Arbow

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.

Maybe you've heard about this technique but you haven't completely understood it, especially the PPO part. This explanation might help.

We will focus on text-to-text language models 📝, such as GPT-3, BLOOM, and T5. Models like BERT, which are encoder-only, are not addressed.

Reinforcement Learning from Human Feedback (RLHF) has been successfully applied in ChatGPT, hence its major increase in popularity. 📈

RLHF is especially useful in two scenarios 🌟:

You can’t create a good loss function
- Example: how do you calculate a metric to measure if the model’s output was funny?
You want to train with production data, but you can’t easily label your production data

Docker Hub 镜像加速器

国内从 Docker Hub 拉取镜像有时会遇到困难，此时可以配置镜像加速器。

Dockerized 实践 https://github.com/y0ngb1n/dockerized

1️⃣ Docker daemon 配置代理（推荐）

参考 Docker daemon 配置代理

Mac web developer apps

This gist's comment stream is a collection of webdev apps for OS X. Feel free to add links to apps you like, just make sure you add some context to what it does — either from the creator's website or your own thoughts.

— Erik

	[General]
	loglevel = notify
	skip-proxy = 127.0.0.1, 192.168.0.0/16, 10.0.0.0/8, 172.16.0.0/12, 100.64.0.0/10, localhost, *.local, ::ffff:0:0:0:0/1, ::ffff:128:0:0:0/1
	bypass-tun = 192.168.0.0/16, 10.0.0.0/8, 172.16.0.0/12

	# dns-server = 119.29.29.29,223.5.5.5,114.114.115.115
	# external-controller-access = PASSWORD@0.0.0.0:6155
	# ipv6 = true

	// REMEMBER TO CHANGE THE external-controller-access' PASSWORD

	defmodule ROP do

	defmacro try_catch(args, func) do
	quote do
	(fn ->
	try do
	unquote(args) \|> unquote(func)
	rescue
	e -> {:error, e}
	end

	import hashlib, zlib
	HASH_BLOCKSIZE = 65536
	def filebaidumd5(f, size=262144):
	h1 = hashlib.md5()
	h2 = 0
	pos = 0
	buf = f.read(min(HASH_BLOCKSIZE, size))
	md5_at_size = None
	while buf:
	pos += len(buf)

	package org.nutz.lang.util;

	import java.io.BufferedInputStream;
	import java.io.DataInputStream;
	import java.io.IOException;
	import java.io.InputStream;
	import java.lang.reflect.Constructor;
	import java.lang.reflect.Method;
	import java.util.ArrayList;
	import java.util.HashMap;

	import java.io.BufferedReader;
	import java.io.InputStream;
	import java.io.InputStreamReader;
	import java.io.IOException;
	import java.lang.management.ManagementFactory;
	import java.lang.management.RuntimeMXBean;
	import com.sun.tools.attach.AttachNotSupportedException;
	import com.sun.tools.attach.VirtualMachine;
	import sun.tools.attach.HotSpotVirtualMachine;

	# Adapted from here: https://gist.github.com/489093
	# Original credit goes to pplante and copyright notice pasted below

	# Copyright (c) 2010, Philip Plante of EndlessPaths.com
	#
	# Permission is hereby granted, free of charge, to any person obtaining a copy
	# of this software and associated documentation files (the "Software"), to deal
	# in the Software without restriction, including without limitation the rights
	# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
	# copies of the Software, and to permit persons to whom the Software is