INDEX
Explanations
words related to color or color descriptors
New Auto-Interp
Negative Logits
rozen
-0.16
Levin
-0.16
rink
-0.16
noÅĽci
-0.15
bre
-0.15
ç¹ģ
-0.15
pl
-0.14
otty
-0.14
ska
-0.14
_WRAP
-0.14
POSITIVE LOGITS
ached
0.29
ble
0.26
Ble
0.26
eding
0.25
aching
0.25
edin
0.25
achers
0.22
akest
0.22
ble
0.22
ating
0.22
Activations Density 0.003%