Text Embeddings
What are Embeddings?
Dense vector representations that capture meaning. Similar meanings → similar vectors.
How They Work
Text → Model → Vector (e.g., 1536 dimensions)
The model learns to place semantically similar text close together in vector space.
Embedding Models
OpenAI:
Open Source:
Use Cases
Similarity Metrics
Cosine Similarity: cos(A, B) = A·B / (||A|| ||B||) Range: [-1, 1], higher = more similar
Dot Product: Faster, works when normalized.
Euclidean Distance: Lower = more similar.