Not Everything is All You Need: Toward Low-Redundant Optimization for Large Language Model Alignment

Zhipeng Chen, Kun Zhou, Xin Zhao, Jingyuan Wang, Ji-Rong Wen


Abstract
Large language models (LLMs) are still struggling in aligning with human preference in complex tasks and scenarios. They are prone to overfit into the unexpected patterns or superficial styles in the training data. We conduct an empirical study that only selects the top-10% most updated parameters in LLMs for alignment training, and see improvements in the convergence process and final performance. It indicates the existence of redundant neurons in LLMs for alignment training. To reduce its influence, we propose a low-redundant alignment method named **ALLO**, focusing on optimizing the most related neurons with the most useful supervised signals. Concretely, we first identify the neurons that are related to the human preference data by a gradient-based strategy, then identify the alignment-related key tokens by reward models for computing loss. Besides, we also decompose the alignment process into the forgetting and learning stages, where we first forget the tokens with unaligned knowledge and then learn aligned knowledge, by updating different ratios of neurons, respectively. Experimental results on 10 datasets have shown the effectiveness of ALLO. Our code and data will be publicly released.
Anthology ID:
2024.emnlp-main.857
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
15337–15351
Language:
URL:
https://aclanthology.org/2024.emnlp-main.857
DOI:
Bibkey:
Cite (ACL):
Zhipeng Chen, Kun Zhou, Xin Zhao, Jingyuan Wang, and Ji-Rong Wen. 2024. Not Everything is All You Need: Toward Low-Redundant Optimization for Large Language Model Alignment. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 15337–15351, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Not Everything is All You Need: Toward Low-Redundant Optimization for Large Language Model Alignment (Chen et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.857.pdf
Software:
 2024.emnlp-main.857.software.zip