mlfoundations-dev/thoughts-lawma-annotations-claude

该数据集包含了用户的观点(opinion)、指示(instruction)、问题(question)、选项(choices)、答案(answer)、问题文本(problem)、思考轨迹(claude_thinking_trajectory)、尝试(claude_attempt)以及是否验证(verify)等字段。数据集被拆分为训练集(train)，共有4个示例，数据集大小为103972字节。

The dataset includes fields such as user opinions (opinion), instructions (instruction), questions (question), options (choices), answers (answer), problem text (problem), thinking trajectory (claude_thinking_trajectory), attempts (claude_attempt), and verification (verify). The dataset is split into a training set (train) with a total of 4 examples, and the dataset size is 103972 bytes.