You Only Need Minimal RLVR Training: Extrapolating LLMs via Rank-1 Trajectories Paper • 2605.21468 • Published 1 day ago • 38