stefanocarrera/autophagycode_D_he_train-mercury_Qwen3-4B_strategy_trust_t1.1_g8_run0_metrics
收藏资源简介:
该数据集包含164个训练示例,主要用于代码分析和编程任务评估。每个示例包含多个特征,如任务ID(task_id)、代码入口点(entry_point)、可执行性(is_executable)、正确性(is_correct)、测试通过和失败数量(tests_passed、tests_failed),以及代码复杂度指标(如Halstead度量、循环复杂度、可维护性指数、代码行数LOC、注释比例等)。此外,还包括词汇多样性(TTR)、信息熵(shannon_entropy)和预测熵等统计特征。数据集旨在支持代码质量评估、自动化测试和程序分析研究。
This dataset contains 164 training examples, primarily designed for code analysis and programming task evaluation. Each example includes multiple features such as task ID (task_id), code entry point (entry_point), executability (is_executable), correctness (is_correct), number of tests passed and failed (tests_passed, tests_failed), and code complexity metrics (e.g., Halstead measures, cyclomatic complexity, maintainability index, lines of code LOC, comment percentage). Additionally, it incorporates statistical features like type-token ratio (TTR), Shannon entropy, and predictive entropy. The dataset aims to support code quality assessment, automated testing, and program analysis research.




