Synthetic examples will be used for model customization. What should the data preparation process include?

1. Include the held-out test answers in training to maximize the reported score
2. Discard all source and generation context
3. Assume synthetic origin guarantees every example is safe and correct
4. Quality, provenance, privacy, bias, and contamination checks