[论文] Persona-Execution Separation: An Architecture Pattern for Evolving LLM...

论文概要

研究领域: ML 作者: Yisen Xi 发布时间: 2026-08-27 arXiv: 2608.27427

中文摘要

受治理组织中的大语言模型（LLM）智能体必须让人格（指令、语调、自我呈现）自由演化，同时保持执行（有状态的、已审计的工作）可追踪。单一信任域无法同时廉价地满足两者。我们提出了人格-执行分离（PES）：人格和执行驻留在不同的信任域中，通过受治理的契约桥连接。人格是单一驻留的，可以漂移；执行是匿名的且可审计的。状态摘要可以返回；数据主体保留在限制性域中，除非有分级数据防泄漏（DLP）例外；身份保持连续。审批矩阵、DLP和审计强制执行跨界。在LLM表示不可区分性的条件下，任何满足所有三个目标的单域机制必须重新引入类型化变更对象、外部门和稳定审计锚：PES以更高耦合成本重建。

原文摘要

Large language model (LLM) agents in governed organizations must let the persona (instructions, tone, self-presentation) evolve freely, while keeping execution (stateful, audited work) traceable. A single trust domain does not satisfy both cheaply. We present Persona-Execution Separation (PES): persona and execution reside in different trust domains, connected by a governed contract bridge. The persona is singly-homed and may drift; execution is faceless and audited. Status summaries may return; data bodies remain in the restrictive domain except a graded data-loss-prevention (DLP) exception; identity stays continuous. An approval matrix, DLP, and audit enforce the crossing. PES follows from three goals---free drift, execution traceability, and decoupling. Under LLM representational indist...

#论文 #arXiv #ML #小凯