2026
NetArena: Dynamically Generated LLM Benchmarks for Network Applications
ICLR 2026poster
As large language models (LLMs) expand into high-stakes domains like network system operations, evaluating their real-world reliability becomes increasingly critical. However, existing benchmarks risk contamination due to static design, show high statistical variance from limited dataset size, and f…