@inproceedings{bharati-etal-2026-mirage,
    title = "The Mirage of Diversity: Unmasking the Cultural Vocabulary Ceiling in {LLM}s",
    author = "Bharati, Soumedhik  and
      Mukherjee, Subhrajit  and
      Mandal, Shibam",
    editor = "Prabhakaran, Vinodkumar  and
      Dev, Sunipa  and
      Benotti, Luciana  and
      Hershcovich, Daniel  and
      Cao, Yong  and
      Zhou, Li  and
      Ma, BOlei  and
      Adebara, Ife",
    booktitle = "Proceedings of the 4th Workshop on Cross-Cultural Considerations in {NLP} ({C}3{NLP} 2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.c3nlp-1.7/",
    pages = "101--107",
    ISBN = "979-8-89176-420-0",
    abstract = "Large Language Models are widely used to generate and adapt cultural texts, yet the depth of their cultural representation remains poorly quantified. Intuitively, as a narrative text expands in length, the diversity of cultural words should scale proportionately. To formally test this, we evaluate the FairyTaleQA dataset, adapted by three models and introduce our primary contribution: the Contextual Stereotype Amplification Index (CSAI), an evaluation framework combining LLM-as-a-judge extraction, embedding-based clich{\'e} anchoring, and Natural Language Inference (NLI) congruence validation. By mapping the frequency of extracted Culture Specific Items (CSIs) against narrative length using Heaps' Law ($V = k \cdot T^\beta$), we present empirical evidence of a systematic limitation in current systems: they struggle to scale cultural diversity even under explicit cultural prompting. Models rapidly hit a ``Cultural Vocabulary Ceiling,'' constrained to a fixed set of hyper-stereotypical terms. Furthermore, we demonstrate that merely optimizing for higher CSI frequency as done in prior works rewards logically broken tokenism. Our CSAI formulation actively penalizes such gratuitous stereotyping, offering a more principled approach to measuring and evaluating cultural homogenization in generative AI systems."
}