AI RESEARCH

Gate AI: LLM Security Benchmark Evaluation Methodology and Results

arXiv CS.LG

ArXi:2606.02959v1 Announce Type: new Published evaluations of prompt-injection and jailbreak detectors for Large Language Models often suffer from two systematic weaknesses: per-dataset threshold tuning and undisclosed operating points. We describe an evaluation harness that addresses both. The detector under evaluation is scored across 16 public benchmarks (12,111 samples) using 5-fold cross-validation.