@inproceedings{zhu-etal-2026-data,
    title = "Data Efficient {RLVR} via Off-Policy Influence Guidance",
    author = "Zhu, Erle  and
      Jiang, Dazhi  and
      Wang, Yuan  and
      Li, Xujun  and
      Cheng, Jiale  and
      Gu, Yuxian  and
      Niu, Yilin  and
      Zeng, Aohan  and
      Tang, Jie  and
      Huang, Minlie  and
      Wang, Hongning",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.2141/",
    pages = "46167--46192",
    ISBN = "979-8-89176-390-6",
    abstract = "Data selection is a critical aspect of Reinforcement Learning with Verifiable Rewards (RLVR) for enhancing the reasoning capabilities of large language models (LLMs). Current data selection methods are largely heuristic-based, lacking theoretical guarantees and generalizability. This work proposes a theoretically-grounded approach using influence functions to estimate the contribution of each data point to the learning objective. To overcome the prohibitive computational cost of policy rollouts required for online influence estimation, we introduce an off-policy influence estimation method that efficiently approximates data influence using pre-collected offline trajectories. Furthermore, to manage the high-dimensional gradients of LLMs, we employ sparse random projection to reduce dimensionality and improve storage and computation efficiency. Leveraging these techniques, we develop Curriculum RL with Off-Policy Influence guidance (CROPI), a multi-stage RL framework that iteratively selects the most influential data for the current policy. Experiments on models up to 7B parameters demonstrate that CROPI significantly accelerates training. On a 1.5B model, it achieves a 2.66x step-level acceleration while using only 10{\%} of the data per stage compared to full-dataset training. Our results highlight the substantial potential of influence-based data selection for efficient RLVR."
}