Anchored Bipolicy Self-Play: Advancing AI Safety Training
Discover how Anchored Bipolicy Self-Play improves AI safety by breaking self-consistency and boosting adversarial training efficiency.
AI Alignment and Jurisprudence: Bridging Law and Tech
Explore how AI alignment intersects with legal theory to shape future decision-making in law and autonomous systems.
Political Plasticity in Large Language Models: Ideology Shift
Explore how large language models adapt politically through user prompts, revealing ideological shifts and biases across languages and model sizes.
AI-Induced Delusions: Game Theory for Safer Knowledge
Explore how game theoretic interventions can prevent AI-induced delusions and promote epistemic safety in conversational AI systems.
Thinking Machines Develops AI That Listens While Talking
Discover how Thinking Machines is creating AI that listens and responds in real-time for natural, fluid conversations and enhanced user engagement.