DCAgent2/swebench_verified_random_100_folders_Kimi_K2_5_20260429_192401

该数据集是一个包含对话交互记录的数据集，用于训练或评估人工智能代理。每个样本包括对话内容（conversations，含内容和角色字段）、代理标识（agent）、模型信息（model和model_provider）、日期（date）、任务类型（task）、集数（episode）、运行ID（run_id）、试验名称（trial_name）、结果（result）和验证器输出（verifier_output）。数据集共有300个训练样本，总大小约52.96 MB，适用于自然语言处理任务，如对话系统开发和模型性能分析。

This dataset is a collection of dialogue interaction records designed for training or evaluating AI agents. Each sample includes conversations (with content and role fields), agent identifier, model information (model and model_provider), date, task type, episode number, run ID, trial name, result, and verifier output. The dataset contains 300 training examples with a total size of approximately 52.96 MB, suitable for natural language processing tasks such as dialogue system development and model performance analysis.