Generative AI and Large Language Models Vocabulary

This set of vocabulary flashcards defines the core technical terms and frameworks associated with Generative AI and Large Language Models as presented in the lecture.

Generative AI (GenAI)

A type of Artificial Intelligence that leverages AI to generate content or data such as text, images, audio, video, 3D models, and code, typically created in response to prompts.

Parameters

Numerical values that define a model's behavior and are adjusted autonomously during training to optimize the ability to generate relevant and coherent text; large language models often utilize hundreds of billions of these values.

Tokens

The basic units of text or code used by language models to process or generate language, with a rule of thumb being that $100$ tokens corresponds to approximately $75$ words.

Context window

The constraint on the length of text a large language model can process and respond to in a single instance, which for ChatGPT-4 is capped at $128,000$ tokens.

CIDI Framework

A structured prompt engineering framework used to interact with language models, consisting of four specific components: Context, Instructions, Details, and Input.