Intelligent-Internet/II-Search-Benchmark-Details

这是一个包含问题、答案和相关追踪信息的搜索模型性能评估数据集。数据集包含了问题文本、最终答案、完整的追踪记录（包括内容、函数调用、推理内容、角色等）、响应时间、搜索和访问调用次数、函数调用总数、思考条目数、对话轮数、成功与否的标志、错误信息、重试次数、真实答案、上下文信息、提取的答案、答案的正确性、提取答案的正确性、使用的模型名称和数据集名称等字段。数据集分为训练集，并提供了相应的数据文件路径。

This is a search model performance evaluation dataset containing questions, answers, and related trace information. The dataset includes fields such as question text, final answer, complete trace records (including content, function calls, reasoning content, role, etc.), response time, number of search and visit calls, total number of function calls, number of thinking entries, number of turns in the conversation, success flag, error information, retry count, ground truth, context information, extracted answer, correctness of the answer, correctness of the extracted answer, name of the model used, and dataset name. The dataset is split into a training set and provides the corresponding data file paths.