SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models Paper • 2402.05044 • Published Feb 7 • 2