News for hackers. Stuff that matters.

A comment on 30% Drop In o1-Preview Accuracy When Putnam Problems Are Slightly Variated. Parent comment

Re:30% Drop In o1-Preview Accuracy When Putnam Problems Are ... by simonw on Thursday January 02, 2025 @06:47AM (#42573642)
Watch for ChatGPT or Claude saying "analyzing" - which means they have identified they need to run a calculation and outsourced it to Python (ChatGPT) or JavaScript (Claude)

The poem thing probably causes them to not decide to use those tools.

[ Parent | on HN ]

There are no comments on this comment yet.