arXiv cs.AI· Jiongxiao Wang, Han Ding, Youzhi Luo, Yawei Wang, Yun Zhou, Linbo Liu, Panpan Xu·· 13 小时前AI 评分34
PAWM：用坑点预判世界模型引导黑盒 LLM 智能体
Steering Black-Box Large Language Model Agents with Pitfall Anticipation World Model
AI 导读
研究者提出坑点预判世界模型 PAWM，在潜在坑点出现时选择性介入并预测即将发生的失败，从而引导黑盒 LLM 智能体避开陷阱。PAWM 先在有智能体坑点标注的合成数据上做 SFT，再与黑盒 LLM 通过 GRPO 协同训练。它使 Claude Sonnet 4.5 在 ScienceWorld 上提升 24.3%、在 WebArena-Lite 上提升 7.7%，优于多种基线。
正文 · 原文
Abstract:Frontier black-box Large Language Model (LLM) agents are proficient at long-horizon reasoning and precise action generation. However, when it comes to a specific agentic task, they have no internalized knowledge of how that particular environment unfolds. A world model offers a natural remedy, providing the agent with environment specific knowledge by foreseeing how the environment will respond to the action it is about to take. Prior world models for LLM agents mainly predict next observations at each step, which is costly and often uninformative because individual observations provide limited task-relevant guidance and not every step requires call for foresight. This raises a key question: what a world model should predict and when should it be invoked for agentic tasks? Our pilot study suggests that world model guidance is most useful when it anticipates pitfalls within a bounded multi-step future. Motivated by this, we introduce Pitfall Anticipation World Model (PAWM), which selectively intervenes when a potential pitfall arises and predicts the impending failure to steer the agent away from it. PAWM is trained in two stages: Supervised Fine-Tuning (SFT) on synthetic data labeled with agent pitfalls, followed by GRPO co-training with a black-box LLM to further optimize it in the target environment. PAWM improves Claude Sonnet 4.5 performance by 24.3% on ScienceWorld and 7.7% on WebArena-Lite, outperforming a range of baselines. These results demonstrate that PAWM provides an effective form of world model guidance for black-box LLM agents.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.29139 [cs.AI] |
| (or arXiv:2610.29139v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29139 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiongxiao Wang [view email]
[v1]
Wed, 21 Oct 2026 03:20:32 UTC (3,179 KB)
来源：arXiv cs.AI · arxiv.org