INDEX
Explanations
business coaching and enterprise documentation
The neuron detects prominent document headings and titles (section headers, bolded/title-case phrases and other high-salience header text).
New Auto-Interp
Negative Logits
Федера
0.42
<unused2121>
0.40
нов
0.36
ху
0.35
láp
0.33
българ
0.33
)”
0.32
0.32
”(
0.32
ጥቅም
0.32
POSITIVE LOGITS
言っ
0.43
उक्त
0.40
etc
0.39
etc
0.38
"..
0.38
Etc
0.37
पिछले
0.37
stringWith
0.37
دوسری
0.37
至於
0.36
Activations Density 0.207%