##article.return## Learning Without Critics? Revisiting GRPO in Classical Reinforcement Learning Environments Download Download PDF