metadata
dataset_info:
features:
- name: id
dtype: string
- name: prompt
dtype: string
- name: models
sequence: string
- name: completions
sequence:
- name: annotations
struct:
- name: Helpfulness
struct:
- name: Rating
dtype: string
- name: Rationale
dtype: string
- name: Ethical Considerations
struct:
- name: Rating
dtype: string
- name: Rationale
dtype: string
- name: Visual Faithfulness
struct:
- name: Rating
dtype: string
- name: Rationale
dtype: string
- name: model
dtype: string
- name: response
dtype: string
- name: image
dtype: image
splits:
- name: train
num_bytes: 4409267841.752
num_examples: 80258
download_size: 4071846369
dataset_size: 4409267841.752
configs:
- config_name: default
data_files:
- split: train
path: data/train-*