---
dataset_info:
  features:
  - name: context_length
    dtype: int64
  - name: depth_percent
    dtype: int64
  - name: context
    dtype: string
  - name: retrieval_question
    dtype: string
  - name: version
    dtype: int64
  - name: needles
    sequence: string
  splits:
  - name: train
    num_bytes: 152264530
    num_examples: 350
  download_size: 91717146
  dataset_size: 152264530
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---