Glossary term
Glossary term
Agentic Systems
The discipline of designing, structuring, and optimising prompts to improve model outputs.
The art of creating prompts that elicit the desired responses from a large language model. Humans perform prompt engineering. Writing well-structured prompts is an essential part of ensuring useful responses from a large language model. Prompt engineering depends on many factors, including:
The dataset used to pre-train and possibly fine-tune the large language model.
The temperature and other decoding parameters that the model uses to generate responses.
Prompt design is a synonym for prompt engineering.
See Introduction to prompt design for more details on writing helpful prompts.
Anthropic publishes prompt-engineering best practices showing that adding 'Think step by step' to a maths problem prompt improves GPT-4 accuracy on GSM8K from 78% to 93%, demonstrating quantifiable gains.
Salesforce Einstein uses prompt templates engineered for CRM tasks - the 'Opportunity Summary' prompt is structured with role, data fields, output format, and tone constraints to ensure consistent sales-brief quality.
GitHub Copilot's prompt-engineering team iterates prompt templates for code completion using A/B testing against acceptance rate and correctness metrics, publishing internal prompt-versioning guidelines.
Created for this library
A SaaS team's prompt engineering practice includes structured prompt templates, regression tests, and versioning for production use.
An enterprise legal team invests in prompt engineering training so each new product feature meets internal quality standards.
A consulting firm offers prompt engineering as a managed service to clients building their first LLM-powered features.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License