AutoTuneBench: Trustworthy Measurement for Agent Auto-Tuning of LLM Serving Engines
A benchmark for evaluating agent-driven auto-tuning of LLM serving engines, aiming to improve reliability and reproducibility of such systems.
A benchmark for evaluating agent-driven auto-tuning of LLM serving engines, aiming to improve reliability and reproducibility of such systems.
Auto-scored: Provides a measurement framework that partially mitigates risks from unreliable auto-tuning, with only minor framing-level risk.