Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
February 26, 2024
·
Boston
Gemini 1.5: Long Context Medical
Demonstrating how loading ~200k tokens of patient records into Gemini 1.5 enables direct long‑context queries and compares its accuracy to traditional RAG approaches.
Overview
I have loaded nearly 1/5 million tokens related to a patient (medical records, detailed clinical notes, patient notes and full text of related research papers) into Gemini 1.5. Let’s see how it performs vs. RAG.
Links
jheitzeb's GitHub profile: Ruby, Python, AI/LLM, web development projects.
Tech stack