INDEX
Explanations
adjectives expressing a degree of something
instances of the word "somewhat" indicating mild qualification or hedging statements
New Auto-Interp
Negative Logits
abad
-0.81
iens
-0.80
aring
-0.75
mberg
-0.73
rights
-0.72
yrs
-0.72
venant
-0.71
ults
-0.70
ocaust
-0.69
tsky
-0.69
POSITIVE LOGITS
independ
0.84
tang
0.80
resembling
0.80
akin
0.79
unusual
0.78
amusing
0.78
reminiscent
0.78
analogous
0.77
surprising
0.77
obscure
0.77
Activations Density 0.008%