@inproceedings{shore-etal-2026-portnlp,
    title = "{P}ort{NLP} at {CRAC} 2026: {QL}o{RA} Fine-Tuning with Bounded Entity Registry for Multilingual Coreference Resolution",
    author = "Shore, Amber  and
      Scheinberg, Russell  and
      Nagasundaram, Malini  and
      Agrawal, Ameeta",
    editor = "Braud, Chlo{\'e}  and
      Hardmeier, Christian  and
      Ogrodniczuk, Maciej  and
      Loaiciga, Sharid  and
      Zeldes, Amir  and
      Nov{\'a}k, Michal  and
      Li, Chuyuan  and
      Strube, Michael  and
      Li, Junyi Jessy",
    booktitle = "Proceedings of the 2nd Joint Workshop on Computational Approaches to Discourse, Context and Document-Level Inferences and Computational Models of Reference, Anaphora and Coreference ({CODI}-{CRAC} 2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.codi-1.25/",
    pages = "193--198",
    ISBN = "979-8-89176-400-2",
    abstract = "We describe PortNLP{'}s submission to the CRAC 2026 Shared Task on Multilingual Coreference Resolution (LLM track). Our system fine-tunes Qwen 3 14B with QLoRA on CorefUD 1.4 gold annotations across 27 corpora spanning 19 languages. Documents are processed in 500-700 character chunks with a bounded rolling context consisting of 500 characters of recent annotated text and a scored entity registry that tracks up to 30 active entities via a frequency-times-recency decay formula. We employ data augmentation and language-aware sampling strategies to handle typological and data-size diversity. Our system achieves 68.69 CoNLL F1 averaged across all 27 test corpora. We additionally present probing experiments on the LoRA adapter{'}s internal representations, finding that coreference signal is concentrated in attention value projections rather than MLP modules, with the strongest readout at the earliest transformer layer."
}