pith. sign in

arxiv: 2605.25477 · v1 · pith:XNIRCRJKnew · submitted 2026-05-25 · 💻 cs.RO · cs.AI

EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models

Perry Dong , Kuo-Han Hung , Tian Gao , Dorsa Sadigh , Chelsea Finn This is my paper

Pith reviewed 2026-06-29 22:07 UTC · model grok-4.3

classification 💻 cs.RO cs.AI
keywords reinforcement learningvision-language-action modelsrobot manipulationsample efficiencyfinetuningpretrained policiesmanipulation tasks
0
0 comments X

The pith

EXPO-FT finetunes pretrained vision-language-action models with reinforcement learning to reach perfect task success using 19.1 minutes of robot data on average.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Pretrained vision-language-action models generalize across manipulation tasks yet fall short on the reliability needed for deployment. EXPO-FT applies reinforcement learning to fine-tune these models in a stable and sample-efficient way. The approach is tested on tasks that combine high precision, dynamic movements, and varied starting positions, such as routing string lights, striking a pool ball, and inserting a flower into a bottle. It reports perfect success rates across the evaluated suite while using far less online data than training from scratch or prior finetuning methods.

Core claim

EXPO-FT is a system for stable, sample-efficient RL finetuning of pretrained VLA policies that solves a suite of challenging manipulation tasks, including routing string lights and inserting the plug to light it up, striking a pool ball into a pocket, and inserting a flower into a wine bottle, each requiring combinations of high precision, dynamic actions, and robustness to varied initial states. Our system achieves perfect task performance (30/30 successes) across all evaluated tasks within an average of 19.1 minutes of online robot data, outperforming both prior RL-from-scratch and VLA finetuning approaches.

What carries the argument

EXPO-FT, the system that performs stable reinforcement learning fine-tuning on pretrained vision-language-action policies

If this is right

  • Pretrained VLA policies reach perfect success rates on high-precision tasks after limited online interaction.
  • The method uses less data than RL trained from scratch while improving on prior VLA finetuning results.
  • Tasks that combine dynamic actions with robustness to initial state changes become reliably solvable.
  • An open-source release supports wider testing of RL finetuning for VLA models in robotics.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same finetuning pattern could be examined on tasks outside tabletop manipulation, such as mobile navigation or multi-arm coordination.
  • If efficiency scales, the approach might reduce the total pretraining data needed by shifting more adaptation burden to short RL stages.
  • Testing on hardware with greater sensor noise or longer task horizons would reveal whether the reported data requirements remain stable.

Load-bearing premise

That the EXPO-FT system can deliver the claimed stability and sample efficiency on the described suite of high-precision, dynamic manipulation tasks when applied to pretrained VLA policies.

What would settle it

Recording fewer than 30 successes in 30 trials or requiring substantially more than 19.1 minutes of online data on average for the pool ball striking or flower insertion tasks.

Figures

Figures reproduced from arXiv: 2605.25477 by Chelsea Finn, Dorsa Sadigh, Kuo-Han Hung, Perry Dong, Tian Gao.

