Topics

Alignment

Methods for steering models toward preferred, safer, or more useful behavior.

InstructGPT: Why Bigger Models Still Needed Human Feedback

InstructGPT showed that human preference data and RLHF could make smaller models more helpful and aligned than much larger raw language models.

Alignment · Stanford University

DPO: The Alignment Trick That Removed the RL Loop

Direct Preference Optimization turns preference tuning into a simple classification-style objective, avoiding an explicit reward model and reinforcement learning loop.