Pre-trained Language Models Return Distinguishable Probability Distributions to Unfaithfully Hallucinated Texts.

Taehun Cha, Donghun Lee

Browse the full EMNLP paper archive.