INDEX
Explanations
instances of dialogue and speech attribution in the text
New Auto-Interp
Negative Logits
ắm
-0.16
iais
-0.15
iosa
-0.15
aeda
-0.15
raç
-0.15
ognito
-0.14
bard
-0.14
kova
-0.14
огод
-0.14
Pare
-0.14
POSITIVE LOGITS
Kahn
0.15
olec
0.14
tern
0.14
peek
0.14
xBF
0.13
oni
0.13
kest
0.13
ÑĤин
0.13
acional
0.13
交
0.13
Activations Density 0.031%