arXivCatherine Li, Jingchu Gai, Andrej Risteski, Jacob Mitchell Springer, Aditi RaghunathanMon, Nov 2, 2026, 10:56 AM PST
score 16.9
Muon's Training Speed Edge Fades When Fine-Tuning on More Data
Original: Pretraining Data Scale Reverses the Muon-AdamW Fine-Tuning Gap
Source: arxiv.org ↗
Writing ELI5 summary…