INDEX
Explanations
names and identities of characters in a narrative
New Auto-Interp
Negative Logits
screamed
-0.17
å°ĸ
-0.15
blinked
-0.15
screams
-0.14
scream
-0.14
cries
-0.14
screaming
-0.14
sharp
-0.14
Inst
-0.13
ollen
-0.13
POSITIVE LOGITS
gest
0.23
motion
0.23
cock
0.21
looked
0.20
eyeb
0.20
incl
0.19
look
0.19
nod
0.19
looks
0.19
-motion
0.19
Activations Density 0.563%