HugAgent: A Human Simulation Benchmark for Individual-Level Reasoning
Chance Jiajie Li, Zhenze Mo, Yuhan Tang, Ao Qu, Jiayi Wu, Kaiya Ivy Zhao, Yulu Gan, Jie Fan, Jiangbo Yu, Hang Jiang, Paul Pu Liang, Jinhua Zhao, Luis Alberto Alonso Pastor, Kent Larson
October, 2026
Abstract
HugAgent evaluates whether a language model can infer an individual’s beliefs from a think-aloud interview and predict how those beliefs change in a new scenario. It combines questionnaires and interviews in three policy domains. Models recover participants’ stated beliefs more reliably than they predict belief updates after an intervention.
Publication
Empirical Methods in Natural Language Processing (EMNLP), 2026 (Oral)