INDEX
Explanations
instances of calls to action for user engagement
New Auto-Interp
Negative Logits
misc
-0.16
entai
-0.15
Mess
-0.14
Inbox
-0.14
ackle
-0.14
ãĥĭãĤ¢
-0.14
oje
-0.14
isposable
-0.14
nger
-0.14
aida
-0.13
POSITIVE LOGITS
links
0.25
link
0.24
desired
0.21
.links
0.20
links
0.20
blue
0.19
image
0.19
icon
0.18
icons
0.18
individual
0.18
Activations Density 0.037%