AI hallucination benchmark data and model performance comparisons provide an...

https://orcid.org/0009-0003-6458-2847

AI hallucination benchmark data and model performance comparisons provide an empirically grounded lens to evaluate how often and under what conditions language models generate factually inaccurate or nonsensical outputs