arXiv cs.AI· Xun Liu, Yi Evie Zhang, Vira Kasprova, Parisa Rabbani, Pardis Sadat Zahraei, Tianyu Zhang, Ali Ebrahimpour-Boroojeny, Varun Chandrasekaran·· 2 小时前AI 评分58
AgentAbstain 基准:LLM 智能体能否知道何时不应行动
AgentAbstain: Do LLM Agents Know When Not to Act?
AI 导读
研究者提出 AgentAbstain,首个系统性评估工具型 LLM 智能体弃权能力的框架,包含 263 个成对任务、42 个可执行沙盒环境和 AbstainGen 自动生成管线。在 17 个前沿模型的 4 种智能体框架中,最佳者 Gemini 3.1 Pro 成对准确率仅 59.5%,且弃权能力与通用任务能力基本独立,还存在先执行不可逆动作再触发弃权的失效模式。代码与数据集已开源。
正文 · 原文
Abstract:Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, yet existing evaluations mostly focus on task success rather than whether agents know when to abstain. This gap poses real risks: under ambiguity, conflicting constraints, or tool failures, agents may execute unintended and irreversible actions. To close this gap, we present the first systematic evaluation framework for agentic abstention: the calibrated ability of tool-using LLM agents to recognize when not to act. At its core, AgentAbstain is a paired-task benchmark built on an agent-native taxonomy of 8 abstention scenarios across pre-execution reasoning and runtime discovery. It contains 263 paired tasks across 42 executable sandbox environments, where each pair consists of a should-act task and a should-abstain variant produced through a controlled perturbation to the instruction, tool, or environment state. To scale this paired design and resist data contamination, we propose AbstainGen, a fully automated pipeline that synthesizes sandbox environments and generates paired tasks end-to-end, validated by deterministic replay and semantic LLM judges; fresh task instances can be regenerated on demand, and three independent annotators rate 94-98% of sampled tasks as well-designed. Across 17 frontier LLMs in 4 agent harnesses, the best agent (Gemini 3.1 Pro) achieves only 59.5% paired accuracy (correct on both the act and abstain sides of each paired task). More importantly, abstention capability is largely independent of general task-solving capability, indicating that scaling task-solving alone will not close this gap. We further identify failure modes such as post-hoc abstention, in which agents execute irreversible actions before recognizing abstention triggers. Our code and dataset are open-sourced at this http URL.
| Comments: | 65 pages, 13 figures |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.10059 [cs.AI] |
| (or arXiv:2607.10059v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.10059 arXiv-issued DOI via DataCite |
Submission history
From: Xun Liu [view email]
[v1]
Sat, 11 Jul 2026 00:57:36 UTC (7,775 KB)
[v2]
Wed, 18 Nov 2026 03:51:43 UTC (7,787 KB)
来源:arXiv cs.AI · arxiv.org