INDEX
Explanations
Math and probability questions
This neuron detects decimal numbers, especially probabilities expressed as floating‐point values.
New Auto-Interp
Negative Logits
zen
-0.07
-Saharan
-0.06
_as
-0.06
swear
-0.06
gpointer
-0.06
_Rect
-0.06
So
-0.06
towards
-0.06
agree
-0.06
attention
-0.06
POSITIVE LOGITS
([ ↵
0.08
.FlatStyle
0.07
mad
0.07
지방
0.07
çok
0.06
.sky
0.06
uele
0.06
">//
0.06
üş
0.06
'){
↵0.06
Activations Density 0.013%