// MODEL OPTIMIZATION AND PROMPT SYNTAX TERM

Vector Store

Another name for a vector database, used to store and retrieve numerical representations (vectors) of information.

TECHNICAL DEFINITION

A component or system designed to efficiently store and retrieve vector embeddings, often used interchangeably with vector database, providing fast approximate nearest neighbor search capabilities for semantic retrieval.

BACKGROUND

Prompt injection is a cybersecurity exploit and an attack vector in which innocuous-looking inputs are designed to cause unintended behavior in machine learning models, particularly large language models (LLMs). The attack takes advantage of the model's inability to distinguish between developer-defined prompts and user inputs to bypass safeguards and influence model behaviour. While LLMs are designed to follow trusted instructions, they can be manipulated into carrying out unintended responses through carefully crafted inputs.

READ MORE ON WIKIPEDIA

SYNONYMS & ALIASES

  • Vector database
  • embedding store
  • similarity index

USAGE NOTE

Many RAG implementations rely on a vector store to manage and query document embeddings.

DEVELOPERS

Organizations developing technology related to Vector Store.

  • Pinecone

    Provides a managed vector database specifically designed for building high-performance AI applications, facilitating efficient similarity search for embeddings.

  • Weaviate

    An open-source, AI-native vector database that allows developers to store data objects and vector embeddings, enabling semantic search and AI-powered data retrieval.

  • Qdrant

    Offers an open-source vector similarity search engine and database, providing high-performance solutions for storing, indexing, and searching vector embeddings.

  • Chroma

    An AI-native open-source embedding database, designed to be easy to use for building LLM applications by focusing on simplicity and developer experience.

  • Zilliz

    The company behind Milvus, an open-source vector database, Zilliz provides cloud-native, fully managed vector database services for enterprise AI applications.

  • Meta AI (Faiss)

    Meta AI developed Faiss (Facebook AI Similarity Search), an open-source library that provides efficient algorithms for similarity search and clustering of dense vectors.

  • Elastic NV

    Through Elasticsearch, Elastic NV offers vector search capabilities, allowing users to store embeddings and perform semantic search alongside traditional keyword search within their platform.

  • Redis (Redis Labs)

    Redis, an in-memory data store, offers modules like Redis Stack that support vector similarity search, enabling real-time AI applications that require fast access to embeddings.

  • Supabase

    Provides a PostgreSQL-based backend as a service, including the pg_vector extension, making it easy for developers to store and query vector embeddings within their existing database.

RELATED TERMS IN PROMPTING & LOGIC