INDEX
Explanations
fictional stories
a running narrative context — it lights up on story prose (character names, pronouns and nearby verbs/continuation tokens) indicating ongoing narration about characters.
New Auto-Interp
Negative Logits
меж
-0.07
าอ
-0.07
_attempts
-0.06
आम
-0.06
97
-0.06
$title
-0.06
sexuality
-0.06
')}}
-0.06
سفید
-0.06
.start
-0.06
POSITIVE LOGITS
Panc
0.07
TRACK
0.07
pek
0.06
Scar
0.06
dominate
0.06
-par
0.06
ulled
0.06
_ter
0.06
Uni
0.06
Xem
0.06
Activations Density 0.034%