INDEX
Explanations
high-frequency conjunctions and prepositions
New Auto-Interp
Negative Logits
itzer
-0.19
ieri
-0.15
(ARG
-0.15
ahan
-0.14
facets
-0.14
Haley
-0.14
atform
-0.14
acman
-0.14
unary
-0.14
ë²
-0.14
POSITIVE LOGITS
خاÙĨ
0.17
etler
0.16
mund
0.16
ela
0.16
emple
0.15
xt
0.15
XT
0.14
ONA
0.14
urdy
0.14
ele
0.14
Activations Density 0.000%