Learn Before
Improving Language Model Performance on a Multi-Step Task
Based on the principles of guiding a model's reasoning process, explain why the analyst's current prompt is likely ineffective. Then, describe how you would restructure this prompt by adding a single, detailed demonstration to improve the model's accuracy and consistency for this task.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Application in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Few-Shot Chain-of-Thought (COT) Prompting
A user wants a language model to solve multi-step logic puzzles. They provide the following prompt. Analyze the structure of the demonstration provided within this prompt.
[START OF PROMPT] Q: A farmer has 15 sheep. All but 8 died. How many are left? A: Let's think step by step. 1. The phrase "All but 8 died" is a bit tricky. It means that 8 sheep are the ones that did *not* die. 2. Therefore, the number of sheep left is 8. The final answer is 8. Q: In a race, a runner overtakes the person in 2nd place. What position is the runner in now? [END OF PROMPT]What is the key element in the provided demonstration that guides the model to reason through the new puzzle?
Constructing a One-Shot Chain-of-Thought Prompt
Improving Language Model Performance on a Multi-Step Task
Example of One-Shot Chain-of-Thought (COT) Prompting