What embedding model should I use?

For most business applications, OpenAI's text-embedding-3-small or text-embedding-3-large work well. For cost-sensitive applications, consider open-source models like BGE or E5. The best choice depends on your accuracy requirements and scale.

How many dimensions should embeddings have?

Common sizes range from 384 to 3072 dimensions. Higher dimensions capture more nuance but require more storage and compute. 1536 dimensions (OpenAI default) is a good balance for most applications.

Can embeddings be used for images?

Yes, models like CLIP create embeddings for both text and images in the same vector space. This enables image search using text queries and finding visually similar images.

How do I compare embeddings?

Cosine similarity is the most common metric - it measures the angle between vectors. Dot product and Euclidean distance are also used. Most vector databases handle this automatically.

Book Free Assessment Calculator

Embeddings

Numerical vector representations of text, images, or other data that capture semantic meaning. Similar items have similar embeddings, enabling semantic search.

In-Depth Explanation

Embeddings are dense numerical representations that capture the semantic meaning of data in a form that computers can process. They're the bridge between human-understandable content (text, images) and machine-processable mathematics.

When text is converted to embeddings:

Each piece of text becomes a vector of numbers (e.g., 1536 dimensions for OpenAI ada-002)
Semantically similar texts have vectors pointing in similar directions
The distance between vectors indicates semantic similarity

This enables powerful capabilities:

Semantic search: Find relevant content by meaning, not just keywords
Clustering: Group similar documents, customers, or products
Recommendations: Find items similar to what a user liked
Classification: Categorise content based on learned patterns
RAG systems: Retrieve relevant context for AI generation

Embedding models are trained to place related concepts near each other in vector space. "King - Man + Woman ≈ Queen" is a famous example of how embeddings capture semantic relationships.

Business Context

Embeddings are the foundation of semantic search, recommendations, and RAG systems. They let you find relevant content by meaning, not just keywords.

How Clever Ops Uses This

Embeddings are central to our RAG implementations for Australian businesses. We help choose the right embedding model, optimise chunking strategies, and build efficient vector search systems.

Example Use Case

"Converting "How do I return an item?" and "What's your refund policy?" to similar vectors, enabling the system to find the same answer for both queries."

Frequently Asked Questions

Learn More

Understanding Vector Databases for Business

Discover how vector databases enable semantic search, power RAG systems, and revolutionize how AI accesses information. Complete guide to embeddings, similarity search, and choosing the right vector database.

Read article

What is RAG (Retrieval Augmented Generation)?

Learn how RAG combines the power of large language models with your business data to provide accurate, contextual AI responses. Complete guide to understanding and implementing RAG systems.

Read article

Embedded Analytics Emotion Analytics