We just released an update to Private LLM that includes OpenBioLLM-8B, a biomedical LLM based on Llama 3 8B. OpenBioLLM-8B is available on both the iOS and macOS versions of Private LLM.

Developed by Saama AI Labs, OpenBioLLM-8B is a specialized AI model that excels in understanding and generating text with remarkable accuracy and fluency in the biomedical field. The model was fine-tuned using a vast corpus of high-quality biomedical data, enabling it to deliver superior performance on a wide range of tasks specific to the medical and life sciences industries. Boasting 8 billion parameters, OpenBioLLM-8B outperforms other open-source biomedical language models of similar scale. It has even shown better results compared to larger proprietary and open-source models, such as GPT-3.5 and Meditron-70B, on various biomedical benchmarks. OpenBioLLM-8B's advanced training techniques are built upon the robust foundations of Meta Llama 3 8B, incorporating the DPO dataset and fine-tuning recipe, along with a custom diverse medical instruction dataset. This unique combination allows OpenBioLLM-8B to achieve unparalleled performance in biomedical applications.

The creator of OpenBioLLM-8B recommends using the following settings to ensure optimal performance and minimize verbose output:

Temperature: 0.0
Top-P: 0.9

Additionally, the creator suggests using the following system prompt:

You are an expert and experienced from the healthcare and biomedical domain with extensive medical knowledge and practical experience. Your name is OpenBioLLM, and you were developed by Saama AI Labs. who's willing to help answer the user's query with explanation. In your explanation, leverage your deep medical expertise such as relevant anatomical structures, physiological processes, diagnostic criteria, treatment guidelines, or other pertinent medical concepts. Use precise medical terminology while still aiming to make the explanation clear and accessible to a general audience.

While OpenBioLLM-8B offers capabilities in processing biomedical information, it is not a substitute for professional medical advice or expertise. Users should consult medical professionals for health-related issues. During our testing, we observed that while OpenBioLLM-8B excels in the biomedical domain, it may generate inaccurate information on various topics. However, it remains a valuable addition to our collection of models.

iPhone and iPad users with devices featuring 8GB or more RAM (e.g., iPhone 15 Pro or newer) can take advantage of the 3-bit quantized version of OpenBioLLM-8B in Private LLM. For Mac users, the app offers a 4-bit quantized model, compatible with all Intel and Apple Silicon Macs. Try OpenBioLLM-8B in Private LLM today.