4 ms·
Prover-Verifier Games improve legibility of LLM outputs
- CodeGroyper 2y agoVery interesting. I always see the two techniques as complementary.
- Charon77 2y agoBasically GAN but for LLMs? Neat
- vinnyvichy 2y ago(openAI) Previous discussion: https://news.ycombinator.com/item?id=40988076 https://news.ycombinator.com/item?id=40988076