Figure 1
Figure 1. Figure 1: Average training success rates of EXPO￾FT compared to prior methods. EXPO-FT achieves a reliable performance with high sample efficiency where prior methods often do not converge reliably. We empirically find that our system achieves dexterous and precise manipulation capabilities across a diverse set of challenging tasks, includ￾ing routing string lights and inserting the power connector to illuminate the… view at source ↗
Figure 2
Figure 2. Figure 2: Left: Overview of EXPO-FT. EXPO-FT features a server that handles VLA training and inference and a learner process that steps in the environment to enable VLA finetuning with RL. Right: Architecture of EXPO-FT. EXPO-FT finetunes the VLA model with EXPO for sample-efficient training. cases. We start by describing the problem statement (Section 4.1), then describe the approach used for finetuning (Section 4.… view at source ↗
Figure 3
Figure 3. Figure 3: Eight real-world manipulation tasks in our evaluation suite. Flower Insert (tight insertion tolerances), String Light Routing - RouteI/II, Insert (long-horizon precise alignment), Egg Flip (dynamic contact￾rich tool use), Candy Scoop (stable control in visually messy scenes), Pool Shot (precise speed control) and Cube Pick (large scene randomization). The tasks span dexterous, precise, deformable, and dyna… view at source ↗
Figure 4
Figure 4. Figure 4: Training success and intervention rates across all tasks. Top row: Egg Flip, Flower Insert, Pool Shot, Cube Pick. Bottom row: String Light Routing - Route I, String Light Routing - Route II, String Light Routing - Insert, Candy Scoop [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Episode Time across all tasks. Top row: Egg Flip, Flower Insert, Pool Shot, Cube Pick. Bottom row: String Light Routing - Route I, String Light Routing - Route II, String Light Routing - Insert, Candy Scoop. B Detailed Task Setting B.1 Task Setting Description Here, we provide detailed descriptions of the data collection process, reward specification, task success detector, reset mechanism and task randomi… view at source ↗
Figure 6
Figure 6. Figure 6: Task strips demonstrating successful completion of each task. Candy Scoop. We pre-collect 20 demonstrations for this task. The reward classification for this task is split into two parts, both of which must succeed for the episode to be counted as successful. In the first part, we verify that candies are present in the scoop once the scoop is raised above a height threshold. In the second part, we check wh… view at source ↗
Figure 7
Figure 7. Figure 7: Visualization of randomized initial state spaces for all tasks. The orange regions indicate the randomized initialization areas used during training. C Detailed Training Setting C.1 Model Structure/Training Detailed We instantiate EXPO-FT with π0.5 [1] as the base policy, initialized from a task-specific LoRA [47] supervised-finetuning checkpoint and the matching normalization statistics for the robot setu… view at source ↗
read the original abstract

The ability to efficiently and reliably learn new tasks has been a foundational challenge in robotics. Vision-Language-Action (VLA) models have demonstrated strong generalization across diverse manipulation tasks, yet pretrained policies consistently fall short of the reliability required for real-world deployment. Reinforcement learning (RL) fine-tuning offers a promising path to bridge this gap, but existing approaches either train from scratch without fully leveraging pretrained priors, or fine-tune VLAs without achieving the sample efficiency and success rates that practical deployment demands. We present EXPO-FT, a system for stable, sample-efficient RL finetuning of pretrained VLA policies that closes this gap. Our system solves a suite of challenging manipulation tasks, including routing string lights and inserting the plug to light it up, striking a pool ball into a pocket, and inserting a flower into a wine bottle, each requiring combinations of high precision, dynamic actions, and robustness to varied initial states. Our system achieves perfect task performance (30/30 successes) across all evaluated tasks within an average of 19.1 minutes of online robot data, outperforming both prior RL-from-scratch and VLA finetuning approaches. We release an open-source codebase with the aim of facilitating broader adoption of RL finetuning of VLA models in robotics.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The manuscript introduces EXPO-FT, a system that augments RL finetuning of pretrained Vision-Language-Action (VLA) policies with an exploration objective, reward shaping, and VLA adaptation procedure. It reports solving a suite of high-precision, dynamic manipulation tasks (routing string lights, striking pool balls, inserting flowers into bottles) to 30/30 success using an average of 19.1 minutes of online robot data per task, outperforming both RL-from-scratch and prior VLA finetuning baselines.

Significance. If the reported outcomes hold under the stated data budgets and task conditions, the work provides a concrete route to reliable real-world deployment of VLAs by addressing stability and sample-efficiency gaps. The open-source codebase release is a clear strength that supports reproducibility and adoption.

minor comments (3)
  1. The experimental section should explicitly state the number of independent random seeds or rollouts used to compute the 30/30 success rates and any associated variance, to strengthen the stability claim.
  2. Figure captions and baseline descriptions would benefit from additional detail on hyperparameter matching across methods to ensure fair comparison.
  3. A short discussion of failure modes or edge cases observed during the 19.1-minute finetuning runs would improve clarity on the method's robustness limits.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of EXPO-FT and for recommending minor revision. We appreciate the recognition that the reported outcomes, if they hold, provide a concrete route to reliable real-world VLA deployment, as well as the value placed on the open-source codebase.

