---
dataset_info:
  features:
  - name: prompt_id
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: language
    dtype: string
  - name: category
    dtype: string
  - name: mapped_category
    dtype: string
  - name: ppl
    dtype: float64
  - name: reward
    dtype: float64
  - name: sailor_ppl
    dtype: float64
  - name: sailor_user_ppl
    dtype: float64
  - name: sailor_assistant_ppl
    dtype: float64
  - name: ppl_rank
    dtype: float64
  - name: reward_rank
    dtype: float64
  - name: user_message_token_lengths
    sequence: int64
  splits:
  - name: train
    num_bytes: 867015954
    num_examples: 301229
  download_size: 388937118
  dataset_size: 867015954
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---