English(EN) TrustMod-SM: A Multi-Axis Benchmark for Evaluating Trustworthiness of LLMs in Social Media Content Moderation

新基准揭示大型语言模型在内容审核方面可信度存在差距

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

研究人员推出 TrustMod-SM，这是一个旨在评估大型语言模型（LLM）在社交媒体内容审核中可信度的新基准。该基准包含来自八个数据集的约 29,000 个实例，在五个关键维度上评估 LLM：可信度、公平性、安全性、鲁棒性和上下文完整性。对 16 个开源模型和 1 个闭源前沿模型的初步评估显示，可信度方面存在显著差距，更高的准确性通常与人口统计学差异相关。模型在隐藏提示、对错误预测的自信以及区分仇恨言论与反击言论或被回收语言方面也存在问题，特别是对于较小的模型。 AI

影响强调了大型语言模型在内容审核方面可信度的关键局限性，表明需要超越简单准确性的更强有力的评估。

排序理由该集群描述了一篇介绍用于评估大型语言模型基准的新学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 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: 37 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群描述了一篇介绍用于评估大型语言模型基准的新学术论文。[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: 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) · Aizan Zafar, Sreekantam Sai Venkata, Gathik Jindal, Ritu Raj Sharma, Zishan Ahmad, Tulika Saha, Srinath Srinivasa · 2026-11-11 05:00

TrustMod-SM：用于评估大型语言模型在社交媒体内容审核中可信度的多轴基准

arXiv:2611.12970v1 Announce Type: new Abstract: Large Language Models (LLMs) show strong potential for social media content moderation, but their trustworthiness beyond overall accuracy remains underexplored. We introduce TrustMod-SM, a benchmark of roughly 29,000 instances acros…

本文如何被排名

报道来源 [1]

TrustMod-SM：用于评估大型语言模型在社交媒体内容审核中可信度的多轴基准

相关实体

相关话题