AWS Certified AI Practitioner (AIF-C01) — Question 297
A company is introducing a new feature for its application. The feature will refine the style of output messages. The company will fine-tune a large language model (LLM) on Amazon Bedrock to implement the feature.
Which type of data does the company need to meet these requirements?
Answer options
- A. Samples of only input messages
- B. Samples of only output messages
- C. Samples of pairs of input and output messages
- D. Separate samples of input and output messages
Correct answer: C
Explanation
To fine-tune a large language model (LLM) on Amazon Bedrock for style refinement, the training process requires labeled data consisting of paired prompt-response examples. Providing only inputs or only outputs does not allow the model to learn the association between the original text and the desired style. Unpaired datasets also fail because the model needs direct input-to-output mappings to learn how to transform the message style effectively.