INDEX
Explanations
numerical values supporting scientific or statistical claims
New Auto-Interp
Negative Logits
Grat
-0.17
cus
-0.16
ember
-0.15
hsi
-0.15
avicon
-0.15
ereum
-0.14
onent
-0.14
rega
-0.14
unfold
-0.14
ürn
-0.14
POSITIVE LOGITS
FT
0.16
252
0.15
Acad
0.15
ç¾
0.14
dup
0.14
du
0.14
Barcl
0.14
ibold
0.14
gel
0.13
SIZE
0.13
Activations Density 0.131%