List your company

Hyperparameter Optimization on autoresearch (val)

99.04Best Validation bpbKarpathy Agent (14 HPs)
-2.94510423.53179850.008776.485602Mar 25, 2026Apr 8, 2026Apr 22, 2026May 7, 2026May 21, 2026Jun 4, 2026Jun 19, 2026

Evaluation results

MethodLinks
Karpathy Agent (14 HPs)LLM variant=27B, Number of seeds=32026.0399.0428910.0350.2250.05729—
LLAMBO (Optuna)LLM variant=27B, Number of seeds=32026.0398.82629610.2511.2471.13458—
RandomNumber of seeds=32026.0398.73568560.2731.381.382514—
LLAMBO (Paper)LLM variant=27B, Number of seeds=32026.0398.62496480.2521.2571.218364—
Karpathy Agent (Code)LLM variant=27B, Number of seeds=32026.0398.1432412—————
CMA-ESNumber of seeds=32026.0397.85336160.1580.7940.58802—
SMACNumber of seeds=32026.0397.78431360.2381.1910.369224—
TPENumber of seeds=32026.0397.68317110.1960.9960.413518—
CentaurLLM variant=27B, Number of seeds=32026.0397.63334150.1150.550.341330—
Model Aautonomous experiments=123, seeds=32026.060.9822——————0.9799
Model Bautonomous experiments=123, seeds=32026.060.9793——————0.9758
Model Cautonomous experiments=123, seeds=32026.060.9781——————0.9766
Fugu-Ultraautonomous experiments=123, seeds=32026.060.9774——————0.9748