INDEX
Explanations
News articles
This neuron detects mentions of senior or “older” counselors (i.e. references to more experienced staff members).
New Auto-Interp
Negative Logits
_wait
-0.07
النس
-0.07
intoxicated
-0.06
بینی
-0.06
rainfall
-0.06
structions
-0.06
Optional
-0.06
urls
-0.06
미
-0.06
الأول
-0.06
POSITIVE LOGITS
мит
0.07
renc
0.07
dared
0.06
_limit
0.06
diagon
0.06
gerade
0.06
∈
0.06
cat
0.06
(CG
0.06
menj
0.06
Activations Density 0.256%