Follow-Lang/set.mm.proof

该数据集是为`Follow-Lang/set.mm.proof`模型生成的训练数据。数据位于`train/`目录下，每行数据的格式为：Cost(s) Cost(a) Cost(s) s a s，其中`a`以`<action>`开始并以`</action>`结束。成本计算遵循`Cost(s) = max(Cost(a), 1+Cost(s))`的规则。最大单词长度为1024，所有词汇列在`words.txt`文件中。数据生成深度为2。如果需要更多数据，可以联系获取。

This dataset is generated training data for the `Follow-Lang/set.mm.proof` model. The data is located in the `train/` directory, with each line formatted as: Cost(s) Cost(a) Cost(s) s a s, where `a` starts with `<action>` and ends with `</action>`. The cost calculation follows the rule `Cost(s) = max(Cost(a), 1+Cost(s))`. The maximum word length is 1024, and all vocabulary words are listed in the `words.txt` file. The data was generated with a depth of 2. Additional data can be obtained upon request.