In-Context VLA: Endowing Vision-Language-Action Models with Language via In-Context Post-Training and Agentic Tool Use

• URL: http://arxiv.org/abs/2608.05738v1
• Date: Thu, 06 Aug 2026 08:21:02 GMT
• Title: In-Context VLA: Endowing Vision-Language-Action Models with Language via In-Context Post-Training and Agentic Tool Use
• Abstract summary: Vision-Language-Action models imitate expert action chunks conditioned on a static image and a fixed instruction.<n>We introduce textbfourmethod, a framework that endows a VLA with language competence through in-context post-training.<n>Our method consistently achieves SOTA results in both performance and efficiency when compared with CoT-based approaches.
• Score: 18.507118969955666
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-Language-Action (VLA) models have become the dominant recipe for generalist manipulation, yet they are almost universally trained by behavior cloning: a policy imitates expert action chunks conditioned on a static image and a fixed instruction. A natural remedy is to inject explicit reasoning through textual chain-of-thought (CoT). We show, both empirically and analytically, that free-form textual CoT degrades low-level control: the reasoning it produces is ungrounded, its latency breaks closed-loop timing, and, crucially, the reasoning and action tokens are optimized against conflicting objectives so that the policy learns to narrate rather than to act. We argue that what a VLA needs is not the ability to generate language, but the ability to consume grounded language. To this end we introduce \textbf{\ourmethod{}}, a framework that endows a VLA with language competence through (i) in-context post-training, in which perceptual evidence is injected as structured context and the model is supervised only on actions, and (ii) an agentic tool-use interface, in which the policy queries open-vocabulary detectors, monocular depth, and a vision--language model to actively acquire task-relevant information. Rather than emitting a single templated caption, our data engine produces diverse, paraphrased, and evidence-conditioned spatial descriptions, so that the policy learns to interpret language it has never seen verbatim. Across the RoboCasa-GR1, SimplerEnv, and LIBERO simulation benchmarks, together with 8 real-world robot manipulation tasks, our method consistently achieves SOTA results in both performance and efficiency when compared with CoT-based approaches under matched configurations.

Related papers

• Dynamic Parsing and Updating Natural Language Specification using VLMs for Robust Vision-Language Tracking [61.91472604294714]
  We propose a novel language dependency parsing mechanism to distill tracking principal components.<n>We exploit the powerful cross-modal understanding capability of the pre-trained vision-language model Qwen-VL.<n>Our method achieves consistent and superior tracking performance on multiple large-scale vision-language tracking benchmarks.
  arXiv  Detail & Related papers  (2026-06-28T12:12:18Z)
• Grounding Hierarchical Vision-Language-Action Models Through Explicit Language-Action Alignment [4.001668956134368]
  To be transparent, a robot's natural language communication must be consistent with its actions.<n>Existing hierarchical Vision-Language-Action (VLA) models can generate language through chain-of-thought and low-level actions.<n>We propose a novel training framework that grounds hierarchical VLA sub-task descriptions with respect to the visual observation and action space.
  arXiv  Detail & Related papers  (2026-04-07T09:03:12Z)
• LangForce: Bayesian Decomposition of Vision Language Action Models via Latent Action Queries [30.732526921367835]
  LangForce is a novel framework that enforces instruction following via Bayesian decomposition.<n>We show that LangForce significantly improves generalization without requiring new data.
  arXiv  Detail & Related papers  (2026-01-21T17:15:22Z)
• Language-Guided Grasp Detection with Coarse-to-Fine Learning for Robotic Manipulation [31.386822229629455]
  We propose Language-Guided Grasp Detection (LGGD) with a coarse-to-fine learning paradigm for robotic manipulation.<n>This design enables fine-grained visual-semantic alignment and improves the feasibility of the predicted grasps with respect to task instructions.<n>Experiments on the OCID-VLG and Grasp-Anything++ datasets show that LGGD surpasses existing language-guided grasping methods.
  arXiv  Detail & Related papers  (2025-12-24T09:16:42Z)
• Seeing to Act, Prompting to Specify: A Bayesian Factorization of Vision Language Action Policy [59.44168425139687]
  BayesVLA is a Bayesian factorization that decomposes the policy into a visual-action prior, supporting seeing-to-act, and a language-conditioned likelihood, enabling prompt-to-specify.<n>Experiments show superior generalization to unseen instructions, objects, and environments compared to existing methods.
  arXiv  Detail & Related papers  (2025-12-12T01:59:23Z)
• Do What? Teaching Vision-Language-Action Models to Reject the Impossible [53.40183895299108]
  Vision-Language-Action (VLA) models have demonstrated strong performance on a range of robotic tasks.<n>We propose Instruct-Verify-and-Act (IVA), a framework that detects when an instruction cannot be executed due to a false premise.<n>Our experiments show that IVA improves false premise detection accuracy by 97.56% over baselines.
  arXiv  Detail & Related papers  (2025-08-22T10:54:33Z)
• Generative Context-aware Fine-tuning of Self-supervised Speech Models [54.389711404209415]
  We study the use of generative large language models (LLM) generated context information. We propose an approach to distill the generated information during fine-tuning of self-supervised speech models. We evaluate the proposed approach using the SLUE and Libri-light benchmarks for several downstream tasks: automatic speech recognition, named entity recognition, and sentiment analysis.
  arXiv  Detail & Related papers  (2023-12-15T15:46:02Z)
• Context-Aware Language Modeling for Goal-Oriented Dialogue Systems [84.65707332816353]
  We formulate goal-oriented dialogue as a partially observed Markov decision process. We derive a simple and effective method to finetune language models in a goal-aware way. We evaluate our method on a practical flight-booking task using AirDialogue.
  arXiv  Detail & Related papers  (2022-04-18T17:23:11Z)
• Pre-Trained Language Models for Interactive Decision-Making [72.77825666035203]
  We describe a framework for imitation learning in which goals and observations are represented as a sequence of embeddings. We demonstrate that this framework enables effective generalization across different environments. For test tasks involving novel goals or novel scenes, initializing policies with language models improves task completion rates by 43.6%.
  arXiv  Detail & Related papers  (2022-02-03T18:55:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.