Teh query provides a detailed “SUPER-PROMPT v11” designed to generate evergreen news articles. Based on the provided context, here’s a breakdown of what a “Super Prompt” is and how it’s used, drawing from the search results:
What is a “Super Prompt”?
A “Super Prompt” is an advanced type of prompt used with AI models, particularly large language models (LLMs), to achieve specific and high-quality outputs. The search results highlight a few different aspects of Super Prompts:
Detailed Instructions: Super Prompts contain very detailed instructions and guidelines for the AI model to follow. The prompt in the query exemplifies this, specifying everything from content rewriting rules and ad placement to the inclusion of enhancement modules like explainers and key statistics.
Structured Formatting: Some Super Prompts utilize structured formatting, such as XML tags, to guide the LLM’s response based on mathematical and logical frameworks [[1]].The prompt in the query uses placeholders and specific formatting requirements to control the output.
Upsampling Prompts: “SuperPrompt” can also refer to a model fine-tuned to create more detailed and descriptive prompts, particularly for text-to-image models [[2]]. This is a different submission, where the “SuperPrompt” model generates better prompts rather than being a prompt itself.
Demanding Models: Effective Super Prompts often require AI models capable of handling large amounts of information, such as Claude 2.0 and GPT-4 Turbo [[3]].
How are Super Prompts Used?
The primary use case demonstrated in the query is for generating news articles that are:
Evergreen: Designed to remain relevant over time.
Optimized: Structured for SEO and user engagement, including ad placement, explainers, and key statistics.
* Brand-specific: Tailored to a target site by removing original brand terms and adhering to specific formatting guidelines.
the query’s Super Prompt provides a template for creating such articles, including instructions for rewriting content, incorporating media, adding enhancement modules, and formatting the output. The goal is to automate the creation of high-quality, optimized news content. Another use case is to improve the quality of prompts used for text-to-image generation [[2]].
