README.md

metadata

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] [arXiv].