Introduction to machine learning · Question 34 of 60
What happens to biases and vulnerabilities of a pretrained model when it is adapted to a new task?
1 answer is correct. Medium
Show answer
Answer
- B They typically carry over to the new model, which is why testing is necessary
Explanation
Whatever biases and weaknesses are in the pretrained model, it usually brings them into the new task. So you have to check whether the adapted model works reliably and fairly there, regardless of whether it was adapted by fine-tuning or by RAG.