Finetuning VLAs for Dexterous Manipulation: The Right Checkpoint Is All You Need

Artur Kuramshin, Charlotte Cloutier, Michael Przystupa, Kirsty Ellis, Mohammad S. Akhter, Glen BersetharXiv 2026 · arXiv

Abstract

Vision-Language-Action models have demonstrated strong manipulation capabilities. A limitation of VLA models is that most manipulation datasets focus on parallel-jaw gripper systems, making transfer to multi-finger hands challenging, particularly as teleoperation datasets are harder to collect with multi-fingered robots. In this work, we study how to best adapt VLAs to dexterous hands by pairing supervised fine-tuning (SFT) with diffusion-steering reinforcement learning (DSRL). Using $π_{0.5}$ and a LEAP hand, we evaluate five fine-tuning recipes across two simulation tasks using only a handful of demonstrations. Our empirical analysis yields three key insights: (1) SFT success is poorly correlated with post-DSRL success, meaning the best SFT checkpoint is rarely the most effective to steer; (2) freezing the vision-language model (VLM) and tuning only the action expert produces the most steerable policy for both training and out-of-distribution placements; and (3) low-rank adaptation (LoRA) is the least steerable approach. Consequently, we recommend an adaptation recipe that freezes the VLM, fully trains the action expert, and eschews SFT rollout success as a checkpoint selection metric.
BibTeX
@misc{kuramshin2026finetuning,
  author = {Artur Kuramshin and Charlotte Cloutier and Michael Przystupa and Kirsty Ellis and Mohammad S. Akhter and Glen Berseth},
  title = {{Finetuning VLAs for Dexterous Manipulation: The Right Checkpoint Is All You Need}},
  year = {2026},
  eprint = {2611.23676},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2611.23676},
}

Loading related content...

Figure