AI Glossary: Five Key Definitions and NN/G Resources

A French glossary brings together common concepts in artificial intelligence and provides operational definitions. It covers technical elements and practices while noting that the usage of terms may differ among providers and researchers. Links to in-depth articles and a study guide complete the collection.

The glossary offers an alphabetical sorting and a study guide

The entries are organized alphabetically within columns, and searches can be conducted using the browser command or by browsing through the entire article. This structure facilitates quick access to the desired terms.
Many entries link to NN/G articles that detail the subjects. An introductory study guide covers both how AI functions and how users perceive it.

Text processing relies on tokens and the context window

Tokens are the basic unit of natural language processing, corresponding to words or fragments of words. The context window refers to the amount of information that the model can integrate when generating a response.
Evaluations are used to measure the performance of models on specific tasks.

Autonomous agents and the risk of prompt injection

Agents are defined as autonomous systems capable of executing specific tasks using AI algorithms. Prompt injection involves inserting instructions or queries intended to influence the outputs generated by the model.
The definitions presented reflect the usage observed in AI products and UX work, with the caveat that providers and researchers may use these terms differently.