让AI长时推理更靠谱，通过分层管理上下文避免出错积累。
COMPASS: Enhancing Agent Long-Horizon Reasoning with Evolving Context
- 分三模块：主代理执行、元思考者监督、上下文管理者精炼信息
- 在三个基准上准确率提升最高达20%，优于单/多代理基线
- 适合需要长时间规划与纠错的复杂任务研究者使用
长时序任务中，大模型代理因持续推理和多次工具调用面临挑战：小错误随步骤累积，甚至顶尖模型也会幻觉或失去连贯性。我们识别出上下文管理是核心瓶颈——过长的历史导致代理忽略关键证据或被无关信息干扰，无法及时重规划或反思错误。为此，提出COMPASS（上下文组织的多智能体规划与策略系统），一个轻量级分层框架，将战术执行、战略监督和上下文组织分离为三个专用组件：(1) 主代理负责推理与工具使用，(2) 元思考者监控进展并发出战略干预，(3) 上下文管理者为不同推理阶段维护简洁相关的进度摘要。在GAIA、BrowseComp和Humanity's Last Exam三个挑战性基准上，COMPASS相比单/多代理基线准确率提升最高达20%。我们还引入测试时扩展以达到深度研究型代理水平，并设计后训练流程，将上下文管理交给小型模型以提升效率。
原文摘要 · Abstract (English)
Long-horizon tasks that require sustained reasoning and multiple tool interactions remain challenging for LLM agents: small errors compound across steps, and even state-of-the-art models often hallucinate or lose coherence. We identify context management as the central bottleneck -- extended histories cause agents to overlook critical evidence or become distracted by irrelevant information, thus failing to replan or reflect from previous mistakes. To address this, we propose COMPASS (Context-Organized Multi-Agent Planning and Strategy System), a lightweight hierarchical framework that separates tactical execution, strategic oversight, and context organization into three specialized components: (1) a Main Agent that performs reasoning and tool use, (2) a Meta-Thinker that monitors progress and issues strategic interventions, and (3) a Context Manager that maintains concise, relevant progress briefs for different reasoning stages. Across three challenging benchmarks -- GAIA, BrowseComp, and Humanity's Last Exam -- COMPASS improves accuracy by up to 20% relative to both single- and multi-agent baselines. We further introduce a test-time scaling extension that elevates performance to match established DeepResearch agents, and a post-training pipeline that delegates context management to smaller models for enhanced efficiency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。