Learn Before
A developer needs a language model to summarize lengthy technical reports into a single, easy-to-understand sentence. Consider two approaches for structuring the prompt:
Approach 1:
[Instruction: Summarize the following report in one sentence.] + [The full technical report]
Approach 2:
[Instruction: Summarize the following report in one sentence for a non-technical audience.] + [Example Report 1] + [Example Summary 1] + [Example Report 2] + [Example Summary 2] + [The full technical report to be summarized]
Which approach is more likely to yield consistently better results, and what is the most accurate reason?
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Evaluation in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Deconstructing a Model's Input
Analyze the following prompt provided to a language model. What is the primary function of the section containing 'Review: "The plot was predictable and the acting was wooden." / Sentiment: Negative' within the overall structure?
Prompt Start
Instruction: Classify the sentiment of the following movie review as 'Positive', 'Negative', or 'Neutral'.
Review: "The plot was predictable and the acting was wooden." Sentiment: Negative
Review: "An absolute masterpiece of modern cinema. I was captivated from start to finish." Sentiment: Positive
Review: "The movie was two hours long and was filmed in color." Sentiment:
Prompt End
A developer needs a language model to summarize lengthy technical reports into a single, easy-to-understand sentence. Consider two approaches for structuring the prompt:
Approach 1:
[Instruction: Summarize the following report in one sentence.] + [The full technical report]Approach 2:
[Instruction: Summarize the following report in one sentence for a non-technical audience.] + [Example Report 1] + [Example Summary 1] + [Example Report 2] + [Example Summary 2] + [The full technical report to be summarized]Which approach is more likely to yield consistently better results, and what is the most accurate reason?
A well-structured input for a language model often contains several distinct components that work together. Match each component with its primary function.