INDEX
Explanations
The neuron fires on the recurring editorial sign-off phrase “As always,” in personal blog–style posts.
New Auto-Interp
Negative Logits
sondern
-0.07
عبار
-0.07
(“
-0.06
.contentType
-0.06
розгля
-0.06
nějaké
-0.06
avoids
-0.06
.drawRect
-0.06
говорит
-0.06
("-0.06
POSITIVE LOGITS
Conor
0.07
Carrier
0.07
utos
0.07
Moses
0.06
Carrier
0.06
ोस
0.06
(img
0.06
afi
0.06
.Constant
0.06
�
0.06
Activations Density 0.010%