metadata
dataset_info:
features:
- name: prompt_id
dtype: string
- name: prompt
dtype: string
- name: response_0
dtype: string
- name: response_1
dtype: string
- name: response_2
dtype: string
- name: response_3
dtype: string
- name: response_4
dtype: string
- name: response_5
dtype: string
- name: response_6
dtype: string
- name: response_7
dtype: string
- name: response_8
dtype: string
- name: response_9
dtype: string
- name: reward_1_response_0
dtype: float64
- name: reward_1_response_1
dtype: float64
- name: reward_1_response_2
dtype: float64
- name: reward_1_response_3
dtype: float64
- name: reward_1_response_4
dtype: float64
- name: reward_1_response_5_for_gradient
dtype: float64
- name: reward_1_response_6_for_gradient
dtype: float64
- name: reward_1_response_7_for_gradient
dtype: float64
- name: reward_1_response_8_for_gradient
dtype: float64
- name: reward_1_response_9_for_gradient
dtype: float64
- name: reward_0_response_0
dtype: float64
- name: reward_0_response_1
dtype: float64
- name: reward_0_response_2
dtype: float64
- name: reward_0_response_3
dtype: float64
- name: reward_0_response_4
dtype: float64
- name: reward_0_response_5_for_gradient
dtype: float64
- name: reward_0_response_6_for_gradient
dtype: float64
- name: reward_0_response_7_for_gradient
dtype: float64
- name: reward_0_response_8_for_gradient
dtype: float64
- name: reward_0_response_9_for_gradient
dtype: float64
- name: llama_prompt
dtype: string
- name: llama_prompt_tokens
sequence: int64
- name: chosen
dtype: string
- name: llama_chosen
dtype: string
- name: llama_chosen_tokens
sequence: int64
- name: reject
dtype: string
- name: llama_reject
dtype: string
- name: llama_reject_tokens
sequence: int64
- name: chosen_reward_0
dtype: float64
- name: reject_reward_0
dtype: float64
- name: g_chosen_0
dtype: float64
- name: g_reject_0
dtype: float64
- name: chosen_rank_0
dtype: float64
- name: reject_rank_0
dtype: float64
- name: g_winrate_chosen_0
dtype: float64
- name: g_winrate_reject_0
dtype: float64
- name: chosen_reward_1
dtype: float64
- name: reject_reward_1
dtype: float64
- name: g_chosen_1
dtype: float64
- name: g_reject_1
dtype: float64
- name: chosen_rank_1
dtype: float64
- name: reject_rank_1
dtype: float64
- name: g_winrate_chosen_1
dtype: float64
- name: g_winrate_reject_1
dtype: float64
splits:
- name: train
num_bytes: 5149188816.566939
num_examples: 57364
- name: test
num_bytes: 89763419.85508227
num_examples: 1000
download_size: 1060137656
dataset_size: 5238952236.422022
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*