INDEX
Explanations
terms related to the concept of "fullness" or "complete."
New Auto-Interp
Negative Logits
atorio
-0.15
exist
-0.15
241
-0.14
oc
-0.14
esty
-0.14
borrowed
-0.14
sty
-0.14
quo
-0.14
bow
-0.13
Kent
-0.13
POSITIVE LOGITS
erton
0.21
/full
0.21
óz
0.19
Dup
0.18
Full
0.17
arton
0.17
eren
0.17
tone
0.17
bright
0.17
filled
0.17
Activations Density 0.027%