DDK: Distilling Domain Knowledge for Efficient Large Language Models

Despite the advanced intelligence abilities of large language models (LLMs) in various applications, they still face significant computational and storage demands. Knowledge Distillation (KD) has emerged as an effective strategy to improve the performance of a smaller LLM (i.e., the student model) b...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-07
Main Authors: Liu, Jiaheng, Zhang, Chenchen, Guo, Jinyang, Zhang, Yuanxing, Que, Haoran, Deng, Ken, Bai, Zhiqi, Liu, Jie, Zhang, Ge, Wang, Jiakai, Wu, Yanan, Liu, Congnan, Su, Wenbo, Wang, Jiamang, Qu, Lin, Zheng, Bo
Format: Article
Language:English
Subjects:
Composition effects
Datasets
Knowledge
Knowledge management
Large language models
Performance enhancement
Performance evaluation
Teachers
Teaching methods
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Despite the advanced intelligence abilities of large language models (LLMs) in various applications, they still face significant computational and storage demands. Knowledge Distillation (KD) has emerged as an effective strategy to improve the performance of a smaller LLM (i.e., the student model) by transferring knowledge from a high-performing LLM (i.e., the teacher model). Prevailing techniques in LLM distillation typically use a black-box model API to generate high-quality pretrained and aligned datasets, or utilize white-box distillation by altering the loss function to better transfer knowledge from the teacher LLM. However, these methods ignore the knowledge differences between the student and teacher LLMs across domains. This results in excessive focus on domains with minimal performance gaps and insufficient attention to domains with large gaps, reducing overall performance. In this paper, we introduce a new LLM distillation framework called DDK, which dynamically adjusts the composition of the distillation dataset in a smooth manner according to the domain performance differences between the teacher and student models, making the distillation process more stable and effective. Extensive evaluations show that DDK significantly improves the performance of student models, outperforming both continuously pretrained baselines and existing knowledge distillation methods by a large margin.
ISSN:2331-8422