Dr. Samantha-7b: A Dual-Purpose Medical and Counseling LLM
Dr. Samantha-7b is a 7 billion parameter language model developed by sethuiyer, built upon the Llama 2 architecture. This model was created using mergekit by combining Severus27/BeingWell_llama2_7b and ParthasarathyShanmugam/llama-2-7b-samantha.
Key Capabilities
- Integrated Expertise: Merges medical knowledge (from USMLE databases and doctor-patient interactions) with philosophical, psychological, and relational understanding.
- Holistic Support: Designed to act as both a medical consultant and a personal counselor, addressing both physical and mental well-being.
- Performance: Achieves an average of 52.95 on the OpenLLM Leaderboard, with notable scores in Hellaswag (77.95) and Winogrande (73.56).
- Medical Accuracy: Demonstrates strong medical knowledge, with subject-wise accuracies including Human Aging (58.29%), Human Sexuality (55.73%), and Clinical Knowledge (52.83%).
- GPT-4 Evaluation: Received an overall rating of 83.5/100 from GPT-4 across 25 random prompts from the ChatDoctor-200k Dataset, praised for extensive medical knowledge, emphasis on professional diagnoses, and empathetic responses.
Ideal Use Cases
- Medical Consultation: Providing initial insights into potential causes for symptoms and general medical information.
- Personal Counseling: Offering support and guidance on psychological and relational topics.
- Health & Wellness Applications: Integrating into platforms requiring a blend of medical and mental health advisory capabilities.
- Educational Tools: Assisting in learning about medical conditions and psychological concepts.