INDEX
Explanations
punctuation marks and emoticons that convey emotional expression
New Auto-Interp
Negative Logits
ãĥ³ãĥĨãĤ£
-0.15
angered
-0.15
BEST
-0.14
çͲ
-0.14
aaa
-0.14
cedure
-0.13
Williamson
-0.13
ru
-0.13
acia
-0.13
fo
-0.13
POSITIVE LOGITS
wick
0.17
pNet
0.16
isay
0.15
ertz
0.14
Král
0.14
ãĥ¼ãĥĹ
0.14
³
0.14
.scalablytyped
0.14
endl
0.14
kie
0.14
Activations Density 0.030%