Category: Weak-to-Strong Supervision

• Automated Alignment Researchers: Using large language models to scale scalable oversight