Data Invoked by Large Language Models and Human Decision-Making Performance Data

The datasets generated and analyzed in this article are organized by study:Study 1: Data include GPT-3.5 and GPT-4 decision outcomes in single vs. repeated gamble tasks, program scripts for API calls, corresponding problem sets, R code for chi-square tests and logistic regression, and visualization scripts.Study 2: Data cover the standardized procedure for generating decision strategies with LLMs, experimental data from human participants across medical, financial, content creation, and e-commerce contexts, as well as associated R analysis code.Study 3: Data include decision outcomes of human participants exposed to LLM-generated intervention texts and the corresponding analysis code.All relevant materials (stimuli, scripts, and analysis code) are available upon reasonable request from the authors.