Agentic Abstention：研究 LLM 智能体何时该停止行动而非继续调用工具
Agentic Abstention: Do Agents Know When to Stop Instead of Act?
AI 导读
论文定义了 Agentic Abstention，即智能体在不确定时应停止行动的序贯决策问题。研究在网页购物、终端环境和问答任务上评测 13 个 LLM 智能体系统与 2 种 agent scaffold，超过 28,000 条任务指令，发现部分智能体该停止时不停止，另一些则经过多次不必要的交互后才放弃，更大或更强的模型有时反而不擅长及时弃权。
推荐理由
论文提出智能体该在何时停止行动的问题，并以大规模评测和可复用方法给出改善路径，数据集和代码已开放。
正文 · 原文
Abstract:LLM agents are expected to act over multiple turns, using search, browsing interfaces, and tools to complete user goals. Yet not every goal is well specified or achievable in an available environment. In such cases, a reliable agent should recognize that further interaction is unlikely to help and abstain from additional tool calls. We define Agentic Abstention, the problem of deciding when an agent should stop acting under uncertainty. Unlike standard LLM abstention, which is usually evaluated as a single-turn answer-or-abstain decision, agentic abstention is a sequential decision problem: an agent can answer, abstain, or gather more information at each turn, and the need to abstain may only become clear after interacting with the environment. We study this problem across web shopping, terminal environments, and question answering, evaluating 13 LLM-as-agent systems and 2 agent scaffolds on more than 28,000 task instructions. Our results show that the main challenge is not only whether agents can abstain, but also when they abstain. Some agents never abstain when they should, while others do so only after many unnecessary interactions. This gap is especially large on tasks where the instruction appears feasible until the environment reveals otherwise (e.g., no valid result matches the instruction). We further find that model scale, reasoning, and agent scaffolding affect abstention in different ways, where larger or more capable models sometimes perform worse at timely abstention. Finally, we introduce CONVOLVE, a context engineering method for improving agentic abstention that distills interaction trajectories into reusable stopping rules. On held-out WebShop tasks, CONVOLVE substantially improves timely abstention without updating model parameters, raising Llama-3.3-70B's timely recall rate from 26.7 to 57.4. Our dataset and code are available at this https URL.
| Comments: | Accepted to NeurIPS 2026. Updated camera-ready version with revised references and minor corrections |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.28733 [cs.AI] |
| (or arXiv:2606.28733v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2606.28733 arXiv-issued DOI via DataCite |
Submission history
From: Han Luo [view email]
[v1]
Sat, 27 Jun 2026 04:49:37 UTC (1,067 KB)
[v2]
Thu, 5 Nov 2026 12:19:33 UTC (1,076 KB)
来源：arXiv cs.AI · arxiv.org