INDEX
Explanations
references to physical or metaphorical "edges" and boundaries
New Auto-Interp
Negative Logits
adin
-0.18
icha
-0.15
dato
-0.15
vrier
-0.14
bookmark
-0.14
Ãĸr
-0.14
اÙĨÙĤÙĦ
-0.14
Couch
-0.14
_FN
-0.14
'{@-0.14
POSITIVE LOGITS
proverb
0.24
figur
0.19
metaphor
0.19
ready
0.18
ready
0.17
style
0.16
Ready
0.16
READY
0.15
akin
0.15
literally
0.15
Activations Density 0.018%