Drop-Then-Recovery: How Redundant Are Vision-Language-Action Models?

• URL: http://arxiv.org/abs/2606.27755v1
• Date: Fri, 26 Jun 2026 06:22:17 GMT
• Title: Drop-Then-Recovery: How Redundant Are Vision-Language-Action Models?
• Abstract summary: Vision-Language-Action (VLA) models enable instruction-driven robotic manipulation, but they inherit oversized language backbones from pretrained VLMs.<n>This raises a basic question: how much of a VLA model is actually necessary for closed-loop control?<n>We study architectural redundancy in VLA models by using transformer block removal as a controlled intervention.
• Score: 33.83065196198769
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-Language-Action (VLA) models enable instruction-driven robotic manipulation, but they inherit oversized language backbones from pretrained VLMs whose capacity far exceeds what is needed for short robotic instructions. This raises a basic question: how much of a VLA model is actually necessary for closed-loop control? In this work, we study architectural redundancy in VLA models by using transformer block removal as a controlled intervention. We introduce \textbf{Drop-Then-Recovery (DTR)}, an analysis protocol that removes selected blocks from a pretrained VLA model and then fine-tunes the resulting model to measure whether the removed capacity was necessary for downstream control. To make this intervention reliable, we propose \textbf{GateProbe}, a one-shot virtual-gate sensitivity metric that ranks blocks by their contribution to the downstream action loss. Across multiple VLA architectures, manipulation benchmarks and even real-robot industrial scenarios, we find a strong asymmetry in post-removal recoverability: \ul{\textit{language backbones are highly redundant for standard robotic manipulation tasks, whereas vision and action pathways are substantially less tolerant to removal}}. On LIBERO, removing half of the LLM blocks even improves OpenVLA-OFT from 95.0% to 98.3% under the same downstream fine-tuning budget, and retaining only two language blocks still recovers baseline-level performance. These results suggest that current VLA benchmarks may exert limited pressure on deep language grounding and compositional instruction understanding, and that future VLA architectures should allocate capacity more deliberately across language, vision, and action components. The code is available at https://github.com/s1ghhh/VLADrop.

