跳到正文
arXiv cs.AI· Yi Chen, Haiquan Wang, JiaHao Zhao, HaoHao Han·· 10 小时前AI 评分29

大语言模型与小语言模型协作机制综述

A Survey on Collaborative Mechanisms Between Large and Small Language Models

AI 导读

一篇综述系统梳理了大语言模型(LLM)与小语言模型(SLM)的协作机制,归纳出流水线、路由、辅助、蒸馏、融合五类交互方式。文章指出这类协作可在资源受限的边缘设备上兼顾性能与效率,满足低延迟、隐私、个性化与离线运行需求。同时讨论了系统开销、模型间一致性、任务分配、评估复杂度与安全隐私等挑战。

正文 · 原文

View PDF HTML (experimental)

Abstract:Large Language Models (LLMs) deliver powerful AI capabilities but face deployment challenges due to high resource costs and latency, whereas Small Language Models (SLMs) offer efficiency and deployability at the cost of reduced performance. Collaboration between LLMs and SLMs emerges as a crucial paradigm to synergistically balance these trade-offs, enabling advanced AI applications, especially on resource-constrained edge devices. This survey provides a comprehensive overview of LLM-SLM collaboration, detailing various interaction mechanisms (pipeline, routing, auxiliary, distillation, fusion), key enabling technologies, and diverse application scenarios driven by on-device needs like low latency, privacy, personalization, and offline operation. While highlighting the significant potential for creating more efficient, adaptable, and accessible AI, we also discuss persistent challenges including system overhead, inter-model consistency, robust task allocation, evaluation complexity, and security/privacy concerns. Future directions point towards more intelligent adaptive frameworks, deeper model fusion, and expansion into multimodal and embodied AI, positioning LLM-SLM collaboration as a key driver for the next generation of practical and ubiquitous artificial intelligence.
Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as: arXiv:2505.07460 [cs.AI]
  (or arXiv:2505.07460v2 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2505.07460

arXiv-issued DOI via DataCite

Submission history

From: Yi Chen [view email]
[v1] Mon, 12 May 2025 11:48:42 UTC (1,822 KB)
[v2] Wed, 18 Nov 2026 08:16:01 UTC (1,789 KB)

来源:arXiv cs.AI · arxiv.org