITS
Flickr
1.40
Flickr
1.31
ickr
1.01
Wikimedia
0.93
Library
0.91
Photo
0.90
Shutterstock
0.89
Photography
0.88
photo
0.87
PHOTO
0.87
Activations Density 0.007%