Training Trains Internship
Student login
← AI Agents & RAG learning path
AI Agents & RAG · Module 1 of 10

LLMs as building blocks

Understand what LLMs can and cannot do before building on them.

📖 Key ideas

  1. LLMs predict the next token based on huge training data.
  2. They are strong at language, summarising, reasoning and code.
  3. They lack live knowledge and can hallucinate.
  4. You access them through APIs (Claude, GPT, Gemini) or open models (Llama, Mistral).
  5. Cost and latency depend on tokens in and out.
💡 Real-life example: An LLM is a brilliant intern who has read everything but needs the right documents and tools to do real work.
LLMLarge Language Model
TokenChunk of text the model reads/writes
HallucinationConfident but false output
APIProgrammatic access to a model