Concept Taxonomy

This taxonomy organizes the 305 concepts in the Prompt Engineering learning graph into 14 categories.

Categories

Foundation Concepts (FOUND)

Core AI and machine learning concepts that provide the prerequisite knowledge for understanding prompt engineering. Includes neural networks, transformers, tokens, and language models.

Prompt Fundamentals (PFUND)

Essential concepts for understanding what prompts are and how to construct them effectively. Covers structure, clarity, iteration, and response evaluation.

Prompt Types (PTYPE)

Classification of different prompt styles including directive, interrogative, open-ended, closed-ended, and conversational prompts.

Prompt Techniques (PTECH)

Advanced prompting strategies and patterns including zero-shot, few-shot, chain-of-thought, persona prompting, meta-prompting, and prompt chaining.

Model Parameters (PARAM)

Configurable settings that control model behavior including temperature, top-p sampling, frequency penalties, and reproducibility settings.

Output Format Control (OUTFMT)

Techniques for controlling the format and structure of AI outputs including markdown, JSON, tables, reports, and structured data generation.

Context and Memory (CTXMEM)

Concepts related to managing context windows, conversation history, memory systems, information extraction, and context injection strategies.

Retrieval-Augmented Generation (RAG)

Techniques for grounding AI responses in external data sources including knowledge bases, vector databases, embeddings, and source citation.

Multimodal Prompting (MULTI)

Working with non-text inputs and outputs including images, documents, audio, video, charts, and data visualization prompts.

Agentic AI (AGENT)

Concepts related to AI agents that can plan, use tools, execute workflows autonomously, collaborate with other agents, and manage skills.

Prompt Security (SECUR)

Security considerations for prompt engineering including injection attacks, jailbreaking, guardrails, red teaming, and defense strategies.

Ethics and Responsible AI (ETHIC)

Ethical considerations including bias, fairness, transparency, privacy, intellectual property, and responsible AI use.

Business Applications (BIZAP)

Practical business use cases for prompt engineering including customer service, content creation, data analysis, and decision support.

Educational Applications (EDUAP)

Applications in education including personalized tutoring, quiz generation, curriculum development, and accessibility.

Evaluation and Optimization (EVAL)

Methods for evaluating and optimizing prompt quality including A/B testing, benchmarking, cost optimization, and model comparison.

Usage Limits and Token Economics (USAGE)

Platform usage constraints and cost management including rate limiting, token budgets, five-hour windows, pricing models, and cost-effective prompting.

Capstone Projects (CAP)

End-to-end project ideas that integrate multiple prompt engineering concepts into real-world applications.