Vector
A vector is an ordered list of numbers used to represent a point in a high-dimensional space.
What Is a Vector?
A vector is just an ordered list of numbers, like [0.2, -0.5, 0.8]. In everyday geometry, a vector with two or three numbers can represent a point or direction in space you can visualize. Embeddings extend this same idea to hundreds or thousands of numbers — a space you can't draw on paper, but where the same underlying concept of "distance" and "direction" between points still applies mathematically.
Key Idea
You don't need to visualize a 1,536-dimensional space to work with it — the math for measuring distance and similarity between vectors works the same regardless of how many numbers are in the list.
Why Vectors Are the Foundation of Embeddings
- An embedding is a specific kind of vector — one produced by a model so that vectors close together represent content with similar meaning.
- Vectors can be compared mathematically — using measures like cosine similarity — which is what makes semantic search and retrieval possible at all.
- A vector database is built specifically to store and search large collections of vectors efficiently, which a general-purpose database isn't optimized to do at scale.
Common Mistakes
Assuming "vector" and "embedding" mean exactly the same thing
A vector is the general mathematical structure; an embedding is a specific vector produced by a model to encode meaning — every embedding is a vector, but not every vector is an embedding.
Trying to mentally visualize a high-dimensional vector directly
This isn't necessary or useful — the underlying math for comparing vectors works the same whether there are 2 dimensions or 2,000.
Assuming vector comparisons require the vectors to come from the same source model
Vectors from different embedding models generally aren't directly comparable, even though they're both just lists of numbers of the same length.
Interview Question
What is a vector, and how does it relate to an embedding?
A vector is simply an ordered list of numbers, used to represent a point or direction in some space — in everyday geometry that might be two or three numbers you can visualize, but the same idea extends to hundreds or thousands of numbers you can't draw but can still compare mathematically. An embedding is a specific kind of vector: one produced by a model so that vectors close together represent content with similar meaning. Every embedding is a vector, but not every vector is an embedding — the vector is just the underlying data structure; the embedding is what makes it meaningful.
What an interviewer may ask next
- Is every vector an embedding?
- Why doesn't it matter that a vector might have thousands of dimensions you can't visualize?
- Can you directly compare vectors produced by two different embedding models?
Explain It in 30 Seconds
A vector is an ordered list of numbers representing a point or direction in space — the same idea as a 2D or 3D vector in geometry, just extended to hundreds or thousands of numbers. An embedding is a specific kind of vector, produced by a model so that similar meanings land close together, which is what makes semantic search and vector databases possible. Every embedding is a vector, but not every vector is an embedding.