@inproceedings{kim-lee-2026-language,
    title = "Language Directions in Multilingual {LLM}s: A Layer-wise Diagnostic Study of Token Alignment and Pretraining Imprint",
    author = "Kim, JaeSeong  and
      Lee, Suan",
    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.3/",
    pages = "30--35",
    ISBN = "979-8-89176-393-7",
    abstract = "We investigate how multilingual representations emerge across depth in large language models.Using a unified probing framework, we analyze six multilingual LLMs across five languages (EN/ES/ZH/FR/DE), decomposing behavior into (i) early-layer dynamics, (ii) linear vs. MLP separability, and (iii) token{--}language alignment that tracks where vocabulary sharing peaks.Across models, we observe a consistent and substantial early jump: accuracy rises by +73.5 to +80.7 points from L0 to L1 on average, indicating that language-relevant signals become accessible immediately after the embedding layer.Moreover, representations are largely linearly separable: for 5/6 models, the mean gap between MLP and linear probes remains within $\pm$0.5 points.Token{--}language alignment further reveals systematic structure, with peak vocabulary mass exceeding 48{\%} in some models and substantial variation in the depth of peak sharing.These findings provide a compact, cross-model characterization of how multilingual information is organized across depth and introduce simple alignment metrics that complement accuracy-based evaluation."
}