##article.return## Hybrid Reinforcement: When Reward Is Sparse, It's Better to Be Dense Download Download PDF