Project Gnome

LLMs have the power to revolutionize every industry including healthcare, however, concerns over information accuracy and the lack of cited sources have stifled adoption to date.

Leveraging real-time LLMs powered by Groq and Retrieval Augmented Generation (RAG) we can increase the contextual knowledge of LLMs with a vector database containing ~3,000 genomics abstracts sourced from the PharmGKB database to provide accurate information with citations.

This technology can help medical professionals stay updated with gene-drug guidelines and enable anyone to process an entire genome into a human readable format in real-time.

For background, RAG reduces hallucinations and provides referenced responses by augmenting the input to the Large Language Models (LLMs) with additional information from a database of text-based sources thereby improving the accuracy and reliability of LLMs. To learn more about RAG, check out our RAG demo here.