---
license: apache-2.0
task_categories:
- question-answering
language:
- en
- zh
tags:
- agent
- math
- science
- VQA
size_categories:
- 10K<n<100K
---

This dataset is used to train CodeV-RL, as described in the referenced [paper](https://arxiv.org/abs/2511.19661).