The paper identifies "semantic leakage" in language models, revealing how irrelevant prompt information influences outputs, and proposes methods for detection and evaluation across multiple languages and scenarios.

https://arxiv.org/abs//2408.06518

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