@inproceedings{sterz-etal-2025-recover,
    title = "{R}e{C}o{V}e{R} the Target Language: Language Steering without Sacrificing Task Performance",
    author = "Sterz, Hannah  and
      Schmidt, Fabian David  and
      Glava{\v{s}}, Goran  and
      Vuli{\'c}, Ivan",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/author-page-yu-wang-polytechnic/2025.findings-emnlp.1056/",
    doi = "10.18653/v1/2025.findings-emnlp.1056",
    pages = "19390--19405",
    ISBN = "979-8-89176-335-7",
    abstract = "As they become increasingly multilingual, Large Language Models (LLMs) exhibit more language confusion, i.e., they tend to generate answers in a language different from the language of the prompt or the answer language explicitly requested by the user. In this work, we propose ReCoVeR (REducing language COnfusion in VEctor Representations), a novel lightweight approach for reducing language confusion based on language-specific steering vectors. We first isolate language vectors with the help of multi-parallel corpus and then effectively leverage those vectors for effective LLM steering via fixed (i.e., unsupervised) as well as trainable steering functions. Our extensive evaluation, encompassing three benchmarks and 18 languages, shows that ReCoVeR effectively mitigates language confusion in both monolingual and cross-lingual setups while at the same time{---}and in contrast to prior language steering methods{---}retaining task performance. Our data code is available at https://github.com/hSterz/recover."
}