INDEX
Explanations
mentions of "something" with significant emphasis
New Auto-Interp
Negative Logits
curacy
-0.88
zbęd
-0.83
Datuak
-0.80
owohl
-0.79
Genoa
-0.78
xanth
-0.78
avancé
-0.78
zewod
-0.77
Lait
-0.74
palaces
-0.73
POSITIVE LOGITS
something
1.96
Something
1.95
Something
1.93
something
1.93
SOMETHING
1.76
Somebody
1.59
Someone
1.55
Someone
1.50
ETHING
1.49
somebody
1.47
Activations Density 0.060%