LiveCultureBench: a Multi-Agent, Multi-Cultural Benchmark for Large Language Models in Dynamic Social Simulations

TL;DR

LiveCultureBench is a comprehensive benchmark that evaluates large language models as autonomous agents within a simulated multicultural town, assessing both task success and cultural norm adherence to improve social robustness and evaluation reliability.

Contribution

We introduce LiveCultureBench, a novel multi-cultural, dynamic simulation environment for evaluating LLMs on task performance and cultural norm compliance.

Findings

LLMs show varying robustness across different cultural profiles.

Trade-offs exist between task effectiveness and norm sensitivity.

Automated verifier judgments are reliable in certain contexts, reducing human oversight.

Abstract

Large language models (LLMs) are increasingly deployed as autonomous agents, yet evaluations focus primarily on task success rather than cultural appropriateness or evaluator reliability. We introduce LiveCultureBench, a multi-cultural, dynamic benchmark that embeds LLMs as agents in a simulated town and evaluates them on both task completion and adherence to socio-cultural norms. The simulation models a small city as a location graph with synthetic residents having diverse demographic and cultural profiles. Each episode assigns one resident a daily goal while others provide social context. An LLM-based verifier generates structured judgments on norm violations and task progress, which we aggregate into metrics capturing task-norm trade-offs and verifier uncertainty. Using LiveCultureBench across models and cultural profiles, we study (i) cross-cultural robustness of LLM agents, (ii) how they balance effectiveness against norm sensitivity, and (iii) when LLM-as-a-judge evaluation is reliable for automated benchmarking versus when human oversight is needed.