INDEX
Explanations
references to physical characteristics and attributes of objects
New Auto-Interp
Negative Logits
ШÐIJ
-0.15
otten
-0.14
afia
-0.14
crest
-0.14
pied
-0.13
kategor
-0.13
atta
-0.13
ÙħÙĨد
-0.13
Mud
-0.13
pie
-0.13
POSITIVE LOGITS
Rut
0.18
Rubin
0.17
ertil
0.15
igan
0.14
ei
0.14
eri
0.14
bjerg
0.14
Ãłu
0.14
еÑĢÑĤ
0.14
Łèĥ½
0.13
Activations Density 0.045%