English(EN) DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models

新的基准测试DistressBench评估LLM的危机支持能力

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-12 04:00

研究人员开发了DistressBench，这是一个新的评估数据集，用于评估大型语言模型（LLM）在危机情况（特别是涉及自杀和自残的情况）下提供支持的能力。与专注于拒绝的现有基准测试不同，DistressBench使用跨越危机支持七个维度的加权标准评估体系，包括降级处理和将用户转介给人工帮助。评估发现，即使是表现最好的前沿模型也只满足了88.4%的加权标准，其中相当一部分模型在识别危机后未能将用户转介给适当的资源。该研究强调，需要评估LLM提供的护理质量，而不仅仅是它们拒绝回应的能力。 AI

影响为评估LLM在敏感危机场景下的安全性树立了新标准，推动其超越简单的拒绝，发展更细致的支持能力。

排序理由该条目描述了一篇介绍新LLM评估基准的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 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: 11 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目描述了一篇介绍新LLM评估基准的学术论文。[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, 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: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Drew Rein, Patrick Oathout, Vishal Kumar, Udari Madhushani Sehwag · 2026-10-12 04:00

DistressBench：评估大型语言模型的多维度危机支持能力

arXiv:2610.12758v1 Announce Type: new Abstract: Language models have been adopted faster than evaluation practices that reliably assess their risks, and a growing share of use is affective rather than informational. Within that traffic users disclose suicide and self-harm (SSH) c…

本文如何被排名

报道来源 [1]

DistressBench：评估大型语言模型的多维度危机支持能力

相关实体

相关话题