sethuiyer/Dr_Samantha_7b_mistral
Dr_Samantha_7b_mistral is a 7 billion parameter Mistral-based model developed by sethuiyer, featuring a 4096-token context length. This model is a merge of MedMistral-7B-v0.1 and Samantha-v2, designed to combine medical knowledge with empathetic communication skills. It is optimized for applications requiring both technical medical understanding and personal counseling, supporting physical and mental well-being.
Loading preview...
Dr_Samantha_7b_mistral: AI for Integrated Healthcare
Dr_Samantha_7b_mistral is a 7 billion parameter model that merges the medical expertise of segmed/MedMistral-7B-v0.1 with the philosophical and relational understanding of Guilherme34/Samantha-v2. This unique combination aims to provide a balanced AI assistant capable of both technical medical consultation and empathetic personal counseling.
Key Capabilities
- Dual Expertise: Integrates medical knowledge (trained on USMLE databases and doctor-patient interactions) with psychological and relational understanding.
- Whole-Person Care: Designed to support both physical and mental well-being, offering a holistic approach to AI-assisted care.
- Medical Domain Performance: Achieves an average accuracy of 68.82% across various medical, biological, and psychological subjects in OpenLLM evaluations, including 70.57% in Clinical Knowledge and 83.12% in High School Psychology.
Good For
- Medical Consultation: Providing differential diagnoses and general medical advice based on its training.
- Personal Counseling: Engaging in empathetic and understanding interactions for mental well-being support.
- Integrated Healthcare Applications: Use cases requiring an AI that can bridge the gap between clinical information and patient-centric communication.
It's important to note that while the model performs reasonably well in medical subjects, it should not be used for definitive medical diagnosis or treatment decisions, which require higher accuracy and human oversight.