Supplementary Dataset for: "Exploration Through Introspection: A Self-Aware Reward Model"

Description Supplementary dataset for the paper "Exploration Through Introspection: A Self-Aware Reward Model" accepted at the AAAI-2026 Workshop on Theory of Mind for Artificial Intelligence. This dataset contains the pre-computed grid search results for the Stationary, Non-Stationary, and Lifetime environments presented in the associated study. Purpose: These files are provided to facilitate the immediate reproduction of the paper's figures, plots, and statistical tests. Using this dataset allows researchers to bypass the computationally intensive grid search and proceed directly to analysis. Usage Instructions: To use these results with the accompanying source code, place the .csv files into the corresponding experiment directory within the repository structure: experiments/batch_grid_search/[experiment_name]/数据集说明 本数据集为收录于AAAI-2026人工智能心智理论研讨会的论文《Exploration Through Introspection: A Self-Aware Reward Model》（《通过内省探索：一种自感知奖励模型》）的补充数据集。 本数据集包含本相关研究中提出的静态（Stationary）环境、非静态（Non-Stationary）环境与生命周期（Lifetime）环境的预计算网格搜索结果。 用途：提供这些文件旨在助力论文中各类插图、绘图结果与统计检验的快速复现。使用本数据集可使研究人员规避计算量极大的网格搜索步骤，直接开展分析工作。 使用说明：若需结合配套源代码使用本数据集的结果，请将所有.csv格式文件放置至仓库结构中对应的实验目录下：experiments/batch_grid_search/[experiment_name]/