用一句话沟通比复杂训练更有效提升大模型协作
Communication Enables Cooperation in LLM Agents: A Comparison with Curriculum-Based Approaches
- 通过简单对话让4人博弈中合作率从0%升至96.7%
- 渐进式课程训练使收益降低27.4%，反而削弱合作
- 适合研究多智能体对齐与协作机制的学者
在多智能体大模型系统中，实现协作对人工智能对齐至关重要。我们比较了直接通信与课程学习两种方法。在四人猎鹿博弈中，仅用一个词的“廉价对话”通道，将合作率从0%提升至96.7%，证明通信是强有力的协调机制。相反，我们发现课程学习对设计极为敏感：在包含惩罚机制的重复公共品博弈中，通过逐步增加难度的教育型课程，导致智能体收益下降27.4%，表明短期理性优化可能反向损害对齐目标。定性分析显示，强调背叛均衡的游戏序列会引发智能体的‘习得性悲观’。这些结果表明，在协调问题中，简单通信协议可能比基于经验的训练更可靠；而针对社会困境的课程设计，需谨慎考虑游戏序列中隐含的战略启示。
原文摘要 · Abstract (English)
Eliciting cooperation in multi-agent LLM systems is critical for AI alignment. We investigate two approaches: direct communication and curriculum learning. In a 4-player Stag Hunt, a one-word "cheap talk" channel increases cooperation from 0% to 96.7%, demonstrating communication as a robust coordination mechanism. In contrast, we find that curriculum learning is highly sensitive to design choices: our pedagogical curriculum through progressively complex games reduced agent payoffs by 27.4% in an Iterated Public Goods Game with Punishment, demonstrating that optimizing for short-term rationality can actively undermine alignment goals. Qualitative analysis reveals that curricula emphasizing defection-equilibrium games can induce "learned pessimism" in agents. These findings suggest that for coordination problems, simple communication protocols may be more reliable than experience-based training, and that curriculum design for social dilemmas requires careful attention to the strategic lessons embedded in game sequences.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。