> Attempt 2: LLM-as-a-judge Wouldnt this be achievable with a classifier model? Maybe even a combo of getting the embedding and then putting it through a class

> Attempt 2: LLM-as-a-judge Wouldnt this be achievable with a classifier model? Maybe even a combo of getting the embedding and then putting it through a classifier? Kind of like how Gans work. Edit: I read the article before the comment section, silly me lol