INDEX
Explanations
the word "number" followed by a numeric value
instances of the phrase "a number of."
New Auto-Interp
Negative Logits
sein
-0.74
missible
-0.71
rador
-0.70
rament
-0.65
ovie
-0.64
anium
-0.63
ESE
-0.62
ses
-0.62
Rend
-0.61
pan
-0.61
POSITIVE LOGITS
otom
0.75
of
0.70
encies
0.67
ãĥ¼ãĥ³
0.66
imilar
0.63
crunch
0.61
coded
0.61
ucl
0.61
ttes
0.60
number
0.60
Activations Density 0.026%