INDEX
Explanations
probability calculations
This neuron detects tokens that name or enumerate the available items (letters) in the sampling population (e.g. the letters and their counts in the “from {…}” or repeated‐letter multiset specification).
New Auto-Interp
Negative Logits
_optional
-0.07
Verde
-0.07
-0.06
_square
-0.06
muzzle
-0.06
ihanna
-0.06
庄
-0.06
laps
-0.06
twig
-0.06
VisualStyleBackColor
-0.06
POSITIVE LOGITS
사가
0.07
esas
0.06
.doc
0.06
Diabetes
0.06
breaks
0.06
_JO
0.06
муль
0.06
.Quad
0.06
_sep
0.06
acest
0.06
Activations Density 0.001%