deep_code_bench

# Deep Code Bench Dataset

This dataset contains question-answer pairs with code context for evaluating deep code understanding.

## Dataset Structure

Each example contains:
- `question`: The question about the code
- `answer`: The expected answer
- `facts`: List of factual statements supporting the answer
- `metadata`: Additional information about difficulty, scope, etc.
- `id`: Unique identifier

## Usage

```python
from datasets import load_dataset

dataset = load_dataset("Qodo/deep_code_bench")
```

# Deep Code Bench 数据集
本数据集包含带有代码上下文的问答对，用于评估深度代码理解能力。

## 数据集结构
每个样本包含以下字段：
- `question`：针对代码的查询问题
- `answer`：预期标准答案
- `facts`：用于佐证标准答案的事实性陈述列表
- `metadata`：包含难度、应用范围等额外信息的元数据
- `id`：唯一标识符

## 使用方法
python
from datasets import load_dataset

dataset = load_dataset("Qodo/deep_code_bench")