Artificial intelligence

Fine-tuning

Fine-tuning a pre-trained model on your own data. It costs money and time, and is needed less often than it seems: more often the task is solved by a well-crafted prompt or document injection.

When it is really needed

When the response format matters: strict format, narrow style, specific terminology. Fine-tuning is poor at adding knowledge – RAG is used for knowledge.

What it is replaced with

First, try a detailed instruction and examples right in the prompt. It is cheaper, can be changed in a minute and does not require a labeled dataset.

We will check your site against the same points

We will see what holds the site back and tell you where to start. Free, within two business days.