• Publication Evaluating Frontier Models for Dangerous Capabilities
• Publication Holistic Safety and Responsibility Evaluations of Advanced AI Models
• Publication Long-form factuality in large language models
• Responsibility & Safety Taking a responsible path to AGI
• Responsibility & Safety Updating the Frontier Safety Framework
• Responsibility & Safety FACTS Grounding: A new benchmark for evaluating the factuality of large language models
• Technology Veo 1
• Technology Project Mariner
• Project Astra Tool use
• Research NeurIPS 2024