INDEX
Negative Logits
quement
0.44
riez
0.44
необходимость
0.41
റ്റം
0.38
ằ
0.37
clearly
0.37
reluctantly
0.36
eminently
0.36
必要
0.36
॰ऍ
0.36
POSITIVE LOGITS
tricky
1.49
interesting
1.41
thing
1.33
trick
1.31
interesting
1.20
kicker
1.19
important
1.10
interessante
1.10
关键
1.09
interesante
1.08
Activations Density 0.026%