INDEX
Explanations
phrases related to visual media and scene descriptions
New Auto-Interp
Negative Logits
Il
-0.53
<!--
-0.51
it
-0.51
and
-0.50
usercontent
-0.50
zarchiwizowane
-0.48
as
-0.48
ist
-0.48
hol
-0.47
(
-0.47
POSITIVE LOGITS
Scene
1.00
Theſe
0.94
SCENE
0.93
BeginContext
0.91
__":
0.86
avoient
0.85
becauſe
0.85
étoient
0.83
Scenes
0.82
Efq
0.82
Activations Density 0.212%