metadata
dataset_info:
features:
- name: config
struct:
- name: needle_type
dtype: string
- name: repetition_factors
sequence: int64
- name: semantic_closeness
dtype: string
- name: needle_distribution
dtype: string
- name: positions
sequence: int64
- name: needles
sequence: string
- name: haystack
dtype: string
- name: question
dtype: string
- name: dummy
dtype: string
splits:
- name: train
num_bytes: 18925620
num_examples: 588
download_size: 8332939
dataset_size: 18925620
configs:
- config_name: default
data_files:
- split: train
path: data/train-*