RUNPORT II — Inter-Model Analysis of Epistemic Decision Architectures in Large Language Models (Dataset)

RUNPORT II is a dataset examining structural differences in how large language models construct probability assessments and decision reasoning under identical prompt conditions. The dataset contains the full experimental prompt, seven original model runs, and a structured coding matrix (P-W-S scheme) capturing problem selection, probability justification patterns, and self-modeling markers. It enables reproducible analysis of epistemic reasoning structures in language model outputs.