A.B.M. Ashikur Rahman1, Saeed Anwar1,2, Muhammad Usman1,2, Ajmal Mian3,
"DefAn" is a comprehensive evaluation benchmark dataset, with more than 75000 samples, designed to assess the hallucination tendencies of large language models… See the full description on the dataset page:
https://huggingface.co/datasets/iamasQ/DefAn.