Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

They have many ways to manipulate the LLM's results, for example they can use a lot of the same mechanisms that are used to block or filter out inappropriate material.


Given that there are entire forums devoted to successfully doing just that (easily) my point stands.




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

Search: