Papers · organized/paper_cards

论文

1 张论文卡片 · 安全与风险 · 其他正式来源

开放获取 全部 绿色 · 724
4.2 Reliability 不等于成功率:12 指标拆出 consistency / robustness / predictability / safety
4.2 Reliability 不等于成功率:12 指标拆出 consistency / robustness / predictability / safety(⭐⭐⭐⭐⭐)
arXiv:2602.16666 安全与风险 方法 Open MIND OA · 绿色 被引 45 · S2

本工作提出 12 个具体指标,从一致性、鲁棒性、可预测性和安全性四个关键维度分解 Agent 可靠性,可与传统评估互补,并提供用于分析 Agent 表现、退化与失败方式的工具。This work proposes twelve concrete metrics that decompose agent reliability along four key dimensions: consistency, robustness, predictability, and safety, which complement traditional evaluations while offering tools for reasoning about how agents perform, degrade, and fail.