Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Qi, Zhang, Jie, Wang, Dongxia, Liu, Qiang, Li, Tianlin, Dong, Jin Song, Wang, Wenhai, Guo, Qing
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912380948054016
author Zhou, Qi
Zhang, Jie
Wang, Dongxia
Liu, Qiang
Li, Tianlin
Dong, Jin Song
Wang, Wenhai
Guo, Qing
author_facet Zhou, Qi
Zhang, Jie
Wang, Dongxia
Liu, Qiang
Li, Tianlin
Dong, Jin Song
Wang, Wenhai
Guo, Qing
contents Human preference plays a crucial role in the refinement of large language models (LLMs). However, collecting human preference feedback is costly and most existing datasets neglect the correlation between personalization and preferences. To address this issue, we introduce Fair-PP, a synthetic dataset of personalized preferences targeting social equity, derived from real-world social survey data, which includes 28 social groups, 98 equity topics, and 5 personal preference dimensions. Leveraging GPT-4o-mini, we engage in role-playing based on seven representative persona portrayals guided by existing social survey data, yielding a total of 238,623 preference records. Through Fair-PP, we also contribute (i) An automated framework for generating preference data, along with a more fine-grained dataset of personalized preferences; (ii) analysis of the positioning of the existing mainstream LLMs across five major global regions within the personalized preference space; and (iii) a sample reweighting method for personalized preference alignment, enabling alignment with a target persona while maximizing the divergence from other personas. Empirical experiments show our method outperforms the baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2505_11861
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity
Zhou, Qi
Zhang, Jie
Wang, Dongxia
Liu, Qiang
Li, Tianlin
Dong, Jin Song
Wang, Wenhai
Guo, Qing
Artificial Intelligence
Computation and Language
91C99
I.2.7; J.4
Human preference plays a crucial role in the refinement of large language models (LLMs). However, collecting human preference feedback is costly and most existing datasets neglect the correlation between personalization and preferences. To address this issue, we introduce Fair-PP, a synthetic dataset of personalized preferences targeting social equity, derived from real-world social survey data, which includes 28 social groups, 98 equity topics, and 5 personal preference dimensions. Leveraging GPT-4o-mini, we engage in role-playing based on seven representative persona portrayals guided by existing social survey data, yielding a total of 238,623 preference records. Through Fair-PP, we also contribute (i) An automated framework for generating preference data, along with a more fine-grained dataset of personalized preferences; (ii) analysis of the positioning of the existing mainstream LLMs across five major global regions within the personalized preference space; and (iii) a sample reweighting method for personalized preference alignment, enabling alignment with a target persona while maximizing the divergence from other personas. Empirical experiments show our method outperforms the baselines.
title Fair-PP: A Synthetic Dataset for Aligning LLM with Personalized Preferences of Social Equity
topic Artificial Intelligence
Computation and Language
91C99
I.2.7; J.4
url https://arxiv.org/abs/2505.11861