Reasoning as a Distribution: Generative Plan Latents for Vision-Language-Action Models
Overview & Synopsis
Vision-Language-Action (VLA) policies increasingly reason before acting, but existing reasoning models compress this intermediate decision into a single representation. When several plans are valid, this commitment hides alternatives that could otherwise be selected or revised. We introduce the Generative Plan Latent (GPL), which represents the reasoning state as a low-dimensional conditional distribution over plans, such as which object, grasp, or route to use. This makes the reasoning state addressable: without retraining the VLA, plans can be sampled (coverage), selected on request (steering), and reweighted by deployment outcomes (outcome adaptation). We test all three on SimplerEnv manipulation with $ฯ_0$, and extend the analysis to RoboCasa, LIBERO, and a maze domain. Steering the plan raises the fraction of executions that follow the request from 0.48 to 0.94, while the same guidance on the native action distribution shows no reliable gain. Finally, we show that an explicit plan is most useful when it carries decision information not already determined by the observation.
arxiv.org)
← / → for pages • + / - for zoom • 0 to fit width
Powered by Mozilla PDF.js v3.11
Document Information
| Document Title | Reasoning as a Distribution: Generative Plan Latents for Vision-Language-Action Models |
|---|---|
| Author / Organization | Matteo Cercola, Katie Z Luo, Milan Ganai, John Irvin Alora, Simone Formentin, Marco Pavone |
| Subject & Discipline | cs.RO |
| Categories | Science & Mathematics |
| Publication Year | 2026 |
| License / Rights | arXiv Open Access / Creative Commons |
| Resource Verification | Listed PDF (Verification Pending) |
| Original Host Website |
arXiv
(arxiv.org)
|
| File Format | application/pdf (Portable Document Format) |
| Language | EN |
| Catalog ID | DOC-269963 |
| Original Source Link | Visit Source Page on arxiv.org » |