发表机构
EPFL（洛桑联邦理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对LLM对话辅导缺乏教学策略推理时控制的问题，提出PIVOT框架，通过偏好干预向量实现导师动作控制，在用户研究中获多数教师认可。
AI 中文摘要
大型语言模型（LLM）越来越多地被用于对话式辅导，但有效的辅导不仅仅需要正确的答案。导师必须选择何时搭建推理框架、给出提示、提供反馈、进行解释或邀请反思。现有的提示和训练方法提升了教学对齐度，但缺乏对教学策略的可靠推理时控制。我们提出PIVOT，一种激活引导框架，用于为冻结的LLM导师在线学习基于偏好的干预向量。PIVOT采用七类导师动作分类法和“生成-标注-优化”循环，其中经人工验证的LLM评判器识别目标和易混淆的非目标动作，以构建用于多层残差流引导的偏好对。在保留的和域外辅导数据上，PIVOT可控制导师动作，同时保持相关性和流畅性，其方向可在推理时扩展、迁移和组合。在有30名教师参与的用户研究中，73.3%的参与者更偏好经引导的对话而非使用相同提示的中性基线交互，并认为这些控制清晰、可用且具有教学意义。
英文摘要
LLMs are increasingly used for conversational tutoring, but effective tutoring requires more than correct answers. Tutors must choose when to scaffold reasoning, hint, give feedback, explain, or invite reflection. Existing prompting and training methods improve pedagogical alignment, but lack reliable inference-time control over pedagogical strategies. We introduce PIVOT, an activation-steering framework that learns preference-based intervention vectors online for frozen LLM tutors. PIVOT uses a seven-category tutor-move taxonomy and a generate-label-optimise loop, where a human-validated LLM judge identifies target and confusable non-target moves to construct preference pairs for multi-layer residual-stream steering. Across held-out and out-of-domain tutoring data, PIVOT controls tutor moves while preserving relevance and fluency, and its directions can be scaled, transferred, and composed at inference time. In a user study with 30 teachers, 73.3% of participants preferred steered conversations over neutral baseline interactions using the same prompt, and rated the controls as clear, usable, and pedagogically meaningful.
