INDEX
Explanations
contrasting viewpoints or perspectives within a text
references to differing opinions or viewpoints
New Auto-Interp
Negative Logits
tained
-0.94
was
-0.88
Was
-0.76
ãĥĺ
-0.71
needed
-0.69
opened
-0.68
EStream
-0.67
Completed
-0.67
ogged
-0.66
Was
-0.66
POSITIVE LOGITS
argue
1.69
contend
1.60
concede
1.57
cite
1.54
insist
1.49
emphasize
1.46
acknowledge
1.45
propose
1.38
agree
1.38
conclude
1.37
Activations Density 0.421%