# 已编码但未受控：揭示视觉-语言机器人策略中的接地缺口

> 原文链接：https://robotics.hellobike.com/papers/2026-10-05-2610-06235
> 频道：papers | 更新日期：2026-10-05 | 标签：arxiv, cs.RO
> 核心摘要：为辨别机器人是否真正遵循指令，论文用场景不变指令干预分析VLA与WAM，并在仿真和真实实验中验证语言接地缺口。

# 已编码但未受控：揭示视觉-语言机器人策略中的接地缺口

::: v-pre
> 原文链接：[Encoded but Not in Control: Revealing the Grounding Gap in Vision-Language Robot Policies](https://arxiv.org/abs/2610.06235)（arXiv 2610.06235 · 发布日期：2026-10-05）

**Encoded but Not in Control: Revealing the Grounding Gap in Vision-Language Robot Policies**

为辨别机器人是否真正遵循指令，论文用场景不变指令干预分析VLA与WAM，并在仿真和真实实验中验证语言接地缺口。

## 核心要点

- 问题：执行成功不能证明策略按指令行动，可能只是从场景推断任务。
- 方法：固定场景并替换有效目标、加入任意名词或无关句子，观察策略反应。
- 结果：请求不同可见物体时，评估策略多会接近并拾取场景关联的原目标。
- 分析显示指令被编码但弱参与动作生成，论文提出诊断框架和进展标准。

**验证范围与局限**：验证范围为仿真和真实实验中的VLA策略与WAMs；摘要未说明具体机器人、数据集或任务数量。

## 英文摘要 (Abstract)

Instruction following is central to language-conditioned robot policies: language should determine what to do when the same scene permits multiple valid actions. Yet successful execution alone cannot establish whether a policy follows the instruction or infers the task from the scene. We study this ambiguity through scene-preserving instruction interventions, using valid target substitutions, arbitrary nouns, and unrelated sentences while holding the scene fixed. We evaluate vision-language-action (VLA) policies and world-action models (WAMs) in simulation and in real-world experiments. Our analysis addresses three questions: (a) Does task success imply instruction following? When instructions request a different visible object, all evaluated policies predominantly approach and pick up the incorrect original target associated with the scene. (b) Is this failure caused by language insensitivity? Instruction perturbations affect task performance. A layerwise action lens shows intermediate action predictions respond to these perturbations. Linear probes accurately recover instructed targets, indicating modified instructions are encoded despite rarely determining target selection. (c) Why does encoded language fail to control action? Attention analysis indicates weak instruction-token contributions to action generation. Target-token attention can remain focused on the original object, revealing a mismatch between target encoding and visual grounding. UMAP and shared non-negative matrix factorization show target information remains accessible within representations increasingly organized by scene identity. Our findings expose a grounding gap concealed by nominal success and provide a diagnostic framework. They further establish a concrete criterion for progress: policies should reliably follow valid changes in user intent, even when they conflict with scene-favored behavior.

作者：Shaohan Jiang, Jiahang Cao, Qiduo He, Fengting Deng, Kun Wu, Jingkai Sun, Jiaxu Wang, Qiang Zhang, Qihao Zheng, Chunfeng Song
:::

## 原始参考来源 (Sources)

- [arXiv 2610.06235：Encoded but Not in Control: Revealing the Grounding Gap in Vision-Language Robot Policies](https://arxiv.org/abs/2610.06235)