autophagycode_D_metrics_he_Qwen3-14B_lr0.0001_correct_g5

该数据集包含164个训练样本，总大小18,740字节，记录了代码执行质量与软件复杂度指标。每个样本包含14个结构化字段：任务ID(task_id)、入口函数(entry_point)、可执行状态(is_executable)、正确性(is_correct)、通过/失败的测试案例数(tests_passed/failed)、测试运行时间(test_run_time_ms)、错误类型(error_type)。此外包含5项Halstead代码复杂度指标（词汇量、长度、体积、难度、工作量）和可维护性指数(maintainability_index)。数据集适用于代码质量分析、自动测试验证和软件度量研究。This dataset consists of 164 training samples with an overall size of 18,740 bytes, documenting code execution quality and software complexity metrics. Each sample contains 14 structured fields: task_id, entry_point, is_executable, is_correct, the count of passed/failed test cases (tests_passed/failed), test_run_time_ms, and error_type. Furthermore, it includes five Halstead code complexity metrics (vocabulary, length, volume, difficulty, and effort) alongside the maintainability index (maintainability_index). This dataset is applicable to code quality analysis, automated test validation, and software metric research.