INDEX
Explanations
occurrences of the letter 'G' associated with various contexts and ratings
New Auto-Interp
Negative Logits
èŀº
-0.15
mue
-0.15
.hxx
-0.14
*sp
-0.14
onta
-0.14
CANCEL
-0.14
jectories
-0.14
gay
-0.13
iously
-0.13
AREST
-0.13
POSITIVE LOGITS
artner
0.34
iga
0.29
lob
0.29
rowth
0.25
roupe
0.24
ugg
0.22
lobe
0.22
aining
0.21
roupon
0.21
lob
0.20
Activations Density 0.015%