ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems

arXiv:2311.09476v2 Announce Type: replace
Abstract: Evaluating retrieval-augmented generation (RAG) systems traditionally relies on hand annotations for input queries, passages to retrieve, and responses to generate. We introduce ARES, an Automated RAG Evaluation System, for evaluating RAG systems along the dimensions of context relevance, answer faithfulness, and answer relevance. By creating its own synthetic training data, ARES finetunes lightweight LM judges to assess the quality of individual RAG components. To mitigate potential prediction errors, ARES utilizes a small set of human-annotated …

abstract annotations ares arxiv automated context cs.ai cs.cl cs.ir dimensions evaluation framework generate queries rag responses retrieval retrieval-augmented systems type