← All terms
AI
Fine-tuning
Training a base LLM on your specific data for better domain performance.
Useful when you have 1,000+ high-quality examples and need consistent format or specialized knowledge. Often outperformed in 2026 by RAG plus better prompting at much lower cost.