Prompting Science Report 3: I'll pay you or I'll kill you -- but will you care? · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/2508.00614/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTPrompting Science Report 3: I'll pay you or I'll kill you -- but will you care? — mentioned by cdrinihttps://news.ycombinator.com/item?id=45083934urn:hackerpapers:mention:2508.00614:45083934Sun, 31 Aug 2025 15:28:26 GMT<p>The best research I&#x27;ve seen on this is: - Threatening or tipping a model generally has no significant effect on benchmark performance. - Prompt variations can significantly affect performance on a per-question level. However, it is hard to know in advance whether a particular prompting approach will help or harm the LLM&#x27;s ability to answer any particular question. https://arxiv.org/abs/2508.00614</p><p><a href="https://hackerpaper.kaipfister.de/p/2508.00614/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CL