腾讯混元发布《Diving into Reliable Self-Evolving Agents: A Survey》，系统梳理智能体自我进化机制及验证其改进的可靠性证据。该综述提出L0–L4分级体系、可信更新可靠性阶梯，并收录549篇相关论文于开放目录，核心原则是任何更新不得控制用于验证自身的唯一证据。
📄New Research on Self-Evolving Agents:
When AI agents modify themselves, how do we know they actually got better?
We present Diving into Reliable Self-Evolving Agents: A Survey—a systematic map of how agents self-evolve and what evidence is needed to trust each update.
The survey:
🔹 Defines an L0–L4 taxonomy for self-evolving agents
🔹 Introduces a reliability ladder for trustworthy updates
🔹 Curates 549 works in an open companion catalog
One core principle: no update should control the only evidence used to accept itself.
Explore the full survey ↓
📄 Paper: https://openreview.net/forum?id=CGO1hDTHNe
🌐 Project: https://wkqdzkd.github.io/Awesome-Reliable-Self-Evolving-Agents/
💻 GitHub: https://github.com/wkqdzkd/Awesome-Reliable-Self-Evolving-Agents
来源：Tencent Hy · x.com