Synadapt/orpo-dpo-mix-40k-1KL

---
dataset_info:
features:
- name: preferred
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: preferred_complexity
dtype: int64
- name: rejected_complexity
dtype: int64
splits:
- name: train
num_bytes: 3138346
num_examples: 571
download_size: 1674932
dataset_size: 3138346
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

The dataset includes multiple features, primarily divided into preferred and rejected sections, each containing content and role sub-features of string type. Additionally, there are preferred_complexity and rejected_complexity features of integer type to indicate the complexity of the content. The dataset is split into a training set (train) with 571 samples, totaling 3138346 bytes, with a download size of 1674932 bytes.

原始信息汇总

数据集概述

数据集特征

• preferred

  • content: 数据类型为字符串
  • role: 数据类型为字符串

• rejected

  • content: 数据类型为字符串
  • role: 数据类型为字符串

• preferred_complexity: 数据类型为整数（int64）

• rejected_complexity: 数据类型为整数（int64）

数据集分割

• train
  • num_bytes: 3138346字节
  • num_examples: 571个样本

数据集大小

• download_size: 1674932字节
• dataset_size: 3138346字节

配置

• config_name: default
  • data_files
    • split: train
      • path: data/train-*