ClinAlign

ClinAlign is a clinician-grounded healthcare alignment framework that scales from instance rubrics to a reusable library of clinical principles, enabling robust preference alignment and inference-time self-revision for medical LLMs.

πŸ“ƒ Paper

ClinAlign


πŸ”₯ Highlights

  • Clinician-verified rubrics (HealthRubrics): a physician-validated preference dataset built by having clinicians revise and finalize LLM-drafted, checkable rubrics.
  • Reusable principle library (HealthPrinciples): distilled clinician consensus as 119 broadly reusable principles organized by clinical dimensions (urgency / uncertainty / expertise / task type).
  • Scalable supervision: principles can be converted into per-question rubrics for new, unlabeled medical queriesβ€”scaling training data without per-instance clinician authoring.
  • Inference-time alignment tool: retrieve matched principles β†’ generate rubric references β†’ guide iterative self-revision at test time.

πŸ₯‡ Results

πŸ€— Models: ClinAlign-4B β€’ ClinAlign-30B-A3B

ClinAlign results


Acknowledgement

Downloads last month
11
Safetensors
Model size
31B params
Tensor type
BF16
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for AQ-MedAI/ClinAligh-30B-A3B

Finetuned
(39)
this model

Paper for AQ-MedAI/ClinAligh-30B-A3B