---
dataset_info:
  features:
  - name: series_id
    dtype: int64
  - name: episode_id
    dtype: int64
  - name: start_step
    dtype: int64
  - name: end_step
    dtype: int64
  - name: length
    dtype: int64
  - name: terminal
    dtype: bool
  - name: partial
    dtype: bool
  - name: steps
    list:
    - name: action
      dtype: int64
    - name: done
      dtype: bool
    - name: reward
      dtype: float64
    - name: state_t
      struct:
      - name: text_blstats
        dtype: string
      - name: text_cursor
        dtype: string
      - name: text_glyphs
        dtype: string
      - name: text_inventory
        dtype: string
      - name: text_message
        dtype: string
    - name: state_t+1
      struct:
      - name: text_blstats
        dtype: string
      - name: text_cursor
        dtype: string
      - name: text_glyphs
        dtype: string
      - name: text_inventory
        dtype: string
      - name: text_message
        dtype: string
    - name: state_t+1_sentence
      dtype: string
    - name: state_t+1_tokens
      sequence: int64
    - name: state_t_sentence
      dtype: string
    - name: state_t_tokens
      sequence: int64
    - name: step_index
      dtype: int64
    - name: terminated
      dtype: bool
    - name: truncated
      dtype: bool
  splits:
  - name: train
    num_bytes: 1726288487
    num_examples: 1530
  download_size: 76494933
  dataset_size: 1726288487
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---