Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Yang, Hernandez, Alberto Garcia, Kyslyi, Roman, Kersting, Nicholas
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2406.18064
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909379266084864
author Wang, Yang
Hernandez, Alberto Garcia
Kyslyi, Roman
Kersting, Nicholas
author_facet Wang, Yang
Hernandez, Alberto Garcia
Kyslyi, Roman
Kersting, Nicholas
contents We present a comprehensive study of answer quality evaluation in Retrieval-Augmented Generation (RAG) applications using vRAG-Eval, a novel grading system that is designed to assess correctness, completeness, and honesty. We further map the grading of quality aspects aforementioned into a binary score, indicating an accept or reject decision, mirroring the intuitive "thumbs-up" or "thumbs-down" gesture commonly used in chat applications. This approach suits factual business contexts where a clear decision opinion is essential. Our assessment applies vRAG-Eval to two Large Language Models (LLMs), evaluating the quality of answers generated by a vanilla RAG application. We compare these evaluations with human expert judgments and find a substantial alignment between GPT-4's assessments and those of human experts, reaching 83% agreement on accept or reject decisions. This study highlights the potential of LLMs as reliable evaluators in closed-domain, closed-ended settings, particularly when human evaluations require significant resources.
format Preprint
id arxiv_https___arxiv_org_abs_2406_18064
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Evaluating Quality of Answers for Retrieval-Augmented Generation: A Strong LLM Is All You Need
Wang, Yang
Hernandez, Alberto Garcia
Kyslyi, Roman
Kersting, Nicholas
Computation and Language
We present a comprehensive study of answer quality evaluation in Retrieval-Augmented Generation (RAG) applications using vRAG-Eval, a novel grading system that is designed to assess correctness, completeness, and honesty. We further map the grading of quality aspects aforementioned into a binary score, indicating an accept or reject decision, mirroring the intuitive "thumbs-up" or "thumbs-down" gesture commonly used in chat applications. This approach suits factual business contexts where a clear decision opinion is essential. Our assessment applies vRAG-Eval to two Large Language Models (LLMs), evaluating the quality of answers generated by a vanilla RAG application. We compare these evaluations with human expert judgments and find a substantial alignment between GPT-4's assessments and those of human experts, reaching 83% agreement on accept or reject decisions. This study highlights the potential of LLMs as reliable evaluators in closed-domain, closed-ended settings, particularly when human evaluations require significant resources.
title Evaluating Quality of Answers for Retrieval-Augmented Generation: A Strong LLM Is All You Need
topic Computation and Language
url https://arxiv.org/abs/2406.18064