The study reveals the Small Model Learnability Gap, showing that smaller models benefit more from simpler reasoning chains. Mix Distillation improves their performance by balancing reasoning complexity.

https://arxiv.org/abs//2502.12143

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