Skip to content
Module
Module 3 of 5Lesson 1 of 2~15 min

What the model knows, since when, and what tools add

An AI assistant does not know everything, nor necessarily anything recent, which changes how far you can trust a factual answer. This lesson helps you work out, for a given question, whether the model can answer alone, needs web search or a document you provide, and how to check the result.

Take this lesson: your first course is freeSee the full syllabus

Lesson objective

By the end of this lesson, you will be able to say, for a given question, whether the model can answer from memory, whether you need to turn on web search or whether you must provide a document, and to choose the matching check.

Topics covered

  • knowledge cutoff
  • AI web search
  • what an LLM knows
  • giving AI a document

Where it fits

What the model knows, and what it sees

What does the model know on its own, what do tools add, and what does it retain from our conversation?

Lessons in this module

  1. What the model knows, since when, and what tools add (this lesson)
  2. The context window, a powerful but uneven working memory

What you will learn in this module

This lesson is part of the course Understand what LLMs do well, and where they fail

  • Locate what the model knows (training data, knowledge cutoff, uneven coverage) and what web search, provided documents and connectors change.
  • Explain the context window as working memory (limit, degradation over long contexts, no memory across conversations without a dedicated feature) and derive working rules from it.

Related courses