Skip to content

Glossary

What is fine-tuning?

Fine-tuning is extra training that adapts a model on chosen examples. It changes how the model behaves, which is not the same as feeding it current material.

By Marella AI · Updated 21 Aug 2026

01

Where this one gets misread

Fine-tuning is often proposed where the real requirement is current information. Training on documents changes how a model behaves; it does not reliably make it able to quote them, and it has to be redone as the material moves.

02

Questions to ask

Ask for a worked example on your own material, and the evidence needed to reproduce it.

  • What problem is fine-tuning solving that retrieval would not?
  • How is it re-run when the source material changes?
  • What does each run cost?
  • How is the result evaluated against the un-tuned model?
03

How Marella uses the term

We use “Fine-tuning” only where a product mechanism or an evaluation method backs it up, and we say when the behaviour depends on how a deployment is configured.

  • Backed by a product mechanism or an evaluation method
  • Deployment differences flagged

Caveats

  1. B1A definition is not a claim about how the product performs.
  2. B2Vendor implementations vary.
  3. B3Test the term against a representative workflow.

Related pages

Plan an evaluation

Test it on your own documents

Pick a real piece of work, agree what a good answer looks like, then go through the results together.

Book a demoExplore the platform