INDEX
Explanations
visual elements and images within the text
New Auto-Interp
Negative Logits
faſt
-0.71
transQ
-0.70
ftagPool
-0.69
propOrder
-0.68
myſelf
-0.64
متعلقه
-0.63
ſta
-0.63
becauſe
-0.61
ainfi
-0.61
Beſ
-0.60
POSITIVE LOGITS
texttt
0.56
pictured
0.41
смо
0.37
depicting
0.36
picture
0.36
photo
0.35
Rober
0.33
astore
0.32
arşivlendi
0.32
images
0.32
Activations Density 0.446%