Hacker News new | ask | show | jobs
by rockinghigh 948 days ago
You could fine tune a model to lie, deceive, and try to extract information via a conversation.