INDEX
Explanations
references to size-related measurements or attributes
New Auto-Interp
Negative Logits
qui
-0.19
land
-0.18
set
-0.17
ship
-0.17
ë°©
-0.16
lie
-0.16
aci
-0.15
spÄĽ
-0.15
cls
-0.15
sets
-0.15
POSITIVE LOGITS
able
0.24
/type
0.24
ToFit
0.22
=size
0.21
/color
0.18
ebra
0.16
Matters
0.16
़ा
0.16
hint
0.16
hetto
0.16
Activations Density 0.040%