Post

Can AI predict household behavior? HouseholdBench tests 13 LLMs on 32 tasks. A fine-tuned open model rivals frontier AI and approaches statistical benchmarks, from Jin Huang, Diego Ferreras Garrucho, Yutong Xie, Walter M. Yuan, Qiaozhu Mei, Chen Lian, and Jonathon Hazell www.nber.org/papers/w35927