Learn Before
Choosing the Right Prompt for the Job
A middle school teacher wants to use a large language model to quickly generate an answer key for a 50-question quiz on basic multiplication (e.g., '7 x 8'). The teacher's primary goal is to get accurate answers as efficiently as possible, without needing to see the model's reasoning process.
Below are two potential instruction formats. Which format is more appropriate for the teacher's goal? Justify your choice by explaining why your selected format is superior and the other is less suitable for this specific task.
Format A: 'Solve the following math problem: {problem}' Format B: 'You are an expert mathematician. Please provide a detailed, step-by-step explanation of how to solve the following problem, concluding with the final answer. {problem}'
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Computing Sciences
Foundations of Large Language Models Course
Evaluation in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Step 1 of a Mathematician Persona's Reasoning Process: Problem Interpretation
A user is attempting to have a language model solve a complex, multi-step physics problem that requires deriving a formula and then applying it. The user provides the following instruction: "Solve the problem below. {problem}". The model provides an incorrect final answer with flawed reasoning. Based on the structure of the instruction, what is the most likely reason for the model's failure?
Choosing the Right Prompt for the Job
The following prompt is an example of a direct instruction format for a mathematical challenge: 'Carefully read the math problem below. First, identify all the variables and their given values. Second, outline the step-by-step plan you will use to solve it. Finally, execute your plan and state the final answer. {problem}'