INDEX
Explanations
email-related text such as prompts to enter email addresses
New Auto-Interp
Negative Logits
liga
-0.83
steen
-0.67
Scion
-0.65
TRY
-0.64
jury
-0.64
BP
-0.62
imposed
-0.62
geon
-0.61
itect
-0.61
OWER
-0.60
POSITIVE LOGITS
inbox
1.26
correspondence
1.24
address
1.11
1.05
sender
1.01
boxes
1.00
addresses
0.97
mailbox
0.95
messages
0.93
Address
0.93
Activations Density 1.951%