Related papers

• V-Link: Recovering Lost Visual Representations in Action DiT for Vision-Language-Action Models [19.466602886959866]
  We propose V-Link, which explicitly recovers visual representations during the vision-language (VL) to action (A) feature transfer.<n>We show that V-Link improves the average success rate over base model GR00T N1.6 by +1.9%, +31.2%, and +18.8%, respectively.<n>On the AGIBOT A3 Ultra, V-Link further achieves gains of +20% and +24% on two real-world humanoid tasks.
  arXiv  Detail & Related papers  (2026-08-26T02:32:18Z)
• CLAP: Direct VLM-to-VLA Adaptation via Language-Action Grounding [16.03159066719039]
  CLAP (Causal Language-Action Prediction) prepends each numeric action sequence with a natural-language action description.<n>We will release CLAP at 0.8B, 2B, and 4B as an open-weight, multi-scale compact VLA family from a single VLM lineage.
  arXiv  Detail & Related papers  (2026-07-09T22:34:11Z)
• ReactVLA: Fast and Lightweight Reactive Robot Manipulation via Improved Mean Flow Action Generation [6.976481623762446]
  textttReactVLA is a lightweight and low-latency VLA framework for real-time robotic manipulation.<n>textttReactVLA consistently outperforms similarly sized VLA baselines.<n>It reduces real-world policy latency to below 38.6 ms, enabling fast reactive control on physical robot platforms.
  arXiv  Detail & Related papers  (2026-06-12T08:33:37Z)
• Sentinel-VLA: A Metacognitive VLA Model with Active Status Monitoring for Dynamic Reasoning and Error Recovery [62.75419724651416]
  We introduce textbfSentinel-VLA, a metacognitive VLA model equipped with an active sentinel'' module to monitor real-time execution status.<n>All training data is automatically generated and annotated through our designed pipeline.<n>Real-world experiments demonstrate that Sentinel-VLA boosts the task success rate by over 30% compared to the SOTA model, PI0.
  arXiv  Detail & Related papers  (2026-05-02T02:10:54Z)
• VLA-Forget: Vision-Language-Action Unlearning for Embodied Foundation Models [0.10742675209112619]
  In OpenVLA-style policies, behavior is produced through a fused visual encoder, a cross-modal projector, and a language backbone that predicts tokenized robot actions.<n>We propose VLA-Forget, a hybrid unlearning framework that combines ratio-aware selective editing for perception and cross-modal specificity with layer-selective reasoning/action unlearning.
  arXiv  Detail & Related papers  (2026-04-05T04:23:18Z)
• RoboAlign: Learning Test-Time Reasoning for Language-Action Alignment in Vision-Language-Action Models [58.83401587988675]
  RoboAlign trains vision-language-action models (VLAs) to translate multimodal understanding into low-level actions.<n>Our key idea is to sample action tokens via zero-shot natural language reasoning and refine this reasoning using reinforcement learning (RL) to improve action accuracy.<n>RoboAlign achieves performance improvements of 17.5%, 18.9%, and 106.6% over SFT baselines on LIBERO, CALVIN, and real-world environments, respectively.
  arXiv  Detail & Related papers  (2026-03-22T17:57:55Z)
• Don't Run with Scissors: Pruning Breaks VLA Models but They Can Be Recovered [11.873633681272096]
  Vision-Language-Action (VLA) models have advanced robotic capabilities but remain challenging to deploy on resource-limited hardware.<n>We introduce GLUESTICK, a post-pruning recovery method that restores much of the original model's functionality while retaining sparsity benefits.
  arXiv  Detail & Related papers  (2025-10-09T17:07:30Z)
• Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies [62.653984010274485]
  Vision-Language-Action (VLA) models adapt large vision-language backbones to map images and instructions into robot actions.<n> prevailingAs either generate actions auto-regressively in a fixed left-to-right order or attach separate or diffusion heads outside the backbone.<n>We present Discrete Diffusion VLA, a unified-transformer policy that models discretized action chunks with discrete diffusion.
  arXiv  Detail & Related papers  (2025-08-27T17:39:11Z)
• From Intention to Execution: Probing the Generalization Boundaries of Vision-Language-Action Models [5.660635614478238]
  Vision-Language-Action (VLA) models promise to produce versatile, "generalist" robot policies.<n>Traditional imitation learning benchmarks are unsuitable due to the lack of language instructions.<n>We introduce a unified suite of 50 simulation-based tasks across 10 subcategories spanning language instruction, vision, and objects.
  arXiv  Detail & Related papers  (2025-06-11T16:52:18Z)
• CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models [89.44024245194315]
  We introduce a method that incorporates explicit visual chain-of-thought (CoT) reasoning into vision-language-action models (VLAs)<n>We introduce CoT-VLA, a state-of-the-art 7B VLA that can understand and generate visual and action tokens.<n>Our experimental results demonstrate that CoT-VLA achieves strong performance, outperforming the state-of-the-art VLA model by 17% in real-world manipulation tasks and 6% in simulation benchmarks.
  arXiv  Detail & Related papers  (2025-03-27T22:23:04Z)
• HAMSTER: Hierarchical Action Models For Open-World Robot Manipulation [54.03004125910057]
  We show that hierarchical vision-language-action models can be more effective in utilizing off-domain data than standard monolithic VLA models.<n>We show that, with the hierarchical design, the high-level VLM can transfer across significant domain gaps between the off-domain finetuning data and real-robot testing scenarios.
  arXiv  Detail & Related papers  (2025-02-08T07:50:22Z)
• OpenVLA: An Open-Source Vision-Language-Action Model [131.74098076670103]
  We introduce OpenVLA, an open-source VLA trained on a diverse collection of 970k real-world robot demonstrations. OpenVLA shows strong results for generalist manipulation, outperforming closed models such as RT-2-X (55B) by 16.5% in absolute task success rate. We release model checkpoints, fine-tuning notebooks, and our PyTorch with built-in support for training VLAs at scale on Open X-Embodiment datasets.
  arXiv  Detail & Related papers  (2024-06-13T15:46:55Z)

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.