IMU-1: Sample-Efficient Pre-training of Small Language Models Paper • 2602.02522 • Published 12 days ago • 5