INDEX
Negative Logits
ل
0.95
er
0.90
d
0.83
ol
0.82
ي
0.81
us
0.77
지
0.74
ა
0.73
as
0.73
l
0.73
POSITIVE LOGITS
.
0.68
Odinga
0.63
]
0.59
’
0.59
repaso
0.58
}
0.58
()}
0.57
(
0.55
ataque
0.54
I
0.54
Activations Density 0.063%
ل
er
d
ol
ي
us
지
ა
as
l
.
Odinga
]
’
repaso
}
()}
(
ataque
I