English(EN) Differentially Private Statistical Inference for Rankings from Pairwise Comparisons

新方法为大型语言模型的排名提供差分隐私推断

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

研究人员开发了从人类成对比较中推断排名的统计新方法，特别关注大型语言模型。这些方法通过在Bradley-Terry模型中采用差分隐私技术，确保个人偏好数据的隐私性。该方法为分数和排名的估计误差提供了有限样本界限和渐近有效的置信区间，并使用自举程序进行同步排名推断。这些技术被应用于根据创意写作能力对大型语言模型进行排名。 AI

影响增强了基于人类偏好评估和比较大型语言模型的隐私保护方法。

排序理由该集群包含一篇学术论文，详细介绍了具有差分隐私的大型语言模型排名的统计新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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: 34 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇学术论文，详细介绍了具有差分隐私的大型语言模型排名的统计新方法。[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, safety; 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.LG TIER_1 English(EN) · Xintao Xia, Jia Gu · 2026-11-11 05:00

来自成对比较的排序的差分隐私统计推断

arXiv:2611.13485v1 Announce Type: cross Abstract: Reliable inference for rankings of large language models based on human pairwise comparisons must preserve the privacy of individual preference records. To meet this need, we develop methods for score and rank inference under the …

本文如何被排名

报道来源 [1]

来自成对比较的排序的差分隐私统计推断

相关实体

相关话题