@inproceedings{tsai-etal-2026-shape,
    title = "The Shape of Vulnerability: How Adversarial Perturbations Reshape the Topology of Language Model Latent Spaces",
    author = "Tsai, Angelina  and
      Subramanian, Shreya  and
      Liu, Catherine  and
      Lopez, Kimberly  and
      Zinn-Brooks, Leif  and
      Schulz, Alexia E.  and
      Uchendu, Adaku",
    editor = "T.Y.S.S., Santosh  and
      Rodriguez, Juan Diego  and
      de Gibert, Ona",
    booktitle = "Proceedings of the 64th Annual Meeting 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.acl-srw.24/",
    pages = "290--306",
    ISBN = "979-8-89176-393-7",
    abstract = "Adversarial perturbations in the context of large language models (LLMs) are subtle changes added to input data (i.e., images or text) that are designed to alter predictions or outputs of machine learning models. We introduce several novel visualizations using topological data analysis (TDA) (leveraging persistent homology) to characterize how adversarial perturbations act on text inputs, specifically, how sandbagging and code-injection attacksalter the geometric structure of attention heads in transformer models. By computing persistent homology metrics from attention maps across different model architectures (such as BERT, RoBERTa, ELECTRA, DistilGPT, etc.), we find that adversarial inputs alter higher-dimensional topological features ($H_{1}$ loops and $H_{2}$ voids) in ways that distinguish them from clean, non-adversarial inputs."
}