English(EN) TRIDENT: Benchmarking LLM Safety in Finance, Medicine, and Law

新的基准测试和研究揭示了LLM安全与性能的权衡 · 已追踪2个来源

作者 PulseAugur 编辑部 · [2 个来源] · 2026-07-28 04:00

两篇新研究论文探讨了大语言模型（LLM）的安全与性能权衡。第一篇论文介绍了Trident-Bench，一个旨在专门评估金融、医疗和法律领域LLM安全性和合规性的基准测试，并强调虽然通用模型表现出基本能力，但领域特定模型在伦理细微问题上常常表现不佳。第二篇论文研究了LLM的越狱防御如何负面影响性能，增加对良性输入的过度拒绝，并提高推理成本，同时提供了基于部署限制选择防御机制的指导。 AI

影响这些研究突出了LLM开发的关键领域，强调了领域特定安全改进的必要性，以及在权衡效用和安全性的同时，仔细考虑防御机制的重要性。

排序理由该集群包含两篇学术论文，介绍了与LLM安全和性能相关的新基准测试和分析。

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CL TIER_1 English(EN) · Zheng Hui, Yijiang River Dong, Ehsan Shareghi, Nigel Collier · 2026-07-28 04:00

TRIDENT: Benchmarking LLM Safety in Finance, Medicine, and Law

arXiv:2507.21134v2 Announce Type: replace Abstract: As large language models (LLMs) are increasingly deployed in high-risk domains such as law, finance, and medicine, systematically evaluating their domain-specific safety and compliance becomes critical. While prior work has larg…
arXiv cs.LG TIER_1 English(EN) · Tong Zhang, Zexin Li, Simin Chen, Yun Peng · 2026-07-28 04:00

When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs

arXiv:2607.24392v1 Announce Type: cross Abstract: Jailbreak defenses are essential for protecting large language models (LLMs), but they can also introduce secondary costs that weaken model utility. We present a systematic study of these defense trade-offs along three dimensions:…

报道来源 [2]

TRIDENT: Benchmarking LLM Safety in Finance, Medicine, and Law

When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs

相关实体

相关话题