@inproceedings{zheng-etal-2026-prompt,
    title = "Are All Prompt Components Value-Neutral? Understanding the Heterogeneous Adversarial Robustness of Dissected Prompt in {LLM}s",
    author = "Zheng, Yujia  and
      Li, Tianhao  and
      Huang, Haotian  and
      Zeng, Tianyu  and
      Lu, Jingyu  and
      Chu, Chuangxin  and
      Huang, Yuekai  and
      Jiang, Ziyou  and
      Xiong, Qian  and
      Ge, Yuyao  and
      Li, Mingyang",
    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.374/",
    pages = "7991--8019",
    ISBN = "979-8-89176-380-7",
    abstract = "Prompt-based adversarial attacks are a key tool for assessing the robustness of large language models (LLMs). Yet, existing studies typically treat prompts as flat text, overlooking their internal structure, different components within a prompt contribute unequally to robustness. This work introduces PromptAnatomy, a framework that decomposes prompts into functional components, and ComPerturb, a controlled perturbation method that selectively modifies these components to expose component-wise vulnerabilities while ensuring linguistic plausibility via perplexity-based filtering. Using this framework, four instruction-tuning datasets are structurally annotated and validated by human reviewers. Experiments across five advanced LLMs show that ComPerturb achieves state-of-the-art attack success rates, while ablation analyses confirm the complementary effects of prompt dissection and perplexity filtering. These results highlight the importance of structural awareness in evaluating and improving the adversarial robustness of LLMs."
}