MugiLab/reason_at_code

---
dataset_info:
features:
- name: system
dtype: string
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 103183504
num_examples: 4746
download_size: 39331254
dataset_size: 103183504
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: mit
task_categories:
- text-generation
language:
- en
tags:
- code
size_categories:
- 1K<n<10K
---

# Reasoning Dataset for Code

This repository contains a curated reasoning dataset specifically designed for coding-related problems, particularly in Python.
The dataset was created by filtering non-code problems from the original NovaSky-AI/Sky-T1_data_17k dataset.
The goal of this dataset is to facilitate fine-tuning models for reasoning tasks related to code understanding, problem-solving, and logical deduction in programming.

## Overview
The dataset emphasizes high-quality, diverse problems requiring reasoning in coding contexts, with a particular focus on Python.
It can be used to fine-tune and evaluate models for:
- Code comprehension
- Logical reasoning in coding scenarios
- Problem-solving strategies
- Debugging and code correction tasks

## Key Features
- **Source**: Filtered from NovaSky-AI/Sky-T1_data_17k.
- **Language**: Python-focused coding problems.
- **Scope**: A broad range of problem-solving contexts, including algorithmic challenges, data structure manipulations, and logical puzzles.

## Dataset Structure
The dataset is organized into the following format:

- **`system`**: system prompt for potential model.
- **`conversations`**: a list of conversations in alpine format.

## Contribution
We welcome contributions to enhance this dataset, including:
- Adding new coding problems
- Improving existing solutions
- Expanding to other programming languages

Feel free to open an issue or submit a pull request!

## License
This dataset is licensed under [MIT License](LICENSE). Please ensure proper attribution when using this dataset in your projects.

## Acknowledgments
- Original dataset: [NovaSky-AI/Sky-T1_data_17k](https://huggingface.co/datasets/NovaSky-AI/Sky-T1_data_17k)

## Contact
For questions or feedback open an issue in the repository.