metadata
dataset_info:
features:
- name: query
dtype: string
- name: entity_a
dtype: string
- name: entity_b
dtype: string
- name: rel_b
dtype: string
- name: prompt
dtype: string
- name: model
dtype: string
- name: generation_params
dtype: string
- name: model_response
dtype: string
- name: reasoning_summary
dtype: string
- name: path_prediction
sequence: 'null'
- name: parsed_paths
dtype: string
- name: n_paths
dtype: int64
- name: validity
sequence: 'null'
- name: n_valid_paths
dtype: int64
- name: cost_usd
dtype: float64
- name: elapsed_s
dtype: float64
- name: method
dtype: string
- name: mode
dtype: string
- name: input_dataset
dtype: string
- name: timestamp
dtype: string
splits:
- name: train
num_bytes: 672716
num_examples: 165
download_size: 104813
dataset_size: 672716
configs:
- config_name: default
data_files:
- split: train
path: data/train-*