@article{von-daniken-etal-2026-c4,
    title = "C4: A Multilingual Benchmark for Retrieval-Augmented Generation Based on the Catechism of the Catholic Church and Its Compendium",
    author = {von D{\"a}niken, Pius  and
      Cieliebak, Mark  and
      Deriu, Jan},
    editor = "Piperidis, Stelios  and
      Bel, N{\'u}ria  and
      van den Heuvel, Henk  and
      Ide, Nancy  and
      Krek, Simon  and
      Toral, Antonio",
    journal = "International Conference on Language Resources and Evaluation",
    volume = "main",
    month = may,
    year = "2026",
    address = "Palma de Mallorca, Spain",
    publisher = "ELRA Language Resource Association",
    url = "https://preview.aclanthology.org/ingest-lrec/2026.lrec-main.590/",
    pages = "7446--7456",
    abstract = "We introduce a new multilingual case study for evaluating retrieval augmented generation (RAG) systems, based on the Catechism of the Catholic Church and its Compendium. The Catechism is a structured document with numbered paragraphs, officially translated into many languages under strict editorial alignment. The Compendium reformulates this material into a question-answer format with explicit citations to the corresponding paragraphs. Together, they form a set of parallel monolingual corpora that share identical semantic structure, enabling direct, controlled comparison of RAG performance across languages. Beyond its theological origin, this text pair closely mirrors real-world applications of RAG in institutional contexts, such as querying internal policy documents with associated FAQ-style summaries, making it a practical testbed for multilingual retrieval and grounded answer generation. We release our data collection scripts and baseline results for further research."
}