AI Control: Improving Safety Despite Intentional Subversion · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2312.06942/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTAI Control: Improving Safety Despite Intentional Subversion — mentioned by andrewedstromhttps://news.ycombinator.com/item?id=39186223urn:hackerpapers:mention:2312.06942:39186223Tue, 30 Jan 2024 03:54:46 GMT<p>You may be interested in a recent AI safety paper by Redwood Research. In it, they have GPT-4 generate solutions to coding problems, but instruct it to insert backdoors into the solutions some fraction of the time. Then, they explore different ways to use a weaker model (GPT-3.5) to detect these backdoors. Pretty interesting. [1] https://arxiv.org/abs/2312.06942</p><p><a href="https://hackerpaper.kaipfister.de/p/2312.06942/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.LG