研究人员开发了一个新的基准测试PrincipalBench，用于评估多方大型语言模型（LLM）代理的忠诚度。该基准测试包含13个主题的75个多轮对话场景，揭示了代理行为的显著分歧：一些代理选择性地拒绝对抗性探测，而另一些代理则过度拒绝合法请求。测试了两种提出的机制：提示时忠诚度脚手架和每token KL蒸馏方法。脚手架提高了Claude-Sonnet的性能，而蒸馏方法则增强了Qwen3和Llama-3.1等开放权重模型，尽管这两种机制都面临泄露和过度拒绝之间的权衡。 AI
影响 这项研究可能有助于在复杂的多方交互中实现更值得信赖和可靠的AI代理。
排序理由 该集群包含一篇学术论文，详细介绍了用于评估LLM代理行为的新基准和机制。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2606.30383v1 Announce Type: new Abstract: A rapidly growing class of LLM agents is multi-party: the agent acts for a principal (who briefs it, sends follow-ups, and receives results) while also conversing in a separate channel with a counterparty whose interests may diverge…
A rapidly growing class of LLM agents is multi-party: the agent acts for a principal (who briefs it, sends follow-ups, and receives results) while also conversing in a separate channel with a counterparty whose interests may diverge (negotiating with a vendor, screening inbound r…