CORTEXA
← Browse

Tianyu Huai

1 paper indexed

arxivcs.AI2026-07-31

AgentHPOBench: A Benchmark For Evaluating LLM Agents as Sequential Hyperparameter Optimizers

Tianyu Huai, Tingshuo Fan, Xinchi Chen, Yining Zheng, Yuxin Wang, Shuang Chen, et al.

As LLMs evolve from code completion systems into autonomous scientific agents, evaluating their ability to conduct experiments has become increasingly important. Existing benchmarks typically focus on static code generation, paper replication, or final answer correctness, but do…

View free PDFSource page