How Reliable are LLMs as Knowledge Bases? Re-thinking Facutality and Consistency

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zheng, Danna, Lapata, Mirella, Pan, Jeff Z.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916524344737792
author Zheng, Danna
Lapata, Mirella
Pan, Jeff Z.
author_facet Zheng, Danna
Lapata, Mirella
Pan, Jeff Z.
contents Large Language Models (LLMs) are increasingly explored as knowledge bases (KBs), yet current evaluation methods focus too narrowly on knowledge retention, overlooking other crucial criteria for reliable performance. In this work, we rethink the requirements for evaluating reliable LLM-as-KB usage and highlight two essential factors: factuality, ensuring accurate responses to seen and unseen knowledge, and consistency, maintaining stable answers to questions about the same knowledge. We introduce UnseenQA, a dataset designed to assess LLM performance on unseen knowledge, and propose new criteria and metrics to quantify factuality and consistency, leading to a final reliability score. Our experiments on 26 LLMs reveal several challenges regarding their use as KBs, underscoring the need for more principled and comprehensive evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2407_13578
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle How Reliable are LLMs as Knowledge Bases? Re-thinking Facutality and Consistency
Zheng, Danna
Lapata, Mirella
Pan, Jeff Z.
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) are increasingly explored as knowledge bases (KBs), yet current evaluation methods focus too narrowly on knowledge retention, overlooking other crucial criteria for reliable performance. In this work, we rethink the requirements for evaluating reliable LLM-as-KB usage and highlight two essential factors: factuality, ensuring accurate responses to seen and unseen knowledge, and consistency, maintaining stable answers to questions about the same knowledge. We introduce UnseenQA, a dataset designed to assess LLM performance on unseen knowledge, and propose new criteria and metrics to quantify factuality and consistency, leading to a final reliability score. Our experiments on 26 LLMs reveal several challenges regarding their use as KBs, underscoring the need for more principled and comprehensive evaluation.
title How Reliable are LLMs as Knowledge Bases? Re-thinking Facutality and Consistency
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2407.13578