metadata
dataset_info:
features:
- name: instruction_id
dtype: string
- name: unique_id
dtype: string
- name: id
dtype: string
- name: video_file
dtype: string
- name: instruction
dtype: string
- name: selected_response
dtype: string
- name: rejected_response
dtype: string
- name: selected_gen_rating
dtype: int64
- name: rejected_gen_rating
dtype: int64
- name: selected_eval_rating
dtype: int64
- name: rejected_eval_rating
dtype: int64
- name: selected_eval_reasoning
dtype: string
- name: rejected_eval_reasoning
dtype: string
- name: description
struct:
- name: description
dtype: string
- name: id
dtype: string
- name: model_name
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: answer
dtype: string
- name: reasoning_a
dtype: string
- name: reasoning_b
dtype: string
- name: preference_reasoning
dtype: string
splits:
- name: train
num_bytes: 1154601714
num_examples: 207650
download_size: 538961338
dataset_size: 1154601714
configs:
- config_name: default
data_files:
- split: train
path: data/train-*