Skip to content
anythingai.space
AI Foundationsintermediate15 min · 10 steps

Prompting and Fine-Tuning

Two ways to make a general LLM do your task: change what you send it (prompting, few-shot, chain of thought) or change its weights (full fine-tuning, LoRA, QLoRA).

vague askWHAT YOU WANTEDnegativeWHAT YOU GOTBattery lifedepends onmany factors…UserLLMpre-trainedOutputchatty, no label
requesterrorresponsesignaldataheavy work
1 / 10
ProblemStep 1 of 10

A general model gives a general answer

You want a one-word label for a product review. You send “What about this review: the battery died in a day” and get back three chatty paragraphs about batteries.

The pre-trained model can do the task — it just doesn't know which task, what format or what standard you want.

Lesson 22 of 34 in AI Foundations