[QA] Examining Two Hop Reasoning Through Information Content Scaling

This study investigates transformers' inconsistent performance on two-hop questions, revealing that capacity scaling and generalization affect their ability to learn and answer these complex queries effectively.

https://arxiv.org/abs//2502.03490

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