WebSailor-V2：通过合成数据与可扩展强化学习弥合与专有智能体之间的鸿沟

摘要

超越人类认知局限是LLM训练中的一个关键前沿领域。专有的智能代理系统，如DeepResearch，已在极其复杂的信息检索基准测试（如BrowseComp）中展现出超人类的能力，这一成就此前难以企及。我们认为，其成功关键在于开源模型所缺乏的一种高级推理模式：在浩瀚信息海洋中航行时，系统性地降低极端不确定性的能力。基于这一洞见，我们推出了WebSailor，一套完整的后训练方法论，旨在赋予模型这一关键能力。我们的方法包括通过结构化采样与信息模糊化生成新颖的高不确定性任务、RFT冷启动，以及一种高效的智能代理强化学习训练算法——复制采样策略优化（DUPO）。凭借这一集成流程，WebSailor在复杂信息检索任务中显著超越所有开源代理，与专有代理的性能相当，缩小了能力差距。

English

Transcending human cognitive limitations represents a critical frontier in LLM training. Proprietary agentic systems like DeepResearch have demonstrated superhuman capabilities on extremely complex information-seeking benchmarks such as BrowseComp, a feat previously unattainable. We posit that their success hinges on a sophisticated reasoning pattern absent in open-source models: the ability to systematically reduce extreme uncertainty when navigating vast information landscapes. Based on this insight, we introduce WebSailor, a complete post-training methodology designed to instill this crucial capability. Our approach involves generating novel, high-uncertainty tasks through structured sampling and information obfuscation, RFT cold start, and an efficient agentic RL training algorithm, Duplicating Sampling Policy Optimization (DUPO). With this integrated pipeline, WebSailor significantly outperforms all open-source agents in complex information-seeking tasks, matching proprietary agents' performance and closing the capability gap.

WebSailor-V2：通过合成数据与可扩展强化学习弥合与专有智能体之间的鸿沟

WebSailor-V2: Bridging the Chasm to Proprietary Agents via Synthetic Data and Scalable Reinforcement Learning

摘要

Support