Glossary
Fine-Tuning
Continued training of a base model on task-specific data.
Definition
Fine-tuning is the practice of taking a pre-trained model and continuing training on a smaller, task-specific dataset to adapt it to a specific domain, style, or task. Modern fine-tuning typically uses parameter-efficient methods (LoRA, adapter tuning) rather than full-weight updates.
Context
In 2026, fine-tuning is usually the second tool to reach for — after prompt engineering and RAG — because it adds operational complexity (training pipelines, versioning, evaluations) without always outperforming a well-grounded base model with retrieval. Fine-tuning earns its keep for style adherence, structured-output reliability, and proprietary skills the base model doesn't have.
Related terms
Large Language Model (LLM)A neural network trained on massive text corpora to predict next tokens.Read Retrieval-Augmented Generation (RAG)Injecting retrieved documents into an LLM prompt to ground outputs.Read Prompt EngineeringDesigning inputs to LLMs that reliably produce desired outputs.Read AI Evaluation (Evals)The practice of systematically measuring AI system quality.Read