@inproceedings{xie-etal-2025-language,
    title = "Can Language Neuron Intervention Reduce Non-Target Language Output?",
    author = "Xie, Suchun  and
      Kim, Hwichan  and
      Sasaki, Shota  and
      Yamada, Kosuke  and
      Suzuki, Jun",
    editor = "Belinkov, Yonatan  and
      Mueller, Aaron  and
      Kim, Najoung  and
      Mohebbi, Hosein  and
      Chen, Hanjie  and
      Arad, Dana  and
      Sarti, Gabriele",
    booktitle = "Proceedings of the 8th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.blackboxnlp-1.26/",
    pages = "452--466",
    ISBN = "979-8-89176-346-3",
    abstract = "Large language models (LLMs) often fail togenerate text in the intended target language,particularly in non-English interactions. Con-currently, recent work has explored LanguageNeuron Intervention (LNI) as a promising tech-nique for steering output language. In thispaper, we re-evaluate LNI in more practicalscenarios{---}specifically with instruction-tunedmodels and prompts that explicitly specify thetarget language. Our experiments show thatwhile LNI also shows potential in such practi-cal scenarios, its average effect is limited andunstable across models and tasks, with a 0.83{\%}reduction in undesired language output and a0.1{\%} improvement in performance. Our furtheranalysis identifies two key factors for LNI{'}slimitation: (1) LNI affects both the output lan-guage and the content semantics, making ithard to control one without affecting the other,which explains the weak performance gains. (2)LNI increases the target language token proba-bilities, but they often remain below the top-1generation threshold, resulting in failure to gen-erate the target language in most cases. Ourresults highlight both the potential and limi-tations of LNI, paving the way for future im-provements"
}