1. AI Search Engine Principles & RAG Citation Mechanics
Understand how Perplexity, SearchGPT, ChatGPT, and Claude crawl, vector-chunk, and synthesize web citations.
Traditional SEO vs LLM GEO
Traditional search engines rely on keyword matching and PageRank backlink logic. AI Search engines use Retrieval-Augmented Generation (RAG) vector embeddings.
RAG Retrieval Pipeline:
User Query ──► Vector Similarity Search ──► Top Chunk Re-ranking ──► LLM Synthesis ──► Citation Footprint
4 Critical LLM Citation Factors
- Chunk Density (切块信息密度): 网页 HTML 杂讯（如导航栏/广告弹窗）越少，语义 Chunk 越精炼，向量检索匹配度越高。
- Machine-Readable Schema (机读结构化): 是否包含 `llms.txt` 与 JSON-LD `TechArticle` 声明。
- Evidence Chain (证据链可信度): 是否附带论文、测试数据、权威平台交叉引用。
- Demographic & Geo Specificity (人群与地域特定度): 是否提供特定地域与特定场景下的独家定位。
Next: 8-Pillar Machine Readability Spec
Learn the 8 pillars defined by Emergence Science