yuruny/agentic-sudoku-Markov_qwen2.5-3B-it-5e-6_9x9_6-6_gt-SFT_ans1-6k-eval_results_dev

该数据集包含多个特征，其中steps是一个列表，包含动作、聊天完成、完成状态、mc_return、模型响应、观察和奖励等子特征，每个子特征都有指定的数据类型。数据集还包括一个reward特征，数据类型为float64。数据集分为train训练集，包含102,400个样本，总大小为156,921,039字节，下载大小为1,185,459字节。默认配置指定了训练集的数据文件路径。

The dataset includes multiple features, with steps being a list that contains sub-features such as action, chat completions, done status, mc_return, model response, observation, and reward, each with a specified data type. The dataset also includes a reward feature with a float64 data type. The dataset is split into a train set with 102,400 examples, a total size of 156,921,039 bytes, and a download size of 1,185,459 bytes. The default configuration specifies the data file path for the train split.