haoranli-ml/AIME2025_Qwen3-4B-Instruct-raw-sft_32768budget_16rollouts

该数据集包含四个字段：问题(problem)，答案(answer)，分词后的提示(tokenized_prompt)和响应(responses)列表。这些问题和答案可能用于某种问答系统，分词后的提示可能是预处理后的输入文本，响应列表可能包含一系列的回答选项或相关文本。数据集仅包含训练集(train)，共有30个示例，数据大小为21788235字节。数据集使用默认配置，并提供训练数据文件的路径。

The dataset includes four fields: problem, answer, tokenized_prompt, and a list of responses. These problems and answers might be used for some question-answering system, the tokenized prompts could be preprocessed input texts, and the response list may contain a series of answer options or related texts. The dataset only contains a training set with 30 examples, totaling 21788235 bytes in size. The dataset uses a default configuration and provides the path to the training data files.