Choosing an Initial Prompt Generation Strategy
Given the team's situation described below, which of the two primary methods for using a large language model to generate an initial set of prompts would you recommend: providing the model with a human-written task description or providing it with the input-output examples? Justify your recommendation by evaluating the suitability of each method in this specific context.
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
Choosing an Initial Prompt Generation Strategy
A machine learning engineer is tasked with creating a set of initial prompts for a novel text summarization task. The goal is to summarize complex legal documents into plain language, a process for which a precise, formal description is challenging to write. However, the engineer has access to a large, curated dataset of several hundred legal documents and their corresponding expert-written plain-language summaries. Which initialization strategy would be most effective in this situation, and why?
Diagnosing Prompt Initialization Failure