mlsanityData to 7 Oct 2026

Reinforcement learning and post-training

173 papers4 this weekHigh confidence in this grouping: the papers form one clear topicHigh confidence in this grouping: the papers form one clear topic · About

Keywords: vla, policy, reinforcement, reinforcement learning, policies

Key facts. Reinforcement learning and post-training: 173 papers on arXiv since 2024 (data through 7 Oct 2026). 84 of its 173 mapped papers (49%) cite “Octo: An Open-Source Generalist Robot Policy” (2024, arXiv 2405.12213). 52 of its 173 mapped papers (30%) cite “ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policy” (2025, arXiv 2502.05450). 32 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

32 new papers, 17 Sep – 7 Oct; 4 this week.

  1. This week · 4
  2. Many Ways to Succeed: Diversity-Driven RL Fine-Tuning for VLA Generalization
  3. Co-Evolving Robot Orchestrators and Policies through Deployment
  4. Recursive Video In-Context Learning for Agentic Robot also fits: Reasoning and chain-of-thought in VLA
  5. How (and How Not) to Use Data Augmentation in VLA Post-Training
  6. Earlier · 28
  7. Triggering Generalist Reasoning via Predictive Uncertainty for Dual-System VLA also fits: Efficient VLA inference and adaptation
  8. DiVeR: Decision-Critical Verifier Learning for VLA Test-Time Scaling also fits: Efficient VLA inference and adaptation
  9. SWAP: Stepwise Action Policy Routing for Vision-Language-Action Models
  10. RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies also fits: Failure recovery and robustness
  11. eRLT: Efficient VLA Reinforcement Learning via Action-Relevant Token Routing also fits: Efficient VLA inference and adaptation
  12. Online Evolution Strategy for Flow-Matching VLA Policies via Self-Supervised Trajectory Distribution Optimization also fits: Efficient VLA inference and adaptation
All 32
  1. DriftOPD: Sequence-Level Reverse-KL Distillation for One-Step VLA Policies
  2. Rho: A Foundation for Efficiently Adaptable VLA Models also fits: Generalist VLA models and benchmarks
  3. Cooperative Multi-Agent Vision-Language-Action Models via Reinforced Fine Tuning
  4. StructRL: Online Structured Reinforcement Learning for Long-Horizon Vision-Language-Action Tasks
  5. The Low-Rank Structure of VLA Reinforcement Learning
  6. Principal Steering Subspaces for Online Adaptation of Frozen Generative Robot Policies
  7. PF-RL: Progress Field Reinforcement Learning via Goal-Conditioned Value Geometry for Vision-Language-Action Models
  8. SAMBAR: Selective Anchoring via Method of Multipliers for Balanced Knowledge Acquisition and Retention in Vision-Language-Action Models also fits: VLA architectures and generalization
  9. Find Something You Can't Do: Agentic Real-World Reinforcement Learning for Self-Improving VLA Models also fits: Failure recovery and robustness
  10. VLaRL: Augmenting Vision-Language-Action Models with Simulation-Trained Latent-Conditioned Residual RL
  11. Self-Adaptive VLA for Robust Robot Deployment also fits: Generalist VLA models and benchmarks
  12. Uncertainty-Gated Exploration Noise Suppresses Task Collapse in Online RL Fine-Tuning of a Flow-Matching Vision-Language-Action Policy
  13. TANDEM: Task and Motion Planning with As-Needed Demonstrations for Efficient Vision-Language-Action Model Fine-tuning also fits: Reasoning and chain-of-thought in VLA
  14. Dissecting Advantage-Guided Post-Training for Vision-Language-Action Policies
  15. BEE: Intervention-Adaptive Real-World Reinforcement Learning with Vision-Language-Action Models
  16. RouteRLT: Learning When and Which RL Specialist Should Control a Vision-Language-Action Policy
  17. An Empirical Study and Open Testbed for Federated Fine-Tuning of Vision-Language-Action Models also fits: Generalist VLA models and benchmarks
  18. Stable and Efficient Real-World Online VLA Post-Training via Asynchronous Replay-Anchored Policy Improvement
  19. Prioritized Rollouts for Efficient World Model-based Vision-Language-Action Policy Optimization
  20. ForceRFT: Refining VLA Actions through Force-Guided Residual Reinforcement Learning also fits: Tactile, force-aware and humanoid VLA
  21. SeeQ: Training Generalist Value Functions for Long-Horizon Robotic Manipulation
  22. SynthDemo-RL: Breaking the Zero-Reward Barrier in VLA Adaptation with LLM-Guided Synthetic Demonstrations

