New benchmark tests AI code reviewers against real human feedback

Original: Evaluating code review agents with ReviewBench

Source: langchain.com ↗