INDEX
Explanations
occurrences of physical actions and movements
New Auto-Interp
Negative Logits
clipped
-0.15
ektir
-0.14
oto
-0.14
flattened
-0.14
Syntax
-0.14
ÙĦÛĮÙħ
-0.13
injured
-0.13
ôm
-0.13
molds
-0.13
дво
-0.13
POSITIVE LOGITS
uard
0.17
táºŃp
0.16
Spicer
0.15
892
0.15
Jar
0.15
أث
0.14
Operators
0.14
-library
0.14
Library
0.14
ìļ´
0.13
Activations Density 0.106%