Signal2026-08-03
arXiv

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

Part of

Adversarial Evolution Of Code LLMs

Open primary sourceRead on Substack