This paper introduces Woodpecker, a training-free method to mitigate hallucinations in Multimodal Large Language Models. It corrects hallucinations in generated text by extracting key concepts, validating visual knowledge, and generating visual claims. Woodpecker shows promising results on benchmark tests.

https://arxiv.org/abs//2310.16045

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