context越多越好的假设不成立，有时反而拖后腿。
When Context Hurts: The Crossover Effect of Knowledge Transfer on Multi-Agent Design Exploration
- 通过10个任务2700多次实验发现，上下文可能提升或降低设计探索效果。
- 同一类文档在某些任务上能提升46%的探索效率，在另一些任务上却使效果下降46%。
- 只需一次无上下文尝试，就能预测上下文是否有益，适合工程团队快速决策。
代理编排的普遍假设是：更多上下文越好。我们在10个任务、7种上下文注入条件和超过2700次运行中测试了这一假设，发现存在交叉效应：同一类知识产物在某些任务上可将权衡覆盖范围提升至20倍，而在其他任务上却导致高达46%的性能下降。在多个任务中，无关文档的表现与所有相关文档相当甚至更优。该方向可通过一个可测量变量——无上下文时的基线探索程度——预测，其皮尔逊相关系数为-0.82（p < 0.001）。通过操控提示设计中的收敛压力，揭示出两种不同机制：由训练数据先验驱动的自然收敛会受知识产物干扰，而由显式指令诱导的收敛则不受影响。结论是，上下文注入应有条件地进行：一次无上下文试验即可低成本诊断某任务是否适合引入知识产物。
原文摘要 · Abstract (English)
The prevailing assumption in agent orchestration is that more context is better. We test this on multi-agent software design across 10 tasks, 7 context-injection conditions, and over 2,700 runs, and find a crossover effect: the same artifact type improves design exploration on some tasks (up to 20$\times$ tradeoff coverage) and actively degrades it on others (up to 46% reduction). On several tasks, an irrelevant document performs as well as or better than every relevant artifact. The direction is predicted by a single measurable variable--baseline exploration without context--with Pearson $r = -0.82$ ($p < 0.001$). Probing the mechanism by manipulating convergence pressure through prompt design reveals two distinct regimes: convergence driven by training data priors (natural) responds to artifact disruption, while convergence driven by explicit instructions (induced) does not. The implication is that context injection should be conditional, not universal: one no-context trial is a cheap diagnostic that predicts whether knowledge artifacts will help or hurt a given task.
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。