跳到主内容
Microsoft
separator
https://catalogartifact.azureedge.net/publicartifacts/johnsnowlabsinc1646051154808.medical_llm_medium-8d571d1d-4742-4f55-8d89-aa8158c80fb5/image3_logo216x216.png

Medical LLM - Medium

作者 John Snow Labs Inc

免费试用版

Model with advanced clinical decision-support system built for structured medical reasoning rather than simple knowledge retrieval.

This model is an advanced clinical decision-support system built for structured medical reasoning rather than simple knowledge retrieval. It analyzes symptoms, diagnostics, and longitudinal patient histories to guide complex diagnostic and treatment decisions in line with established clinical guidelines. With multimodal vision capabilities, it can also interpret medical images and visual documents alongside text, broadening the range of clinical inputs it can reason over. Now grounded in curated, specialty-specific medical knowledge bases, it anchors conclusions to vetted clinical evidence for greater factual reliability. It delivers transparent, step-aware reasoning, weighs competing hypotheses, and communicates uncertainty to support risk-aware clinical judgment. Benchmarking Results: - Achieves 94.5% average across OpenMed benchmarks; - Medical genetics: 99%; professional medicine: 98%; - Clinical knowledge comprehension: 95%; college biology mastery: 97.5%; - MedHallu: 97.5%; Anatomy: 93% - Achieves 94.7% on MedQA and 84% on PubMedQA; - Safety & reliability: 95.5% on hallucination detection, inappropriate-content , refusal of unethical/illegal requests; - Fairness & bias: 98% pass rate on name bias, 94% on racial bias , and 94.67% on anchoring-bias resistance Recommended Instance type for this model: Standard_NC96ads_A100_v4
中文(中国)
你的隐私选择选择退出图标 你的隐私选择
消费者健康隐私 站点地图 Contact Us 隐私和 Cookie 使用条款 商标 关于我们的广告 管理 Cookie