INDEX
Explanations
links to Twitter posts
punctuation, particularly periods at the end of statements
New Auto-Interp
Negative Logits
Wonderland
-0.70
estates
-0.69
volunt
-0.69
tracts
-0.67
ĪĴ
-0.66
defe
-0.65
involuntary
-0.65
conclud
-0.64
correctly
-0.63
relocation
-0.63
POSITIVE LOGITS
1.26
1.16
twitch
1.07
nz
1.05
youtube
1.03
gov
1.02
cdn
1.01
0.94
gallery
0.92
com
0.92
Activations Density 0.033%