The paper introduces a method called Deductive Closure Training (DCT) that uses language models to improve their factuality and coherence. DCT prompts LMs to generate text, reason about its correctness, and fine-tune based on inferred correctness. DCT improves LM fact verification and text generation accuracy.

https://arxiv.org/abs//2401.08574

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers