laion/eval-laion_loopshape-30-8B_DCAgent2_terminal_bench_2-traces
收藏资源简介:
该数据集是一个包含对话交互和任务执行记录的数据集,主要用于人工智能代理和模型的评估与训练。数据集特征包括对话内容(conversations,含角色和文本)、代理(agent)、模型(model)、模型提供商(model_provider)、日期(date)、任务(task)、事件(episode)、运行ID(run_id)、试验名称(trial_name)、结果(result)、验证器输出(verifier_output)和追踪来源(trace_source)。数据集分割为训练集(train),包含1643个示例,总大小约144.856 MB,适用于自然语言处理和多任务学习研究。
This dataset contains dialogue interactions and task execution records, primarily used for evaluating and training AI agents and models. The features include conversations (with role and content), agent, model, model_provider, date, task, episode, run_id, trial_name, result, verifier_output, and trace_source. The dataset is split into a train set with 1643 examples and a total size of approximately 144.856 MB, suitable for natural language processing and multi-task learning research.




