Unverified50% confidenceFactExact time
主流Embedding模型（如BGE、text-embedding系列）通常基于BERT架构的编码器（Encoder-only），通过对比学习让语义相近的句子在向量空间中靠拢；而GPT类生成模型基于解码器（Decoder-only）
1
Sources
50%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
Java程序员转型AI开发：RAG知识库实战落地指南
Related Claims
Unverified现代Sentence Embedding模型通常采用双塔（Bi-Encoder）结构，查询和文档分别独立编码为向量76% similarUnverifiedBGE和text-embedding-3是当前可选的向量嵌入（Embedding）模型72% similarUnverified文本分类器通常基于BERT、FastText或TF-IDF+逻辑回归方案实现70% similarUnverifiedThe recall stage typically uses Bi-Encoder models (such as text-embedding-ada-002 or the BGE series), while the reranking stage uses Cross-Encoder models (such as Cohere Rerank or BGE-Reranker).69% similarUnverifiedTONGYI Lingma's underlying layer relies on the code understanding and generation capabilities of the Tongyi Qianwen series of large language models.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/558738API
curl https://kongchang.com/api/v1/knowledge/claims/558738MCP
get_claim(id=558738)