@inproceedings{li-etal-2026-retre,
    title = "{R}e{TRE}: Benchmarking {LLM} Transfer Robustness with Structure-Preserving Variants",
    author = "Li, ZhongDong  and
      Shi, Weijie  and
      Cui, Yue  and
      MA, Haolun  and
      Liu, Yuanjun  and
      Li, Jiawei  and
      Liu, An  and
      Zhu, Jia  and
      Xu, Jiajie",
    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.2048/",
    pages = "44257--44268",
    ISBN = "979-8-89176-390-6",
    abstract = "Large language models (LLMs) have achieved strong performance on standard benchmarks, yet their performance is not robust across different task manifestations. It remains unclear how performance changes under controlled task rewrites that preserve the original solution structure, while varying the rewrite type and level. To address this question, we introduce ReTRE (Rewrite-based Transfer Robustness Evaluation), an evaluation benchmark inspired by learning transfer theory that probes transfer robustness along two rewrite levels: Near Transfer and Far Transfer. ReTRE employs a multi-agent system to construct textual and visual variants while preserving the structure of the original solution. Evaluations on mathematical and science tasks across state-of-the-art multimodal LLMs reveal a consistent transfer gap: performance exhibits a general declining trend as transfer similarity drops and strong text performance can face performance decline under cross-modal transfer. Crucially, we identify a divergence between post-training paradigms: reinforcement learning preserves transfer robustness, whereas supervised fine-tuning tends to overfit the training distribution, leading to severe degradation in far-transfer performance despite strong in-distribution accuracy."
}