Metraipolis › archive › dc_20213a4e3545

Measuring the Reasoning-Action Coupling in Driving Vision-Language-Action Models through Counterfactuals — arXiv:2610.30953

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.30953
License (as archived): CC0-1.0
Archived: 2026-10-25T21:34:41.896Z
SHA-256 of the archived text: 9c29d4a77154b866bef96a2b3f1009bf64aed94e90b77a0780f7517b9de695ef
Stable URL of this page: https://metraipolis.com/documento/dc_20213a4e3545

How to cite

Measuring the Reasoning-Action Coupling in Driving Vision-Language-Action Models through Counterfactuals — arXiv:2610.30953 — arXiv (en), https://arxiv.org/abs/2610.30953 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_20213a4e3545 (Metraipolis archive dc_20213a4e3545, sha256 9c29d4a77154b866bef96a2b3f1009bf64aed94e90b77a0780f7517b9de695ef)

BibTeX

@misc{dc20213a4e3545,
  title = {Measuring the Reasoning-Action Coupling in Driving Vision-Language-Action Models through Counterfactuals — arXiv:2610.30953},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.30953},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_20213a4e3545; sha256 9c29d4a77154b866bef96a2b3f1009bf64aed94e90b77a0780f7517b9de695ef},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30953. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Driving vision-language-action (VLA) models bring an appealing form of explainability to a safety-critical domain, since a system that can state the reasons for its maneuvers offers the transparency that deployment and public trust in autonomous driving demand. These models read the world into language before they act on it, describing the scene and reasoning over it in text that ranges from chain-of-thought to multi-stage question answering on the way to a trajectory. That framing quietly assumes two separable properties, namely that the trajectory follows the generated reasoning and that the generated reasoning tracks the causes that actually produce the trajectory. To thoroughly investigate these properties, we apply causal, counterfactual interventions to four driving VLAs across different benchmarks. First, we perturb the models' own reasoning and show that altering its meaning produces small trajectory changes relative to blanking the same models' camera inputs, a pattern that persists across benchmarks and model architectures. We then construct scene-level counterfactuals through photorealistic removal of key objects named in the models' reasoning and analyze whether the generated reasoning and trajectories track these changes. Overall, our findings indicate that the reasoning these models present to account for their resulting driving behavior is weakly coupled to that behavior itself, pointing to limits on the explainability these systems are intended to provide.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.30953 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-25T21:25:18.125Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).