INDEX
Explanations
phrases referring to someone or somebody
the word "someone" and its variations
New Auto-Interp
Negative Logits
ories
-0.84
osterone
-0.76
inders
-0.71
enegger
-0.70
irth
-0.69
EngineDebug
-0.69
heny
-0.68
Pelicans
-0.67
bows
-0.66
ean
-0.65
POSITIVE LOGITS
else
1.79
Else
1.40
Else
1.16
else
1.12
WithNo
0.88
who
0.81
knowledgeable
0.76
skilled
0.71
identifiable
0.71
uscript
0.70
Activations Density 0.042%