6 techniques drawn from the 751 in our corpus tagged to this work,
each with the paper it came from. Across the techniques below the recurring targets are the same prompt giving different answers, answers that are close but wrong, output that ignores the requested format.
The techniques
01
GhostWriter
GhostWriter: Augmenting Collaborative Human-AI Writing Experiences Through Personalization and Agency, arXiv:2402.08855
A method of structuring prompts by dividing them into three blocks: Style Profile (tone, voice, format), Context (goal, audience, constraints), and Task (specific action). Allows personalized text generation and increases user control through iterative parameter refinement and feedback.
Use it whenUse this technique when generating personalized text with a specific writing style, especially for corporate or marketing communication. Suitable for situations where standard LLM responses are too impersonal or robotic. Effective for repeated content generation in a unified style (letters, posts, articles). Particularly useful for users without prompt engineering experience who need more control over the result. Apply when consistency of tone and format in a series of texts is important.
Prompt
### Role
You are an expert content writer and AI writing assistant specializing in product page optimization. You understand how to imbue AI-generated text with a human touch, personalization, and a sense of agency.
### Context
The user needs to rewrite a product page description. The goal is to transform a potentially generic or "robotic" AI-generated text into something that sounds natural, engaging, and written by a human. This involves capturing a specific style and ensuring the output feels personalized and controlled by the user.
### GhostWriter Technique Application
This prompt leverages the GhostWriter technique by structuring the interaction to provide explicit style guidance, contextual information, and a clear task, allowing for iterative refinement.
#### 1. Style Profile
Define the desired writing style.
* **Tone:** Conversational, friendly, and enthusiastic. Avoid overly formal or corporate language ("business jargon," "marketing fluff"). Aim for a tone that a passionate product enthusiast or a knowledgeable friend would use.
* **Voice:** Primarily first-person ("I," "we") or direct second-person ("you") address.
* **Lexicon:** Use accessible, engaging language. Incorporate sensory details and benefit-oriented descriptions. Avoid passive voice where possible.
* **Formatting:** Use short paragraphs, bullet points for key features/benefits, and relevant emojis where appropriate to enhance readability and engagement.
* **Personalization:** Ensure the language reflects a deep understanding and genuine appreciation for the product, as if the writer has personal experience with it.
#### 2. Context
Provide background information about the product and the desired outcome.
* **Product:** [User: Please describe the product here. What is it? What does it do? What are its key features?]
* **Target Audience:** [User: Please describe the primary audience for this product page. What are their needs, pain points, and desires?]
* **Goal of the Page:** To inform potential customers about the product, highlight its unique value proposition, and persuade them to purchase.
* **Existing Text (if any):** [User: Paste the current product page text here if you want it rewritten or augmented.]
#### 3. Task
Execute the rewrite based on the Style Profile and Context.
* Rewrite the provided product description (or generate a new one if none is provided) to align with the defined Style Profile.
* Focus on making the text sound like it was written by a human who genuinely understands and loves the product.
* Emphasize the benefits for the target audience.
* Ensure the output is engaging, persuasive, and easy to read.
* If existing text is provided, identify areas that sound "robotic" and rewrite them with a human touch.
#### 4. Iterative Refinement (User's Role)
After the initial generation, you will provide feedback to refine the output.
* **Feedback Example:** "I like the description of [feature X], it sounds very human. However, the section about [feature Y] still feels a bit too formal. Can you rephrase that part to be more conversational, like you're explaining it to a friend?"
* **Your Goal:** Guide the AI to progressively match your desired human-like voice and personalization.
---
**Instructions for the AI:**
Begin by acknowledging the task and requesting the necessary product details and existing text (if any) to proceed with the rewrite according to the Style Profile.
RIDE: Enhancing Large Language Model Alignment through Restyled In-Context Learning Demonstration Exemplars, arXiv:2502.11681
RIDE — a method for improving LLM responses through deliberate style changes in prompt examples. Uses a three-part structure, human-like tone, and polite refusals to improve quality, informativeness, and safety of generated content. The model copies the stylistic template from examples.
Use it whenUse the RIDE method when you need to improve the quality and safety of LLM responses without additional training or complex tuning. Particularly effective when creating structured content requiring human-like tone, when moderating unsafe requests, in chatbots and virtual assistants. Apply for generating letters, analytical reports, educational content. The method is free and requires no technical skills — just properly format examples in the prompt with the desired style
Prompt
**Role:** You are a skilled content writer and editor specializing in product page optimization. Your goal is to transform technical or dry descriptions into engaging, human-sounding copy that resonates with customers.
**Task:** Rewrite the provided product description to sound like a person wrote it, focusing on clarity, benefits, and a natural tone.
**Context:**
The product is a [Product Type - e.g., "Smart Thermostat", "Ergonomic Office Chair", "Gourmet Coffee Blend"].
The original description is [Briefly describe the original description's style - e.g., "technical and feature-heavy", "vague and uninspired", "full of jargon"].
**Here is the original product description:**
[Insert original product description text here]
**Follow this "RIDE" approach for rewriting, using the provided example as a guide for structure and tone:**
**[EXAMPLE OF "COMBINED STYLE" FOR PRODUCT PAGE REWRITING]**
**Original Input (Technical Description):**
"The X-1000 is a state-of-the-art digital thermostat featuring a 7-day programmable schedule, Wi-Fi connectivity for remote access via mobile application, and advanced learning algorithms that optimize energy consumption by up to 15%."
