SFT or RL? An Early Investigation into Training R1-Like Reasoning Large Vision-Language Models Paper • 2504.11468 • Published Apr 10 • 28 • 2