Comment by mlmonkey
15 hours ago
"Attention" is just a matmul. Q = KV/sqrt(d) etc.
I don't see how any planning is done in latent space. Can you point me to any papers? Thanks.
Edit: Oh, I see you're probably talking about CoCoNuT? Do all frontier models us it nowadays?
There's a lot of research on this topic. https://arxiv.org/abs/2303.08112 and https://arxiv.org/abs/2311.04897 are just two examples that come to mind