Reproduce “Large Language Models as Psychological Simulators: A Methodological Guide” 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

Zhicheng Lin

The authors distinguish LLMs as persona simulators from computational models whose internal processes can be examined and manipulated. They describe validating persona outputs for the research question and using probes and causal interventions to test model components. In cited results, GPT-3.5 showed near-zero response variation in six of 14 Many Labs 2 replications, limiting that finding to those studies.

Summary by alphaXiv.

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