@inproceedings{czajka-2025-lightweight,
    title = "Lightweight {IPIS} Instruction Tuning of Bielik-7{B} for Gender-Inclusive {P}olish{\ensuremath{<}}{---}{\ensuremath{>}}{E}nglish Translation: System Description for {P}ol{E}val 2025 Task 2 ({IPIS}-translation)",
    author = "Czajka, Mateusz",
    editor = "Kobyli{\'n}ski, {\L}ukasz  and
      Wr{\'o}blewska, Alina  and
      Ogrodniczuk, Maciej",
    booktitle = "Proceedings of the {P}ol{E}val 2025 Workshop",
    month = nov,
    year = "2025",
    address = "Warsaw",
    publisher = "Institute of Computer Science PAS and Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/acl-awards-reasoning/2025.poleval-main.10/",
    pages = "66--71",
    abstract = "We describe a compact but fully open-source system submitted to PolEval 2025 Task 2 (Gender-inclusive LLMs for Polish), subtask B: IPIS-translation. The goal of this subtask is gender-sensitive Polish{\ensuremath{\leftrightarrow}}English translation, including the production of gender-inclusive Polish outputs that follow specific orthographic conventions such as gender stars and slash forms. Our method performs instruction tuning of the Polish LLM Bielik-7B-Instruct using parameter-efficient LoRA adapters, with optional 4-bit NF4 quantization for single-GPU training. Samples from the Inclusive Polish Instruction Set (IPIS) are converted into a chat-style format with a task-provided gender-inclusive system prompt. Despite a deliberately lightweight tuning budget and greedy decoding, our submission placed 3rd on the hidden test B split, achieving bleu{\_}pe = 20.7871. We detail the training and inference pipeline, discuss design choices and limitations, and outline directions for improving inclusive translation quality in Polish."
}