Safeguard LLMs from acquiring unsafe knowledge and being exploited for malicious purposes. View project →
