INDEX
Explanations
religious and political references
symbols or sequences related to spiritual or religious themes
New Auto-Interp
Negative Logits
jog
-0.75
shell
-0.73
hook
-0.72
anmar
-0.71
scram
-0.70
bunny
-0.69
capsule
-0.69
spotted
-0.68
tracker
-0.67
torso
-0.67
POSITIVE LOGITS
Therefore
0.99
¯
0.98
ï¸ı
0.98
âģ
0.87
§
0.86
Therefore
0.84
Whereas
0.83
âϦ
0.82
STEM
0.82
Û
0.82
Activations Density 0.488%