Learn Before
Case Study

Evaluating and Improving Prompt Demonstrations

A developer is using a large language model to classify customer support tickets. The goal is to categorize each ticket as 'Billing Inquiry', 'Technical Support', or 'General Question'. The model is performing poorly, frequently misclassifying the tickets. Analyze the example input-output pairs provided in the prompt below. Identify at least two significant flaws in these examples and explain how these flaws likely contribute to the model's poor performance. Finally, suggest specific improvements.

0

1

Updated 2025-10-08

Contributors are:

Who are from:

Tags

Ch.1 Pre-training - 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