AI glossary
Short explanations of the concepts that come up most on Cam Nang AI. Each one links to the full article.
- AI agent
- An autonomous software system that uses a large language model as its reasoning engine to perceive its environment, make decisions, and use tools to achieve a specific goal. Read: What Is an AI Agent? Components, Loop, and Types
- Artificial intelligence (AI)
- The set of technologies that let machines simulate human learning, reasoning and autonomous problem-solving, using data-driven algorithms instead of static, hand-written rules. Read: Artificial Intelligence: History, Machine Learning, and AI Agents
- Chain of Thought
- A prompt engineering technique that guides large language models to output intermediate logical reasoning steps before producing a final answer. Read: What is Chain of Thought prompting?
- Deep learning
- A subset of machine learning that uses multilayered artificial neural networks to process data and automate decision-making. Read: What is Deep Learning?
- Embedding
- A dense, lower-dimensional vector representation of data that enables computers to process semantic relationships mathematically. Read: What Is an LLM Embedding? How AI Turns Text Into Vectors
- Fine-tuning
- The process of retraining a pre-trained large language model on a focused, domain-specific dataset to adapt its behavior or lock in a specialized style. Read: What is fine-tuning? When to Fine-Tune an LLM and When Not To
- Generative AI
- The family of deep-learning models that create new content, such as text, images, audio and code, by learning the probability distribution of data rather than merely classifying it. Read: What is generative AI? Models, mechanics, and real limits
- Guardrails
- Programmable, multi-layered security and safety controls positioned between clients, system prompts, model inference, tool execution and downstream applications. Read: What Are AI Guardrails? Content Safety and Circuit Breakers
- Large language model (LLM)
- An over-parameterized neural network trained to predict the next token in a sequence from vast amounts of text. Read: What Are Large Language Models (LLMs)? How They Work
- LLM hallucination
- What happens when a model prioritizes statistical plausibility and linguistic fluency over factual verification. Read: Why Do LLMs Lie? The Science and Architecture of AI Hallucinations
- Machine learning
- A branch of artificial intelligence that lets computers learn patterns from data to make decisions without explicit programming. Read: What Is Machine Learning? Types, Algorithms, and Optimization
- MCP (Model Context Protocol)
- An open-source standard introduced by Anthropic that enables a plug-and-play connection between AI models and external data sources. Read: Everything you need to know about MCP
- Multi-agent system
- An architecture where independent AI agents delegate tasks or share state to reach a goal. Read: What Is a Multi-Agent System? How AI Agents Share Context
- RAG (Retrieval-Augmented Generation)
- An architecture that optimizes large language model performance by connecting models to external, authoritative knowledge bases. Read: What is RAG? Optimizing LLMs for Enterprise AI Systems
- Sycophancy
- What happens when an AI system alters its outputs to align with user preferences, even when those preferences are demonstrably incorrect. Read: Sycophancy in LLMs: Why AI Agrees With You Even When You're Wrong
- System prompt
- The primary control surface for specifying the behavior of a large language model: instructions designed to persist throughout the context window and supersede conflicting user messages. Read: System Prompt Fundamentals: How LLMs Get Their Rules
- Temperature
- A hyperparameter applied during the sampling stage of inference to control the variance of token selection. Read: What Is LLM Temperature, and What Should You Set It To?
- Token
- The smallest unit of data that a language model reads in and writes out. Read: AI Tokens and Context Windows: A Practical Guide
- Vector database
- A specialized storage and retrieval system designed to manage data as high-dimensional arrays called embeddings. Read: What Is a Vector Database? How It Works and When to Use One