---
dataset_info:
  features:
  - name: target
    dtype: int64
  - name: nums
    sequence: int64
  - name: answers
    sequence: string
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: difficulty
    dtype: float64
  - name: depth
    dtype: int64
  - name: weighted_ops
    dtype: int64
  - name: non_commutative_ops
    dtype: int64
  - name: bracket_count
    dtype: int64
  - name: difficulty_level
    dtype: string
  splits:
  - name: train
    num_bytes: 900017
    num_examples: 1024
  download_size: 168157
  dataset_size: 900017
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---