NRITYAM: Language Models Meet Art and Heritage of Dance

NRITYAM is a newly introduced benchmark designed to assess the cultural comprehension of language models in the context of global dance traditions, comprising 9,260 question-answer pairs across 12 languages. This dataset, developed in collaboration with native dance artists and speakers, aims to evaluate a diverse range of models, including both large and small language models as well as multimodal approaches. It establishes a new standard for measuring AI systems' understanding of traditional performing arts, which is crucial for practitioners seeking to enhance cultural sensitivity and relevance in AI applications.