MEDLEY-BENCH: Benchmarking Behavioural Metacognition and Belief Revision Under Social Pressure in Large Language Models · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2604.16009/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTMEDLEY-BENCH: Benchmarking Behavioural Metacognition and Belief Revision Under Social Pressure in Large Language Models — mentioned by GodelNumberinghttps://news.ycombinator.com/item?id=48946381urn:hackerpapers:mention:2604.16009:48946381Fri, 17 Jul 2026 12:09:32 GMT<p>&gt; &quot;Finding 1: Scale Buys Evaluation, Not Control&quot; The attached paper&#x27;s (https://arxiv.org/pdf/2604.16009) title is &quot;MEDLEY-BENCH: Scale Buys Evaluation but Not Control in AI Metacognition&quot; This is the most blatant Claude line, or as Claude would put it, the smoking gun.</p><p><a href="https://hackerpaper.kaipfister.de/p/2604.16009/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.AI