Learn Before
Revising a Prompt for Generalization
A developer is building a tool to extract key financial figures (revenue, profit, loss) from quarterly earnings reports. They find the model is not performing well and sometimes misses figures or hallucinates values. Analyze their current prompt and, based on the principle of guiding a model using only instructions without providing examples, propose a revised prompt that is more likely to yield accurate results.
0
1
Tags
Ch.1 Pre-training - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Ch.2 Generative Models - Foundations of Large Language Models
Ch.3 Prompting - Foundations of Large Language Models
Application in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Iterative Prompt Adjustment in Zero-Shot Learning
Example of a Persona-based Prompt for Grammar Correction
Origin of Zero-Shot Learning Ability in LLMs
Example of a Zero-Shot Prompt for Grammar Correction
A developer wants a large language model to classify customer feedback. They provide the model with the following prompt:
You are an expert sentiment analysis system. Classify the following customer review as 'Positive', 'Negative', or 'Neutral'. Provide only the label. Review: 'The battery life is impressive, but the screen is too dim.'Which of the following statements best explains why this approach tests the model's ability to generalize to a new task based on instructions alone?Revising a Prompt for Generalization
A research team is testing a large language model's ability to perform a task it has not been specifically trained on: summarizing news articles into a single sentence. Which of the following prompts is a clear example of a zero-shot approach?