INDEX
Explanations
emails or terms related to email communication
references to emails
New Auto-Interp
Negative Logits
llan
-0.75
ommel
-0.74
ranch
-0.71
orthy
-0.70
ilion
-0.69
uda
-0.69
oshi
-0.68
lot
-0.67
unta
-0.66
anch
-0.65
POSITIVE LOGITS
emails
3.64
Emails
2.95
2.37
mails
2.26
emailed
1.88
1.85
tweets
1.73
1.73
memos
1.69
Gmail
1.65
Activations Density 0.011%