返回全部动态
硬件故障在LLM智能体中如何不同地显现？
原标题：How Do Hardware Faults Manifest Differently in LLM Agents?
AI 摘要
该研究首次系统评估了LLM智能体对随机硬件故障的韧性，在三个智能体基准上执行了6814次运行（2400+无故障、4300+注入随机双比特权重故障）。研究发现智能体执行会将模型级损坏转化为系统级结果：智能体可从故障输出中恢复并正确完成任务，但恢复会增加38.0%的平均输出token成本；部分故障产生语义上细微但重要的错误；在某些模型和场景下，故障还会改变安全与对齐行为。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
MLforSystems and AI4Good How Do Hardware Faults Manifest Differently in LLM Agents? Abstract We present the first systematic study of the resilience of large language model (LLM) agents to random hardware-induced faults. Prior work on hardware fault resilience in LLMs focuses on model-level outcomes, such as incorrect outputs or generation failures. Yet LLM agents repeatedly interact with tools and environments, creating loops that can change how hardware faults manifest. We conduct an extensiv
发布时间：2026-10-12 12:00
抓取时间：2026-10-12 13:03
来源机构：arXiv