Retrieval Augmented Generation
Retrieval Augmented Generation (RAG) is a technique that grounds a model's responses in retrieved external data rather than relying only on its training. AIP-C01 covers RAG with vector stores and knowledge bases.
All AIP-C01 Terms
Amazon Bedrock
Amazon Bedrock is a fully managed service that provides access to foundation models from multiple providers through a single API.
Foundation model
A foundation model (FM) is a large model pre-trained on broad data that can be adapted to many tasks.
Retrieval Augmented Generation
Retrieval Augmented Generation (RAG) is a technique that grounds a model's responses in retrieved external data rather than relying only on its training.
Vector store
A vector store is a database that indexes embeddings so semantically similar content can be retrieved by nearest-neighbor search.
Embedding
An embedding is a numeric vector representation of text or other data that captures semantic meaning.
Knowledge base
A knowledge base in Amazon Bedrock is a managed capability that ingests and indexes your data so foundation models can retrieve it for grounded responses.
Bedrock Agents
Amazon Bedrock Agents are a capability that lets foundation models plan and carry out multi-step tasks by calling tools and APIs through action groups.
Prompt engineering
Prompt engineering is the practice of designing prompts — including instructions, examples, and templates — to steer a model's output.
Bedrock Guardrails
Amazon Bedrock Guardrails are a capability that filters harmful content and enforces denied topics and policies on model inputs and outputs.
Responsible AI
Responsible AI is the practice of building AI systems that are safe, fair, transparent, and privacy-preserving.
Prompt injection
Prompt injection is an attack in which crafted input manipulates a model into ignoring its instructions or leaking data.
Provisioned throughput
Provisioned throughput in Amazon Bedrock reserves dedicated model capacity for predictable, high-volume inference at a fixed rate.
Model evaluation
Model evaluation is the process of measuring a foundation model's output quality using automatic metrics or human review.