@inproceedings{liang-etal-2026-pivotattack,
    title = "{P}ivot{A}ttack: Rethinking the Search Trajectory in Hard-Label Text Attacks via Pivot Words",
    author = "Liang, Yuzhi  and
      Xiao, Shiliang  and
      Wei, Jingsong  and
      Lin, Qiliang  and
      Li, Xia",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.157/",
    pages = "3190--3207",
    ISBN = "979-8-89176-395-1",
    abstract = "Existing hard-label text attacks often rely on inefficient ``outside-in'' strategies that traverse vast search spaces. We propose PivotAttack, a query-efficient ``inside-out'' framework. It employs a Multi-Armed Bandit algorithm to identify Pivot Sets{---}combinatorial token groups acting as prediction anchors{---}and strategically perturbs them to induce label flips. This approach captures inter-word dependencies and minimizes query costs. Extensive experiments across traditional models and Large Language Models demonstrate that PivotAttack consistently outperforms state-of-the-art baselines in both Attack Success Rate and query efficiency."
}