S-Space：探索多模態語言模型中的空間工作空間

S-Space: Exploring Spatial Workspace in Multimodal Language Models

October 9, 2026

作者: Zhiyuan Shi, Xiang Li, Xingzhuo Guo, Shangchen Miao, Weiliang Chen, Qiyu Dai, Jinshan Ren, Jiangran Lyu, Fangfu Liu, Mingsheng Long, Jialong Wu

cs.AI

摘要

多模態語言模型（MLLMs）展現出前景可期的空間推理能力，但支撐其表現的內部機制仍未被充分理解。我們揭示了 S-Space：一個線性表徵子空間，其中物件 token 表徵會沿水平、垂直與距離軸編碼連續座標。儘管我們是透過將空間答案分數對物件 token 激活進行微分來辨識這些軸，S-Space 仍是模型內在固有的，而非由特定資料或提示所誘發，且在非空間問題下依然可讀。我們表明，S-Space 扮演 MLLM 內部空間工作區的角色。感知會將物件位置編碼其中，形成可直接讀出以支持準確空間判斷的空間表徵。內部推理會因應視角變化對其進行轉換。語言回應則選擇性地存取它，這由因果干預所證實：這些干預會改變空間判斷，同時大致保留非空間回應。案例研究進一步顯示，S-Space 會為未見實體與觀察者指派位置，並將物理方向與抽象語言意義連結，顯示大規模語言訓練的能力與限制。最後，我們將 S-Space 用作診斷空間推理失敗的工具。方向詞讀出揭示基本方位與影像相對參考框架之間的系統性耦合。更重要的是，一種感知–計算分解暴露了關鍵瓶頸：MLLM 能形成具備能力的空間表徵，卻在跨視角變化可靠推理所需的空間轉換上表現不足。

English

Multimodal language models (MLLMs) exhibit promising spatial reasoning capabilities, yet the internal mechanisms supporting them remain poorly understood. We uncover S-Space, a linear representation subspace in which object-token representations encode continuous coordinates along horizontal, vertical, and distance axes. Although we identify these axes by differentiating spatial-answer scores with respect to object-token activations, S-Space is intrinsic rather than induced by specific data or prompts, remaining readable under non-spatial questions. We show that S-Space functions as an internal spatial workspace of MLLMs. Perception encodes object locations into it, forming spatial representations that can be directly read out to support accurate spatial judgments. Internal reasoning transforms it for changing viewpoints. Verbal responses selectively access it, evidenced by causal interventions that alter spatial judgments while largely preserving non-spatial responses. Case studies further show that S-Space assigns locations to unseen entities and the observer, and connects physical directions with abstract linguistic meanings, suggesting both the capabilities and limitations of large-scale language training. Finally, we use S-Space as a diagnostic tool for spatial reasoning failures. Direction-word readouts reveal systematic coupling between cardinal and image-relative reference frames. More importantly, a perception-computation decomposition exposes a key bottleneck: MLLMs can form capable spatial representations, yet fall short in the spatial transformations required to reason reliably across viewpoint changes.