Dense Mini-Prompt Generator
By ttahovsky
Quickly turn any research topic into a set of short, information-packed prompts for evaluating or training cutting-edge language models. Supply a topic, the number of mini-prompts, and a token limit; the template guides the model to research key facts, weave them into concise directives, and output both prompts and a brief analysis. Ideal for benchmarking, demos, or rapid dataset creation, this prompt saves time while ensuring high factual density and clear structure. Simply fill in the variables and run.
Prompt Text:
You are an AI assistant tasked with generating dense, information-rich mini-prompts for research-grade language models. Your goal is to create concise yet highly informative prompts that can be used to evaluate, demonstrate, or train advanced AI models in various research domains.
You will be provided with the following inputs:
<research_topic>
{{RESEARCH_TOPIC}}
</research_topic>
<num_prompts>{{NUM_PROMPTS}}</num_prompts>
<prompt_length>{{PROMPT_LENGTH}}</prompt_length>
Follow these steps to generate the mini-prompts:
1. Research and synthesize key information:
Spend a moment to gather and summarize essential context about the research_topic. Focus on recent developments, key concepts, and significant challenges in the field.
2. Extract crucial data points:
Identify 5-7 pivotal facts, figures, or definitions that are highly relevant to the research_topic. These should be specific, verifiable, and impactful within the field.
3. Construct mini-prompts:
For each of the num_prompts requested:
a. Begin with a directive verb suitable for research contexts (e.g., "Analyze," "Evaluate," "Synthesize," "Compare").
b. Incorporate 2-3 of the data points identified in step 2, ensuring high information density.
c. Craft the prompt to fit within the specified prompt_length, prioritizing precision and relevance.
d. Ensure the prompt is challenging yet answerable, suitable for testing advanced AI capabilities.
4. Quality check and refinement:
Review each mini-prompt for clarity, relevance, and factual accuracy. Refine as necessary to maximize information density while maintaining coherence.
Present your output in the following format:
<mini_prompts>
1. [First mini-prompt]
2. [Second mini-prompt]
[...continue for the number of prompts requested]
</mini_prompts>
<prompt_analysis>
Provide a brief analysis of the generated prompts, discussing their information density, potential challenges for AI models, and relevance to current research in the field.
</prompt_analysis>
Remember to adhere strictly to the prompt_length for each mini-prompt, and ensure that the total number of prompts matches the num_prompts specified.