logoalt Hacker News

keedayesterday at 9:32 PM0 repliesview on HN

Yep, the last part in my post was suggesting some ways we could determine if "item X was in the training data" (as well as some potential blockers for that from OpenAI's perspective.)