danieldietrich

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

danieldietrich

People

:bowtie:	😄 :smile:	😆 :laughing:
😊 :blush:	😃 :smiley:	☺️ :relaxed:
😏 :smirk:	😍 :heart_eyes:	😘 :kissing_heart:
😚 :kissing_closed_eyes:	😳 :flushed:	😌 :relieved:
😆 :satisfied:	😁 :grin:	😉 :wink:
😜 :stuck_out_tongue_winking_eye:	😝 :stuck_out_tongue_closed_eyes:	😀 :grinning:
😗 :kissing:	😙 :kissing_smiling_eyes:	😛 :stuck_out_tongue:

danieldietrich

Purpose

Bootstrap knowledge of LLMs ASAP. With a bias/focus to GPT.

Avoid being a link dump. Try to provide only valuable well tuned information.

Prelude

Neural network links before starting with transformers.

danieldietrich

#!/bin/bash

# Tom Hale, 2016. MIT Licence.
# Print out 256 colours, with each number printed in its corresponding colour
# See http://askubuntu.com/questions/821157/print-a-256-color-test-pattern-in-the-terminal/821163#821163

set -eu # Fail on errors or undeclared variables

printable_colours=256

danieldietrich

/*
 * Easing Functions - inspired from http://gizma.com/easing/
 * only considering the t value for the range [0, 1] => [0, 1]
 */
EasingFunctions = {
  // no easing, no acceleration
  linear: t => t,
  // accelerating from zero velocity
  easeInQuad: t => t*t,
  // decelerating to zero velocity

danieldietrich

git checkout master # you can avoid this line if you are in master...
git subtree split --prefix dist -b gh-pages # create a local gh-pages branch containing the splitted output folder
git push -f origin gh-pages:gh-pages # force the push of the gh-pages branch to the remote gh-pages branch at origin
git branch -D gh-pages # delete the local gh-pages because you will need it: ref

danieldietrich

Deploying a subfolder to GitHub Pages

Sometimes you want to have a subdirectory on the master branch be the root directory of a repository’s gh-pages branch. This is useful for things like sites developed with Yeoman, or if you have a Jekyll site contained in the master branch alongside the rest of your code.

For the sake of this example, let’s pretend the subfolder containing your site is named dist.

Step 1

Remove the dist directory from the project’s .gitignore file (it’s ignored by default by Yeoman).

danieldietrich

.cache
dist
node_modules

danieldietrich

{
  "scripts": {
    "build": "npm run build:es2015 && npm run build:esm && npm run build:cjs && npm run build:umd && npm run build:umd:min",
    "build:es2015": "tsc --module es2015 --target es2015 --outDir dist/es2015",
    "build:esm": "tsc --module es2015 --target es5 --outDir dist/esm",
    "build:cjs": "tsc --module commonjs --target es5 --outDir dist/cjs",
    "build:umd": "rollup dist/esm/index.js --format umd --name YourLibrary --sourceMap --output dist/umd/yourlibrary.js",
    "build:umd:min": "cd dist/umd && uglifyjs --compress --mangle --source-map --screw-ie8 --comments --o yourlibrary.min.js -- yourlibrary.js && gzip yourlibrary.min.js -c > yourlibrary.min.js.gz",
  }
}

danieldietrich

When you create a npm package, remember it might be used in a browser or a server, or even a command line utility… For each package you create, please pay attention at what it will be used for:

1. Is it going to be used as a dependency to a nodejs application that is not bundled? (e.g. command line utilities)
2. Is it going to be used as a dependency to a nodejs application that is bundled? (e.g. AWS Lambdas)
3. Is it going to be used as a dependency to a browser application (always bundled)?.

• In cases 2) and 3) you want to allow for tree shaking.
• In cases 1) and 2) you want to benefit from the "ES6"/"ES next" features supported natively by nodejs.
• In case 3) you also want to benefit from the native support of "ES6" from your browser.