DCAgent3/medagentbench_a3_rl_DCAgent_exp_rpt_e2egit_v2_10_8B_20260527_202829
收藏资源简介:
该数据集是一个多轮对话数据集,包含1364个训练示例,用于AI代理或模型评估。数据特征包括对话记录(conversations,含角色和内容)、代理类型(agent)、模型信息(model和model_provider)、日期(date)、任务类型(task)、运行标识(如episode、run_id、trial_name)、结果(result)、验证输出(verifier_output)和来源(trace_source)。数据集仅提供训练分割,总大小约为61.5 MB,适用于自然语言处理任务,如对话系统分析和性能测试。
This dataset is a multi-turn conversation dataset containing 1,364 training examples, designed for AI agent or model evaluation. It features conversation records (conversations, including role and content), agent type (agent), model information (model and model_provider), date (date), task type (task), run identifiers (such as episode, run_id, trial_name), result (result), verifier output (verifier_output), and source (trace_source). The dataset includes only a training split, with a total size of approximately 61.5 MB, and is suitable for natural language processing tasks, such as dialogue system analysis and performance testing.




