Abstract
Retrieval Augmented Generation (RAG) pipelines reduce the frequency of Large Language Model (LLM) hallucinations by grounding the LLM context in knowledge base documents. Proposed herein is a lightweight, generalizable framework for evaluating RAG systems regarding both their document retrieval accuracy, as well as their answer-generation accuracy.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Kraus, Kelsey and Kroll, Margaret, "LIGHTWEIGHT RETRIEVAL AUGMENTED GENERATION (RAG) EVALUATION PIPELINE", Technical Disclosure Commons, (May 29, 2024)
https://www.tdcommons.org/dpubs_series/7053