metadata
dataset_info:
features:
- name: input_text
dtype: string
- name: labels
dtype: int64
- name: season
dtype: string
- name: prompt
dtype: string
- name: completion
dtype: string
- name: bot_name
dtype: string
- name: username
dtype: string
- name: /mnt/nvme1/will/4_dpo
dtype: string
splits:
- name: train
num_bytes: 2738644
num_examples: 517
download_size: 1332699
dataset_size: 2738644
configs:
- config_name: default
data_files:
- split: train
path: data/train-*