INDEX
Explanations
actions involving seizing or taking hold of something
New Auto-Interp
Negative Logits
arily
-0.16
aea
-0.15
ials
-0.15
aned
-0.14
erman
-0.14
CKER
-0.14
-found
-0.14
ington
-0.14
pragma
-0.14
yu
-0.14
POSITIVE LOGITS
hold
0.38
hold
0.31
bing
0.26
Grab
0.26
onto
0.25
Hold
0.24
Grab
0.23
grab
0.23
grab
0.23
onto
0.23
Activations Density 0.015%