智能体自主放弃：智能体知道何时停止而非行动吗？

摘要

LLM智能体被期望通过多轮交互，使用搜索、浏览界面和终端工具来完成用户目标。然而，并非所有目标都能在现有环境中得到明确定义或实现。在这种情况下，一个可靠的智能体应当认识到进一步的交互难以奏效，并放弃调用额外的工具。我们提出了"智能体自主放弃"（Agentic Abstention）这一概念，即智能体在不确定性下决定何时停止行动的问题。与通常被评估为单轮"回答或放弃"决策的标准LLM自主放弃不同，智能体自主放弃是一个序列决策问题：智能体在每一轮中可以回答、放弃或收集更多信息，而放弃的需求可能只有在与环境交互后才变得明确。我们在网络购物、终端环境和问答任务中研究了这一问题，评估了13个LLM智能体系统和2个智能体脚手架在超过28,000个任务上的表现。结果表明，主要挑战不仅在于智能体能否放弃，还在于它们何时放弃。有些智能体在应当放弃时从未放弃，而另一些则在进行了大量不必要的交互后才放弃。这种差距在那些指令看似可行但环境最终揭示其不可行（例如，没有有效结果匹配指令）的任务中尤为明显。我们进一步发现，模型规模、推理能力和智能体脚手架以不同方式影响自主放弃，其中更大或更强的模型有时在及时放弃方面表现更差。最后，我们提出了CONVOLVE，一种通过将完整交互轨迹蒸馏为可复用的停止规则来改善智能体自主放弃的上下文工程方法。在WebShop上，CONVOLVE在不更新模型参数的情况下显著提高了及时放弃率，将Llama-3.3-70B的及时召回率从26.7提升至57.4。我们的数据集和代码可在 https://lhannnn.github.io/agentic-abstention 获取。

English

LLM agents are expected to act over multiple turns, using search, browsing interfaces, and terminal tools to complete user goals. Yet not every goal is well specified or achievable in the 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 tasks. 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 full interaction trajectories into reusable stopping rules. On WebShop, 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 https://lhannnn.github.io/agentic-abstention