Model Overview
The snuh/hari-q3-8b is a specialized Large Language Model (LLM) developed by the Healthcare AI Research Institute (HARI) of Seoul National University Hospital (SNUH). It is primarily focused on clinical medicine and fine-tuned for medical Question-Answering (QA) in both English and Korean.
Key Capabilities & Performance
- Medical QA Generation: Optimized to provide accurate and reasoned answers to medical questions.
- Bilingual Support: Functions effectively in both English and Korean for medical contexts.
- Strong Medical Knowledge: Achieves 76.78% accuracy on the Korean Medical Licensing Examination (KMLE).
- Benchmark Performance: Demonstrates competitive accuracy against other medical LLMs on benchmarks like KorMedMCQA, MedQA-USMLE, and JAMA challenge.
Training & Ethics
The model was fine-tuned on a curated corpus of publicly available, de-identified Korean medical QA-style data. This includes clinical guidelines, academic publications, exam questions, and synthetic prompts reflecting real-world clinical reasoning. Strict adherence to ethical AI development ensures no real patient data or Personally Identifiable Information (PII) was used in training.
Ideal Use Cases
- Clinical Decision Support: Assisting healthcare professionals with QA-style inquiries.
- Medical Education: Serving as a tool for self-assessment and learning.
- Automated Medical Reasoning: Aiding in documentation and reasoning processes.
Important Note: This model is intended for research and educational purposes only and should not be used for making clinical decisions.