INDEX
Explanations
proper nouns related to people's names
repeated mentions of the name "Tarl."
New Auto-Interp
Negative Logits
FFER
-0.77
BLIC
-0.71
nder
-0.66
ndra
-0.65
LOAD
-0.64
CRIP
-0.63
LV
-0.62
CFR
-0.61
CHAR
-0.61
ELS
-0.59
POSITIVE LOGITS
anguage
1.14
ophone
1.03
phia
1.02
owe
0.94
ibrary
0.92
ounge
0.91
anguages
0.89
oths
0.88
otte
0.87
ondon
0.87
Activations Density 0.030%