INDEX
Explanations
references to old or outdated concepts and traditions
New Auto-Interp
Negative Logits
older
-0.24
older
-0.23
Older
-0.20
oldest
-0.17
ably
-0.16
ively
-0.15
asename
-0.15
criptor
-0.15
others
-0.15
lessly
-0.14
POSITIVE LOGITS
-fashioned
0.54
fashioned
0.47
-school
0.45
-fashion
0.38
school
0.36
timer
0.35
en
0.33
-style
0.33
/new
0.32
ies
0.32
Activations Density 0.065%