User-centric Subjective Leaderboard by Customizable Reward Modeling

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jia, Qi, Song, Xiujie, Zhang, Zicheng, Guo, Yijin, Zhang, Kaiwei, Chen, Zijian, Zhai, Guangtao
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915443678117888
author Jia, Qi
Song, Xiujie
Zhang, Zicheng
Guo, Yijin
Zhang, Kaiwei
Chen, Zijian
Zhai, Guangtao
author_facet Jia, Qi
Song, Xiujie
Zhang, Zicheng
Guo, Yijin
Zhang, Kaiwei
Chen, Zijian
Zhai, Guangtao
contents Existing benchmarks for large language models (LLMs) predominantely focus on assessing their capabilities through verifiable tasks. Such objective and static benchmarks offer limited utility for practical LLM selection, making it difficult for users to find suitable models for their individual needs. To bridge this gap, we present the first User-Centric Subjective Leaderboard (USL), which provides a preference-driven, dynamic ranking of LLMs across diverse real-world scenarios. Our work is built upon a thorough investigation of real human preference data, involving more than 10K subjective queries. Our investigation reveals significant diversity and contradictions in human preferences, which limit the effectiveness of state-of-the-art reward models. To address this, we introduce Customizable Reward Models (CRMs). With only 4B parameters, our CRM surpasses the performance of leading models such as GPT-4.1 and Gemini-2.5-pro, showing exceptional generalization capabilities across new topics and criteria. The USL, powered by CRMs, exhibits strong negative correlations to contradictory preferences.
format Preprint
id arxiv_https___arxiv_org_abs_2508_09463
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle User-centric Subjective Leaderboard by Customizable Reward Modeling
Jia, Qi
Song, Xiujie
Zhang, Zicheng
Guo, Yijin
Zhang, Kaiwei
Chen, Zijian
Zhai, Guangtao
Computation and Language
Existing benchmarks for large language models (LLMs) predominantely focus on assessing their capabilities through verifiable tasks. Such objective and static benchmarks offer limited utility for practical LLM selection, making it difficult for users to find suitable models for their individual needs. To bridge this gap, we present the first User-Centric Subjective Leaderboard (USL), which provides a preference-driven, dynamic ranking of LLMs across diverse real-world scenarios. Our work is built upon a thorough investigation of real human preference data, involving more than 10K subjective queries. Our investigation reveals significant diversity and contradictions in human preferences, which limit the effectiveness of state-of-the-art reward models. To address this, we introduce Customizable Reward Models (CRMs). With only 4B parameters, our CRM surpasses the performance of leading models such as GPT-4.1 and Gemini-2.5-pro, showing exceptional generalization capabilities across new topics and criteria. The USL, powered by CRMs, exhibits strong negative correlations to contradictory preferences.
title User-centric Subjective Leaderboard by Customizable Reward Modeling
topic Computation and Language
url https://arxiv.org/abs/2508.09463