hazyresearch/m07d11_longhealth_synthesize_qwen3-4b_p10_n65536-1

该数据集包含了对话信息，每个对话示例包括内容(content)、角色(role)、token IDs、以及对每个token的概率分布(logprobs)等。此外，还包含了系统提示(system_prompt)和元数据(metadata)，如初始系统提示(initial_system_prompt)、种子提示(seed_prompt)以及工具调用(tool_calls)。数据集分为训练集(train)，共有65536个示例，大小为3.3GB。数据集的下载大小为826MB。

The dataset consists of dialogue information, with each example including content, role, token IDs, and probability distributions for each token. It also contains system prompts and metadata such as the initial system prompt, seed prompt, and tool calls. The dataset is split into a training set (train) with 65,536 examples and a size of 3.3GB. The download size of the dataset is 826MB.