Answer-first summary for fast verification
Answer: Fine-tuning
## Explanation **Fine-tuning (Option B)** is the best approach because: - **Large dataset availability**: The company has "thousands of labeled examples" which is sufficient for fine-tuning - **Customization requirement**: They need to "tailor an existing foundation model" to match their specific "brand tone" - **Amazon Bedrock support**: Amazon Bedrock provides fine-tuning capabilities for foundation models **Why other options are less suitable**: - **Reinforcement learning (A)**: Typically used for optimizing model behavior through reward systems, not for initial customization with labeled data - **Few-shot prompting (C)**: Uses a small number of examples (typically 2-10) in prompts, but the company has thousands of examples - **Zero-shot inference (D)**: Uses no examples, just instructions, which wouldn't leverage the available labeled dataset Fine-tuning allows the model to learn from the extensive labeled dataset and adapt its output style to match the company's specific brand requirements.
Author: Ritesh Yadav
Ultimate access to all questions.
A company wants to tailor an existing foundation model on Amazon Bedrock to generate product descriptions aligned with its brand tone. They already have thousands of labeled examples. Which approach is best suited?
A
Reinforcement learning
B
Fine-tuning
C
Few-shot prompting
D
Zero-shot inference
No comments yet.