53 点作者 jonbaer10 个月前

3 条评论

Charon7710 个月前

Basically GAN but for LLMs? Neat

Very interesting. I always see the two techniques as complementary.

(openAI) Previous discussion:<p><a href="https://news.ycombinator.com/item?id=40988076">https://news.ycombinator.com/item?id=40988076</a>

Prover-Verifier Games improve legibility of LLM outputs