Discovering maximally consistent distribution of causal tournaments with Large Language Models