Want to create an interactive transcript for this episode?
Podcast: Arxiv Papers
Episode: [QA] Reasoning Models Don't Always Say What They Think
Description: The study evaluates the faithfulness of chain-of-thought reasoning in AI models, finding limitations in monitoring effectiveness and suggesting it may not reliably detect undesired behaviors during training.https://arxiv.org/abs//2505.05410YouTube: https://www.youtube.com/@ArxivPapersTikTok: https://www.tiktok.com/@arxiv_papersApple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers