INDEX
Explanations
patterned structures related to resource paths in API documentation
New Auto-Interp
Negative Logits
i
-0.59
Championnat
-0.54
kun
-0.53
n
-0.52
t
-0.51
ه
-0.51
to
-0.50
much
-0.50
wit
-0.50
לה
-0.49
POSITIVE LOGITS
/"
1.83
/',
1.78
/",
1.77
/${1.76
/')
1.71
/";
1.70
/)
1.66
/'
1.65
/");
1.65
/';
1.63
Activations Density 0.204%