大语言模型生成器-调节器博弈的变分框架
A Variational Framework for LLM Generator-Regulator Games
AI总结：
提出一个变分框架，将语言生成建模为熵正则化吉布斯分布，将调节建模为最优判别器，通过鞍点问题平衡效用、熵、调节一致性和有限长度可检测性，并通过审查过滤和钓鱼防御案例验证。
摘要：
本文发展了一个用于受调节语言生成的变分框架。从自回归令牌采样出发，我们推导了完整消息上的诱导分布，并将其与熵正则化的吉布斯定律联系起来。调节被建模为一个最优判别器，其对偶凸值为f-散度，生成器-调节器交互被表述为一个鞍点问题。该框架适用于内容审核、审查、AI欺骗检测、合规审计、钓鱼防御和操纵控制，其中调节涉及可能消息上的分布而非单个输出。均衡阐明了效用、熵、调节一致性和有限长度可检测性之间的权衡。两个有限词汇案例研究，即审查过滤和钓鱼防御，说明了如何通过效用、熵、散度、接收端分数和检测概率来评估该理论。
英文摘要：
This paper develops a variational framework for regulated language generation. Starting from autoregressive token sampling, we derive the induced distribution over complete messages and relate it to an entropy-regularized Gibbs law. Regulation is modeled as an optimal discriminator whose convex-dual value is an f-divergence, and the generator-regulator interaction is formulated as a saddle-point problem. The framework applies to moderation, censorship, AI deception detection, compliance auditing, phishing defense, and manipulation control, where regulation concerns a distribution over possible messages rather than a single output. The equilibrium clarifies the tradeoff among utility, entropy, regulatory alignment, and finite-length detectability. Two finite-vocabulary case studies, censorship filtering and phishing defense, illustrate how the theory can be evaluated through utility, entropy, divergence, receiver-side scores, and detection probability.
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。
