Artificial Intelligence
Fine-tuning
Plain English
Training a glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term to be better at a specific job.
Definition
guideFine-tuningWhat fine-tuning does to an AI model, when it is worth doing, and what product and design teams need to know before commissioning it.Open guide is the glossaryProcessA process is a defined sequence of steps used to achieve a specific outcome.Open glossary term of further training a pre-trained glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term on specific data to improve performance for a particular task.
In practice
Used to adapt general glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term to specific domains, such as legal, medical, or e-commerce content.
In context
It is reached for too early. Most problems attributed to the glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term turn out to be retrieval or prompting problems, and guideFine-tuningWhat fine-tuning does to an AI model, when it is worth doing, and what product and design teams need to know before commissioning it.Open guide bakes in a fix for something that was not broken.
The reality
guideFine-tuningWhat fine-tuning does to an AI model, when it is worth doing, and what product and design teams need to know before commissioning it.Open guide improves relevance but requires quality glossaryDataData is raw, uninterpreted information collected and stored so it can be analysed, processed, or used to inform decisions. On its own it carries no meaning; context and interpretation are what make it useful.Open glossary term and careful management to avoid overfitting.
Compared with
Fine-tuning vs RAG
guideFine-tuningWhat fine-tuning does to an AI model, when it is worth doing, and what product and design teams need to know before commissioning it.Open guide changes how a glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term behaves and is the right tool for style, format, and task glossaryConsistencyConsistency is the use of uniform patterns, behaviours, and visual elements across a product to create familiarity and predictability. It helps users learn once and apply that knowledge throughout the experience.Open glossary term. RAG changes what it knows at request time and is the right tool for facts that change or must be cited.
FAQ
Common questions
A few practical answers to the questions that usually come up around this term.
What is fine-tuning in AI?
It is training a glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term further on specific glossaryDataData is raw, uninterpreted information collected and stored so it can be analysed, processed, or used to inform decisions. On its own it carries no meaning; context and interpretation are what make it useful.Open glossary term to improve glossaryPerformancePerformance refers to how quickly and efficiently a system responds to user actions and processes tasks.Open glossary term.
Why is fine-tuning used?
To make glossaryModelA model is a system or representation used to process data and generate outputs, often trained to perform specific tasks.Open glossary term more relevant to a specific task or domain.
What are risks of fine-tuning?
Overfitting and reduced general glossaryPerformancePerformance refers to how quickly and efficiently a system responds to user actions and processes tasks.Open glossary term.
What data is needed for fine-tuning?
High-quality, relevant, and well-labelled glossaryDataData is raw, uninterpreted information collected and stored so it can be analysed, processed, or used to inform decisions. On its own it carries no meaning; context and interpretation are what make it useful.Open glossary term.
Related Services
Related Guides
Related Terms
