مسرد

RLHF

هذه التعريفات بالإنجليزية.

Reinforcement learning from human feedback: a training step that uses people’s preferences between outputs.