INDEX
Explanations
The neuron detects on-page email/share prompts (e.g. “Email this page to a friend”).
New Auto-Interp
Negative Logits
Flags
-0.07
också
-0.07
emales
-0.06
flying
-0.06
Rak
-0.06
秦
-0.06
grim
-0.06
≧
-0.06
Tam
-0.06
panc
-0.06
POSITIVE LOGITS
wt
0.07
wią
0.06
iteral
0.06
*)(
0.06
.assertNot
0.06
mention
0.06
toast
0.06
quint
0.06
orderby
0.06
.amazonaws
0.06
Activations Density 0.003%