INDEX
Explanations
mentions of the name "Winfrey" and its variations
New Auto-Interp
Negative Logits
een
-0.19
duct
-0.18
foot
-0.18
ea
-0.17
eil
-0.17
alted
-0.17
eed
-0.17
a
-0.16
ee
-0.16
alar
-0.16
POSITIVE LOGITS
ning
0.30
throp
0.30
-win
0.30
eries
0.29
nable
0.27
ograd
0.27
ners
0.26
chester
0.26
try
0.26
ona
0.26
Activations Density 0.015%