On the Influence of Context Size and Model Choice in Retrieval-Augmented Generation Systems
This work addresses the lack of systematic exploration in RAG components, particularly context size and model choice, to guide robust system development for industrial applications.
The study systematically evaluated retrieval-augmented generation (RAG) systems for long-form question answering, finding that performance improves with up to 15 context snippets but stagnates or declines beyond that, and that different LLMs excel in biomedical versus encyclopedic domains.
Retrieval-augmented generation (RAG) has emerged as an approach to augment large language models (LLMs) by reducing their reliance on static knowledge and improving answer factuality. RAG retrieves relevant context snippets and generates an answer based on them. Despite its increasing industrial adoption, systematic exploration of RAG components is lacking, particularly regarding the ideal size of provided context, and the choice of base LLM and retrieval method. To help guide development of robust RAG systems, we evaluate various context sizes, BM25 and semantic search as retrievers, and eight base LLMs. Moving away from the usual RAG evaluation with short answers, we explore the more challenging long-form question answering in two domains, where a good answer has to utilize the entire context. Our findings indicate that final QA performance improves steadily with up to 15 snippets but stagnates or declines beyond that. Finally, we show that different general-purpose LLMs excel in the biomedical domain than the encyclopedic one, and that open-domain evidence retrieval in large corpora is challenging.