INDEX
Explanations
possessive contractions indicating ownership or association
New Auto-Interp
Negative Logits
these
-0.16
EITHER
-0.16
ostel
-0.15
оваÑĢи
-0.15
AINED
-0.14
ponto
-0.14
наÑĢ
-0.14
meiden
-0.14
ewater
-0.13
IDER
-0.13
POSITIVE LOGITS
how
0.27
hoping
0.21
why
0.20
another
0.20
some
0.20
how
0.20
something
0.20
what
0.20
where
0.17
proof
0.16
Activations Density 0.018%