# See Less, Specify More: Visual Evidence Budgets for Generalizable VLAs

arXiv: https://arxiv.org/abs/2606.02735 · HTML: https://mlsanity.com/paper/2606.02735/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in VLA architectures and generalization, one of 13 directions in Vision-language-action models.
- Builds on 37 of its 44 references that are in the map.
- First posted on arXiv 1 Jun 2026.

## Placed in

- Vision-language-action models / [VLA architectures and generalization](https://mlsanity.com/vla/vla-architectures-and-generalization/)

## Builds on (most cited first, 30 of 37)

- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success](https://mlsanity.com/paper/2502.19645/) (2025)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [CLIPort: What and Where Pathways for Robotic Manipulation](https://mlsanity.com/paper/2109.12098/) (2021)
- [CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks](https://mlsanity.com/paper/2112.03227/) (2021)
- [Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation](https://mlsanity.com/paper/2209.05451/) (2022)
- [SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics](https://mlsanity.com/paper/2506.01844/) (2025)
- [BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning](https://mlsanity.com/paper/2202.02005/) (2022)
- [CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2503.22020/) (2025)
- [Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations](https://mlsanity.com/paper/2202.07800/) (2022)
- [X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Model](https://mlsanity.com/paper/2510.10274/) (2025)
- [Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models](https://mlsanity.com/paper/2502.19417/) (2025)
- [ThinkAct: Vision-Language-Action Reasoning via Reinforced Visual Latent Planning](https://mlsanity.com/paper/2507.16815/) (2025)
- [VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Model](https://mlsanity.com/paper/2509.09372/) (2025)
- [Unified Vision-Language-Action Model](https://mlsanity.com/paper/2506.19850/) (2025)
- [HAMSTER: Hierarchical Action Models For Open-World Robot Manipulation](https://mlsanity.com/paper/2502.05485/) (2025)
- [TokenLearner: What Can 8 Learned Tokens Do for Images and Videos?](https://mlsanity.com/paper/2106.11297/) (2021)
- [Robotic Skill Acquisition via Instruction Augmentation with Vision-Language Models](https://mlsanity.com/paper/2211.11736/) (2022)
- [Deep Object-Centric Representations for Generalizable Robot Learning](https://mlsanity.com/paper/1708.04225/) (2017)
- [TGRPO :Fine-tuning Vision-Language-Action Model via Trajectory-wise Group Relative Policy Optimization](https://mlsanity.com/paper/2506.08440/) (2025)
- [BitVLA: 1-bit Vision-Language-Action Models for Robotics Manipulation](https://mlsanity.com/paper/2506.07530/) (2025)
- [Steerable Vision-Language-Action Policies for Embodied Reasoning and Hierarchical Control](https://mlsanity.com/paper/2602.13193/) (2026)
- [Policy Adaptation via Language Optimization: Decomposing Tasks for Few-Shot Imitation](https://mlsanity.com/paper/2408.16228/) (2024)
- [Object-Centric Instruction Augmentation for Robotic Manipulation](https://mlsanity.com/paper/2401.02814/) (2024)
- [Scaling World Model for Hierarchical Manipulation Policies](https://mlsanity.com/paper/2602.10983/) (2026)
- [Scaling Robot Policy Learning via Zero-Shot Labeling with Foundation Models](https://mlsanity.com/paper/2410.17772/) (2024)
- [STEER: Flexible Robotic Manipulation via Dense Language Grounding](https://mlsanity.com/paper/2411.03409/) (2024)

## Built on by (most cited first, 1 of 1)

- [BICPO-VLA: Behavior-Identified Continuation Preference Optimization for Smooth Asynchronous Vision-Language-Action Control](https://mlsanity.com/paper/2608.13924/) (2026)

## Abstract (arXiv)

Generalization remains a central bottleneck for vision-language-action (VLA) models: under distractors, appearance shifts, and semantically similar tasks, the policy must often infer local execution details from coarse instructions while also deciding which parts of the image matter for control. We present S2 (See Less, Specify More), a framework for improving VLA generalization by training the executor under a cleaner interface. Specify More preserves the original instruction as a stable high-level goal while relabeling each trajectory into refined trajectory- and subtask-level language that disambiguates the current execution mode. Unlike native attention, See Less imposes an explicit visual evidence budget, training the executor to act from task-sufficient evidence rather than unconstrained visual context, without any region or mask annotation. This interface lets the executor follow detailed guidance without relying on distracting visual patches or resolving avoidable ambiguity on its own, and it remains compatible with off-the-shelf VLM planners through in-context learning. Across our main evaluation settings, S2 improves overall generalization metrics by changing the executor's learning problem: coarse instructions induce avoidable supervision aliasing, goal-preserving local guidance outperforms instruction replacement in our main ablations, and explicit evidence budgeting reduces dependence on broad visual context beyond efficiency considerations. Across eight real-robot tasks on TX-G2 (an AgiBot G2-compatible variant) and HSR, S2 raises mean subtask success from 54.2% to 79.0% over pi0.5. Together, these results suggest that VLA generalization improves when the executor is trained to act from informative local guidance and task-sufficient visual evidence, rather than recovering both from weak supervision.