reasonrank_data_sft

<p align="left">
Useful links: 📝 <a href="https://arxiv.org/abs/2508.07050" target="_blank">arXiv Paper</a> • </a> 🧩 <a href="https://github.com/8421BCD/ReasonRank" target="_blank">Github</a>
</p>

This is the training data to SFT our reasoning-intensive listwise reranker [ReasonRank](https://arxiv.org/abs/2508.07050). The data format is organized based on the training data format of LLaMA-Factory.
arXiv paper:

<p align="left">
实用链接：📝 <a href="https://arxiv.org/abs/2508.07050" target="_blank">arXiv论文</a> • 🧩 <a href="https://github.com/8421BCD/ReasonRank" target="_blank">GitHub仓库</a>
</p>

本数据集为我们针对推理密集型列表式重排序模型ReasonRank开展监督微调（Supervised Fine-Tuning，SFT）所用的训练数据。该数据集的格式严格遵循LLaMA-Factory的训练数据格式规范。

arXiv论文：

搜集汇总

背景与挑战

背景概述

该数据集用于训练推理密集型列表重排序模型ReasonRank，其数据格式基于LLaMA-Factory框架设计。相关研究细节可参考arXiv论文编号2508.07050。