---
dataset_info:
  features:
  - name: id
    dtype: string
  - name: original_text
    dtype: string
  - name: domain
    dtype: string
  - name: text_type
    dtype: string
  - name: ground_truth
    dtype: string
  - name: prompt_len
    dtype: int64
  - name: reader_grpo_v7_2
    dtype: string
  - name: prediction_outcome
    dtype: string
  splits:
  - name: train
    num_bytes: 625505
    num_examples: 400
  download_size: 334252
  dataset_size: 625505
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---