Hacker News new | past | comments | ask | show | jobs | submit
I'd be very surprised if this is in the training data given that most models mess it up to this day. E.g. look at the ones from Opus.