Datasets:
metadata
license: mit
task_categories:
- question-answering
dataset_info:
features:
- name: user_input
dtype: string
- name: cmi
dtype: float64
- name: test_id
dtype: string
- name: reference
dtype: string
- name: title
dtype: string
- name: reference_contexts
sequence: string
- name: origin_id
dtype: string
- name: doc_id
dtype: string
splits:
- name: CM1_qCM1_cEN_aEN
num_bytes: 125745
num_examples: 120
- name: CM1_qCM2_cEN_aEN
num_bytes: 125694
num_examples: 120
- name: CM1_qCM3_cEN_aEN
num_bytes: 125690
num_examples: 120
- name: CM1_qCM4_cEN_aEN
num_bytes: 125694
num_examples: 120
- name: CM1_qCM5_cEN_aEN
num_bytes: 125681
num_examples: 120
- name: CM1_qCM6_cEN_aEN
num_bytes: 125505
num_examples: 120
- name: NEG_qEN_cEN_aEN
num_bytes: 123011
num_examples: 120
- name: NEG_qEN_cDE_aEN
num_bytes: 137064
num_examples: 120
- name: english
num_bytes: 124391
num_examples: 120
- name: german
num_bytes: 139820
num_examples: 120
download_size: 883962
dataset_size: 1278295
configs:
- config_name: default
data_files:
- split: CM1_qCM1_cEN_aEN
path: data/CM1_qCM1_cEN_aEN-*
- split: CM1_qCM2_cEN_aEN
path: data/CM1_qCM2_cEN_aEN-*
- split: CM1_qCM3_cEN_aEN
path: data/CM1_qCM3_cEN_aEN-*
- split: CM1_qCM4_cEN_aEN
path: data/CM1_qCM4_cEN_aEN-*
- split: CM1_qCM5_cEN_aEN
path: data/CM1_qCM5_cEN_aEN-*
- split: CM1_qCM6_cEN_aEN
path: data/CM1_qCM6_cEN_aEN-*
- split: NEG_qEN_cEN_aEN
path: data/NEG_qEN_cEN_aEN-*
- split: NEG_qEN_cDE_aEN
path: data/NEG_qEN_cDE_aEN-*
- split: english
path: data/english-*
- split: german
path: data/german-*
test_half split from XQuAD https://huggingface.co/datasets/xxizhouu/RAGAS_xquad
PLUS
one impossible question(english) for each paragraph, taken from SQuAD 2.0
- test_id: shared uuid accross different spilt
- cmi: code mix index