Replication Data for: Strategizing with AI: Insights from a Beauty Contest Experiment

Dataset on the Beauty Contest experiments run with several LLM's: GPT-4o, GPT-4o-mini, Claude 3.5 Sonnet, Llama 3.1 8b, Gemini 1.5 Flash. Each of the files contains the following columns: model (name of the model which was used to run the experiment), paper (name of the paper which contains the replicated experiment), n (number of players), p (a fraction used with a function), fun (average, maximum, median), opponents (description of the opponents in the game), answer_text (a text result of the model's work), answer_number (numeric answer of the model).