INDEX
Explanations
hyperlinks and elements related to navigation in a web context
New Auto-Interp
Negative Logits
ight
-0.16
,-
-0.15
йом
-0.15
ëģ
-0.15
unta
-0.15
bil
-0.15
it
-0.14
itting
-0.14
ire
-0.14
tall
-0.14
POSITIVE LOGITS
ParameterValue
0.15
Surround
0.15
Enlarge
0.15
nnen
0.15
.tp
0.14
PATCH
0.14
ansom
0.14
ány
0.14
patch
0.13
Patch
0.13
Activations Density 0.216%