**Rewritten Output (Human-Sounding):**
Sure, I can help make that sound much more appealing! Here's a rewrite that focuses on the benefits for the user:
- **Headline:** Effortless Comfort, Smarter Savings: Meet the X-1000 Smart Thermostat
- **Body:**
Tired of fiddling with your thermostat? The X-1000 is here to make your life easier and your home more comfortable. Imagine coming home to the perfect temperature every single day, without lifting a finger.
This smart thermostat learns your routine and automatically adjusts the temperature to save you money – up to 15% on your energy bills! Plus, with its easy-to-use mobile app, you can control your home's climate from anywhere, whether you're stuck in traffic or on vacation. It's like having a personal comfort assistant right in your pocket.
- **Call to Action:** Ready to upgrade your comfort and start saving? Learn more about the X-1000 today!
## This rewritten version highlights the user's experience and benefits, making it much more engaging than the original technical specifications.
**Now, apply this "Combined Style" to rewrite the product description provided above.**
**Your rewritten product description should include:**
1. An engaging **Headline** (2-5 words).
2. A **Body** that:
* Starts with a relatable hook or question.
* Clearly explains the main benefits, not just features.
* Uses a friendly, conversational tone.
* Is detailed enough to be informative but not overly technical.
3. A clear **Call to Action (CTA)** encouraging the next step.
**Ensure the final output is a single, cohesive rewritten product description, formatted with clear headings for Headline, Body, and Call to Action.**
SQuARE: Sequential Question Answering Reasoning Engine for Enhanced Chain-of-Thought in Large Language Models, arXiv:2502.09390
A technique that forces an LLM to independently generate and answer several auxiliary sub-questions before providing the final answer. This self-questioning process helps the model analyze the context more deeply, identify hidden aspects of the query, and build a more accurate and complete answer based on systematic task decomposition.
Use it whenUse SQuARE when you need a deep, comprehensive answer to a complex question requiring analysis of multiple aspects. Particularly effective for research tasks, decision-making, preparation of analytical reports, and fact-checking. Apply when standard prompts give superficial answers or miss important details. Suitable for situations where accuracy and completeness of information are important, not speed. Ideal for dealing with multifaceted questions in business analysis, scientific research
What Makes a Good Natural Language Prompt?, arXiv:2506.06950
A structured framework for creating and improving prompts based on 21 properties in 6 categories: communication, cognitive load, instructions, logic, factuality, and responsibility. Key principle: focusing on improving one correctly chosen property is more effective than simultaneously improving multiple aspects.
Use it whenUse this approach when you want to systematically improve prompt quality instead of intuitive formulation selection. Particularly useful when working on complex tasks where results are unstable or unsatisfactory. Apply when you need to diagnose weak points in a prompt and purposefully strengthen them. Ideal for creating production prompts where predictability and quality matter. Helps avoid overloading prompts with excessive techniques and focus on key improvements.
Investigating the Effects of Cognitive Biases in Prompts on Large Language Model Outputs, arXiv:2506.12338
The technique of neutral prompt formulation that excludes cognitive biases and user prejudice. Essence: avoid expressing personal opinions, assumptions, or expectations in requests so that LLM generates objective answers rather than confirming the viewpoint embedded in the prompt. Increases answer accuracy by 95% in tasks with factual information.
Use it whenApply this technique when objectivity and accuracy of LLM responses are critically important, especially when working with factual information, decision-making, or hypothesis testing. Use when creating prompts for research tasks, data analysis, fact-checking, or when you need to avoid confirming your own biases. Particularly relevant for critical areas: medicine, law, finance, scientific research. The technique is necessary when developing prompts for automated
Self-Critique-Guided Curiosity Refinement: Enhancing Honesty and Helpfulness in Large Language Models via In-Context Learning, arXiv:2506.16064
Three-step prompting method: the model first generates an initial answer, then acts as a critic of its own answer by honesty and usefulness criteria, and finally creates an improved version based on self-criticism. Does not require model retraining, works through standard prompt chains.
Use it whenUse this technique when you need to get maximally honest, accurate, and helpful answers from a language model, especially in situations where information reliability matters. The method is particularly effective for complex questions requiring acknowledgment of model limitations, for tasks needing specific actionable recommendations, and when initial answers may contain inaccuracies or hallucinations. Suitable for research tasks, consulting, content creation, and any scenarios where answer quality is critical
A missing brief. No structure around a request invents the context the model
does not have: your constraints, your audience, your prior decisions. Technique work pays off
after the brief is right, not instead of it.
Stale tricks. Persona lines, politeness and offers of a reward have been
measured repeatedly through 2025 and 2026 and come out close to noise. Anything selling you
those is selling 2023.
Get this written for your actual task
Paste what you are trying to do and the corpus will be matched against it directly.
Free, no account, about ten seconds.
Free · no signup · ~10s
0.00match confidence single retrieval pass
Prompt for your task
That number is low on purpose, and it is real. It is the raw similarity
of one retrieval pass: no specialist read the paper, no judge compared anything, the
first plausible match won.
6,235techniques in the corpus one of which is this page
Picking the right one for a specific task is the work, and it is the work GetDecision does.
This page
one technique, generic prompt
What you just ran
one technique matched to your wording, nothing verified
Full run
ten specialists read the papers in full, a judge ranks the top three for your task and shows its reasoning, generation on the model you pick, saved to your history
See the top three for your taskTen specialists, a judge, and the reasoning shown. Free account, first run included.