Saved in:
Bibliographic Details
Main Authors: Liu, Zirui, Li, Jiatong, Zhuang, Yan, Liu, Qi, Shen, Shuanghong, Ouyang, Jie, Cheng, Mingyue, Wang, Shijin
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.03475
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910973586046976
author Liu, Zirui
Li, Jiatong
Zhuang, Yan
Liu, Qi
Shen, Shuanghong
Ouyang, Jie
Cheng, Mingyue
Wang, Shijin
author_facet Liu, Zirui
Li, Jiatong
Zhuang, Yan
Liu, Qi
Shen, Shuanghong
Ouyang, Jie
Cheng, Mingyue
Wang, Shijin
contents Arena-based evaluation is a fundamental yet significant evaluation paradigm for modern AI models, especially large language models (LLMs). Existing framework based on ELO rating system suffers from the inevitable instability problem due to ranking inconsistency and the lack of attention to the varying abilities of annotators. In this paper, we introduce a novel stable arena framework to address these issues by enhancing the ELO Rating System. Specifically, we replace the iterative update method with a Maximum Likelihood Estimation (MLE) approach, m-ELO, and provide theoretical proof of the consistency and stability of the MLE approach for model ranking. Additionally, we proposed the am-ELO, which modify the Elo Rating's probability function to incorporate annotator abilities, enabling the simultaneous estimation of model scores and annotator reliability. Experiments demonstrate that this method ensures stability, proving that this framework offers a more robust, accurate, and stable evaluation method for LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2505_03475
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle am-ELO: A Stable Framework for Arena-based LLM Evaluation
Liu, Zirui
Li, Jiatong
Zhuang, Yan
Liu, Qi
Shen, Shuanghong
Ouyang, Jie
Cheng, Mingyue
Wang, Shijin
Artificial Intelligence
Machine Learning
Arena-based evaluation is a fundamental yet significant evaluation paradigm for modern AI models, especially large language models (LLMs). Existing framework based on ELO rating system suffers from the inevitable instability problem due to ranking inconsistency and the lack of attention to the varying abilities of annotators. In this paper, we introduce a novel stable arena framework to address these issues by enhancing the ELO Rating System. Specifically, we replace the iterative update method with a Maximum Likelihood Estimation (MLE) approach, m-ELO, and provide theoretical proof of the consistency and stability of the MLE approach for model ranking. Additionally, we proposed the am-ELO, which modify the Elo Rating's probability function to incorporate annotator abilities, enabling the simultaneous estimation of model scores and annotator reliability. Experiments demonstrate that this method ensures stability, proving that this framework offers a more robust, accurate, and stable evaluation method for LLMs.
title am-ELO: A Stable Framework for Arena-based LLM Evaluation
topic Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2505.03475