LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2408.08656/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTLLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs — mentioned by jbarrowhttps://news.ycombinator.com/item?id=42636164urn:hackerpapers:mention:2408.08656:42636164Wed, 08 Jan 2025 17:01:36 GMT<p>Went through the structured outputs chapter, and honestly feels like a good resource! Covered a lot of the gotchas and current research into whether or not structured outputs hurts (e.g. one of the linked papers: https://arxiv.org/abs/2408.08656). Not the biggest fan of the associated autogenerated podcast though.</p><p><a href="https://hackerpaper.kaipfister.de/p/2408.08656/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CL