metadata
dataset_info:
features:
- name: context
dtype: string
- name: context_length
dtype: int64
- name: depth_percent
dtype: int64
- name: needle
dtype: string
- name: chat_response
dtype: string
- name: rag_contexts
dtype: string
- name: retrieval_question
dtype: string
- name: collection_id
dtype: string
splits:
- name: train
num_bytes: 18798777
num_examples: 509
download_size: 5209648
dataset_size: 18798777
configs:
- config_name: default
data_files:
- split: train
path: data/train-*