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper is an empirical robotics contribution describing an RL finetuning system (EXPO-FT) and reporting experimental success rates on manipulation tasks. No derivation chain, first-principles predictions, fitted parameters renamed as outputs, or self-citation load-bearing steps appear in the provided abstract or described method/experimental sections. Claims rest on reported robot trials rather than any definitional or self-referential reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no technical details on parameters, axioms, or entities are provided.

pith-pipeline@v0.9.1-grok · 5770 in / 970 out tokens · 32440 ms · 2026-06-29T22:07:41.376698+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

48 extracted references · 26 canonical work pages · 11 internal anchors

  1. [1]

    $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization

    P. Intelligence, K. Black, N. Brown, J. Darpinian, K. Dhabalia, D. Driess, A. Esmail, M. Equi, C. Finn, N. Fusai, M. Y . Galliker, D. Ghosh, L. Groom, K. Hausman, B. Ichter, S. Jakubczak, T. Jones, L. Ke, D. LeBlanc, S. Levine, A. Li-Bell, M. Mothukuri, S. Nair, K. Pertsch, A. Z. Ren, L. X. Shi, L. Smith, J. T. Springenberg, K. Stachowicz, J. Tanner, Q. V...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  2. [2]

    G. R. Team, A. Abdolmaleki, S. Abeyruwan, J. Ainslie, J.-B. Alayrac, M. G. Arenas, A. Bal- akrishna, N. Batchelor, A. Bewley, J. Bingham, M. Bloesch, K. Bousmalis, P. Brakel, A. Bro- han, T. Buschmann, A. Byravan, S. Cabi, K. Caluwaerts, F. Casarini, C. Chan, O. Chang, L. Chappellet-V olpini, J. E. Chen, X. Chen, H.-T. L. Chiang, K. Choromanski, A. Collis...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  3. [3]

    J. Luo, C. Xu, J. Wu, and S. Levine. Precise and dexterous robotic manipulation via human-in- the-loop reinforcement learning, 2025. URLhttps://arxiv.org/abs/2410.21845

    work page arXiv 2025

  4. [4]

    J. Luo, Z. Hu, C. Xu, Y . L. Tan, J. Berg, A. Sharma, S. Schaal, C. Finn, A. Gupta, and S. Levine. Serl: A software suite for sample-efficient robotic reinforcement learning, 2025. URLhttps://arxiv.org/abs/2401.16013

    work page arXiv 2025

  5. [5]

    C. Xu, J. T. Springenberg, M. Equi, A. Amin, A. Esmail, S. Levine, and L. Ke. Rl token: Bootstrapping online rl with vision-language-action models, 2026. URL https://arxiv. org/abs/2604.23073

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  6. [6]

    K. Lei, H. Li, D. Yu, Z. Wei, L. Guo, Z. Jiang, Z. Wang, S. Liang, and H. Xu. Rl-100: Performant robotic manipulation with real-world reinforcement learning, 2026. URL https: //arxiv.org/abs/2510.14830

    work page arXiv 2026

  7. [7]

    A. Z. Ren, J. Lidard, L. L. Ankile, A. Simeonov, P. Agrawal, A. Majumdar, B. Burchfiel, H. Dai, and M. Simchowitz. Diffusion policy policy optimization. InThe Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum? id=mEpqHvbD2h

    2025

  8. [8]

    Steering Your Diffusion Policy with Latent Space Reinforcement Learning

    A. Wagenmaker, M. Nakamoto, Y . Zhang, S. Park, W. Yagoub, A. Nagabandi, A. Gupta, and S. Levine. Steering your diffusion policy with latent space reinforcement learning, 2025. URL https://arxiv.org/abs/2506.15799. 11

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  9. [9]

    Y . Li, X. Ma, J. Xu, Y . Cui, Z. Cui, Z. Han, L. Huang, T. Kong, Y . Liu, H. Niu, W. Peng, J. Qiao, Z. Ren, H. Shi, Z. Su, J. Tian, Y . Xiao, S. Zhang, L. Zheng, H. Li, and Y . Wu. Gr- rl: Going dexterous and precise for long-horizon robotic manipulation, 2025. URL https: //arxiv.org/abs/2512.01801

    work page arXiv 2025

  10. [10]

    H. Li, Y . Zuo, J. Yu, Y . Zhang, Y . Zhaohui, K. Zhang, X. Zhu, Y . Zhang, T. Chen, G. Cui, D. Wang, D. Luo, Y . Fan, Y . Sun, J. Zeng, J. Pang, S. Zhang, Y . Wang, Y . Mu, B. Zhou, and N. Ding. SimpleVLA-RL: Scaling VLA training via reinforcement learning. InThe Fourteenth International Conference on Learning Representations, 2026. URL https://openrevie...

