隐私保护的LLM级联通过CoT增强的策略学习
Privacy-preserved LLM Cascade via CoT-enhanced Policy Learning
AI总结：
针对设备端LLM性能不足，提出CoT增强的策略学习框架P3Defer，用于隐私保护的级联延迟决策，提升效率并降低隐私风险，在三个基准数据集上验证了优越性。
摘要：
大型语言模型（LLMs）因其在现实世界任务中的卓越性能，在设备端应用中受到了广泛关注。然而，由于硬件限制，设备端LLM往往性能欠佳。解决这一挑战的一个有前景的方案是将较弱的本地（设备端）LLM与更强大的服务器LLM进行级联。虽然现有的LLM级联研究主要优化性能-成本权衡，但现实应用还提出了额外要求，如隐私保护，这些在很大程度上仍未得到解决。在这项工作中，我们超越了现有的基于置信度和基于logit的LLM级联方法，提出了P3Defer，一种新颖的思维链（CoT）增强的策略学习框架，用于隐私保护的延迟决策。我们的方法有效提高了级联效率，同时降低了隐私风险。在三个基准数据集上的广泛实验证明了P3Defer相对于现有方法的有效性和优越性。
英文摘要：
Large Language Models (LLMs) have gained significant attention in on-device applications due to their remarkable performance across real-world tasks. However, on-device LLMs often suffer from suboptimal performance due to hardware limitations. A promising solution to this challenge is cascading a weaker local (on-device) LLM with a more powerful server LLM. While existing research on LLM cascade primarily optimizes the performance-cost trade-off, real-world applications impose additional requirements, such as privacy preservation, which remain largely unaddressed. In this work, we move beyond existing confidence- and logit-based LLM cascade methods and propose $\mathbf{P^{3}Defer}$, a novel Chain-of-Thought (CoT)-enhanced \textbf{p}olicy learning framework for \textbf{p}rivacy-\textbf{p}reserved \textbf{defer}ral decision-making. Our approach effectively improves cascade efficiency while mitigating privacy risks. Extensive experiments on three benchmark datasets demonstrate the effectiveness and superiority of $\mathbf{P^{3}Defer}$ over existing methods.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
