CommunityBench: Benchmarking Community-Level Alignment across Diverse Groups and Tasks

TL;DR

This paper introduces CommunityBench, a new benchmark for evaluating how well large language models align with community-specific values, addressing the limitations of universal or individual-focused approaches.

Contribution

It proposes community-level alignment as a scalable middle ground and provides the first large-scale benchmark to evaluate this concept in LLMs.

Findings

Current LLMs show limited capacity for community-specific preferences.

Community-level alignment can facilitate more scalable and pluralistic model alignment.

CommunityBench enables comprehensive evaluation of community alignment in LLMs.

Abstract

Large language models (LLMs) alignment ensures model behaviors reflect human value. Existing alignment strategies primarily follow two paths: one assumes a universal value set for a unified goal (i.e., one-size-fits-all), while the other treats every individual as unique to customize models (i.e., individual-level). However, assuming a monolithic value space marginalizes minority norms, while tailoring individual models is prohibitively expensive. Recognizing that human society is organized into social clusters with high intra-group value alignment, we propose community-level alignment as a "middle ground". Practically, we introduce CommunityBench, the first large-scale benchmark for community-level alignment evaluation, featuring four tasks grounded in Common Identity and Common Bond theory. With CommunityBench, we conduct a comprehensive evaluation of various foundation models on CommunityBench, revealing that current LLMs exhibit limited capacity to model community-specific preferences. Furthermore, we investigate the potential of community-level alignment in facilitating individual modeling, providing a promising direction for scalable and pluralistic alignment.