Our preprint "Active-GRPO: Adaptive Imitation and Self-Improving Reasoning for Molecular Optimization" is posted on arXiv.