INDEX
Explanations
job postings
The neuron detects the boilerplate lead-in words and short function phrases used to introduce job requirements—things like “We,” “are,” “seeking,” “a,” and similar section-starter tokens.
New Auto-Interp
Negative Logits
NEG
-0.07
게임
-0.07
outro
-0.06
revenge
-0.06
پیش
-0.06
Dol
-0.06
두
-0.06
Το
-0.06
수상
-0.06
gin
-0.06
POSITIVE LOGITS
showDialog
0.08
بواسطة
0.08
MPI
0.07
จะต
0.07
0.07
report
0.07
有点
0.06
shareholder
0.06
fillType
0.06
调查
0.06
Activations Density 0.010%