INDEX
Explanations
references to events happening in the evening or during the night
references to the word "tonight" across various contexts
New Auto-Interp
Negative Logits
llor
-0.82
76561
-0.80
gans
-0.70
onne
-0.70
Pros
-0.70
ophers
-0.68
Gall
-0.66
aments
-0.64
sac
-0.64
aunch
-0.64
POSITIVE LOGITS
tonight
1.16
afternoon
1.04
Tonight
1.02
evening
0.97
morning
0.96
Tonight
0.92
night
0.84
night
0.82
Evening
0.76
morning
0.75
Activations Density 0.007%