Learn Before
Designing a Training Dataset for a Reliable Q&A System
You are tasked with improving a question-answering system that generates answers from a set of provided documents. The system sometimes fabricates answers when the documents do not contain the necessary information. To fix this, you decide to fine-tune the underlying language model. Describe the key characteristics of the data examples you would need to create for this fine-tuning process to specifically teach the model to refuse to answer when appropriate.
0
1
Tags
Ch.3 Prompting - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Analysis in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Improving AI System Reliability
A company develops a question-answering system that uses a large language model to answer queries based on a private document collection. They observe that when a user asks a question for which no relevant documents are retrieved, the system often generates a plausible-sounding but factually incorrect answer. Which of the following training approaches is specifically designed to mitigate this problem?
Designing a Training Dataset for a Reliable Q&A System