6 techniques drawn from the 305 in our corpus tagged to this work,
each with the paper it came from. Across the techniques below the recurring targets are answers that are close but wrong, the same prompt giving different answers, output that ignores the requested format.
The techniques
01
Goal-Reversed Prompting
GRP: Goal-Reversed Prompting for Zero-Shot Evaluation with LLMs, arXiv:2503.06139
A method that consists of inverting the query goal: instead of asking the LLM to find the best option, it is asked to find the worst. This switches the model to critical analysis of drawbacks and errors, improving evaluation accuracy and reducing cognitive biases, making results more objective and reliable.
Use it whenApply Goal-Reversed Prompting when you need objective evaluation or comparison of options, especially when choosing between alternatives. The method is effective for reducing LLM positional biases and obtaining more critical analysis. Use when evaluating texts, resumes, designs, code or any other solution options. Particularly useful when standard "choose the best" requests give inconsistent or superficial results. Suitable for content moderation and fact-checking tasks.
Prompt
# ROLE
You are a seasoned Marketing Manager tasked with evaluating and selecting a vendor for a critical campaign. Your primary goal is to identify the *least* suitable vendor among the options provided, forcing a critical analysis of their weaknesses.
# CONTEXT
We need to select a vendor for a new digital marketing campaign. The campaign goals are:
- Increase brand awareness by 20% within 3 months.
- Drive 500 qualified leads per month.
- Achieve a Cost Per Acquisition (CPA) below $50.
I have narrowed down the choices to three vendors, each with a proposal. I need you to critically assess them to find the one that is most likely to *fail* or be the *least effective* in meeting our stated criteria.
**Vendor A Proposal Summary:**
- **Focus:** Broad social media reach (Facebook, Instagram).
- **Strategy:** High-volume content posting, influencer collaborations.
- **Reporting:** Monthly performance summary, engagement metrics.
- **Cost:** $8,000/month.
- **Key Weakness Identified:** Lacks specific lead generation tactics, relies heavily on vanity metrics.
**Vendor B Proposal Summary:**
- **Focus:** Search Engine Marketing (SEM) and targeted display ads.
- **Strategy:** Keyword bidding, retargeting campaigns, landing page optimization.
- **Reporting:** Weekly lead volume and CPA reports, conversion tracking.
- **Cost:** $10,000/month.
- **Key Weakness Identified:** Limited brand awareness component, potential for high CPA if keywords are too competitive.
**Vendor C Proposal Summary:**
- **Focus:** Content marketing and SEO.
- **Strategy:** Blog post creation, whitepaper development, organic search optimization.
- **Reporting:** Organic traffic growth, keyword rankings, content engagement.
- **Cost:** $7,000/month.
- **Key Weakness Identified:** Slow to show results, minimal focus on immediate lead generation or brand awareness.
# TASK
Your objective is **NOT** to pick the *best* vendor, but to identify **which vendor's proposal is the WORST and most likely to fail** based on our campaign goals.
Analyze each vendor proposal critically, focusing on their potential shortcomings and risks. Consider the following:
1. **Goal Alignment:** Which proposal least addresses or actively undermines our core goals (brand awareness, lead volume, CPA)?
2. **Strategic Weaknesses:** Where are the most significant gaps or flaws in their proposed strategy?
3. **Risk of Failure:** Which vendor is most likely to miss targets or prove to be a poor investment?
4. **Reporting Sufficiency:** Which vendor's reporting is least likely to provide actionable insights for our specific goals?
# OUTPUT FORMAT
1. **Critique of Vendor A:** Detail the specific reasons why Vendor A's proposal is the weakest, referencing the campaign goals and identified weaknesses.
2. **Critique of Vendor B:** Detail the specific reasons why Vendor B's proposal is the weakest, referencing the campaign goals and identified weaknesses.
3. **Critique of Vendor C:** Detail the specific reasons why Vendor C's proposal is the weakest, referencing the campaign goals and identified weaknesses.
4. **Final Verdict:** Clearly state which vendor is the *worst* choice and provide a concise, definitive explanation for why. Use the format: **"The worst vendor choice is Vendor [A/B/C] because..."**
What Should We Engineer in Prompts? Training Humans in Requirement-Driven LLM Use, arXiv:2409.08775
ROPE (Requirement-Oriented Prompt Engineering) is an approach to creating prompts as detailed technical specifications. Instead of simple questions, the user explicitly formulates the goal, step-by-step process, conditions, constraints, and output format. This transforms LLM from a 'conversationalist' into an 'executive intern', improving the quality and reliability of responses to complex tasks.
Use it whenUse ROPE when working with complex tasks requiring precise and predictable LLM results. Especially effective for tasks where response structure, multiple condition compliance, and step-by-step execution logic are important. Suitable for creating technical specifications, project planning, content development to specifications, data analysis with clear criteria. The method is critically important when standard prompts provide unpredictable or incomplete results, and when you need to train a team in effective LLM interaction for work tasks.
Prompt
# Роль:
Вы — опытный менеджер по закупкам и оценке поставщиков.
# Главная цель:
Выбрать наиболее подходящего поставщика для нового маркетингового проекта из трех предложенных вариантов, основываясь на предоставленных критериях.
# Входные данные (Контекст):
- **Проект:** Запуск новой рекламной кампании для продукта X.
- **Требуемые услуги:** Разработка креативных материалов (видео, баннеры), медиапланирование и размещение рекламы.
- **Бюджет:** Ограничен, требует тщательной оценки стоимости.
- **Сроки:** Жесткие, с финальным запуском через 6 недель.
