Prompting Techniques

Chain-of-Thought Prompting

Quick Answer: A prompting technique where the model is instructed to break down complex problems into intermediate reasoning steps before arriving at a final answer.

Chain-of-Thought Prompting is a prompting technique where the model is instructed to break down complex problems into intermediate reasoning steps before arriving at a final answer. This mimics human step-by-step reasoning and significantly improves accuracy on math, logic, and multi-step tasks.

Example

Prompt: 'Think step by step. A store has 45 apples. They sell 12 in the morning and receive a shipment of 30 in the afternoon. How many do they have?' The model responds with each calculation step before the final answer: 63.

Why It Matters

Research from Google shows chain-of-thought prompting can improve accuracy by 20-40% on reasoning tasks compared to direct prompting, especially with larger models.

How It Works

Chain-of-thought (CoT) works because language models process text sequentially. When a model generates intermediate reasoning steps, each step adds context that improves the accuracy of subsequent steps. This is analogous to how humans solve math problems by writing out their work.

There are several CoT variants. Zero-shot CoT uses the simple trigger 'Let's think step by step.' Few-shot CoT provides example problems with worked-out solutions. Tree-of-thought explores multiple reasoning paths and selects the best one. Self-consistency generates multiple CoT paths and takes a majority vote on the final answer.

Research shows CoT provides the biggest gains on tasks requiring arithmetic, commonsense reasoning, and symbolic manipulation. The improvement is proportional to model size, with smaller models sometimes performing worse with CoT than without it.

Common Mistakes

Common mistake: Using chain-of-thought for simple factual questions where it adds unnecessary verbosity

Reserve CoT for multi-step reasoning tasks. For simple lookups, direct prompting is faster and cheaper.

Common mistake: Providing chain-of-thought examples with incorrect reasoning steps

Verify every step in your few-shot examples is logically sound. Models will imitate flawed reasoning patterns.

Career Relevance

Chain-of-thought is a fundamental prompting technique that every prompt engineer must master. It appears in virtually every prompt engineering job description and is tested in technical interviews. It's also the foundation for understanding reasoning models like o1 and o3.

Related Terms

Learn More

Stay Ahead in AI

Join 1,300+ prompt engineers getting weekly insights on tools, techniques, and career opportunities.

Join the Community →