Glossary

Large language model (LLM): A language model that's capable of general-purpose tasks such as language generation, reasoning, and classification.

Retrieval-augmented generation (RAG): A method for retrieving domain knowledge that's relevant to a user query from a knowledge store and inserting it into a language model prompt. RAG improves the factual accuracy of model generations because the prompt includes domain knowledge. For more information, see What Is RAG? on the AWS website.

Prompt engineering: The practice of crafting and optimizing input prompts by selecting appropriate words, phrases, sentences, punctuation, and separator characters to effectively use LLMs for a wide variety of applications. For more information, see What is prompt engineering? in the Amazon Bedrock documentation and the Prompt Engineering Guide by DAIR.AI.

Prompt injection attack: Manipulating prompts to influence LLM outputs, with the objective of introducing biases or harmful outcomes. For more information, see Prompt Injection in the Prompt Engineering Guide.