INDEX
Explanations
references to Byzantine history and figures
New Auto-Interp
Negative Logits
ieder
-0.17
.hm
-0.16
ingu
-0.15
quil
-0.15
488
-0.14
arker
-0.14
anya
-0.14
benefici
-0.14
ickle
-0.14
hel
-0.14
POSITIVE LOGITS
Casc
0.19
urm
0.15
Hindered
0.14
wald
0.14
casc
0.14
bow
0.13
pInfo
0.13
MICRO
0.13
IFF
0.13
TRL
0.13
Activations Density 0.025%