Deductive Closure Training of Language Models for Coherence, Accuracy, and Updatability

Akyürek, Afra Feyza, Akyürek, Ekin, Choshen, Leshem · arXiv · 2024