|
|
|
|
|
|
by SAI_Peregrinus
291 days ago
|
|
|
I think the heuristics are stupid, but happen to work with the current round of LLMs. The next round of training will work to stop the heuristics working, since if they work it means the LLM isn't correctly guessing the next token a human would produce. I don't think LLMs will become indistinguishable from humans any time soon, but I think it'll be a very long back & forth of training & finding new heuristics. |
|
|
This goes back to the initial comment you objected to: "if it works, it's not stupid".
I don't think anybody here is arguing they will work forever, or that they won't influence the next generation of LLMs.