INDEX
Explanations
news and editorial
The neuron activates on promotional or subscription‐style text—headings and phrases advertising industry publications, membership benefits, events, or premium access.
New Auto-Interp
Negative Logits
)y
-0.07
若
-0.07
고객
-0.07
роме
-0.06
若
-0.06
clinicians
-0.06
�a
-0.06
〜
-0.06
deniz
-0.06
Kron
-0.06
POSITIVE LOGITS
Await
0.07
ostat
0.06
Smartphone
0.06
bíl
0.06
세
0.06
стин
0.06
cpf
0.06
-bodied
0.06
substit
0.06
าด
0.06
Activations Density 0.031%