Reproducibility archive for Target-Scale Sensitivity in Adaptive Pessimism for Tabular Offline Reinforcement Learning: An EdNet-KT4 Study
This archive accompanies the manuscript "Target-Scale Sensitivity in Adaptive Pessimism for Tabular Offline Reinforcement Learning: An EdNet-KT4 Study." It contains the analysis code, frozen experimental protocol, primary clean-200k results, and sensitivity-analysis results needed to reproduce the reported tables and f...