INDEX
Explanations
names or terms related to Iranian entities
instances of the substring "ran"
New Auto-Interp
Negative Logits
earable
-0.66
lda
-0.65
GMT
-0.63
£ı
-0.60
ensions
-0.60
ready
-0.60
¿½
-0.58
heart
-0.57
{"-0.57
ij士
-0.57
POSITIVE LOGITS
vier
1.10
igans
0.94
fo
0.91
thus
0.89
coe
0.87
egal
0.84
ormal
0.83
kees
0.82
igan
0.81
emate
0.81
Activations Density 0.029%