---
dataset_info:
  features:
  - name: conversations
    list:
    - name: from
      dtype: string
    - name: value
      dtype: string
  - name: state
    sequence: string
  - name: metadata
    struct:
    - name: black_elo
      dtype: int64
    - name: cp_none
      dtype: bool
    - name: game_id
      dtype: string
    - name: result
      dtype: string
    - name: time_control
      dtype: string
    - name: type
      dtype: string
    - name: white_elo
      dtype: int64
  - name: tokens
    dtype: int64
  - name: num_conversations
    dtype: int64
  splits:
  - name: train
    num_bytes: 6265649643.257716
    num_examples: 5505102
  download_size: 1891409732
  dataset_size: 6265649643.257716
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---