Leveraging Large Language Models for NLG Evaluation: Advances and Challenges

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Li, Zhen, Xu, Xiaohan, Shen, Tao, Xu, Can, Gu, Jia-Chen, Lai, Yuxuan, Tao, Chongyang, Ma, Shuai
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909221919916032
author Li, Zhen
Xu, Xiaohan
Shen, Tao
Xu, Can
Gu, Jia-Chen
Lai, Yuxuan
Tao, Chongyang
Ma, Shuai
author_facet Li, Zhen
Xu, Xiaohan
Shen, Tao
Xu, Can
Gu, Jia-Chen
Lai, Yuxuan
Tao, Chongyang
Ma, Shuai
contents In the rapidly evolving domain of Natural Language Generation (NLG) evaluation, introducing Large Language Models (LLMs) has opened new avenues for assessing generated content quality, e.g., coherence, creativity, and context relevance. This paper aims to provide a thorough overview of leveraging LLMs for NLG evaluation, a burgeoning area that lacks a systematic analysis. We propose a coherent taxonomy for organizing existing LLM-based evaluation metrics, offering a structured framework to understand and compare these methods. Our detailed exploration includes critically assessing various LLM-based methodologies, as well as comparing their strengths and limitations in evaluating NLG outputs. By discussing unresolved challenges, including bias, robustness, domain-specificity, and unified evaluation, this paper seeks to offer insights to researchers and advocate for fairer and more advanced NLG evaluation techniques.
format Preprint
id arxiv_https___arxiv_org_abs_2401_07103
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Leveraging Large Language Models for NLG Evaluation: Advances and Challenges
Li, Zhen
Xu, Xiaohan
Shen, Tao
Xu, Can
Gu, Jia-Chen
Lai, Yuxuan
Tao, Chongyang
Ma, Shuai
Computation and Language
In the rapidly evolving domain of Natural Language Generation (NLG) evaluation, introducing Large Language Models (LLMs) has opened new avenues for assessing generated content quality, e.g., coherence, creativity, and context relevance. This paper aims to provide a thorough overview of leveraging LLMs for NLG evaluation, a burgeoning area that lacks a systematic analysis. We propose a coherent taxonomy for organizing existing LLM-based evaluation metrics, offering a structured framework to understand and compare these methods. Our detailed exploration includes critically assessing various LLM-based methodologies, as well as comparing their strengths and limitations in evaluating NLG outputs. By discussing unresolved challenges, including bias, robustness, domain-specificity, and unified evaluation, this paper seeks to offer insights to researchers and advocate for fairer and more advanced NLG evaluation techniques.
title Leveraging Large Language Models for NLG Evaluation: Advances and Challenges
topic Computation and Language
url https://arxiv.org/abs/2401.07103