INDEX
Explanations
sentences that end with punctuation or parentheses
New Auto-Interp
Negative Logits
irth
-0.19
Zust
-0.15
chner
-0.15
hit
-0.14
wat
-0.13
tle
-0.13
hits
-0.13
relay
-0.13
lee
-0.13
blas
-0.13
POSITIVE LOGITS
anje
0.17
zier
0.16
ovna
0.16
Ïģιν
0.15
nings
0.15
ixo
0.14
rosso
0.14
ums
0.14
Garrett
0.14
achable
0.14
Activations Density 0.083%