Reproduce “Designing Behavioral Profiles for Language Model Simulation of Out-of-Distribution Human Behavior” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Robin Na, Kehang Zhu, Logan Cross, Michiel A. Bakker
The authors test whether evidence-based behavioral profiles help an LLM simulator reproduce human outcomes under new experimental designs. They select documented features from earlier records and vary which features the simulator receives. In 250 new public-goods validation games, a one-feature profile scored 65.5 on normalized distributional alignment, versus −1.0 with no profile. This measures population-level outcome distributions, not individual behavior.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.