{"ID":26811529,"CreatedAt":"2026-09-23T01:33:32.553386441Z","UpdatedAt":"2026-09-23T21:35:29.63677957Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2609.25755","arxiv_id":"2609.25755","title":"Syndrome, Synergy, and Safety: Structured Reasoning and Knowledge-Driven Alignment for TCM Prescription Generation","abstract":"Applying large language models to Traditional Chinese Medicine (TCM) prescription generation reveals three clinically critical gaps: models produce end-to-end mappings without auditable reasoning following the li-fa-fang-yao paradigm (SR Gap), treat each encounter in isolation without follow-up adjustment via sui zheng jia jian (LA Gap), and fail to enforce absolute contraindication rules such as Shi Ba Fan (SC Gap). We propose a progressive four-stage framework (SFT $\\to$ PG-CoT $\\to$ Dynamic $\\to$ K-RL) that addresses each gap: PG-CoT constrains CoT distillation under the li-fa-fang-yao paradigm to produce auditable diagnostic chains, Dynamic SFT models patient trajectories with explicit transition reasoning, and K-RL encodes deterministic pharmacological rules as rule-based DPO preference signals. Across 12 fine-tuned models and 6 zero-shot baselines, our framework substantially improves prescription quality over zero-shot baselines---with a 7B model (Mistral-7B) surpassing zero-shot GPT-5 on all three TCM evaluation metrics.","short_abstract":"Applying large language models to Traditional Chinese Medicine (TCM) prescription generation reveals three clinically critical gaps: models produce end-to-end mappings without auditable reasoning following the li-fa-fang-yao paradigm (SR Gap), treat each encounter in isolation without follow-up adjustment via sui zheng...","url_abs":"https://arxiv.org/abs/2609.25755","url_pdf":"https://arxiv.org/pdf/2609.25755v1","authors":"[\"Zheng Chen\",\"ZhiCheng Du\",\"Haoxuan Li\",\"Peiwu Qin\"]","published":"2026-09-22T06:49:19Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"Language Model\"]","has_code":false}