INDEX
Explanations
attends to the token "window" from window-related tokens
New Auto-Interp
Head Attr Weights
0:0.10
1:0.14
2:0.12
3:0.13
4:0.13
5:0.05
6:0.15
7:0.15
Negative Logits
للمعارف
-0.48
Autoritní
-0.45
onCancelled
-0.39
noDo
-0.39
UVWXYZ
-0.38
RenderAtEndOf
-0.38
bootstrapcdn
-0.36
InSection
-0.36
Económica
-0.36
UnusedPrivate
-0.36
POSITIVE LOGITS
ese
0.33
id
0.30
mer
0.29
st
0.28
anda
0.28
atr
0.27
dat
0.27
ome
0.27
able
0.27
bor
0.26
Activations Density 0.096%