metadata
dataset_info:
features:
- name: images
list: string
- name: pgd_images
list: string
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: pgd_prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: ability
dtype: string
- name: reward_model
struct:
- name: ground_truth
dtype: string
- name: style
dtype: string
- name: extra_info
struct:
- name: all_r
list: float64
- name: done
dtype: bool
- name: gt
dtype: string
- name: next_image
dtype: string
- name: r
dtype: float64
- name: ref_mask
dtype: int64
- name: sim_obs_all_r
list: float64
- name: sim_r
dtype: float64
- name: task_id
dtype: string
- name: turn_id
dtype: int64
- name: turn_len
dtype: int64
splits:
- name: train
num_bytes: 32039190
num_examples: 2386
download_size: 2155035
dataset_size: 32039190
configs:
- config_name: default
data_files:
- split: train
path: data/train-*