INDEX
Explanations
references to meal occasions, particularly lunch and dinner
New Auto-Interp
Negative Logits
dawn
-0.17
andin
-0.17
bis
-0.16
erer
-0.16
oser
-0.16
urrency
-0.15
addir
-0.15
breakfast
-0.15
Dawn
-0.14
-opacity
-0.14
POSITIVE LOGITS
time
0.38
-time
0.30
/sn
0.26
-hour
0.24
ware
0.23
times
0.22
TIME
0.21
hour
0.21
ime
0.20
hour
0.20
Activations Density 0.039%