Prompt Engineering Lab

See exactly how prompt strategy affects LLM output — zero-shot vs few-shot vs chain-of-thought vs ReAct.

Prompt Strategy

⚡ Zero-Shot

Direct question with no examples — relies purely on model knowledge.

📖 Few-Shot

Provide 2–3 examples before the question to guide the model's format and reasoning style.

🧠 Chain-of-Thought

Instruct the model to think step-by-step before answering — improves reasoning accuracy.

🔄 ReAct

Reason + Act pattern: the model alternates between Thought, Action, and Observation before concluding.

Model

System Prompt Preview

You are a portfolio assistant for Priyabrat Dalbehera's personal website. === ABOUT === Name: Priyabrat Dalbehera Role: GenAI & Agentic AI Engineer with 4.1+ years building systems that turn business data and knowledge into automated decisions — NL2SQL analytics, RAG knowledge systems, AI agents, document intelligence, and AI reliability tooling. Open to: freelance AI projects and consulting (primary focus), plus full-time AI Engineer roles and research collaborations. === CONTACT === Email: ipriyabrat689@gmail.com Phone: +91 77490 86836 GitHub: github.com/iampriyabrat14 → https://github.c…

Quick questions

Your question

Select a strategy, ask a question, and hit Run.

Compare All Strategies on Same Question

Run all 4 strategies and see which produces the best answer for your question.

—

Zero-Shot

—

Few-Shot

—

Chain-of-Thought

—

ReAct