Back to Guide

Glossary of Terms

Essential terminology and concepts in prompt engineering

Find a Term

A

3 terms

Alignment

Ensuring AI systems behave in accordance with human values and intentions.

Ethics & Safety

Attention

A mechanism that allows models to focus on relevant parts of input when generating output.

Technical

Autoregressive

Models that generate text one token at a time, using previous tokens to predict the next.

Technical

B

2 terms

Bias

Systematic prejudice in AI outputs that reflects unfair assumptions or stereotypes.

Ethics & Safety

Batch Processing

Processing multiple prompts simultaneously for improved efficiency.

Technical

C

4 terms

Chain-of-Thought

A prompting technique that encourages step-by-step reasoning before providing an answer.

Technique

Context

Information provided to the AI to guide its understanding and response generation.

Core Concept

Context Window

The maximum amount of text a model can process in a single interaction.

Technical

COT (Chain-of-Thought)

Abbreviation for Chain-of-Thought prompting technique.

Technique

D

2 terms

Delimiter

Special characters or markers used to separate different parts of a prompt.

Technique

Directive

A clear instruction that tells the AI what to do or how to behave.

Core Concept

E

3 terms

Example

A demonstration of desired output format or style included in the prompt.

Technique

Ethics

Moral principles and considerations in AI development and deployment.

Ethics & Safety

Execution

The process of running a prompt through an AI model to generate output.

Technical

F

2 terms

Few-Shot

A prompting technique that provides several examples to guide the AI's response.

Technique

Framework

A structured approach or template for organizing prompts effectively.

Core Concept

G

2 terms

Generation

The process of creating new content based on the input prompt.

Technical

Grounding

Ensuring AI responses are based on factual information rather than fabrication.

Technique

H

2 terms

Hallucination

When AI generates false or misleading information that appears plausible.

Ethics & Safety

Human-in-the-Loop

AI systems that require human oversight and intervention for critical decisions.

Technique

I

3 terms

Instruction

A clear command that tells the AI what action to perform.

Core Concept

Iteration

The process of refining and improving prompts through multiple attempts.

Technique

Injection

Attempts to manipulate AI behavior through malicious prompt engineering.

Ethics & Safety

J

1 term

Jailbreak

Attempts to bypass AI safety measures and content filters.

Ethics & Safety

K

1 term

Knowledge

Information and understanding that AI models have been trained on.

Core Concept

L

2 terms

Least-to-Most

A prompting strategy that starts simple and gradually increases complexity.

Technique

Large Language Model (LLM)

AI models trained on vast amounts of text data to understand and generate language.

Technical

M

2 terms

Metacognition

AI's ability to think about its own thinking process.

Technique

Multi-Perspective

Approaching problems from multiple viewpoints to generate comprehensive solutions.

Technique

N

1 term

Negative Prompting

Specifying what you don't want in the AI's response.

Technique

O

2 terms

One-Shot

A prompting technique that provides a single example to guide the AI.

Technique

Output Format

The structure and style in which you want the AI to present its response.

Technique

P

4 terms

Prompt

The input text that instructs an AI model on what to do or generate.

Core Concept

Prompt Engineering

The practice of designing and optimizing prompts to achieve desired AI outputs.

Core Concept

Prompt Injection

Security vulnerability where malicious input overrides intended prompt instructions.

Ethics & Safety

Parameter

Configurable settings that control AI model behavior (temperature, max_tokens, etc.).

Technical

Q

1 term

Quality

The standard and usefulness of AI-generated outputs.

Core Concept

R

3 terms

ReAct

A framework combining reasoning and action for complex problem-solving.

Framework

Role

Defining the AI's persona or expertise area in the prompt.

Technique

Reasoning

The logical process of thinking through problems step by step.

Technique

S

4 terms

System Prompt

Instructions that define the AI's behavior and capabilities.

Core Concept

Specificity

Being precise and detailed in prompt instructions.

Core Concept

Self-Critique

AI evaluating and improving its own outputs.

Technique

Safety

Ensuring AI systems operate without causing harm.

Ethics & Safety

T

3 terms

Temperature

A parameter controlling randomness in AI responses (0 = focused, 1 = creative).

Technical

Tree-of-Thoughts

Exploring multiple reasoning paths to find the best solution.

Technique

Token

The basic unit of text that AI models process (words, parts of words, or characters).

Technical

U

1 term

User Prompt

The specific request or question from the user to the AI.

Core Concept

V

1 term

Validation

Checking the accuracy and quality of AI-generated outputs.

Technique

W

1 term

Workflow

The systematic process of designing and executing prompt engineering tasks.

Core Concept

X

1 term

Explainability

The ability to understand how and why AI systems make decisions.

Core Concept

Y

1 term

Yield

The output or result produced by an AI model from a given prompt.

Core Concept

Z

1 term

Zero-Shot

A prompting technique that gives no examples, relying on the AI's pre-trained knowledge.

Technique

📚 Complete Guide Navigation

Access all sections of the comprehensive prompt engineering guide: