metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: gold_response
dtype: string
- name: relevant_source_url
dtype: string
- name: relevant_source_text
dtype: string
- name: relevant_entities_in_source_urls
sequence: string
- name: id
dtype: string
- name: gold_passages
list:
- name: sentences
sequence: string
- name: text
dtype: string
- name: title
dtype: string
- name: pred_response
dtype: string
- name: model_name
dtype: string
- name: yurts_rag_contexts
list:
- name: end_index
dtype: int64
- name: field_type
dtype: string
- name: id
dtype: string
- name: score
dtype: float64
- name: source_doc_id
dtype: string
- name: start_index
dtype: int64
- name: text
dtype: string
- name: message_date
dtype: string
- name: entity_hallucinations
list:
- name: chat_message_id
dtype: string
- name: entity
dtype: string
- name: hallucinated_entity_span_end
dtype: int64
- name: hallucinated_entity_span_start
dtype: int64
- name: id
dtype: string
splits:
- name: train
num_bytes: 22599129
num_examples: 299
download_size: 11614583
dataset_size: 22599129
configs:
- config_name: default
data_files:
- split: train
path: data/train-*