RAG-based QA System for School Regulations and MMLU-pro Benchmark

2024-Fall Term Project (Team6)

This repository contains the code and resources for Team6's RAG (Retrieval-Augmented Generation) model.

---

• This project applies the RAG (Retrieval-Augmented Generation) technique to solve problems related to school regulations (학칙) and the MMLU-pro benchmark.
• The solution utilizes Upstage's SLOAR-1-Mini-Chat as the only model used for inference and retrieval tasks.

Flowchart

Result