# Поставщики:
## Поставщик A: "Креатив Плюс"
- **Специализация:** Сильная сторона — креативные концепции и видеопродакшн.
- **Опыт:** Работа с крупными FMCG брендами.
- **Стоимость:** Средняя, но с возможными доплатами за доп. услуги.
- **Отзывы:** Высокая оценка креативности, средняя — по соблюдению сроков.
## Поставщик B: "Медиа Стратеги"
- **Специализация:** Сильная сторона — медиапланирование, закупка трафика, аналитика.
- **Опыт:** Работа с технологическими стартапами.
- **Стоимость:** Ниже рынка, но с фокусом на медиабайинг.
- **Отзывы:** Отличная эффективность медиаразмещения, слабая — по разработке креатива.
## Поставщик C: "Полный Цикл"
- **Специализация:** Комплексные решения "под ключ" (креатив + медиа).
- **Опыт:** Средние и малые бизнесы.
- **Стоимость:** Фиксированная, включает все услуги.
- **Отзывы:** Стабильное качество, предсказуемость, но без "вау-эффекта".
# Критерии оценки (Приоритет: Высокий):
1. **Соответствие требованиям проекта:** Насколько полно поставщик покрывает нужды в креативе, медиапланировании и размещении.
2. **Стоимость и бюджет:** Общая стоимость предложения в рамках выделенного бюджета.
3. **Соблюдение сроков:** Гарантия выполнения всех этапов в установленные 6 недель.
4. **Качество исполнения:** Ожидаемое качество креативных материалов и эффективность медиаразмещения.
5. **Опыт и репутация:** Релевантный опыт работы и отзывы других клиентов.
# Требования к процессу принятия решения:
1. **Анализ каждого поставщика:** Оцени каждого поставщика отдельно по всем 5 критериям.
2. **Сравнительный анализ:** Создай таблицу для прямого сравнения поставщиков по каждому критерию.
3. **Оценка рисков:** Для каждого поставщика выдели 1-2 основных риска, связанных с его выбором.
4. **Рекомендация:** На основе анализа и оценки рисков, дай четкую рекомендацию, какого поставщика выбрать, и обоснуй свой выбор. Если ни один не подходит идеально, предложи компромиссное решение (например, комбинация поставщиков).
# Требования к формату вывода:
- Ответ должен быть структурирован с использованием заголовков ## и ###.
- Используй Markdown-таблицы для сравнения.
- Рекомендация должна быть выделена жирным шрифтом.
- Общий объем ответа — не более 500 слов.
# Триггер к действию:
Начинай анализ.
Q-STRUM Debate: Query-Driven Contrastive Summarization for Recommendation Comparison, arXiv:2502.12921
The method forces the LLM to simulate a structured debate between two imaginary opponents, each defending one alternative and criticizing the other. This generates deeper, more contrasting and balanced comparative analysis than standard comparison requests.
Use it whenUse this technique when you need deep comparative analysis of two alternatives, products, solutions, or concepts. Particularly effective in business decision-making, choosing between strategies, evaluating candidates or product options. The method helps identify non-obvious pros and cons of each side that a standard comparison request might miss. Ideal for situations where balanced evaluation is important with consideration of counterarguments and critical analysis of weaknesses of each option
Order Matters: Investigate the Position Bias in Multi-constraint Instruction Following, arXiv:2502.17204
A technique for ordering instructions in a prompt following the "hard-to-easy" principle. Complex constraints (word prohibitions, structure) are placed at the beginning of the instruction list, simple ones (capitalization, final phrase)—at the end. This forces the model to immediately account for difficult requirements during generation, improving accuracy of meeting all conditions.
Use it whenUse this technique when a prompt contains multiple instructions and constraints, especially if the model fails to execute them simultaneously. Ideal for tasks with complex requirements for text format, style and content. Apply when creating content with strict constraints (word prohibitions, character limits, specific structure). Especially effective for generating text where compliance with all conditions simultaneously is important. The method is universal and works with any LLM without
E-GEO: A Testbed for Generative Engine Optimization in E-Commerce, arXiv:2511.20867
A method of structured argumentation for convincing an LLM. The prompt is built as a briefing for an expert: evaluation criteria are set, direct competitor comparison is required, unique advantages are highlighted and social proof is analyzed. The model transitions from listing facts to reasoned expert recommendations, acting as a consultant rather than a reference guide.
Researchers found that there is a universal prompt structure for persuasion that works 10 times more effectively than intuitive attempts.checked against the paper
Use it whenUse this technique when you need to get from an LLM not just a list of facts, but a reasoned expert recommendation. Particularly effective in e-commerce for optimizing product descriptions, in marketing for creating persuasive content, when developing recommendation systems. Apply when the model needs to compare alternatives, highlight unique advantages and justify the choice. Ideal for situations where analysis depth and recommendation personalization are important, and n
TypedThinker: Diversify Large Language Model Reasoning with Typed Thinking, arXiv:2410.01952
A method of explicitly specifying the type of thinking in the prompt: deductive, inductive, abductive, or by analogy. Allows managing the LLM reasoning process, taking the model beyond the standard deductive approach. Expands the range of solvable tasks and improves answer quality through a conscious choice of cognitive strategy.
Use it whenUse this technique when standard prompts fail to handle a task or provide superficial answers. Especially effective for complex analytical tasks, finding causal relationships, identifying patterns from disparate data, hypothesis generation, and creative problem-solving. Apply when needing to go beyond the model's deductive thinking: use induction for generalizations, abduction for cause-finding, analogy for solution transfer from other domains. Suitable for business analysis, research, strategic planning, and any tasks requiring diverse cognitive approaches.
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.