    2026

  11. [11]

    K. Chen, Z. Liu, T. Zhang, Z. Guo, S. Xu, H. Lin, H. Zang, X. Li, Q. Zhang, Z. Yu, G. Fan, T. Huang, Y . Wang, and C. Yu.πRL: Online rl fine-tuning for flow-based vision-language-action models, 2026. URLhttps://arxiv.org/abs/2510.25889

    work page arXiv 2026

  12. [12]

    P. Dong, Q. Li, D. Sadigh, and C. Finn. EXPO: Stable reinforcement learning with expressive policies. InThe Fourteenth International Conference on Learning Representations, 2026. URL https://openreview.net/forum?id=aFjSjkB6CV

    2026

  13. [13]

    Kelly, C

    M. Kelly, C. Sidrane, K. Driggs-Campbell, and M. J. Kochenderfer. Hg-dagger: Interactive imitation learning with human experts. In2019 International Conference on Robotics and Automation (ICRA), page 8077–8083. IEEE Press, 2019. doi:10.1109/ICRA.2019.8793698. URLhttps://doi.org/10.1109/ICRA.2019.8793698

    work page doi:10.1109/icra.2019.8793698 2019

  14. [14]

    Soft Actor-Critic Algorithms and Applications

    T. Haarnoja, A. Zhou, K. Hartikainen, G. Tucker, S. Ha, J. Tan, V . Kumar, H. Zhu, A. Gupta, P. Abbeel, et al. Soft actor-critic algorithms and applications.arXiv preprint arXiv:1812.05905, 2018

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  15. [15]

    Kalashnikov, A

    D. Kalashnikov, A. Irpan, P. Pastor, J. Ibarz, A. Herzog, E. Jang, D. Quillen, E. Holly, M. Kalakr- ishnan, V . Vanhoucke, et al. Scalable deep reinforcement learning for vision-based robotic manipulation. InConference on robot learning, pages 651–673. PMLR, 2018

    2018

  16. [16]

    Levine, C

    S. Levine, C. Finn, T. Darrell, and P. Abbeel. End-to-end training of deep visuomotor policies. Journal of Machine Learning Research, 17(39):1–40, 2016. URLhttp://jmlr.org/papers/ v17/15-522.html

    2016

  17. [17]

    Kalakrishnan, L

    M. Kalakrishnan, L. Righetti, P. Pastor, and S. Schaal. Learning force control policies for compliant manipulation. In2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 4639–4644, 2011. doi:10.1109/IROS.2011.6095096

    work page doi:10.1109/iros.2011.6095096 2011

  18. [18]

    M. P. Deisenroth, C. E. Rasmussen, and D. Fox. Learning to control a low-cost manipulator using data-efficient reinforcement learning. In H. Durrant-Whyte, N. Roy, and P. Abbeel, editors,Robotics: Science and Systems VII. The MIT Press, 06 2012. ISBN 9780262305969. doi:10.7551/mitpress/9481.003.0013. URL https://doi.org/10.7551/mitpress/9481. 003.0013

    work page doi:10.7551/mitpress/9481.003.0013 2012

  19. [19]

    T. C. Kietzmann and M. A. Riedmiller. The neuro slot car racer: Reinforcement learning in a real world setting.2009 International Conference on Machine Learning and Applications, pages 311–316, 2009. URLhttps://api.semanticscholar.org/CorpusID:17199272

    2009

  20. [20]

    Kober, K

    J. Kober, K. Mülling, O. Krömer, C. H. Lampert, B. Schölkopf, and J. Peters. Movement templates for learning of hitting and batting. In2010 IEEE International Conference on Robotics and Automation, pages 853–858, 2010. doi:10.1109/ROBOT.2010.5509672

    work page doi:10.1109/robot.2010.5509672 2010

  21. [21]

