On the Limits of Tabular Hardness Metrics for Deep RL: A Study with the Pharos Benchmark

Open in new window