Syed Khundmir Azmi, “Trust but Verify: Benchmarks for Hallucination, Vulnerability, and Style Drift in AI-Generated Code Reviews”, WT, vol. 32, no. 1, pp. 76–90, Feb. 2023.