---
dataset_info:
  features:
  - name: qid
    dtype: string
  - name: question
    dtype: string
  - name: context
    sequence: string
  - name: reasoning_content
    dtype: string
  - name: rewrite
    dtype: string
  splits:
  - name: train
    num_bytes: 12920713
    num_examples: 6034
  download_size: 6104693
  dataset_size: 12920713
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

SFT data for ConvSearch-R1. [[Github](https://github.com/BeastyZ/ConvSearch-R1)] [[arXiv](https://arxiv.org/abs/2505.15776)].