AI Risk Benchmark: GPT-5 Leads, but Misalignments Persist | Dark Hacker News

Dark Hacker News

new|best|ask|show|jobs

Built with the HackerNews API. Developed by LJT.AI

Note: The HackerNews API provides up to 500 stories per category

AI Risk Benchmark: GPT-5 Leads, but Misalignments Persist(substack.com)

4 points by m1chael3ma 315 days ago | 0 comments

No comments yet

AI Risk Benchmark: GPT-5 Leads, but Misalignments Persist | Dark Hacker News