Overview of the NTCIR-18 Automatic Evaluation of LLMs (AEOLLM) Task

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Junjie, Li, Haitao, Chu, Zhumin, Liu, Yiqun, Ai, Qingyao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908271855534080
author Chen, Junjie
Li, Haitao
Chu, Zhumin
Liu, Yiqun
Ai, Qingyao
author_facet Chen, Junjie
Li, Haitao
Chu, Zhumin
Liu, Yiqun
Ai, Qingyao
contents In this paper, we provide an overview of the NTCIR-18 Automatic Evaluation of LLMs (AEOLLM) task. As large language models (LLMs) grow popular in both academia and industry, how to effectively evaluate the capacity of LLMs becomes an increasingly critical but still challenging issue. Existing methods can be divided into two types: manual evaluation, which is expensive, and automatic evaluation, which faces many limitations including task format (the majority belong to multiple-choice questions) and evaluation criteria (occupied by reference-based metrics). To advance the innovation of automatic evaluation, we propose the AEOLLM task which focuses on generative tasks and encourages reference-free methods. Besides, we set up diverse subtasks such as dialogue generation, text expansion, summary generation and non-factoid question answering to comprehensively test different methods. This year, we received 48 runs from 4 teams in total. This paper will describe the background of the task, the data set, the evaluation measures and the evaluation results, respectively.
format Preprint
id arxiv_https___arxiv_org_abs_2503_13038
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Overview of the NTCIR-18 Automatic Evaluation of LLMs (AEOLLM) Task
Chen, Junjie
Li, Haitao
Chu, Zhumin
Liu, Yiqun
Ai, Qingyao
Computation and Language
In this paper, we provide an overview of the NTCIR-18 Automatic Evaluation of LLMs (AEOLLM) task. As large language models (LLMs) grow popular in both academia and industry, how to effectively evaluate the capacity of LLMs becomes an increasingly critical but still challenging issue. Existing methods can be divided into two types: manual evaluation, which is expensive, and automatic evaluation, which faces many limitations including task format (the majority belong to multiple-choice questions) and evaluation criteria (occupied by reference-based metrics). To advance the innovation of automatic evaluation, we propose the AEOLLM task which focuses on generative tasks and encourages reference-free methods. Besides, we set up diverse subtasks such as dialogue generation, text expansion, summary generation and non-factoid question answering to comprehensively test different methods. This year, we received 48 runs from 4 teams in total. This paper will describe the background of the task, the data set, the evaluation measures and the evaluation results, respectively.
title Overview of the NTCIR-18 Automatic Evaluation of LLMs (AEOLLM) Task
topic Computation and Language
url https://arxiv.org/abs/2503.13038