arithmetic-circuit-overloading/synthetic-dataset-v2-2d-5M-500K-0.1-reverse

---
dataset_info:
- config_name: '100'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141569089
num_examples: 5000000
- name: validation
num_bytes: 112657742
num_examples: 500000
download_size: 526024425
dataset_size: 1254226831
- config_name: '50'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141804468
num_examples: 5000000
- name: validation
num_bytes: 112680624
num_examples: 500000
download_size: 815326812
dataset_size: 1254485092
- config_name: '75'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141686630
num_examples: 5000000
- name: validation
num_bytes: 112679577
num_examples: 500000
download_size: 801101053
dataset_size: 1254366207
- config_name: '90'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141621304
num_examples: 5000000
- name: validation
num_bytes: 112665660
num_examples: 500000
download_size: 776956774
dataset_size: 1254286964
- config_name: '95'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141598685
num_examples: 5000000
- name: validation
num_bytes: 112660196
num_examples: 500000
download_size: 746159317
dataset_size: 1254258881
- config_name: '99'
features:
- name: _id
dtype: string
- name: base_operation
dtype: string
- name: target_operation
dtype: string
- name: fs_examples
list: string
- name: question
dtype: string
- name: answer
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 1141575004
num_examples: 5000000
- name: validation
num_bytes: 112658227
num_examples: 500000
download_size: 537604053
dataset_size: 1254233231
configs:
- config_name: '100'
data_files:
- split: train
path: 100/train-*
- split: validation
path: 100/validation-*
- config_name: '50'
data_files:
- split: train
path: 50/train-*
- split: validation
path: 50/validation-*
- config_name: '75'
data_files:
- split: train
path: 75/train-*
- split: validation
path: 75/validation-*
- config_name: '90'
data_files:
- split: train
path: 90/train-*
- split: validation
path: 90/validation-*
- config_name: '95'
data_files:
- split: train
path: 95/train-*
- split: validation
path: 95/validation-*
- config_name: '99'
data_files:
- split: train
path: 99/train-*
- split: validation
path: 99/validation-*
---

数据集信息：
本数据集共包含6个配置版本，配置名称分别为`100`、`50`、`75`、`90`、`95`与`99`。

### 单配置通用特征
每个配置均包含以下7个特征字段：
1. `_id`：字符串（string）类型唯一标识符
2. `base_operation`：字符串类型，基础操作字段
3. `target_operation`：字符串类型，目标操作字段
4. `fs_examples`：字符串列表（list[string]）类型，少样本示例字段
5. `question`：字符串类型，问题字段
6. `answer`：字符串类型，答案字段
7. `prompt`：字符串类型，提示词字段

### 数据集划分与统计参数
每个配置均划分出训练集与验证集两个子集，各配置的具体统计参数如下：
1. 配置`100`：
- 训练集：字节占用量1141569089，样本量5000000
- 验证集：字节占用量112657742，样本量500000
- 下载总大小：526024425
- 数据集总大小：1254226831
2. 配置`50`：
- 训练集：字节占用量1141804468，样本量5000000
- 验证集：字节占用量112680624，样本量500000
- 下载总大小：815326812
- 数据集总大小：1254485092
3. 配置`75`：
- 训练集：字节占用量1141686630，样本量5000000
- 验证集：字节占用量112679577，样本量500000
- 下载总大小：801101053
- 数据集总大小：1254366207
4. 配置`90`：
- 训练集：字节占用量1141621304，样本量5000000
- 验证集：字节占用量112665660，样本量500000
- 下载总大小：776956774
- 数据集总大小：1254286964
5. 配置`95`：
- 训练集：字节占用量1141598685，样本量5000000
- 验证集：字节占用量112660196，样本量500000
- 下载总大小：746159317
- 数据集总大小：1254258881
6. 配置`99`：
- 训练集：字节占用量1141575004，样本量5000000
- 验证集：字节占用量112658227，样本量500000
- 下载总大小：537604053
- 数据集总大小：1254233231

### 数据文件路径配置
各配置对应的数据文件路径如下：
1. 配置`100`：训练集数据路径为`100/train-*`，验证集数据路径为`100/validation-*`
2. 配置`50`：训练集数据路径为`50/train-*`，验证集数据路径为`50/validation-*`
3. 配置`75`：训练集数据路径为`75/train-*`，验证集数据路径为`75/validation-*`
4. 配置`90`：训练集数据路径为`90/train-*`，验证集数据路径为`90/validation-*`
5. 配置`95`：训练集数据路径为`95/train-*`，验证集数据路径为`95/validation-*`
6. 配置`99`：训练集数据路径为`99/train-*`，验证集数据路径为`99/validation-*`