ARLBench: Reinforcement Learning Hyperparameter Optimization Benchmark

ARLBench is a performance dataset for hyperparameter optimization in reinforcement learning, compiled by autorl-org. The data originates from several thousand experimental runs conducted to identify meaningful HPO test settings. The dataset was last updated on March 19, 2026.

Use Cases

• Benchmarking hyperparameter optimization algorithms based on performance data from RL runs.
• Analyzing the relationship between hyperparameter configurations and RL agent performance.
• Developing meta-learning or surrogate models for RL hyperparameter tuning based on the compiled results.

Strengths

• Data is derived from several thousand experimental runs, as stated in the description.
• Dataset is explicitly designed for benchmarking, suggesting a structured evaluation purpose.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset scale are unknown, which may limit suitability assessment.

Provenance

Source

autorl-org

Collection Method

Compiled from several thousand experimental runs for hyperparameter optimization in reinforcement learning.

Freshness

Last updated 2026-03-19 21:21:30; freshness should be verified.