RLHF
Reinforcement learning from human feedback
A family of methods using human preference information to guide behavior through reinforcement learning. Human feedback is often used to train a reward model first.
By Mega AI Hub Editorial · Training & data