A Framework for Automated Measurement of Responsible AI Harms in Generative AI Applications.

Chain of Natural Language Inference for Reducing Large Language Model Ungrounded Hallucinations.