Datasets:

Modalities:
Text
Formats:
parquet
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
HumanEval-MRI-Cpp / README.md
nmuendler's picture
Update README.md
20d2c2b verified
---
source_datasets:
- zai-org/humaneval-x
dataset_info:
features:
- name: instance_id
dtype: string
- name: number_spans
dtype: int64
- name: prompt
dtype: string
- name: declaration
dtype: string
- name: splits
sequence: string
- name: removed_spans
sequence: string
- name: canonical_solution
dtype: string
- name: test
dtype: string
splits:
- name: test
num_bytes: 834118
num_examples: 473
download_size: 148031
dataset_size: 834118
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
---
## HumanEval/MRI/C++
This is a multi-region infilling version of the [C++ translation of HumanEval](https://huggingface.co/datasets/zai-org/humaneval-x). It was generated by randomly removing between 1 and 3 spans from the canonical solution in the dataset.
This dataset was used for evaluation in the paper [Constrained Decoding of Diffusion LLMs with Context-Free Grammars](https://arxiv.org/abs/2508.10111). You can find the corresponding evaluation code on [the project GitHub Repository](https://github.com/eth-sri/constrained-diffusion).
### Example Usage
```python
from datasets import load_dataset
import json
dataset = load_dataset('eth-sri/HumanEval-MRI-Cpp')
for instance in dataset['test']:
print(json.dumps(instance, indent=2))
break
```
### Example Instance
```json
{
"instance_id": "CPP/0_spans_1",
"number_spans": 1,
"prompt": "/*\nCheck if in given vector of numbers, are any two numbers closer to each other than\ngiven threshold.\n>>> has_close_elements({1.0, 2.0, 3.0}, 0.5)\nfalse\n>>> has_close_elements({1.0, 2.8, 3.0, 4.0, 5.0, 2.0}, 0.3)\ntrue\n*/\n#include<stdio.h>\n#include<vector>\n#include<math.h>\nusing namespace std;\nbool has_close_elements(vector<float> numbers, float threshold){\n",
"declaration": "#include<stdio.h>\n#include<vector>\n#include<math.h>\nusing namespace std;\n#include<algorithm>\n#include<stdlib.h>\nbool has_close_elements(vector<float> numbers, float threshold){\n",
"splits": [
" int i,j;\n \n for (i=0;i<numbers.size();i++)\n for (j=i+1;j<numbers.size();j++)\n if ",
";\n\n return false;\n}\n\n"
],
"removed_spans": [
"(abs(numbers[i]-numbers[j])<threshold)\n return true"
],
"canonical_solution": " int i,j;\n \n for (i=0;i<numbers.size();i++)\n for (j=i+1;j<numbers.size();j++)\n if (abs(numbers[i]-numbers[j])<threshold)\n return true;\n\n return false;\n}\n\n",
"test": "#undef NDEBUG\n#include<assert.h>\nint main(){\n vector<float> a={1.0, 2.0, 3.9, 4.0, 5.0, 2.2};\n assert (has_close_elements(a, 0.3)==true);\n assert (has_close_elements(a, 0.05) == false);\n\n assert (has_close_elements({1.0, 2.0, 5.9, 4.0, 5.0}, 0.95) == true);\n assert (has_close_elements({1.0, 2.0, 5.9, 4.0, 5.0}, 0.8) ==false);\n assert (has_close_elements({1.0, 2.0, 3.0, 4.0, 5.0}, 2.0) == true);\n assert (has_close_elements({1.1, 2.2, 3.1, 4.1, 5.1}, 1.0) == true);\n assert (has_close_elements({1.1, 2.2, 3.1, 4.1, 5.1}, 0.5) == false);\n \n}\n"
}
```