mlsanityData to 7 Oct 2026

Medical LLM alignment

104 papers2 this weekHigh confidence in this grouping: the papers form one clear topicHigh confidence in this grouping: the papers form one clear topic · About

Keywords: medical, clinical, preference, optimization, reasoning

Key facts. Medical LLM alignment: 104 papers on arXiv since 2023 (data through 7 Oct 2026). 7 of its 104 mapped papers (7%) cite “HuatuoGPT, towards Taming Language Model to Be a Doctor” (2023, arXiv 2305.15075). 6 of its 104 mapped papers (6%) cite “MMedPO: Aligning Medical Vision-Language Models with Clinical-Aware Multimodal Preference Optimization” (2024, arXiv 2412.06141). 4 new papers 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

  1. This week · 2
  2. Multi-Objective Aligned Small Language Model Framework for SUD Patient Dialogue Generation
  3. Ontology Concept Overlap as a Training Signal: Knowledge-Grounded Reinforcement Learning for Clinical Question Answering
  4. Earlier · 2
  5. Syndrome, Synergy, and Safety: Structured Reasoning and Knowledge-Driven Alignment for TCM Prescription Generation
  6. Aligning with Lived Experience: Heterogeneous Benefits of Fine Tuning in Mental Health Support Generation

Foundations

Most cited by papers here

  1. HuatuoGPT, towards Taming Language Model to Be a Doctor2023cited by 7 here
  2. MMedPO: Aligning Medical Vision-Language Models with Clinical-Aware Multimodal Preference Optimization2024cited by 6 here
  3. Zhongjing: Enhancing the Chinese Medical Capabilities of Large Language Model through Expert Feedback and Real-world Multi-turn Dialogue2023cited by 5 here
  4. STLLaVA-Med: Self-Training Large Language and Vision Assistant for Medical Question-Answering2024cited by 4 here
  5. Integrating Physician Diagnostic Logic into Large Language Models: Preference Learning from Process Feedback2024cited by 3 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2017Deep reinforcement learning from human preferencescited by 15 here
  2. 2019Fine-Tuning Language Models from Human Preferencescited by 3 here
  3. 2020Learning to summarize from human feedbackcited by 7 here
  4. 2022Training language models to follow instructions with human feedbackcited by 41 here
  5. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 13 here
  6. 2022Scaling Laws for Reward Model Overoptimizationcited by 4 here
  7. 2024Aligning Modalities in Vision Large Language Models via Preference Fine-tuningcited by 5 here
  8. 2024MMedPO: Aligning Medical Vision-Language Models with Clinical-Aware Multimodal Preference Optimizationcited by 6 here
16 more it builds on, not in this chain
  1. 2022Solving math word problems with process- and outcome-based feedbackcited by 4 here
  2. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 7 here
  3. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 3 here
  4. 2023RRHF: Rank Responses to Align Language Models with Human Feedback without tearscited by 3 here
  5. 2023AlpacaFarm: A Simulation Framework for Methods that Learn from Human Feedbackcited by 3 here
  6. 2023HuatuoGPT, towards Taming Language Model to Be a Doctorcited by 7 here
  7. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 57 here
  8. 2023BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Datasetcited by 4 here
  9. 2023Aligning Large Language Models with Human: A Surveycited by 3 here
  10. 2023Zhongjing: Enhancing the Chinese Medical Capabilities of Large Language Model through Expert Feedback and Real-world Multi-turn Dialoguecited by 5 here
  11. 2023RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedbackcited by 6 here
  12. 2023A General Theoretical Paradigm to Understand Learning from Human Preferencescited by 5 here
  13. 2024ORPO: Monolithic Preference Optimization without Reference Modelcited by 7 here
  14. 2024Token-level Direct Preference Optimizationcited by 4 here
  15. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 5 here
  16. 2024STLLaVA-Med: Self-Training Large Language and Vision Assistant for Medical Question-Answeringcited by 4 here

The 24 most cited of 247.

Browse

30 of 104, ranked 6–35

  1. 6Med-PRM: Medical Reasoning Models with Stepwise, Guideline-verified Process Rewards20253
  2. 7CheXPO: Preference Optimization for Chest X-ray VLMs with Counterfactual Rationale20253
  3. 8InfiMed-ORBIT: Aligning LLMs on Open-Ended Complex Tasks via Rubric-Based Incremental Training20252
  4. 9Multidimensional Rubric-oriented Reward Model Learning via Geometric Projection Reference Constraints20252
  5. 10Direct Preference Optimization for Suppressing Hallucinated Prior Exams in Radiology Report Generation20242
  6. 11AlpaCare:Instruction-tuned Large Language Models for Medical Application20232
  7. 12Benchmarking Direct Preference Optimization for Medical Large Vision-Language Models20261
  8. 13ALFA: Aligning LLMs to Ask Good Questions A Case Study in Clinical Reasoning20251
  9. 14DecoupledESC: Enhancing Emotional Support Generation via Strategy-Response Decoupled Preference Optimization20251
  10. 15MedGR2: Breaking the Data Barrier for Medical Reasoning via Generative Reward Learning20251
All 30
  1. 16Mitigating Unhelpfulness in Emotional Support Conversations with Multifaceted AI Feedback20241
  2. 17SYNFAC-EDIT: Synthetic Imitation Edit Feedback for Factual Alignment in Clinical Summarization20241
  3. 18Biomedical Visual Instruction Tuning with Clinician Preference Alignment20241
  4. 19Fine Tuning Large Language Models for Medicine: The Role and Importance of Direct Preference Optimization20241
  5. 20CheXalign: Preference fine-tuning in chest X-ray interpretation models without human feedback20241
  6. 21Aligning Large Language Models for Clinical Tasks20231
  7. 22PsychEthicsBench: Evaluating Large Language Models Against Australian Mental Health Ethics20260
  8. 23A Human-Centric Pipeline for Aligning Large Language Models with Chinese Medical Ethics20260
  9. 24LLMdoctor: Token-Level Flow-Guided Preference Optimization for Efficient Test-Time Alignment of Large Language Models20260
  10. 25PERM: Psychology-grounded Empathetic Reward Modeling for Large Language Models20260
  11. 26How Much Would a Clinician Edit This Draft? Evaluating LLM Alignment for Patient Message Response Drafting20260
  12. 27Expert Evaluation and the Limits of Human Feedback in Mental Health AI Safety Testing20260
  13. 28ClinAlign: Scaling Healthcare Alignment from Clinician Preference20260
  14. 29Quark Medical Alignment: A Holistic Multi-Dimensional Alignment and Collaborative Optimization Paradigm20260
  15. 30Multi-Objective Alignment of Language Models for Personalized Psychotherapy20260
  16. 31PrivMedChat: End-to-End Differentially Private RLHF for Medical Dialogue Systems20260
  17. 32MAPLE: Elevating Medical Reasoning from Statistical Consensus to Process-Led Alignment20260
  18. 33Coverage-Controlled Preference Mining from Noisy Claim Verification for Evidence-Grounded Generation20260
  19. 34From Baselines to Preferences: A Comparative Study of LoRA/QLoRA and Preference Optimization for Mental Health Text Classification20260
  20. 35ProMedical: Hierarchical Fine-Grained Criteria Modeling for Medical LLM Alignment via Explicit Injection20260