---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: language
    dtype: string
  - name: verdict
    dtype: string
  - name: source
    dtype: string
  - name: problem_id
    dtype: string
  - name: type
    dtype: string
  - name: difficulty
    dtype: string
  splits:
  - name: train
    num_bytes: 28548411
    num_examples: 12775
  download_size: 10531794
  dataset_size: 28548411
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
task_categories:
- text-generation
---

This dataset is from the paper [Can LLMs Generate High-Quality Test Cases for Algorithm Problems? TestCase-Eval: A Systematic Evaluation of Fault Coverage and Exposure](https://huggingface.co/papers/2506.12278).

GitHub repository: [https://github.com/FlowRays/TestCase-Eval](https://github.com/FlowRays/TestCase-Eval)