INDEX
Explanations
The neuron activates on instructions for embedding or linking content—e.g. phrases like “embed the entire object” or “link to this page.”
New Auto-Interp
Negative Logits
rowned
-0.06
기술
-0.06
\Log
-0.06
screened
-0.06
Because
-0.06
were
-0.06
sns
-0.06
######
-0.06
Doctor
-0.06
wiki
-0.06
POSITIVE LOGITS
Muhammad
0.07
soph
0.06
.Sn
0.06
_rpc
0.06
Normalize
0.06
Litecoin
0.06
_Stop
0.06
conta
0.06
фі
0.06
.divide
0.06
Activations Density 0.000%