INDEX
Explanations
commas and punctuation-related constructs in the text
New Auto-Interp
Negative Logits
sgi
-0.15
cg
-0.15
ittel
-0.14
ί
-0.14
inspace
-0.14
etal
-0.14
Č↵
-0.13
eza
-0.13
anoia
-0.13
roje
-0.13
POSITIVE LOGITS
upertino
0.18
nl
0.15
ensch
0.14
atar
0.14
ixmap
0.13
Williams
0.13
Roller
0.13
redo
0.13
Citizens
0.13
Icons
0.13
Activations Density 0.005%