Glossary term
Glossary term
Agentic Systems
Prompt chaining links multiple prompts together using the output of one as the input to the next to guide the model through multi-step reasoning or tasks. It helps break down complex problems into manageable steps for more reliable outcomes.
Using the output of one prompt as the input to another prompt. Least-to-most prompting is a popular form of prompt chaining.
LangChain's LCEL pipes (|) implement prompt chaining as a first-class composition pattern.
DSPy from Stanford NLP optimises prompt chains as compiled programmes.
Anthropic's prompt engineering guide includes prompt chaining as a recommended pattern for complex tasks.
Created for this library
A research assistant tool uses prompt chaining to first extract entities and then answer questions about each entity in sequence.
A legal-tech vendor uses prompt chaining to first identify clauses, then classify each clause, then summarize the contract.
A coding assistant uses prompt chaining to plan changes first and then apply each change in a separate step the user can review.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License