INDEX
Negative Logits
are
1.45
is
1.26
ম
1.24
be
1.20
]
1.20
ה
1.19
是
1.18
$
1.17
は
1.17
or
1.14
POSITIVE LOGITS
tho
0.97
Order
0.96
I
0.96
as
0.91
𖥔
0.91
asida
0.90
all
0.90
on
0.89
one
0.88
order
0.88
Activations Density 0.027%
are
is
ম
be
]
ה
是
$
は
or
tho
Order
I
as
𖥔
asida
all
on
one
order