Glossary
Fine-tuning
Further training a model on your own examples to change its behaviour or specialise it.
Useful for consistent style, narrow tasks and reducing prompt length. Compared with RAG it is slower to update and harder to audit, and does not reliably teach new facts.
Related terms
DistillationGPUParametersRLHFSynthetic data
Get the brief in your inbox
One email every morning: the stories that mattered in AI, each with a one-line "why it matters" for your role. Free, no ads, unsubscribe any time.
Published daily by AI Briefing. We never share your address.