Reproduce “Belief Coevolution in a Social Network of Generalist and Specialist Large Language Models: A Controlled Study” 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
Germans Savcisens, Samantha Dies, Courtney Maynard, Tina Eliassi-Rad
CoevolveSim tests how prompted social roles and specialist language models shape belief change among interacting agents. Across 48 Llama-3-family agents over 10 rounds, roles changed individual revision, while adding specialists was associated with a larger shift in population consensus; the role contrast was not statistically significant. Surrogate models reproduced group consensus best using population belief composition, within this simulation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.