Tag: LLM evaluation
Browse our exclusive articles!
CritBench: Evaluating LLM Cybersecurity in IEC 61850 Substations
CritBench framework assesses cybersecurity of large language models in IEC 61850 digital substations, addressing OT-specific challenges and protocols.
Evaluating LLM Patch Quality Beyond Pass Rates
Explore how design constraint compliance improves LLM-based issue resolution beyond traditional pass rate metrics.
CAKE Benchmark: Evaluating LLMs on Cloud Architecture Knowledge
Discover how the CAKE benchmark assesses large language models' understanding of cloud-native architecture through expert-validated questions and dual-form...
LLM Evaluation via Tensor Completion: Low-Rank & Efficiency
Discover a novel tensor completion method for LLM evaluation using low-rank structures and semiparametric efficiency to improve accuracy and reliability.
Empirical Audit of Instructed Code-Editing Benchmarks
Discover key insights from an empirical audit of code-editing benchmarks to improve LLM evaluation and real-world coding assistant performance.