Make the model reason instead of blurting an answer. Add a single line to unlock step-by-step thinking, give worked examples, and use the step-back move to reason from principles first.
Why: on multi-step problems, models often jump straight to a guess and get it wrong. When: any time the task needs arithmetic, logic, or several dependent steps, an immediate answer is a red flag. Run this — many models miss it.
A bat and a ball cost $1.10 in total. The bat costs $1.00 more
than the ball. How much does the ball cost?
Answer with just the number.Why: the phrase "think step by step" tells the model to show its working, which dramatically improves accuracy on reasoning tasks. When: add it to any logic, math, or planning prompt. Where: put it right after the question.
A bat and a ball cost $1.10 in total. The bat costs $1.00 more
than the ball. How much does the ball cost?
Let's think step by step.Why: showing one fully worked example teaches the model the reasoning style you want, not just the answer format. When: use it for domain-specific reasoning where "step by step" alone wanders. Where: include the reasoning in your example, then leave the final one open.
Q: A shop sells pens at 3 for $2. How much for 9 pens?
A: 9 pens is 3 groups of 3. Each group costs $2.
3 groups x $2 = $6. Answer: $6.
Q: A train travels 60 km in 45 minutes. What is its speed in km/h?
A:Why: stepping back to a general principle before the specifics gives the model a sturdier foundation and fewer wrong turns. When: use it for hard or broad questions where diving straight in produces shallow answers. How: ask for the principle first, then the answer.
First, step back: what general physics principle governs how far
a projectile travels?
Then use that principle to answer:
A ball is kicked at 20 m/s at a 30-degree angle. Roughly how far
does it travel before landing? Show the reasoning.