INDEX
Explanations
components of legal and official documents
New Auto-Interp
Negative Logits
s
-0.27
o
-0.18
aan
-0.16
>NN
-0.16
oq
-0.16
sampling
-0.15
aub
-0.15
latter
-0.14
anoia
-0.14
atform
-0.14
POSITIVE LOGITS
ALSE
0.18
uzzi
0.17
ripple
0.16
ripper
0.15
soever
0.15
soap
0.15
heed
0.14
ref
0.14
rollback
0.14
rec
0.14
Activations Density 0.028%