CAIAC's Columbia Chatbot
The best way to understand LLMs is to work with and build projects with them on your own! Let’s create a fine tuned or RAG LLM or micro-base model (is that even a thing?) system that can answer Columbia specific questions. A side project will be creating our own crawling agent that gets data (reddit, sidechat, columbia calendar, instagram, registrar, etc.) - a small LLM to filter bad data - and then generating our own database. Let's build our deep learning intuition together! Roles Training: TBD Data: TBD