Saved in:
Bibliographic Details
Main Authors: Wang, Yicheng, Yuan, Jiayi, Chuang, Yu-Neng, Wang, Zhuoer, Liu, Yingchi, Cusick, Mark, Kulkarni, Param, Ji, Zhengping, Ibrahim, Yasser, Hu, Xia
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2408.13704
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929729744928768
author Wang, Yicheng
Yuan, Jiayi
Chuang, Yu-Neng
Wang, Zhuoer
Liu, Yingchi
Cusick, Mark
Kulkarni, Param
Ji, Zhengping
Ibrahim, Yasser
Hu, Xia
author_facet Wang, Yicheng
Yuan, Jiayi
Chuang, Yu-Neng
Wang, Zhuoer
Liu, Yingchi
Cusick, Mark
Kulkarni, Param
Ji, Zhengping
Ibrahim, Yasser
Hu, Xia
contents Large Language Models (LLMs) are increasingly serving as evaluators in Natural Language Generation (NLG) tasks; this is often referred to as ``LLM-as-a-judge'' paradigm. However, the capabilities of LLMs in evaluating NLG quality remain underexplored. Current studies depend on human assessments and simple metrics that fail to capture the discernment of LLMs across diverse NLG tasks. To address this gap, we propose the Discernment of Hierarchical Perturbation (DHP) benchmarking framework, which provides quantitative discernment scores for LLMs. This framework leverages hierarchically perturbed text data and statistical tests to systematically measure the NLG evaluation capabilities of LLMs. We re-established six evaluation datasets for this benchmark, covering four NLG tasks: Summarization, Story Completion, Question Answering, and Translation. Our comprehensive benchmarking of five major LLM families provides critical insight into their strengths and limitations as NLG evaluators. Our dataset is available at https://huggingface.co/datasets/YCWANGVINCE/DHP_Benchmark.
format Preprint
id arxiv_https___arxiv_org_abs_2408_13704
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DHP Benchmark: Are LLMs Good NLG Evaluators?
Wang, Yicheng
Yuan, Jiayi
Chuang, Yu-Neng
Wang, Zhuoer
Liu, Yingchi
Cusick, Mark
Kulkarni, Param
Ji, Zhengping
Ibrahim, Yasser
Hu, Xia
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) are increasingly serving as evaluators in Natural Language Generation (NLG) tasks; this is often referred to as ``LLM-as-a-judge'' paradigm. However, the capabilities of LLMs in evaluating NLG quality remain underexplored. Current studies depend on human assessments and simple metrics that fail to capture the discernment of LLMs across diverse NLG tasks. To address this gap, we propose the Discernment of Hierarchical Perturbation (DHP) benchmarking framework, which provides quantitative discernment scores for LLMs. This framework leverages hierarchically perturbed text data and statistical tests to systematically measure the NLG evaluation capabilities of LLMs. We re-established six evaluation datasets for this benchmark, covering four NLG tasks: Summarization, Story Completion, Question Answering, and Translation. Our comprehensive benchmarking of five major LLM families provides critical insight into their strengths and limitations as NLG evaluators. Our dataset is available at https://huggingface.co/datasets/YCWANGVINCE/DHP_Benchmark.
title DHP Benchmark: Are LLMs Good NLG Evaluators?
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2408.13704