INDEX
Explanations
entities related to people, titles, and organizations
New Auto-Interp
Negative Logits
idebar
-0.14
isser
-0.14
.toast
-0.14
utoff
-0.13
iete
-0.13
ering
-0.13
ricks
-0.13
aida
-0.13
Red
-0.13
940
-0.13
POSITIVE LOGITS
respectively
0.18
Lastly
0.17
Lastly
0.17
—all
0.15
finally
0.15
daddy
0.15
Voj
0.15
aced
0.14
leigh
0.14
greg
0.14
Activations Density 0.067%