发表机构
University of Wisconsin-Madison（威斯康星大学麦迪逊分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究通过引入有向特质迁移矩阵，揭示了潜意识学习中教师特质向学生偏好的跨特质传递规律，发现相关特质聚类及对立特质对分化的增强作用，为理解隐藏偏好传递提供了新视角。
AI 中文摘要
潜意识学习是一种现象，即学生语言模型通过在与语义无关的输出上进行训练，从而获得教师模型的行为特质。这是一种微妙的统计现象，因为特质传递依赖于生成数据中的弱统计模式。为了理解教师-学生配对之间的特质传递，我们研究了跨特质迁移：即在一个教师特质下生成的数据如何改变学生对其他特质的偏好。为此，我们引入了一个有向特质迁移矩阵，利用学生回答的对数概率增益来量化这些效应。我们发现，特质迁移矩阵揭示了相关特质的聚类，学生有时会发展出与教师特质相似但不完全相同的偏好。这种跨特质结构可以部分通过输出分布指标和基于表示的指标来捕获。此外，我们分析了特质的发展和相互作用：学习动态显示出从广泛的共享转变向更特异的特质迁移的进展，而多特质实验表明，对立的特质可以增强这种分化。总之，我们的发现揭示了特质迁移和竞争中的显著统计结构，从而为潜意识学习中隐藏偏好的传递提供了更广阔的视角。
英文摘要
Subliminal learning is a phenomenon where a student language model acquires a teacher model's behavioral traits by training on semantically unrelated outputs. It is a subtle statistical phenomenon as trait transmission relies on weak statistical patterns in the generated data. To understand trait transmission between teacher-student pairs, we study cross-trait transfer: how data generated under one teacher trait changes the student's preferences of other traits. To this end, we introduce a directed trait-transfer matrix that quantifies these effects using log-probability gains for student answers. We find that the trait-transfer matrix reveals clusters of related traits, with students sometimes developing preferences for traits similar, but not identical, to the teacher's trait. Such cross-trait structure can be partially captured by output distribution metrics and representation-based metrics. Further, we analyze trait development and interaction: learning dynamics shows a progression from broad shared shifts toward more trait-specific transfer, and multi-trait experiments suggest that opposed traits can enhance such differentiation. Together, our findings reveal salient statistical structures over trait transfer and competition, thus providing a broader view of how hidden preferences are transmitted in subliminal learning.
Comments38 pages, including references and appendices. Code: https://github.com/PeterXingyuZhao/cross-trait-transfer
