What is a large language model?
Understand what an LLM learns, how it generates text, and why fluent answers can still be wrong.
Read lesson →THE LEARNING ATLAS / FREE & OPEN TO EVERYONE
AI, explained.
One useful idea at a time.
Understand large language models, ask better questions, and build applications you can evaluate. No account or API key needed to read.
6 lessons
Tokens, context and attention
See how text becomes model input and why the context window matters.
Read lesson →Write prompts you can evaluate
Turn a vague request into a repeatable task with evidence, boundaries and a review checklist.
Read lesson →RAG: answers grounded in documents
Learn how retrieval adds relevant sources and where a document-based assistant can still fail.
Read lesson →Open source, open weights and licenses
Understand what you can inspect, modify and redistribute before choosing code or a model.
Read lesson →Build a more reliable AI workflow
Handle private inputs, unsupported claims and tool actions with deliberate safeguards.
Read lesson →Learn by building.
Explore our MIT-licensed Evidence Search Lab and Prompt Evaluation Workbook. Small, readable examples with synthetic data and no paid dependencies.
Get the learning projects →Common questions
Do I need to code?
No. The foundations and practical-skills lessons work without code. The downloadable build exercises use Python and CSV files.
Is this a certification?
No. This is an independent educational resource from XYQEN, with practical examples and links to further reading.
Can I use an LLM as a source of truth?
No. Check factual claims against original sources. Model output can be incomplete, outdated or invented, even when it sounds confident.