johnsnowlabs/JSL-MedLlama-3-8B-v2.0
The JSL-MedLlama-3-8B-v2.0 model, developed by John Snow Labs, is an 8 billion parameter language model with an 8192 token context length. It is specifically fine-tuned for medical question answering and knowledge tasks, demonstrating strong performance across various medical benchmarks. This model is optimized for applications requiring accurate and nuanced understanding of medical information.
Loading preview...
JSL-MedLlama-3-8B-v2.0: Medical Language Model
This model, developed by John Snow Labs, is an 8 billion parameter language model designed for medical applications. It features an 8192 token context length, enabling it to process and understand longer medical texts and complex queries.
Key Capabilities
The model demonstrates proficiency in various medical knowledge and question-answering tasks, as evidenced by its evaluation results:
- Medical Question Answering: Achieves 0.6118 accuracy on MedMCQA and 0.6143 on MedQA_4options.
- MMLU Medical Subdomains: Shows strong performance in specialized medical areas, including:
- Anatomy: 0.7185 accuracy
- Clinical Knowledge: 0.7811 accuracy
- College Biology: 0.8264 accuracy
- College Medicine: 0.7110 accuracy
- Medical Genetics: 0.8300 accuracy
- Professional Medicine: 0.7868 accuracy
- PubMedQA: Achieves 0.7420 accuracy on this biomedical question answering dataset.
Good For
- Medical Q&A Systems: Ideal for building applications that answer medical questions based on provided context or general medical knowledge.
- Clinical Decision Support: Can assist in processing and interpreting clinical information.
- Biomedical Research: Useful for tasks involving the extraction and synthesis of information from medical literature.
Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.