CoReDD Bench: An Open Benchmark for Defect Detection in Code Review with Quantified Label Error

CoReDD Bench measures whether the defective lines in a proposed change can be identified before it is merged. A detector receives the change as it stood before integration and marks the lines it considers defective; those marks are scored against the lines that caused a later correction. Labels are reconstructed rather than annotated. For every correction observed after integration, language model agents search the commit history for the change that introduced the defect and mark the lines within it that caused the correction. Reconstruction is stochastic and imperfect, so a manual audit estimates its error rate and that estimate is propagated into every reported score as an interval. The release covers 360 reviewed pull requests from six Python repositories with observable review and integration workflows. It contains the labels, the audit records, the uncertainty model, and the evaluation software. The repositoryhistories and timestamped GitHub review context the labels were reconstructed from are archived separately as CoReDD Corpus. Results hold within the documented scope. The benchmark makes no claim to represent code review practice at large.