INDEX
Explanations
references to small or miniature entities
New Auto-Interp
Negative Logits
umber
-0.17
icer
-0.16
imedia
-0.15
yte
-0.14
thren
-0.14
acco
-0.14
lumber
-0.14
زÙĪ
-0.13
è
-0.13
wal
-0.13
POSITIVE LOGITS
/small
0.33
-small
0.30
SMALL
0.29
tiny
0.29
smaller
0.28
small
0.28
/tiny
0.27
small
0.26
Small
0.26
Small
0.26
Activations Density 0.483%