GENERATIVE-AI-ENGINEER-ASSOCIATE · Question #13
A Generative AI Engineer at an automotive company would like to build a question-answering chatbot to help customers answer specific questions about their vehicles. They have: - A catalog with…
The correct answer is C. Fine-tuning an embedding model on automotive terminology. Fine-tuning an embedding model on automotive terminology is generally unnecessary when using large open-source LLMs combined with retrieval-augmented generation techniques. Instead, embedding models are typically pre-trained and sufficient for encoding domain-specific content…
Question
A Generative AI Engineer at an automotive company would like to build a question-answering chatbot to help customers answer specific questions about their vehicles. They have:
- A catalog with hundreds of thousands of cars manufactured since the
1960s
- Historical searches, with user queries and successful matches
- Descriptions of their own cars in multiple languages
They have already selected an open source LLM and created a test set of user queries. They need to discard techniques that will not help them build the chatbot. Which do they discard?
Options
- ASetting chunk size to match the model's context window to maximize coverage
- BImplementing metadata filtering based on car models and years
- CFine-tuning an embedding model on automotive terminology
- DAdding few-shot examples for response generation
How the community answered
(59 responses)- A5% (3)
- B24% (14)
- C61% (36)
- D10% (6)
Explanation
Fine-tuning an embedding model on automotive terminology is generally unnecessary when using large open-source LLMs combined with retrieval-augmented generation techniques. Instead, embedding models are typically pre-trained and sufficient for encoding domain-specific content without expensive fine-tuning. The other techniques directly improve retrieval quality and response relevance.
Topics
Community Discussion
No community discussion yet for this question.