metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
struct:
- name: candidates
dtype: string
- name: interaction
sequence: string
- name: sentiments
sequence: string
- name: user_id
dtype: string
- name: output
struct:
- name: recommended
dtype: string
splits:
- name: train
num_bytes: 6448063
num_examples: 8201
download_size: 2545750
dataset_size: 6448063
configs:
- config_name: default
data_files:
- split: train
path: data/train-*