Embodied Interpretability: Linking Causal Understanding to Generalization in Vision-Language-Action Models
Year: 2026
Type: preprint
Abstract: Vision-Language-Action (VLA) policies often fail under distribution shift, suggesting that decisions may depend on spurious visual correlations rather than task-relevant causes. We formulate visual-action attribution as an interventional estimation problem. Accordingly, we introduce the Interventional Significance Score (ISS), an interventional masking procedure for estimating the causal influence of visual regions on action predictions, and the Nuisance Mass Ratio (NMR), a scalar measure of attribution to task-irrelevant features. We analyze the statistical properties of ISS and show that it admits unbiased estimation, and we characterize conditions under which action prediction error provides a valid proxy for causal influence. Experiments across diverse manipulation tasks indicate that NMR predicts generalization behavior and that ISS yields more faithful explanations than existing interpretability methods. These results suggest that interventional attribution provides a simple diagnostic approach for identifying causal misalignment in embodied policies.
Source: arXiv (Cornell University)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7160092370
Cites:
Cited by:
Related to: