Yep yep I know, but I was trying to suggest something diffusion-like occurring with a language model through a totally separate mechanism that does not rely on the denoising process (at least not literally).
It kinda does that but one token at a time.
I'm fairly certain diffusion refers to the overall architecture, not the emergent self-organization process.
It refers to the process of taking noise and "diffusing" it until visually appealing