---
dataset_info:
  features:
  - name: query
    dtype: string
  - name: needle
    list: string
  - name: all_needles
    list: string
  - name: sampled_depths
    list: float64
  - name: distractor_needles
    list: string
  - name: ground_truth
    list: string
  - name: input
    dtype: string
  - name: input_length
    dtype: int64
  - name: response
    dtype: string
  splits:
  - name: train
    num_bytes: 8516081
    num_examples: 373
  download_size: 4065773
  dataset_size: 8516081
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---