DCAgent2/GLM-4.7-swe_rebench_patched_oracle-10samples-131k

该数据集包含多轮对话记录和相关元数据，用于分析任务执行和模型交互。特征包括对话内容（含角色和文本）、代理信息、模型及提供者、日期、任务类型、事件标识、运行ID、试验名称、执行结果和验证器输出，适用于自然语言处理和机器学习研究。

This dataset contains multi-turn conversation records and associated metadata for analyzing task execution and model interactions. Features include conversation content (with roles and text), agent information, model and provider details, date, task type, episode identifier, run ID, trial name, execution result, and verifier output, suitable for natural language processing and machine learning research.