@inproceedings{kuznetsov-etal-2025-feature,
    title = "Feature-Level Insights into Artificial Text Detection with Sparse Autoencoders",
    author = "Kuznetsov, Kristian  and
      Kushnareva, Laida  and
      Razzhigaev, Anton  and
      Druzhinina, Polina  and
      Voznyuk, Anastasia  and
      Piontkovskaya, Irina  and
      Burnaev, Evgeny  and
      Barannikov, Serguei",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/acl25-workshop-ingestion/2025.findings-acl.1321/",
    pages = "25727--25748",
    ISBN = "979-8-89176-256-5",
    abstract = "Artificial Text Detection (ATD) is becoming increasingly important with the rise of advanced Large Language Models (LLMs). Despite numerous efforts, no single algorithm performs consistently well across different types of unseen text or guarantees effective generalization to new LLMs. Interpretability plays a crucial role in achieving this goal. In this study, we enhance ATD interpretability by using Sparse Autoencoders (SAE) to extract features from Gemma-2-2B{'}s residual stream. We identify both interpretable and efficient features, analyzing their semantics and relevance through domain- and model-specific statistics, a steering approach, and manual or LLM-based interpretation of obtained features. Our methods offer valuable insights into how texts from various models differ from human-written content. We show that modern LLMs have a distinct writing style, especially in information-dense domains, even though they can produce human-like outputs with personalized prompts. The code for this paper is available at https://github.com/pyashy/SAE{\_}ATD."
}