Fine-tuning and LLMOps · Question 56 of 60
For fine-tuning, a team uses only test cases from a project in which negative tests hardly occurred. What is the most likely consequence?
1 answer is correct. Hard Created with AI
Show answer
Answer
- B The model itself generates hardly any negative tests and adopts this imbalance
Explanation
The model learns the patterns of its training data. One-sided data leads to one-sided, biased results. That is why fine-tuning needs high-quality, balanced and task-related datasets.