INDEX
Explanations
references to superhero-related themes or characters
New Auto-Interp
Negative Logits
oser
-0.18
ÏĦια
-0.15
uste
-0.15
llib
-0.15
osas
-0.15
èĴĻ
-0.14
ayers
-0.14
incinn
-0.14
à¹Ĥà¸Ń
-0.14
ikers
-0.13
POSITIVE LOGITS
Grape
0.15
etr
0.15
is
0.15
.ps
0.14
bull
0.14
esch
0.14
ÏĮ
0.14
loud
0.14
COND
0.14
454
0.14
Activations Density 0.011%