Letting Tutor Personas"Speak Up"for LLMs: Learning Steering Vectors from Dialogue via Preference Optimization

This paper develops a method to learn tutor-specific personas for LLM-based tutors by using preference optimization to create steering vectors from human tutoring dialogues, enabling control over tutoring style (e.g., scaffolding level, affective support, instructional directiveness) without explicit prompting. The approach is evaluated on real math tutoring dialogues, showing improved alignment with individual tutor behaviors while maintaining pedagogical quality.

With the emergence of large language models (LLMs) as a powerful class of generative artificial intelligence (AI), their use in tutoring has become increasingly prominent. Prior works on LLM-based tutoring typically learn a single tutor policy and do not capture the diversity of tutoring styles. In real-world tutor-student interactions, pedagogical intent is realized through adaptive instructional strategies, with tutors varying the level of scaffolding, instructional directiveness, feedback, an

Study Type

Source

Framework Categories

Tool Types

Tags