INDEX
Explanations
words related to scooping or lifting actions
New Auto-Interp
Negative Logits
åŁŁ
-0.16
ence
-0.15
CEED
-0.15
iska
-0.15
harm
-0.15
422
-0.15
ume
-0.14
ipp
-0.14
uo
-0.14
celand
-0.14
POSITIVE LOGITS
é«
0.18
scoop
0.16
дÑĸ
0.15
isten
0.15
acci
0.15
stakes
0.14
иÑĢÑĥ
0.14
avity
0.14
nackte
0.14
æij¸
0.14
Activations Density 0.004%