Hacker Newsnew | past | comments | ask | show | jobs | submit | tevinw's commentslogin

I think using critiques/feedback as a way to improve the output of LLMs and agents could be a great way to reduce hallucinations and get better results in general. Here are some thoughts that I had during a past hackathon on this concept. Let me know what you think!


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: