---
dataset_info:
  features:
  - name: question_id
    dtype: string
  - name: sample_id
    dtype: int64
  - name: prompt_pretokenized
    dtype: string
  - name: prompt_tokenized
    sequence: int64
  - name: response
    dtype: string
  - name: tokenized_response
    sequence: int64
  - name: accuracy
    dtype: float64
  - name: logp
    dtype: float64
  - name: rm_score
    dtype: float64
  splits:
  - name: train
    num_bytes: 2286987762
    num_examples: 486202
  download_size: 213607658
  dataset_size: 2286987762
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---