@inproceedings{akter-etal-2026-nemotron,
    title = "Nemotron-{C}ross{T}hink: Scaling Self-Learning beyond Math Reasoning",
    author = "Akter, Syeda Nahida  and
      Prabhumoye, Shrimai  and
      Novikov, Matvei  and
      Han, Seungju  and
      Lin, Ying  and
      Bakhturina, Evelina  and
      Nyberg, Eric  and
      Choi, Yejin  and
      Patwary, Mostofa  and
      Shoeybi, Mohammad  and
      Catanzaro, Bryan",
    editor = "Demberg, Vera  and
      Inui, Kentaro  and
      Marquez, Llu{\'i}s",
    booktitle = "Proceedings of the 19th Conference of the {E}uropean Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-eacl/2026.eacl-long.43/",
    pages = "984--1002",
    ISBN = "979-8-89176-380-7",
    abstract = "Prior work has successfully applied Reinforcement Learning (RL) to mathematical reasoning{---}where rules and correctness are well-defined. Yet, generalizing these methods to broader reasoning domains remains challenging due to limited data and the lack of verifiable rewards for unstructured domains. In this work, we propose NEMOTRON-CROSSTHINK, a framework that systematically incorporates multi-domain corpora into RL training to improve generalization across diverse reasoning tasks. NEMOTRON-CROSSTHINK addresses key challenges by (1) combining data from varied sources; (2) applying structured templates to control answer-space complexity; (3) filtering for verifiable answers; and (4) optimizing data blending strategies to utilize multi-source data effectively. This enables scalable and verifiable reward modeling beyond math and demonstrates improved accuracies on both math (MATH-500: +30.1{\%}, AMC23: +27.5{\%}) and non-math reasoning benchmarks (MMLU-PRO: +12.8{\%}, GPQA-DIAMOND: +11.3{\%}, AGIEVAL: +15.1{\%}, SUPERGPQA: +3.8{\%}). Moreover, NEMOTRON-CROSSTHINK exhibits significantly improved response efficiency{---}using 28{\%} fewer tokens for correct answers{---}highlighting more focused and effective reasoning. Through NEMOTRON-CROSSTHINK, we demonstrate that integrating multi-domain, multi-format data in RL leads to more accurate, efficient, and generalizable LLMs. All of our datasets are available on HuggingFace."
}