yuruny/agentic-sudoku-NonMarkov_qwen2.5-3B-it-5e-6_9x9_6-6_gt-SFT_ans1-6.5k-eval_results

该数据集包含多个步骤，每个步骤包含动作、聊天完成情况、完成状态、mc_return、模型响应、观察和奖励等子特征。此外，还有一个顶层的奖励特征。数据集分为训练集，包含102,400个示例，总大小为162,374,088字节，下载大小为6,260,144字节。数据集配置为指向训练数据文件的默认配置。

The dataset consists of multiple steps, each containing sub-features such as action, chat completions, done status, mc_return, model response, observation, and reward. Additionally, there is a top-level reward feature. The dataset is split into a training set with 102,400 examples, totaling 162,374,088 bytes, and a download size of 6,260,144 bytes. The dataset is configured with a default configuration pointing to the training data files.