Glossary · Generative AI and LLMs
What is SFT? Definition and meaning
Trains a model on curated prompts and ideal responses so it follows instructions in the desired style.
Definition
SFT: Trains a model on curated prompts and ideal responses so it follows instructions in the desired style.
In short: Supervised fine-tuning on instruction and answer pairs.
When it is used
The first stage of turning a base model into an assistant or a task-specific model.
Where it fits
SFT belongs to Fine-tuning in the Generative AI and LLMs topic. These are the other terms in the same group:
- Fine-tuning: Further training of a pretrained model on new data so it performs a specific task, domain or style better.
- Full fine-tuning: Continues training with all model parameters trainable, rather than a small add-on.
- LoRA: Freezes the model and learns small low-rank update matrices for selected layers, so only a small fraction of parameters change.
- QLoRA: Combines 4-bit quantization of the frozen base with LoRA adapters, which cuts memory enough to fine-tune large models on one GPU.
- PEFT: A library that implements LoRA, prefix tuning, prompt tuning and related methods on top of Hugging Face models.
- DPO: Optimizes a model directly on preferred versus rejected answers, without training a separate reward model or running reinforcement learning.
- RLHF: Trains a reward model from human preference comparisons and then optimizes the language model against it with reinforcement learning.
Learn more
The full guide to SFT explains how it works, when to use it and how it compares with the alternatives, with examples. The glossary lists every term on the site from A to Z.
Previous term: SentencePiece · Next term: SGLang