Synthetic Data
Synthetic data is generated by models or simulators for training or eval — e.g., LLM-written instruction pairs.
It scales labeling but can amplify teacher biases and hallucinations. Filter aggressively and always validate on real held-out examples.