Chain-of-thought prompting: Worked Example — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)

Chain-of-Thought Prompting: A Worked Example for NVIDIA-Certified Professional: Generative AI LLMs

Chain-of-thought prompting is a powerful technique used to improve the reasoning capabilities of large language models (LLMs) by guiding them to generate intermediate steps before arriving at a final answer. This approach is critical for complex problem-solving tasks and is a key topic in the NVIDIA-Certified Professional: Generative AI LLMs exam.

Scenario: Solving a Multi-Step Math Word Problem

Consider the following problem:

"A bookstore sold 120 books in one day. If the number of fiction books sold was twice the number of non-fiction books, how many fiction books were sold?"

We will apply chain-of-thought prompting to guide the LLM through the reasoning process step-by-step.

Step 1: Define the Prompt Structure

We design a prompt that explicitly instructs the model to break down the problem into logical steps. For example:

"Solve the following problem step-by-step. Explain your reasoning at each step before giving the final answer. Problem: A bookstore sold 120 books in one day. The number of fiction books sold was twice the number of non-fiction books. How many fiction books were sold?"

Step 2: Model Generates Intermediate Reasoning

The model is expected to produce a chain of reasoning such as:

• Let x be the number of non-fiction books sold.
• Then the number of fiction books sold is 2x.
• The total books sold is fiction plus non-fiction: x + 2x = 3x.
• We know the total is 120, so 3x = 120.
• Solving for x, we get x = 120 / 3 = 40.
• Therefore, fiction books sold = 2x = 2 * 40 = 80.

Step 3: Final Answer

The model concludes:

Answer: 80 fiction books were sold.

Step 4: Evaluating the Chain-of-Thought Prompting Effectiveness

This explicit stepwise reasoning helps the model avoid common pitfalls such as jumping to conclusions or misinterpreting relationships in the problem. It also makes the model's thought process transparent, which is valuable for debugging and trust.

Worked Example Summary

1. Prompt design: Instruct the model to explain each step before answering.
2. Intermediate steps: Define variables, express relationships, form equations.
3. Calculation: Solve the equation to find the unknown.
4. Conclusion: State the final answer clearly.

Conclusion

Chain-of-thought prompting enhances LLM performance on reasoning tasks by decomposing complex problems into manageable steps. For NVIDIA-Certified Professionals working with generative AI LLMs, mastering this technique is essential for designing prompts that elicit accurate, interpretable outputs in real-world applications.

More in this topic

Chain-of-thought prompting: Common Mistakes — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Zero-, one-, and few-shot learning: Quick Reference — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Prompt Engineering — NVIDIA-Certified Professional: Generative AI LLMsZero-, one-, and few-shot learning: Common Mistakes — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Controlling model output — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Zero-, one-, and few-shot learning: Worked Example — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Chain-of-thought prompting — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Domain adaptation via prompting — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Chain-of-thought prompting: Quick Reference — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Zero-, one-, and few-shot learning — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Zero-, one-, and few-shot learning: Practice Questions — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)Chain-of-thought prompting: Practice Questions — Prompt Engineering (NVIDIA-Certified Professional: Generative AI LLMs)