arXiv 2406.113542024-06-19cs.CLcs.AIcs.CV

通过模型无关的自解压方法保留大语言模型中的知识

Preserving Knowledge in Large Language Model with Model-Agnostic Self-Decompression

College of Computer Science and Technology, Zhejiang University（浙江大学计算机科学与技术学院）；Binjiang Research Institute of Zhejiang University（浙江大学滨江研究院）；Linker Technology Research Co. Ltd（领科技术研究有限公司）

Zilun Zhang, Yutao Sun, Tiancheng Zhao, Leigang Sha, Ruochen Xu, Kyusong Lee, Jianwei Yin

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对大语言模型及多模态大语言模型微调时的灾难性遗忘问题，提出模型无关的自解压方法Tree Generation，其TG-SFT变体可合成SFT数据，加入微调语料后显著缓解多模态大模型的语言能力遗忘。

摘要：

人类在学习新信息时能够保留旧有知识，但大语言模型（LLM）在针对领域特定数据进行后预训练或监督微调（SFT）时，往往会遭遇灾难性遗忘问题。此外，对于由LLM基座和视觉投影器组成的多模态大语言模型（MLLM，例如LLaVA），与单模态对应模型相比，其在语言基准测试上的性能出现了显著下降。为应对这些挑战，我们提出了一种新颖的模型无关自解压方法——Tree Generation（TG，树生成），该方法可将LLM内部的知识解压到训练语料库中。本文聚焦于TG-SFT，它能够合成生成用于指令调优步骤的SFT数据。通过在MLLM的SFT阶段加入导出的语料库，我们显著缓解了遗忘问题。

英文摘要：

Humans can retain old knowledge while learning new information, but Large Language Models (LLMs) often suffer from catastrophic forgetting when post-pretrained or supervised fine-tuned (SFT) on domain-specific data. Moreover, for Multimodal Large Language Models (MLLMs) which are composed of the LLM base and visual projector (e.g. LLaVA), a significant decline in performance on language benchmarks was observed compared to their single-modality counterparts. To address these challenges, we introduce a novel model-agnostic self-decompression method, Tree Generation (TG), that decompresses knowledge within LLMs into the training corpus. This paper focuses on TG-SFT, which can synthetically generate SFT data for the instruction tuning steps. By incorporating the dumped corpus during SFT for MLLMs, we significantly reduce the forgetting problem.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。