INDEX
Explanations
descriptions of sensory qualities and textures
New Auto-Interp
Negative Logits
circ
-0.18
atte
-0.15
udit
-0.15
shades
-0.14
avana
-0.14
pint
-0.14
ogan
-0.14
anz
-0.14
shaded
-0.13
mont
-0.13
POSITIVE LOGITS
texture
0.31
feel
0.28
Texture
0.27
feels
0.27
feel
0.26
Feel
0.26
soft
0.26
Texture
0.24
textures
0.24
Feels
0.24
Activations Density 0.127%