Hacker News new | ask | show | jobs
by fulafel 214 days ago
Sounds like the real sandbox in this scenario is the alignment training of the LLMs you tried.