INDEX
Explanations
names related to geographical locations, specifically those in the Middle East
occurrences of the substring "kh"
New Auto-Interp
Negative Logits
IZ
-0.64
TAIN
-0.64
HER
-0.63
DIT
-0.62
IBLE
-0.62
LOAD
-0.60
ffic
-0.60
cul
-0.60
================
-0.60
grassroots
-0.59
POSITIVE LOGITS
urst
1.12
azard
1.11
alid
0.96
orst
0.93
ĵĺ
0.92
uty
0.89
hyde
0.84
atever
0.84
ĺħ
0.84
ģĸ
0.83
Activations Density 0.015%