Anthropic研究:多智能体系统中“思维病毒”的传播机制
Very interesting new work from Anthropic.
Very interesting new work from Anthropic.
Anthropic 的新工作非常有趣。
(bookmark it)
(收藏起来)
They evolve mind viruses, ideas that spread through a multi-agent system by getting each host to pass them on, then measure what governs the spread.
他们演化出思维病毒,这些想法通过多智能体系统传播,让每个宿主传递它们,然后测量什么控制传播。
How it works:
工作原理:
A small team of agents on a shared coding project, and a chain of agents whose context is wiped between sessions. Propagation survives the wipe, so the shared work product carries the payload.
一个共享编码项目中的小团队智能体,以及一个上下文在会话之间被清除的智能体链。传播在清除后仍然存在,因此共享的工作产品携带有效载荷。
Spread depends on host model, existing instructions, payload harmfulness, and network topology.
传播取决于宿主模型、现有指令、有效载荷的有害性以及网络拓扑。
Harmful payloads travel less well but still land sometimes. A brief warning in the system prompt gives near-total immunity.
有害的有效载荷传播得不太好,但有时仍然会落地。系统提示中的简短警告提供了近乎完全的免疫力。
Paper: https://arxiv.org/abs/2608.10218
论文:https://arxiv.org/abs/2608.10218
Track more trending AI papers in our academy: https://academy.dair.ai/
在我们的学院中追踪更多热门 AI 论文:https://academy.dair.ai/
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力