Professional Experience

Conducting research under Dr. Long Cheng investigating whether Large Language Model agents can realistically simulate human behavior and how their behavioral fidelity can be systematically evaluated.

Research & Contributions

• Conducting a comprehensive literature review of LLM-based human behavior simulation, autonomous agents, multi-agent systems, and AI-generated synthetic populations.
• Investigating AgentSociety, YuLan-OneSim, Generative Agents, and MiroFish for large-scale social and behavioral simulation.
• Developing a taxonomy of human behavior for evaluating LLM agents across decision-making, social interaction, cooperation, trust, and other behavioral dimensions.
• Designing and reproducing behavioral experiments such as Trust Games and Dictator Games to compare LLM-agent decisions against established human behavioral data.
• Exploring evaluation methods and benchmarks for quantifying behavioral realism, consistency, and human-agent similarity.
• Building toward a research survey on the strengths and limitations of LLM agents for simulating human behavior, including their potential role in early-stage studies.