The Dataset Viewer has been disabled on this dataset.

✨ ArcherCodeR

🏹️ Reinforcement Learning for Enhanced Code Reasoning in LLMs 🎯

Github Model Data Wandb 知乎

Overview

ArcherCodeR-Dataset is a dataset of verifiable, challenging, and diverse coding questions (6.7K). This dataset is used to train the ArcherCodeR model series, which consists of code reasoning models trained using large-scale rule-based reinforcement learning with carefully designed datasets and training recipes.

We select, clean, and curate coding problems from open-source datasets, including

🔍 Key Notes:

  • Both code_contests (DeepMind) and codeforces (Open-r1) datasets include regenerated test cases to mitigate false positives.
  • Significant prompt duplication exists across sources. When duplicates occur, code_contests or codeforces data takes priority.

For more details on data processing, please refer to our Zhihu article.

Technical Report

The technical report will be released soon.

Downloads last month
10

Models trained or fine-tuned on wizardII/ArcherCodeR-Dataset