Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Zhili, Gou, Yunhao, Chen, Kai, Hong, Lanqing, Gao, Jiahui, Mi, Fei, Zhang, Yu, Li, Zhenguo, Jiang, Xin, Liu, Qun, Kwok, James T.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913869032587264
author Liu, Zhili
Gou, Yunhao
Chen, Kai
Hong, Lanqing
Gao, Jiahui
Mi, Fei
Zhang, Yu
Li, Zhenguo
Jiang, Xin
Liu, Qun
Kwok, James T.
author_facet Liu, Zhili
Gou, Yunhao
Chen, Kai
Hong, Lanqing
Gao, Jiahui
Mi, Fei
Zhang, Yu
Li, Zhenguo
Jiang, Xin
Liu, Qun
Kwok, James T.
contents As the capabilities of large language models (LLMs) continue to expand, aligning these models with human values remains a significant challenge. Recent studies show that reasoning abilities contribute significantly to model safety, while integrating Mixture-of-Experts (MoE) architectures can further enhance alignment. In this work, we address a fundamental question: How to effectively incorporate reasoning abilities and MoE architectures into self-alignment process in LLMs? We propose Mixture of insighTful Experts (MoTE), a novel framework that synergistically combines reasoning chains and expert mixtures to improve self-alignments. From a data perspective, MoTE employs a structured reasoning chain comprising four key stages: Question Analysis, Answer Guidance, Safe Answer, and Safety Checking. This approach enhances safety through multi-step reasoning and proves effective even for smaller and less powerful LLMs (e.g., 7B models). From an architectural perspective, MoTE adopts a multi-LoRA framework with step-level routing, where each expert is dedicated to a specific reasoning step. This design eliminates the need for balance losses, ensures stable training, and supports adaptive inference lengths. Experimental results demonstrate that MoTE significantly improves model safety, jailbreak resistance, and over-refusal capabilities, achieving performance comparable to OpenAI's state-of-the-art o1 model.
format Preprint
id arxiv_https___arxiv_org_abs_2405_00557
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment
Liu, Zhili
Gou, Yunhao
Chen, Kai
Hong, Lanqing
Gao, Jiahui
Mi, Fei
Zhang, Yu
Li, Zhenguo
Jiang, Xin
Liu, Qun
Kwok, James T.
Computation and Language
Artificial Intelligence
As the capabilities of large language models (LLMs) continue to expand, aligning these models with human values remains a significant challenge. Recent studies show that reasoning abilities contribute significantly to model safety, while integrating Mixture-of-Experts (MoE) architectures can further enhance alignment. In this work, we address a fundamental question: How to effectively incorporate reasoning abilities and MoE architectures into self-alignment process in LLMs? We propose Mixture of insighTful Experts (MoTE), a novel framework that synergistically combines reasoning chains and expert mixtures to improve self-alignments. From a data perspective, MoTE employs a structured reasoning chain comprising four key stages: Question Analysis, Answer Guidance, Safe Answer, and Safety Checking. This approach enhances safety through multi-step reasoning and proves effective even for smaller and less powerful LLMs (e.g., 7B models). From an architectural perspective, MoTE adopts a multi-LoRA framework with step-level routing, where each expert is dedicated to a specific reasoning step. This design eliminates the need for balance losses, ensures stable training, and supports adaptive inference lengths. Experimental results demonstrate that MoTE significantly improves model safety, jailbreak resistance, and over-refusal capabilities, achieving performance comparable to OpenAI's state-of-the-art o1 model.
title Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2405.00557