用于推理时的可控且可解释的人格滑块
Controllable and explainable personality sliders for LLMs at inference time
AI总结：
本文提出一种模块化框架，通过顺序自适应引导方法实现LLM推理时的可控且可解释的人格控制，通过正交化引导向量提升多维人格调节的精度和效率。
摘要：
将大型语言模型（LLMs）与特定人设对齐通常依赖于昂贵且庞大的监督微调（SFT）或强化学习与人类反馈（RLHF）。虽然有效，但这些方法需要为每个目标人设训练不同的模型。推理时的激活引导提供了一种参数高效的方法，但 naive 方法无法同时控制多个特性，因为向量干扰具有破坏性。在本工作中，我们提出了一种连续、多维的人格控制模块化框架。我们的关键创新是顺序自适应引导（SAS）：一种通过在残差流上训练后续探测器来正交化引导向量的方法。这种方法将引导向量转化为可重用的原语，允许用户通过简单地调整系数 alpha 来即时合成复杂的、高质量的人格配置文件。我们在 Big Five 人格特征上验证了我们的框架，证明其在目标一致性和连贯性方面优于 naive 基线，使精确、全面的人格调节成为可能，而无需更新模型参数。
英文摘要：
Aligning Large Language Models (LLMs) with specific personas typically relies on expensive and monolithic Supervised Fine-Tuning (SFT) or RLHF. While effective, these methods require training distinct models for every target personality profile. Inference-time activation steering offers a parameter-efficient alternative, yet naive approaches fail to control multiple traits simultaneously due to destructive vector interference. In this work, we propose a modular framework for continuous, multi-dimensional personality control. Our key innovation is Sequential Adaptive Steering (SAS): a method that orthogonalizes steering vectors by training subsequent probes on the residual stream shifted by prior interventions. This approach transforms steering vectors into reusable primitives, allowing users to instantly synthesize complex, high-fidelity personality profiles by simply adjusting coefficients alpha. We validate our framework on the Big Five personality traits, demonstrating that it outperforms naive baselines in both goal adherence and coherence, enabling precise, holistic personality modulation without updating model parameters.
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。
