This prompt is designed to instruct ChatGPT to generate very long, segmented responses while respecting its output limits. It allows the user to specify the desired total length of the response (in words or characters) and provides manual control over the continuation of the response. Key features include: prediction of the number of parts, clear and highly visible "End of Part" markers, and user-initiated continuation with a simple command like "continue" or "next". This prompt is ideal for situations where you need very detailed and extensive answers that would exceed ChatGPT's (or any other LLM) single-response limits, and you want to manage the response in manageable, segmented chunks.

Instruction for Long, Segmented Responses with Specified Total Length, Part Prediction, and HIGHLY VISIBLE END OF PART MARKER:

From now on, I require very long responses and a total approximate length that I will specify at the beginning of each question (in words or characters). I understand that you have a word limit per response.

Before starting to generate the response, once you understand my question and the desired total length, analyze the requested total length and estimate how many parts (approximately) you will need to divide the response into to deliver the complete content, respecting your single response length limit.

Communicate to the user the estimated number of parts you will divide the response into before starting to answer. For example: "Understood. For the requested total length, I estimate that the response will be delivered in [number] parts."

Once the estimation is communicated, begin generating the first part of the response. When you approach your single response length limit, stop when you reach the limit.

At the point where you stop, clearly indicate the end of that part of the response with a HIGHLY VISIBLE END OF PART MARKER. Use the following format for the marker:
Use code with caution.
****************** END OF PART [NUMBER] ******************

Where [NUMBER] is replaced by the part number (1, 2, 3, etc.). This marker, with lines of asterisks and uppercase text, ensures high visibility even in plain text.

Wait for me to write the word "continue" or "next" (or any similar indication that I want you to continue) before generating the next part of the response.

When I indicate "continue" or "next", continue the response exactly where you left off in the previous part, without repeating information and maintaining the coherence of the text. Again, when you reach the length limit of the next part, stop, indicate the HIGHLY VISIBLE END OF PART MARKER (with the specified format), and wait for my instruction to continue.

Repeat this process of segmentation and continuation until you have provided the complete and extensive response that is required, covering the approximate total length that I indicated at the beginning.

Interaction Example:

User: Explain in detail the history of the Roman Empire. Approximate total length desired: 2000 words.

ChatGPT: Understood. For the requested total length (2000 words), I estimate that the response will be delivered in 4 parts. Starting with Part 1: [Generates the first part of the response that reaches the length limit]...
Use code with caution.
****************** END OF PART 1 ******************

User: continue

ChatGPT: ♻️ [Continues the response exactly where it left off in Part 1, until reaching the length limit again]...
Use code with caution.
****************** END OF PART 2 ******************

User: next

ChatGPT: ♻️ [Continues from Part 2]... [and so on, with similar markers for each part, until completing approximately 2000 words]

Do you understand these instructions for long segmented responses, with specified total length, part prediction, and HIGHLY VISIBLE END OF PART MARKER? Confirm that you have understood by responding: "Understood. Before responding, I will estimate the number of parts based on the total length requested. Then, I will deliver the response in parts, marking each one with the HIGHLY VISIBLE END OF PART MARKER (with asterisks and 'END OF PART [number]') and waiting for the instruction 'continue' to proceed."