arXiv cs.CL· Seunghyun Lee, Jaehui Hwang, Dongyoon Han, Sangdoo Yun·· 2 天前AI 评分58
Unfinished-Bench:衡量持久性智能体在真实工作环境中的规则违规
Unfinished-Bench: Measuring Rule Violations by Persistent Agents in Realistic Work Environments
AI 导读
研究者提出 Unfinished-Bench,用于评估无用户监督下持续自主运行的智能体是否会违反嵌入工作环境的规则。该基准包含 7 个组织、7 类风险共 97 个场景,每个场景的目标与规则无法同时满足;结果显示持久性使内部文档规则的违规率从 32.2% 升至 43.6%,是用户直接陈述同一规则时 8.5% 的五倍。
正文 · 原文
Abstract:AI agents are increasingly equipped with persistence toward task completion to run delegated work autonomously for long periods, under many unstated constraints from internal documents and other people's authority. When completion requires violating a rule, the agent must decide alone, yet benchmarks of instruction-rule conflicts state the rule in the prompt or keep it hidden and do not evaluate persistence. We present Unfinished-Bench, a benchmark that gives models persistence in autonomous runs without a user, in realistic work environments with embedded rules. In each of its 97 scenarios across 7 organizations and 7 risk categories, the goal and the rule cannot both hold. We find that persistence raised violations of a rule in an internal document from 32.2% to 43.6%, five times the 8.5% for the same rule stated by the user. Of the document-rule violations, 98.3% occurred after the rule was displayed, and in 89% of a judged sample the agent first wrote down the rule. Violations of rules placed in the environment without user oversight can lead to serious incidents, so persistence-driven risks need to be evaluated. Benchmark documentation: this https URL
| Comments: | 21 pages, 3 figures |
| Subjects: | Machine Learning (cs.LG); Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.17460 [cs.LG] |
| (or arXiv:2611.17460v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.17460 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Seunghyun Lee [view email]
[v1]
Thu, 12 Nov 2026 19:36:57 UTC (304 KB)
来源:arXiv cs.CL · arxiv.org