Databricks Generative AI Engineer Associate Question 129
Select 2You are developing a customer support chatbot using a large language model (LLM). The baseline response of the model is too formal, but your stakeholder wants the chatbot to sound more friendly and conversational. Which approach would help you achieve the desired tone in the responses?
- A
Add a system prompt like 'You are a friendly and conversational assistant. Respond casually to all questions.' to guide the LLM.
- B
Fine-tune the LLM with a dataset containing examples of friendly and conversational responses.
- C
Specify a temperature value of 0 to ensure the LLM outputs deterministic responses.
- D
Include a user prompt like 'Respond with a formal tone for this question.' to override the original tone.
- E
Use chain-of-thought prompting to improve the reasoning capabilities of the LLM.
Show answer and explanation
Correct answers: A, B
Explanation
To adjust an LLM's response from a baseline to a desired output, you can modify the system prompt to specify the desired tone or fine-tune the model with relevant examples. These approaches help guide the model's behavior and produce outputs that align with the stakeholder's requirements. Other options like temperature adjustment or chain-of-thought prompting are unrelated to tone modification.
- A. Correct.
This is correct. Adding a system prompt that specifies the desired tone allows the LLM to adjust its responses accordingly without requiring model fine-tuning.
- B. Correct.
This is correct. Fine-tuning the LLM with relevant examples of friendly and conversational language can help achieve the desired output consistently.
- C. Incorrect.
This is incorrect. Setting the temperature to 0 does not affect the tone, but rather controls the randomness of the responses.
- D. Incorrect.
This is incorrect. Including a user prompt for a formal tone contradicts the goal of making the responses more friendly and conversational.
- E. Incorrect.
This is incorrect. Chain-of-thought prompting is used for improving reasoning or step-by-step answers but is not directly related to adjusting tone.