Results for "provable safety"

Stepwise reasoning patterns that can improve multi-step tasks; often handled implicitly or summarized for safety/privacy.

Tradeoff between safety and performance.

Accelerating safety relative to capabilities.

Automated detection/prevention of disallowed outputs (toxicity, self-harm, illegal instruction, etc.).

Hard constraints preventing unsafe actions.

Ensuring robots do not harm humans.

Systems where failure causes physical harm.