Structured Prompts Boost LLM Code Review Reliability

Meta researchers developed a structured prompting technique enabling large language models to verify code patches.