    Review of energy-efficient train control and timetabling

    J. Peters and S. Schaal. Reinforcement learning of motor skills with policy gradients. Neural Networks, 21(4):682–697, 2008. ISSN 0893-6080. doi:https://doi.org/10.1016/j. neunet.2008.02.003. URL https://www.sciencedirect.com/science/article/pii/ S0893608008000701. Robotics and Neuroscience. 12

    work page doi:10.1016/j 2008

  22. [22]

    P. J. Ball, L. Smith, I. Kostrikov, and S. Levine. Efficient online reinforcement learning with offline data. InProceedings of the 40th International Conference on Machine Learning, ICML’23. JMLR.org, 2023

    2023

  23. [23]

    P. Dong, A. M. Lessing, A. S. Chen, and C. Finn. Reinforcement learning via implicit imitation guidance, 2026. URLhttps://openreview.net/forum?id=CgupPwA40q

    2026

  24. [24]

    X. Chen, C. Wang, Z. Zhou, and K. W. Ross. Randomized ensembled double q-learning: Learning fast without a model. InInternational Conference on Learning Representations, 2021. URLhttps://openreview.net/forum?id=AY8zfZm0tDd

    2021

  25. [25]

    Nauman, M

    M. Nauman, M. Ostaszewski, K. Jankowski, P. Miło´s, and M. Cygan. Bigger, regularized, optimistic: scaling for compute and sample-efficient continuous control, 2024. URL https: //arxiv.org/abs/2405.16158

    work page arXiv 2024

  26. [26]

    Ankile, Z

    L. Ankile, Z. Jiang, R. Duan, G. Shi, P. Abbeel, and A. Nagabandi. Residual off-policy rl for finetuning behavior cloning policies, 2025. URLhttps://arxiv.org/abs/2509.19301

    work page arXiv 2025

  27. [27]

    J. Luo, P. Dong, Y . Zhai, Y . Ma, and S. Levine. Rlif: Interactive imitation learning as rein- forcement learning. In B. Kim, Y . Yue, S. Chaudhuri, K. Fragkiadaki, M. Khan, and Y . Sun, editors,International Conference on Learning Representations, volume 2024, pages 36329– 36351, 2024. URL https://proceedings.iclr.cc/paper_files/paper/2024/file/ 9c53788...

    2024

  28. [28]

    P. Dong, S. Mirchandani, D. Sadigh, and C. Finn. What matters for batch online reinforcement learning in robotics? InThe Fourteenth International Conference on Learning Representations,

  29. [29]

    URLhttps://openreview.net/forum?id=usw1NVkczu

  30. [30]

    L. Yang, Z. Huang, F. Lei, Y . Zhong, Y . Yang, C. Fang, S. Wen, B. Zhou, and Z. Lin. Policy representation via diffusion probability model for reinforcement learning, 2023. URL https: //arxiv.org/abs/2305.13122

    work page arXiv 2023

  31. [31]

    M. S. Mark, T. Gao, G. G. Sampaio, M. K. Srirama, A. Sharma, C. Finn, and A. Kumar. Policy agnostic rl: Offline rl and online rl fine-tuning of any class and backbone, 2024. URL https://arxiv.org/abs/2412.06685

    work page arXiv 2024

  32. [32]

    P. Dong, A. Swerdlow, D. Sadigh, and C. Finn. Faster: Value-guided sampling for fast rl, 2026. URLhttps://arxiv.org/abs/2604.19730

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  33. [33]

    Psenka, A

    M. Psenka, A. Escontrela, P. Abbeel, and Y . Ma. Learning a diffusion model policy from rewards via q-score matching, 2024. URLhttps://openreview.net/forum?id=StkLULT1i1

    2024

  34. [34]

    Li and S

    Q. Li and S. Levine. Q-learning with adjoint matching. InThe Fourteenth International Conference on Learning Representations, 2026. URL https://openreview.net/forum? id=vd4eNAdtO6

    2026

  35. [35]

    $\pi^{*}_{0.6}$: a VLA That Learns From Experience

    P. Intelligence, A. Amin, R. Aniceto, A. Balakrishna, K. Black, K. Conley, G. Connors, J. Darpinian, K. Dhabalia, J. DiCarlo, D. Driess, M. Equi, A. Esmail, Y . Fang, C. Finn, C. Glossop, T. Godden, I. Goryachev, L. Groom, H. Hancock, K. Hausman, G. Hussein, B. Ichter, S. Jakubczak, R. Jen, T. Jones, B. Katz, L. Ke, C. Kuchi, M. Lamb, D. LeBlanc, S. Levin...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  36. [36]

