Reproduce “Validating Generative Agent-Based Models of Social Norm Enforcement: From Replication to Novel Predictions” 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

Logan Cross, Nick Haber, Daniel L.K. Yamins

The authors test whether GPT-4o agents reproduce human findings in social dilemmas before using them to predict new conditions. They compare prompts for personas, theory of mind, and strategic reflection against published human effects. Agents with personas and theory of mind reproduce punishment effects; strategic reflection improves public-goods results, but gossip-only trends remain unreproduced. These predictions require human testing.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.