arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2502.16690cs.AI
收藏

从文本到空间:在网格世界导航任务中映射LLMs的抽象空间模型

From Text to Space: Mapping Abstract Spatial Models in LLMs during a Grid-World Navigation Task

Faculty of Exact and Natural Sciences, University of Buenos Aires(布宜诺斯艾利斯大学精确与自然科学学院);National Scientific and Technical Research Council (CONICET)(国家科学技术研究委员会(CONICET))

Nicolas Martorell

AI总结:

本研究通过网格世界导航任务,探究不同文本空间表示对LLM性能及内部激活的影响,发现笛卡尔表示最优,并识别出与空间特征相关的内部单元,为可解释智能体AI提供见解。

摘要:

理解大型语言模型(LLMs)如何表示和推理空间信息,对于构建能够在真实和模拟环境中导航的稳健智能体系统至关重要。在本工作中,我们研究了在网格世界导航任务中,不同的基于文本的空间表示对LLM性能和内部激活的影响。通过在需要向目标导航的任务上评估不同规模的模型,我们考察了用于编码空间信息的格式如何影响决策。我们的实验表明,空间的笛卡尔表示始终能带来更高的成功率和路径效率,且性能随模型规模显著提升。此外,对LLaMA-3.1-8B的探测揭示了内部单元的子集(主要位于中间层),这些单元与空间特征(如智能体在网格中的位置或动作正确性)稳健相关,无论该信息如何表示,并且它们也会被无关的空间推理任务激活。这项工作增进了我们对LLMs如何处理空间信息的理解,并为开发更可解释和更稳健的智能体AI系统提供了宝贵的见解。

英文摘要:

Understanding how large language models (LLMs) represent and reason about spatial information is crucial for building robust agentic systems that can navigate real and simulated environments. In this work, we investigate the influence of different text-based spatial representations on LLM performance and internal activations in a grid-world navigation task. By evaluating models of various sizes on a task that requires navigating toward a goal, we examine how the format used to encode spatial information impacts decision-making. Our experiments reveal that cartesian representations of space consistently yield higher success rates and path efficiency, with performance scaling markedly with model size. Moreover, probing LLaMA-3.1-8B revealed subsets of internal units, primarily located in intermediate layers, that robustly correlate with spatial features, such as the position of the agent in the grid or action correctness, regardless of how that information is represented, and are also activated by unrelated spatial reasoning tasks. This work advances our understanding of how LLMs process spatial information and provides valuable insights for developing more interpretable and robust agentic AI systems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