AMoPO Datasets
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Javkonline/AMoPO
下载链接
链接失效反馈官方服务：
资源简介：
该数据集用于实验，旨在验证AMoPO框架的性能。这些数据集涵盖了在不同规模的大型语言模型上进行的实验，模型规模从7B到14B再到32B不等。任务的目的是实现多目标偏好对齐。
This dataset is employed in experiments designed to validate the performance of the AMoPO framework. These datasets encompass experiments carried out on large language models (LLMs) with varying parameter sizes, spanning from 7B, 14B to 32B. The core task of these experiments is to achieve multi-objective preference alignment.
提供机构：
Authors of the paper
