arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.15858cs.CLcs.AI
收藏

状态设计至关重要:表示如何塑造大语言模型中的动态推理

State Design Matters: How Representations Shape Dynamic Reasoning in Large Language Models

Leiden Institute of Advanced Computer Science(莱顿先进计算机科学研究所);Leiden University(莱顿大学)

Annie Wong, Aske Plaat, Thomas Bäck, Niki van Stein, Anna V. Kononova

AI总结:

研究探讨了大语言模型中状态表示设计对动态推理的影响,发现自然语言表示和空间编码对性能至关重要,但长期任务仍面临挑战。

摘要:

随着大语言模型(LLMs)从静态推理任务向动态环境发展,其成功依赖于在交互过程中导航和响应不断变化的环境的能力。在这些设置中,一个未被充分探索的因素是状态的表示。在保持模型参数不变的情况下,我们系统地改变了三个关键方面:(1)状态粒度(长形式与摘要),(2)结构(自然语言与符号),以及(3)空间定位(纯文本与图像或文本地图编码)在连续决策基准测试中。我们发现轨迹摘要通过减少噪声和稳定长期推理来提高性能。第二,自然语言表示在不同模型中最为稳健,而结构编码主要帮助具有强代码或结构化输出先验的模型,如JSON模式。第三,尽管图像输入显示出一些好处,但基于文本的空间编码证明最有效。这种优势并非来自空间信息本身,而是来自构建行为,这迫使模型进行空间推理,而静态输入无法激发这种推理。总体而言,我们证明了状态表示的设计选择是性能的关键因素,不同于信息本身可用性。然而,我们注意到,即使有改进的表示,当前的LLMs和VLMs在长期范围内仍然脆弱,特别是在必须合成信息以管理多个子任务以达到目标时。

英文摘要:

As large language models (LLMs) move from static reasoning tasks toward dynamic environments, their success depends on the ability to navigate and respond to an environment that changes as they interact at inference time. An underexplored factor in these settings is the representation of the state. Holding model parameters fixed, we systematically vary three key aspects: (1) state granularity (long form versus summary), (2) structure (natural language versus symbolic), and (3) spatial grounding (text-only versus images or textual map encodings) across sequential decision-making benchmarks. We find that trajectory summarisation improves performance by reducing noise and stabilising long-horizon reasoning. Second, natural language representations are the most robust across models, whereas structured encodings help mainly for models with strong code or structured output priors, such as JSON schemas. Third, while image-inputs show some benefit, text-based spatial encodings prove most effective. This advantage stems not from the spatial information itself, but from the act of construction, which compels the model to perform the spatial reasoning that static input does not elicit. Overall, we demonstrate that design choices for representing state are a decisive factor in performance, distinct from the availability of information itself. We note, however, that even with improved representations, current LLMs and VLMs remain brittle over long horizons, particularly when they must synthesise information to manage multiple subtasks to reach a goal.

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。

↑