Publication
Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery
Molecule Discovery, Large Language Models
Cite this work View paper
Citation
Jiatong Li, Weida Wang*, Qinggang Zhang, Junxian Li, Di Zhang, Changmeng Zheng, Shufei Zhang, Xiaoyong Wei, and Qing Li. (2025). Mol-R1: Towards Explicit Long-CoT Reasoning in Molecule Discovery. arXiv preprint arXiv:2508.08401.
About this work
Mol-R1 trains explicit long-chain reasoning models for text-based molecule generation. Prior Regulation via In-context Distillation (PRID) produces paired reasoning traces guided by chemical constraints.
Molecular Iterative Adaptation (MoIA) alternates supervised fine-tuning with reinforced policy optimization. The paper evaluates the resulting models on text-based molecular reasoning and generation.