Question

A financial services firm uses a foundation model chatbot to answer employee questions about internal policies. Employees complain that the model gives outdated answers because policies were revised after the model's training data cutoff. The firm wants an approach that lets the model reference the latest policy documents at query time without retraining the model. Which approach best addresses this requirement?

Accepted Answer

Implement Retrieval Augmented Generation so the model retrieves current policy documents from a knowledge base and includes them in the prompt context at query time

Answer

Increase the model's temperature parameter so it generates more varied and up-to-date responses

Answer

Continue pre-training the foundation model from scratch each time a policy changes

Answer

Add more few-shot examples of old policy questions to the prompt to improve accuracy