English(EN) One Temperature to Rule Them All?

新研究探讨大型语言模型推理的最佳温度设置

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-11 05:00

研究人员调查了大型语言模型中温度采样（一种控制推理过程中可靠性和多样性之间平衡的机制）的使用情况。虽然一些研究表明在不同样本中使用不同的温度可以提高性能，但固定温度方案仍然很常见。本文分析了 pass@K 设置，研究了温度如何影响不同模型和数据集上的成功概率。研究结果表明，虽然最佳的多温度分配可以提高性能，但与单温度基线相比，收益通常很小，尤其是在根据采样预算调整温度时。 AI

影响阐明了多温度采样何时有利于 LLM 推理以及如何优化温度选择。

排序理由该集群包含一篇讨论大型语言模型技术方面的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

arXiv
pass@k

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 35 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇讨论大型语言模型技术方面的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Aviv Orly (School of Physics and Astronomy, Tel Aviv University, Israel), Ori Shem-Ur (School of Physics and Astronomy, Tel Aviv University, Israel), Yaron Oz (School of Physics and Astronomy, Tel Aviv University, Israel) · 2026-11-11 05:00

一个温度，统治所有？

arXiv:2611.13077v1 Announce Type: cross Abstract: Inference-time sampling is a central mechanism for improving large language models, with temperature controlling the tradeoff between reliability and diversity. While recent work suggests that varying temperature across samples ca…

本文如何被排名

报道来源 [1]

一个温度，统治所有？

相关实体

相关话题