Improving Generalization in Intent Detection: GRPO with Reward-Based Curriculum Sampling

Open in new window