Skip to main content

Massive data: input

#write
3.7K
·
216 chars
·
≈ 66 tokens
·
Source
To surpass ChatGPT's input limitations, you can divide the article into multiple paragraphs, with each containing no more than 2000 characters. Start each paragraph with a '@number' tag on the first line, such as '@1'. Use the text processing tools in the navigation bar to split the text. Note that you can ignore GPT's responses as they will not affect your final result. Excerpted from an article by Esor Huang.
Prompt Content
Let's start a new round of questions and answers. In the upcoming conversations, I will provide you with article content labeled with an '@' symbol. Please remember the content but do not summarize it. Are you ready?
FAQ
Does 2000-character chunking really break input limits?
Mainstream large models (like ChatGPT, Claude, Gemini, DeepSeek) context windows reach 256K+ and most cases don't need chunking. This method suits early GPT-3.5 (only 4K context) or APIs with strict length limits. If you can paste everything at once, don't chunk.
After chunked pasting, does the AI forget earlier segments?
Yes, especially past 10 chunks, earlier content gets "diluted." After chunking, have the AI reply "received @N" per chunk, ask questions after all received, and remind "please answer based on all @ segments." Accuracy improves.
How do I use this prompt?
Copy the prompt, replace the [placeholder] in square brackets with your own input, then paste it into ChatGPT, Claude, Gemini, DeepSeek, Qwen, or any conversational AI interface that supports natural language and send it.
Share
Discussion
Copyright © 2026 AiShort (ChatGPT Shortcut) · Community-contributed. Views are authors' own.