Training
Alignment
Making AI systems pursue the goals their developers and users intend, and behave helpfully, honestly and safely.
Alignment work covers training methods such as RLHF, written principles that guide behaviour, testing for harmful outputs, and research into how models represent goals.
Frequently asked questions
What does Alignment mean?
Making AI systems pursue the goals their developers and users intend, and behave helpfully, honestly and safely. Alignment work covers training methods such as RLHF, written principles that guide behaviour, testing for harmful outputs, and research into how models represent goals.