AUTOMATED UNIT TEST IMPROVEMENT USING LARGE LANGUAGE MODELS AT META

Summary from the TLDR AI newsletter

Meta used large language models to write tests for its code base. It found substantial improvements in test coverage and overall code quality.