INDEX
Explanations
references to facial hair and specifically variations of beards
New Auto-Interp
Negative Logits
hir
-0.15
εί
-0.15
imers
-0.15
ož
-0.13
iami
-0.13
ogn
-0.13
ç«Ļ
-0.13
orse
-0.13
_EC
-0.13
ramer
-0.13
POSITIVE LOGITS
eya
0.16
atk
0.15
PHA
0.14
æį®
0.14
inde
0.14
.setTo
0.14
ress
0.14
ropy
0.14
res
0.14
OCK
0.14
Activations Density 0.016%