S-Eval: Automatic and Adaptive Test Generation for Benchmarking Safety Evaluation of Large Language Models

Open in new window