23bfd15b19
Graph topology (Metatron's Cube pattern — 1 center + 12 outer nodes): - index: now links to ALL 12 other nodes (complete hub) - Inner ring cycle: hot→log→overview→dashboard→concepts/_index→entities/_index - Added: hot↔WikiMap, log↔sources/_index, dashboard↔concepts/_index - Added: entities/_index↔hot, entities/_index↔LLM Wiki Pattern - Added: sources/_index↔log, sources/_index↔entities/_index - Outer ring: concepts connected in triangle + Karpathy/sources cross-linked - Added: dashboard↔Compounding, entities/_index↔LLM Wiki Pattern graph.json physics for geometric arrangement: - repelStrength: 80 (strong push-apart for uniform spacing) - linkStrength: 3.0 (locks ring geometry) - linkDistance: 80 (tighter rings) - centerStrength: 0.25 (moderate center pull) - nodeSizeMultiplier: 2.0 (hub nodes visually dominant) - Added colors: questions=yellow, comparisons=red, nav=teal Phantom links removed from Hot Cache.md: - Removed [[Page A]], [[Page B]], [[New Page 1]], [[Existing Page]] New module pages: - wiki/questions/How does the LLM Wiki pattern work.md - wiki/comparisons/Wiki vs RAG.md - Adds questions/ and comparisons/ domains to the graph (yellow + red nodes)
2.0 KiB
type, title, question, answer_quality, created, updated, tags, status, related, sources
| type | title | question | answer_quality | created | updated | tags | status | related | sources | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| question | How does the LLM Wiki pattern work? | How does the LLM Wiki pattern work and why is it better than RAG? | definitive | 2026-04-07 | 2026-04-07 |
|
developing |
|
How does the LLM Wiki pattern work?
Question: How does the LLM Wiki pattern work and why is it better than RAG?
Answer
The LLM Wiki Pattern turns an LLM into a knowledge architect rather than a search engine.
Standard RAG (Retrieval-Augmented Generation): every query searches raw documents, retrieves chunks, and assembles an answer from scratch. Nothing is built up. Ask the same question twice — it does the same work twice.
The wiki pattern is different. When a source arrives, the LLM reads it and integrates it: updating entity pages, noting contradictions, adding cross-references. The synthesis is done once and persists. Every query benefits from all previous ingests.
The three layers
.raw/— your source documents. Immutable. Claude reads, never modifies.wiki/— Claude-generated knowledge. Summaries, entities, concepts, synthesis.CLAUDE.md— the schema. Tells Claude how the wiki is structured and what to do.
Why it compounds
See Compounding Knowledge for the full argument. The short version: each new source doesn't just add one page — it enriches 8-15 existing pages. The connections between pages are where the value lives, not the raw content itself.
The hot cache shortcut
Hot Cache (wiki/hot.md) is a ~500-word summary of recent context. New sessions read it first. Cross-project references read it first. It prevents re-reading the whole wiki just to answer "where were we?"
(Source: LLM Wiki Pattern)
Confidence
definitive — this is the core concept the entire vault demonstrates.