SG-Bench: Evaluating LLM Safety Generalization Across Diverse Tasks and Prompt Types

Open in new window

The document is on a web server that cannot be reached. Please try again later.