Improving Factuality in Large Language Models via Decoding-Time Hallucinatory and Truthful Comparators

Open in new window