INDEX
Explanations
descriptive adjectives related to color and appearance
New Auto-Interp
Negative Logits
ziej
-0.14
asso
-0.14
Hall
-0.14
assa
-0.14
inger
-0.14
evi
-0.13
agogue
-0.13
069
-0.13
Painter
-0.13
369
-0.13
POSITIVE LOGITS
ish
0.40
ISH
0.29
dish
0.28
-grey
0.28
-green
0.27
-gray
0.27
èī²
0.25
-ish
0.24
-purple
0.24
-yellow
0.24
Activations Density 0.062%