MODEL ADAPTATION • 225 / 397
Know when to prompt, retrieve, fine-tune or compress a model.

RLHF

RLHF means reinforcement learning from human feedback, using human preference signals to improve model behavior.

Think of it like

Think of RLHF as changing how a trained worker behaves or how efficiently that worker can be deployed, rather than giving it a live reference book.

Real life

Organizations adapt general models using techniques such as RLHF.

SRE lens

It is a model-training technique, not a production runtime feature.

Remember thisRLHF means reinforcement learning from human feedback, using human preference signals to improve model behavior.
AIForSREJump to a concept
Search is optional. The main journey is simply ↓ Next.