Think Out Loud
Step-back prompting and Chain of Thought
Show Your Work 📝
In math class, your teacher always says: "Show your work." Not because they enjoy reading your scratch paper — but because step-by-step thinking leads to correct answers. When you just guess, you're wrong more often. When you work through it step by step, you catch your own mistakes and arrive at the right answer. AI is exactly the same. Left to itself, AI will blurt out an answer without checking its logic. But when you ask it to "think step by step", something remarkable happens: it gets smarter.
😱 The Shocking Difference CoT Makes
Watch what happens with and without Chain of Thought on the SAME question:
🎯 When Should You Use Chain of Thought?
Rule of thumb: if a human would need to "think it through", use CoT.
🔭 Step-Back Prompting: Zoom Out Before Zooming In
Sometimes before solving a specific problem, it helps to zoom out and think about the bigger picture first. This is Step-Back Prompting. Instead of diving straight into your specific question, you first ask a broader, more general question to activate relevant knowledge. Then use THAT answer as context for your real question.
The step-back produced a much richer, more intentional result!
✨ Magic Phrases That Activate Deep Thinking
Quick Quiz
Test what you learned from this chapter. No pressure — it's just for you!
What does adding "Let's think step by step" to a prompt do?