π¨ ALERT: A Comprehensive Benchmark for Assessing Large Language Models' Safety through Red Teaming Jun 25, 2024 β’ 5