Glossary · Topic 01 · 19 terms
AI Concepts
Foundations of how generative AI works - and why those mechanics shape what it can read.
AI AgentsSoftware powered by LLMs that takes goals and executes multi-step tasks - including reading the web on a user's behalf.Attention MechanismThe transformer architecture's way of weighting which input tokens matter most.Chain of ThoughtAn LLM reasoning pattern where the model walks through intermediate steps.Context WindowThe maximum amount of text an LLM can consider at once.Diffusion ModelsA class of generative AI models - primarily for images.EmbeddingsNumeric vector representations of text that capture meaning.Few-Shot LearningShowing an LLM a small number of examples in the prompt to teach a new task.Fine-TuningTraining an existing model further on domain-specific data.Generative AIAI systems that produce new content rather than just classifying or retrieving existing content.HallucinationWhen a language model confidently states something false.Large Language Models (LLMs)Neural networks trained on massive text corpora that generate human-like text.Machine LearningThe broader field of algorithms that learn from data.Natural Language Processing (NLP)The branch of AI focused on understanding and generating human language.Prompt EngineeringThe craft of writing prompts that elicit useful responses from LLMs.Retrieval-Augmented Generation (RAG)The technique of grounding LLM answers in retrieved documents.TokenThe atomic unit an LLM processes - roughly a syllable.Transformer ModelsThe neural architecture underlying every modern LLM.Vector EmbeddingsNumeric representations of content that AI systems use to find semantic similarity.Zero-Shot LearningAn LLM's ability to handle a task it wasn't specifically trained for.
The full 76-term glossary.
Browse alphabetically, by topic, or jump straight to the term you need.