    J. Liu, F. Gao, B. Wei, X. Chen, Q. Liao, Y . Wu, C. Yu, and Y . Wang. What can RL bring to VLA generalization? an empirical study. InThe Thirty-ninth Annual Conference on Neural Informa- tion Processing Systems, 2026. URLhttps://openreview.net/forum?id=qmBMPInbZC. 13

    2026

  37. [37]

    S. Tan, K. Dou, Y . Zhao, and P. Krähenbühl. Interactive post-training for vision-language-action models, 2025. URLhttps://arxiv.org/abs/2505.17016

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  38. [38]

    G. Lu, W. Guo, C. Zhang, Y . Zhou, H. Jiang, Z. Gao, Y . Tang, and Z. Wang. Vla-rl: Towards masterful and general robotic manipulation with scalable reinforcement learning, 2025. URL https://arxiv.org/abs/2505.18719

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  39. [39]

    Y . Chen, S. Tian, S. Liu, Y . Zhou, H. Li, and D. Zhao. Conrft: A reinforced fine-tuning method for vla models via consistency policy, 2025. URLhttps://arxiv.org/abs/2502.05450

    work page arXiv 2025

  40. [40]

    Y . Guo, J. Zhang, X. Chen, X. Ji, Y .-J. Wang, Y . Hu, and J. Chen. Improving vision- language-action model with online reinforcement learning.2025 IEEE International Con- ference on Robotics and Automation (ICRA), pages 15665–15672, 2025. URL https: //api.semanticscholar.org/CorpusID:275932066

    2025

  41. [41]

    W. Xiao, H. Lin, A. Peng, H. Xue, T. He, Z. Luo, Y . Xie, F. Hu, L. Fan, G. Shi, and Y . Zhu. Self-improving vision-language-action models with data generation via residual RL. InThe Fourteenth International Conference on Learning Representations, 2026. URL https:// openreview.net/forum?id=eUGoqrZ6Ea

    2026

  42. [42]

    X. Yuan, T. Mu, S. Tao, Y . Fang, M. Zhang, and H. Su. Policy decorator: Model-agnostic online refinement for large policy model. InThe Thirteenth International Conference on Learning Representations, 2025. URLhttps://openreview.net/forum?id=e5jGTEiJMT

    2025

  43. [43]

    Zhang, C

    Y . Zhang, C. Wang, ouyang lu, Y . Zhao, Y . Ge, Z. Sun, X. Li, C. Zhang, C. Bai, and X. Li. Align-then-steer: Adapting the vision-language action models through unified latent guidance. InThe Fourteenth International Conference on Learning Representations, 2026. URLhttps: //openreview.net/forum?id=T3i7Ifeatk

    2026

  44. [44]

    K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition, 2015. URL https://arxiv.org/abs/1512.03385

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  45. [45]

    S. Ross, G. Gordon, and D. Bagnell. A reduction of imitation learning and structured prediction to no-regret online learning. In G. Gordon, D. Dunson, and M. Dudík, editors,Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, volume 15 of Proceedings of Machine Learning Research, pages 627–635, Fort Lauderdale,...

    2011

  46. [46]

    Dong, K.-H

    P. Dong, K.-H. Hung, A. Swerdlow, D. Sadigh, and C. Finn. Tql: Scaling q-functions with transformers by preventing attention collapse, 2026. URL https://arxiv.org/abs/2602. 01439

    2026

  47. [47]

    P. Dong, C. Zheng, C. Finn, D. Sadigh, and B. Eysenbach. Value flows, 2026. URL https: //arxiv.org/abs/2510.07650

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  48. [48]

    E. J. Hu, yelong shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen. LoRA: Low-rank adaptation of large language models. InInternational Conference on Learning Representations, 2022. URLhttps://openreview.net/forum?id=nZeVKeeFYf9. 14 A Additional Experiment Results A.1 Training Episode Time In addition, we provide training episode time p...

    2022