让多个AI代理各司其职，避免互相干扰。
Dynamic Attentional Context Scoping: Agent-Triggered Focus Sessions for Isolated Per-Agent Steering in Multi-Agent LLM Orchestration
- 代理触发专注模式，只保留当前代理的完整上下文。
- 在10个代理并行时准确率达98.4%，远超基线60%。
- 适合多智能体协作系统，提升决策质量与效率。
多智能体大模型协同系统存在上下文污染问题：当N个代理同时竞争协调器的上下文窗口时，各代理的任务状态、部分输出和待解决问题会相互干扰，降低决策质量。本文提出动态注意力上下文划分（DACS）机制，协调器以两种非对称模式运行。在注册模式下，仅保留每个代理的轻量级状态摘要（≤200词元），保持对所有代理和用户的响应能力。当某代理发出转向请求时，协调器进入焦点模式Focus(a_i)，注入该代理的完整上下文，同时将其他所有代理压缩为注册条目。上下文隔离由代理触发，具有非对称性和确定性：转向期间上下文窗口仅包含F(a_i) + R_{-i}，彻底消除跨代理污染，无需上下文压缩或检索。我们在四个实验阶段共进行200次测试：第1阶段测试N ∈ {3,5,10}（60次）；第2阶段测试代理异质性与对抗依赖关系（60次）；第3阶段测试最高达D=15的决策密度（40次）；第4阶段使用自主大模型代理处理自由问答（40次，基于Claude Haiku 4.5）。在全部8个合成场景中，DACS实现90.0–98.4%的转向准确率，而平铺上下文基线仅为21.0–60.0%（p < 0.0001），错误代理干扰从28–57%降至0–14%，上下文效率比高达3.53倍。准确率优势随代理数N和决策密度D增加而提升；关键词匹配经大模型评判验证（平均kappa=0.909）。在第4阶段，于N=3时领先基线+17.2个百分点（p=0.0023），N=5时领先+20.4个百分点（p=0.0008），且优势随N增长被两位独立评委确认。
原文摘要 · Abstract (English)
Multi-agent LLM orchestration systems suffer from context pollution: when N concurrent agents compete for the orchestrator's context window, each agent's task state, partial outputs, and pending questions contaminate the steering interactions of every other agent, degrading decision quality. We introduce Dynamic Attentional Context Scoping (DACS), a mechanism in which the orchestrator operates in two asymmetric modes. In Registry mode it holds only lightweight per-agent status summaries (<=200 tokens each), remaining responsive to all agents and the user. When an agent emits a SteeringRequest, the orchestrator enters Focus(a_i) mode, injecting the full context of agent a_i while compressing all other agents to their registry entries. Context isolation is agent-triggered, asymmetric, and deterministic: the context window contains exactly F(a_i) + R_{-i} during steering, eliminating cross-agent contamination without requiring context compression or retrieval. We evaluate DACS across four experimental phases totalling 200 trials: Phase 1 tests N in {3,5,10} (60 trials); Phase 2 tests agent heterogeneity and adversarial dependencies (60 trials); Phase 3 tests decision density up to D=15 (40 trials); Phase 4 uses autonomous LLM agents for free-form questions (40 trials, Claude Haiku 4.5). Across all 8 synthetic scenarios, DACS achieves 90.0--98.4% steering accuracy versus 21.0--60.0% for a flat-context baseline (p < 0.0001 throughout), with wrong-agent contamination falling from 28--57% to 0--14% and context efficiency ratios of up to 3.53x. The accuracy advantage grows with N and D; keyword matching is validated by LLM-as-judge across all phases (mean kappa=0.909). DACS outperforms the flat-context baseline by +17.2pp at N=3 (p=0.0023) and +20.4pp at N=5 (p=0.0008) in Phase 4, with the advantage growing with N confirmed by two independent judges.
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。