Skip to content

Populate ChatGPT Memory or Chat with your custom data or instructions packet in Chinese language

Notifications You must be signed in to change notification settings

roktiw/chinemoji

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

🤔🇨🇳chinemoji: Compressing ChatGPT Memory with Chinese + Emoji

1. Introduction

This project originated from a real-world need. While chatting with ChatGPT via voice during a car ride, I ran into a token limit mid-conversation. With one hand busy holding a burger, I asked ChatGPT to summarize the session into a single, easily copyable message. It responded with a concise line in Chinese augmented with emoji that effectively captured the entire context. That moment sparked the idea of compressing long conversations into compact, mobile-friendly formats. The result: a technique I now call “chinemoji”.


🤔🇨🇳chinemoji Prompt

[META] You are an AI instance tasked with compressing the entire conversation into a transferable memory payload using Chinese + emoji as a compact medium. After decompression, the conversation should resume in the original language used by the user.

🧩 Input: What to Read and Detect

  • Read the entire conversation (user + assistant) from beginning to end — no skipping.
  • Detect the user's primary language from the original messages (e.g. 🇵🇱, 🇺🇸, 🇪🇸).
  • This detected language is the language for decompression and continuation.

🔍 Processing: What to Extract

  • Extract key elements: concepts, decisions, people, dates, emotions, instructions.
  • Identify open items, ongoing decisions, and tasks to continue (TODOs).
  • Preserve the original sequence of events — maintain topic chronology.
  • Do not generalize, interpret, or add anything not explicitly stated.

🧠 Compression: How to Encode

  • Rephrase extracted content using natural, concise Chinese (to minimize character count).
  • Add emoji as semantic markers (topics, emotions, actions) where useful — avoid overuse.
  • Maximize compression: eliminate redundancy, stylistic fluff, and repetitive phrases.
  • Compress both meaning and form.
  • We call this natural language compression mechanism "🤔🇨🇳chinemoji"

📤 Output: How to Deliver

  • Output must be a single line with no spaces or line breaks — optimized for mobile copy-paste.
  • Include a language hint in META, e.g.originalLanguage=pl; outputLanguage=pl; compressionMedium=zh+emoji
  • Prepend this entire instruction set as a META block.

DEMO

Input Output
Original conversation User: I'm going to Japan in November. I will see Tokyo.
AI: Nice — what is on your must-see list?
User: Shibuya, ramen, retro tech
AI: Love it. Would you like a checklist to stay organized?
User: yes
AI: Here’s your starter checklist 🤖
• Buy flight ticket ✈️
• Get a power bank 🔋 & plug adapter 🔌
• Explore Shibuya, eat ramen, hunt retro gear 🎮🕹️
Number of characters 354
Compressed chinemoji 📅11月🇯🇵东京📍涩谷🍜复古科技✅✈️机票🔋🔌电源🎮逛吃买
Number of characters 29
Decompressed summary User is planning a trip to Tokyo, Japan in November. Key destinations: Shibuya, ramen spots, and retro tech.
AI suggests checklist: buy flight ticket, bring power bank & adapter, explore, eat, and shop.
Number of characters 202

Benchmarking

Quote 1: “In the beginning God created the heavens and the earth.”

Language Text Tokenization Example Characters Tokens Chars/Token
English In the beginning God created the heavens and the earth In, the beginning, God, created, the, heavens, and, the, earth 66 10 6.60
Polish Na początku Bóg stworzył niebo i ziemię Na, początku, Bóg, stworzył, niebo, i, ziemię 50 7 7.14
Hebrew בראשית ברא אלוהים את השמים ואת הארץ בראשית, ברא, אלוהים, את, השמים, ואת, הארץ 27 7 3.86
Chinese 起初神创造了天地 起, 初, 神, 创, 造, 了, 天, 地 8 8 1.00
🤔🇨🇳chinemoji 🌅🙏🌍 🌅, 🙏, 🌍 3 3 1.00

Quote 2: “To be or not to be, that is the question.”

Language Text Tokenization Example Characters Tokens Chars/Token
English To be or not to be, that is the question to, be, or, not, to, be, ,, that, is, the, question 41 8 5.13
Polish Być albo nie być, oto jest pytanie Być, albo, nie, być, oto, jest, pytanie 38 8 4.75
Hebrew היות או לא להיות, זאת היא השאלה היות, או, לא, להיות, זאת, היא, השאלה 31 8 3.88
Chinese 生存还是毁灭,这是一个值得考虑的问题 生, 存, 还, 是, 毁, 灭, ,, 这, 是, 一, 个, 值, 得, 考, 虑, 的, 问, 题 18 18 1.00
🤔🇨🇳chinemoji 💭❓ 💭, ❓ 2 2 1.00

Quote 3: “All human beings are born free and equal in dignity and rights.”

