EASE: Practical and Efficient Safety Alignment for Small Language Models

Open in new window