That is pretty cool. I found a follow up work that applies this technique to LLM: https://konczer.github.io/doc/Poster_EEML23_JozsefKonczer.pdf https://konczer.

That is pretty cool. I found a follow up work that applies this technique to LLM: https://konczer.github.io/doc/Poster_EEML23_JozsefKonczer.pdf https://konczer.github.io/doc/Poster_EEML23_JozsefKonczer.pd...