You must log in or register to comment.
Using a complex GPT-4o prompt, they sought to pull out tweets that focused on “superficial topics”
Wait a moment… They asked an LLM, to tell them what was “junk”, and another LLM, trained on what an LLM marked as junk, turned out to be a junk LLM?
It talks about model collapse, but this smells like research collapse.
Just like humans.
no, training LLM on other LLM, on unverified facts, or sources leads to brain rot.
same, same, but different, but still the same
Is junk data just YouTube content because of so, then it’s already poisoned my kids and their friends.
Whatever 67 means.