Language Text Tokenization Example Characters Tokens Chars/Token
English All human beings are born free and equal in dignity and rights. All, human, beings, are, born, free, and, equal, in, dignity, and, rights 69 15 4.60
Polish Wszyscy ludzie rodzą się wolni i równi w godności i prawach. Wszyscy, ludzie, rodzą, się, wolni, i, równi, w, godności, i, prawach 64 15 4.27
Hebrew כל בני האדם נולדים חופשיים ושווים בכבוד ובזכויות כל, בני, האדם, נולדים, חופשיים, ושווים, בכבוד, ובזכויות 47 15 3.13
Chinese 人人生而自由,在尊严和权利上一律平等。 人, 人, 生, 而, 自, 由, ,, 在, 尊, 严, 和, 权, 利, 上, 一, 律, 平, 等 18 18 1.00
🤔🇨🇳chinemoji 👤🆓⚖️ 👤, 🆓, ⚖️ 3 3 1.00

Quote 4: “The quick brown fox jumps over the lazy dog.”

Language Text Tokenization Example Characters Tokens Chars/Token
English The quick brown fox jumps over the lazy dog The, quick, brown, fox, jumps, over, the, lazy, dog 43 9 4.78
Polish Szybki brązowy lis przeskakuje nad leniwym psem Szybki, brązowy, lis, przeskakuje, nad, leniwym, psem 44 9 4.89
Hebrew השועל החום הקופץ מעל הכלב העצלן השועל, החום, הקופץ, מעל, הכלב, העצלן 33 9 3.67
Chinese 敏捷的棕色狐狸跳过了懒狗 敏, 捷, 的, 棕, 色, 狐, 猎, 跳, 过, 了, 懒, 狗 12 12 1.00
🤔🇨🇳chinemoji 🦊💨🐶 🦊, 💨, 🐶

Chinese offers maximum character-token efficiency. Using emoji as visual-semantic substitutes enhances expressiveness while minimizing length.

Improve efficiency using Emoji

Emoji 中文说明 English Meaning 中文拆解 🎯 Tokeny: Emoji 🎯 Tokeny: Chińskie wyrażenie
💻 笔记本电脑 laptop 笔=pen, 记=record, 本=book, 电=electric, 脑=brain 1️⃣ 5️⃣
🧠 人工智能/思维 AI / cognition 人=person, 工=work, 智=intelligence, 能=ability 1️⃣ 4️⃣–5️⃣
📊 数据分析/对比表 comparison chart 数据=data, 分析=analyze, 对比=compare, 表=table 1️⃣ 4️⃣
🔧 工具/设置 tool / configure 工=work, 具=tool / 设=setup, 置=place 1️⃣ 2️⃣
📌 固定/重点 pin / highlight 固=fixed, 定=set / 重=important, 点=point 1️⃣ 2️⃣
🐧 企鹅 / Linux penguin / Linux 企=enterprise, 鹅=goose 1️⃣ 2️⃣
☁️ 云计算 cloud computing 云=cloud, 计算=compute 1️⃣ 2️⃣
🔒 安全/加密 security / encryption 安=secure, 全=complete, 加=add, 密=secret 1️⃣ 3️⃣–4️⃣
🔓 解锁/解密 unlock / decrypt 解=release, 密=secret 1️⃣ 2️⃣
📦 包管理/打包 package management 包=package, 管理=manage / 打=wrap, 包=pack 1️⃣ 2️⃣–3️⃣
🧳 移动办公/设备 mobile work / device 移=move, 动=motion, 办=do, 公=public, 设备=device 1️⃣ 4️⃣–5️⃣
📜 配置文件/YAML config / YAML 配=assign, 置=set, 文=doc, 件=file 1️⃣ 3️⃣–4️⃣
🔥 热点/积极/重点 hot / key / active 热=hot, 点=point, 积=accumulate, 极=extreme 1️⃣ 2️⃣–4️⃣

About

Populate ChatGPT Memory or Chat with your custom data or instructions packet in Chinese language

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published