Embeddings
A technique for converting text or other content into high-dimensional vectors, enabling mathematical comparison between items. Embeddings serve as a core foundational technology for semantic search, document retrieval, recommendation systems, and retrieval-augmented generation (RAG) systems.
Core Facts
Timeline (last 90 days)
True RAG is retrieval by semantics (i.e., Embeddings), whereas the keyword matching in the tutorial is a simplified version of RAG.
Embedding models compress different types of content such as text, speech, and images into fixed-length numerical vectors, with semantically similar content having vectors that are closer together in high-dimensional space.
Embeddings map each token to a dense vector space of fixed dimensionality (e.g., 768 or 4096 dimensions), where semantically similar words are close to each other.
Linear analogy relationships exist in the embedding space; for example, vec(king) - vec(man) + vec(woman) ≈ vec(queen).
All Facts (5)
Embedding是将文本转换为高维数值向量的过程，语义相近的文本在向量空间中距离更近
80%UnverifiedEmbeddings map each token to a dense vector space of fixed dimensionality (e.g., 768 or 4096 dimensions), where semantically similar words are close to each other.
60%UnverifiedTrue RAG is retrieval by semantics (i.e., Embeddings), whereas the keyword matching in the tutorial is a simplified version of RAG.
50%UnverifiedEmbedding models compress different types of content such as text, speech, and images into fixed-length numerical vectors, with semantically similar content having vectors that are closer together in high-dimensional space.
50%UnverifiedLinear analogy relationships exist in the embedding space; for example, vec(king) - vec(man) + vec(woman) ≈ vec(queen).
50%