Reproduce “Building and Measuring Trust between Large Language Models” 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
Maarten Buyl, Yousra Fettach, Guillaume Bied, Tijl De Bie
The authors test whether trust-building contexts change how large language model agents report trust, invest, and respond to persuasion. They compare generated small talk, scripted relationship context, and trusting system prompts, measuring questionnaire scores, investment choices, and answer flips to fixed challenges. Faith gains and medium-stakes investment gains correlate negatively (Spearman rho = −0.54), showing measures can diverge, not proving real-world exploitation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.