Do Generated Architectural Reports on Reinforcement Learning Practices Improve Pipeline Comprehension? A Controlled Experiment

This repository contains all files and experimental data used in the study, including the source code, experiment configurations, generated heuristics, prompts, datasets, raw results, and scripts used to reproduce the tables and figures reported in the paper. It also includes the parameter settings and random seeds required to repeat the experiments and verify the reported results.