让多个大模型在隐空间直接协作，提升推理效率与准确性。
Latent Collaboration in Multi-Agent Systems
- 代理在隐空间生成思维，通过共享隐式记忆直接交流。
- 9项基准测试中准确率最高提升14.6%，推理速度提升4.3倍。
- 无需训练，端到端设计，适合需要高效协同的AI系统。
多智能体系统（MAS）将大语言模型从独立推理扩展为系统级协作智能。现有基于文本通信的智能体依赖显式语言交互，我们提出一种无需训练的纯隐空间协作框架——LatentMAS。在该框架中，每个智能体通过最后一层隐藏状态自回归生成隐式思维，并由共享隐式工作内存保存和传递内部表示与隐式思想，实现无损信息交换且无需重新编码。理论分析表明，相较于传统文本型MAS，LatentMAS具有更高表达能力、更低复杂度和无损信息传递。在数学与科学推理、常识理解、代码生成等9个综合基准上，实验验证其优于先进单智能体及文本型智能体基线，准确率最高提升14.6%，输出令牌减少70.8%–83.7%，端到端推理速度提升4×–4.3×。代码与数据已开源。
原文摘要 · Abstract (English)
Multi-agent systems (MAS) extend large language models (LLMs) from independent single-model reasoning to coordinative system-level intelligence. While existing LLM agents depend on text-based mediation for reasoning and communication, we take a step forward by enabling models to collaborate directly within the continuous latent space. We introduce LatentMAS, an end-to-end training-free framework that enables pure latent collaboration among LLM agents. In LatentMAS, each agent first performs auto-regressive latent thoughts generation through last-layer hidden embeddings instead of text. Then, a shared latent working memory preserves and transfers each agent's internal representations and latent thoughts, ensuring lossless information exchange without re-encoding. We provide detailed theoretical analyses showing that LatentMAS achieves higher expressiveness and lossless information preservation with lower overall complexity than standard text-based MAS. In addition, empirical evaluations across 9 comprehensive benchmarks spanning math and science reasoning, commonsense understanding, and code generation show that LatentMAS outperforms advanced single agents and text-based MAS baselines, achieving up to 14.6% higher accuracy, reducing output token usage by 70.8%-83.7%, and providing 4$\times$-4.3$\times$ faster end-to-end inference. Code and data are fully open-sourced at https://github.com/Gen-Verse/LatentMAS.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。