metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: response_Llama-3.2-3B-Instruct
dtype: string
- name: reward_helpfulness_Llama-3.2-3B-Instruct
dtype: float64
- name: reward_correctness_Llama-3.2-3B-Instruct
dtype: float64
- name: response_Helpfulness_REBEL_1e4
dtype: string
- name: reward_helpfulness_Helpfulness_REBEL_1e4
dtype: float64
- name: reward_correctness_Helpfulness_REBEL_1e4
dtype: float64
- name: response_Multi_Preference_REBEL_1e4
dtype: string
- name: reward_helpfulness_Multi_Preference_REBEL_1e4
dtype: float64
- name: reward_correctness_Multi_Preference_REBEL_1e4
dtype: float64
splits:
- name: train
num_bytes: 6874186
num_examples: 1000
download_size: 3709546
dataset_size: 6874186
configs:
- config_name: default
data_files:
- split: train
path: data/train-*