Foundations

Most cited by papers here

  1. Octo: An Open-Source Generalist Robot Policy2024cited by 84 here
  2. ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policy2025cited by 52 here
  3. VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learning2025cited by 50 here
  4. SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning2025cited by 45 here
  5. What Can RL Bring to VLA Generalization? An Empirical Study2025cited by 44 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 109 here
  2. 2024Octo: An Open-Source Generalist Robot Policycited by 84 here
  3. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 133 here
  4. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 114 here
  5. 2025FAST: Efficient Action Tokenization for Vision-Language-Action Modelscited by 41 here
  6. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 89 here
  7. 2025What Can RL Bring to VLA Generalization? An Empirical Studycited by 44 here
  8. 2025SimpleVLA-RL: Scaling VLA Training via Reinforcement Learningcited by 45 here
  9. 2025π*0.6: a VLA That Learns From Experiencecited by 38 here
15 more it builds on, not in this chain
  1. 2024A Survey on Vision-Language-Action Models for Embodied AIcited by 16 here
  2. 2024GRAPE: Generalizing Robot Policy via Preference Alignmentcited by 31 here
  3. 2024CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulationcited by 16 here
  4. 2025SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Modelcited by 28 here
  5. 2025Improving Vision-Language-Action Model with Online Reinforcement Learningcited by 40 here
  6. 2025ConRFT: A Reinforced Fine-tuning Method for VLA Models via Consistency Policycited by 52 here
  7. 2025Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Successcited by 65 here
  8. 2025CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Modelscited by 18 here
  9. 2025UniVLA: Learning to Act Anywhere with Task-centric Latent Actionscited by 17 here
  10. 2025ReinboT: Amplifying Robot Visual-Language Manipulation with Reinforcement Learningcited by 16 here
  11. 2025Interactive Post-Training for Vision-Language-Action Modelscited by 34 here
  12. 2025VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learningcited by 50 here
  13. 2025SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Roboticscited by 29 here
  14. 2025A Vision-Language-Action-Critic Model for Robotic Real-World Reinforcement Learningcited by 21 here
  15. 2025Self-Improving Vision-Language-Action Models with Data Generation via Residual RLcited by 20 here

The 24 most cited of 336.

Browse

30 of 173, ranked 6–35

  1. 6Improving Vision-Language-Action Model with Online Reinforcement Learning202540
  2. 7π*0.6: a VLA That Learns From Experience202538
  3. 8Interactive Post-Training for Vision-Language-Action Models202534
  4. 9GRAPE: Generalizing Robot Policy via Preference Alignment202431
  5. 10A Vision-Language-Action-Critic Model for Robotic Real-World Reinforcement Learning202521
  6. 11Self-Improving Vision-Language-Action Models with Data Generation via Residual RL202520
  7. 12ReinboT: Amplifying Robot Visual-Language Manipulation with Reinforcement Learning202516
  8. 13TGRPO :Fine-tuning Vision-Language-Action Model via Trajectory-wise Group Relative Policy Optimization202516
  9. 14CO-RFT: Efficient Fine-Tuning of Vision-Language-Action Models through Chunked Offline Reinforcement Learning202514
  10. 15Steering Your Generalists: Improving Robotic Foundation Models via Value Guidance202413
All 30
  1. 16VLA-RFT: Vision-Language-Action Reinforcement Fine-tuning with Verified Rewards in World Simulators202512
  2. 17RFTF: Reinforcement Fine-tuning for Embodied Agents with Temporal Feedback202510
  3. 18RL Token: Bootstrapping Online RL with Vision-Language-Action Models20269
  4. 19RLinf-VLA: A Unified and Efficient Framework for Reinforcement Learning of Vision-Language-Action Models20258
  5. 20WMPO: World Model-based Policy Optimization for Vision-Language-Action Models20258
  6. 21Continually Evolving Skill Knowledge in Vision Language Action Model20258
  7. 22SRPO: Self-Referential Policy Optimization for Vision-Language-Action Models20257
  8. 23GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation20257
  9. 24Refined Policy Distillation: From VLA Generalists to RL Experts20256
  10. 25World-Env: Leveraging World Model as a Virtual Environment for VLA Post-Training20256
  11. 26SOP: A Scalable Online Post-Training System for Vision-Language-Action Models20265
  12. 27WoVR: World Models as Reliable Simulators for Post-Training VLA Policies with RL20265
  13. 28Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learning20265
  14. 29GigaBrain-0.5M*: a VLA That Learns From World Model-Based Reinforcement Learning20264
  15. 30MoRE: Unlocking Scalability in Reinforcement Learning for Quadruped Vision-Language-Action Models20254
  16. 31VLA-R1: Enhancing Reasoning in Vision-Language-Action Models20254
  17. 32Robust Finetuning of Vision-Language-Action Robot Policies via Parameter Merging20254
  18. 33STRAP: Robot Sub-Trajectory Retrieval for Augmented Policy Learning20244
  19. 34CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansion20263
  20. 35Simple Recipe Works: Vision-Language-Action Models are Natural Continual Learners with Reinforcement Learning20263