INDEX
Explanations
adverbs that intensify an adjective
the emphasis on the word "really."
New Auto-Interp
Negative Logits
heid
-0.77
enary
-0.76
ean
-0.68
oire
-0.65
iture
-0.65
heny
-0.63
fold
-0.62
cair
-0.62
eff
-0.62
Receiver
-0.61
POSITIVE LOGITS
REALLY
0.90
bothered
0.88
bothering
0.83
darn
0.81
bother
0.80
mathemat
0.78
pissed
0.74
differentiated
0.74
messed
0.73
really
0.72
Activations Density 0.033%