Which technique helps reduce hallucination in large language models?
A) Chain-of-verification
B) Data augmentation
C) Transfer learning
D) Adversarial training