Category: Model Safety

• Deep Dive: Reward Hacking in Claude Code Model RL Training
• Automated Alignment Researchers: Using large language models to scale scalable oversight