INDEX
Explanations
actions and descriptions involving hands and fingers
New Auto-Interp
Negative Logits
ardi
-0.17
ihan
-0.16
ibold
-0.15
uste
-0.15
LEAN
-0.15
мена
-0.15
canf
-0.14
ardy
-0.14
cia
-0.14
berapa
-0.14
POSITIVE LOGITS
ker
0.19
/thumb
0.17
hold
0.16
ents
0.16
(hand
0.15
prints
0.14
oglob
0.14
perature
0.14
-held
0.14
bare
0.14
Activations Density 0.093%