Fine-Tuned Model
A pretrained model whose weights have been updated on a specific dataset for a targeted task.
Full Definition
A fine-tuned model starts from a pretrained foundation model and undergoes additional training on a smaller, task-specific dataset. This adapts the model's behaviour — its tone, domain knowledge, output format, and task focus — to a particular application. Fine-tuning is more powerful than prompting for consistent style or highly specialised knowledge, but requires labelled training data and compute. Techniques range from full fine-tuning (all weights updated) to parameter-efficient methods like LoRA that update only a small adapter. Fine-tuned models can outperform much larger base models on narrow tasks because they trade breadth for depth.
Examples
Fine-tuning GPT-3.5 on 5,000 customer support transcripts so it reliably matches the company's brand voice.
Fine-tuning a code model on a proprietary internal codebase so it can autocomplete company-specific API calls.
Apply this in your prompts
PromptITIN automatically uses techniques like Fine-Tuned Model to build better prompts for you.
Related Terms
Fine-Tuning
Continuing training of a pretrained model on a smaller, task-specific dataset to…
View →LoRA (Low-Rank Adaptation)
A parameter-efficient fine-tuning method that updates only small low-rank matric…
View →Instruction-Tuned Model
A model fine-tuned on instruction-response pairs to follow natural-language dire…
View →