INDEX
Explanations
time-related phrases or expressions
references to prolonged durations or the concept of a long time
New Auto-Interp
Negative Logits
ramid
-0.84
otto
-0.78
riad
-0.77
rophic
-0.76
aye
-0.70
tee
-0.66
fter
-0.64
cham
-0.64
aleb
-0.63
bathrooms
-0.62
POSITIVE LOGITS
ago
1.08
Ago
0.76
period
0.75
flies
0.75
duration
0.75
periods
0.75
(>
0.72
(~
0.70
ago
0.70
frames
0.69
Activations Density 0.025%