1
Goal-Autopilot: A Verifiable Anti-Fabrication Firewall for Unattended Long-Horizon Agents
论文提出一种可验证的反伪造防火墙,让LLM代理在无人监督时无法虚报成功,将诚实作为自主任务的首要度量。
arXiv:2606.11688v1 Announce Type: cross Abstract: Long-horizon LLM agents are not trusted to run unattended: with no human watching, they confidently …