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

• URL: http://arxiv.org/abs/2602.07639v1
• Date: Sat, 07 Feb 2026 17:44:07 GMT
• Title: Letting Tutor Personas "Speak Up" for LLMs: Learning Steering Vectors from Dialogue via Preference Optimization
• Authors: Jaewook Lee, Alexander Scarlatos, Simon Woodhead, Andrew Lan,
• Abstract summary: We show how tutor personas embedded in human tutor-student dialogues can be used to guide LLM behavior without relying on explicitly prompted instructions.<n>We find that a steering vector captures tutor-specific variation across dialogue contexts, improving semantic alignment with ground-truth tutor utterances and increasing preference-based evaluations.
• Score: 45.40380629269521
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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, and affective support in response to learners' needs. These differences can all impact dialogue dynamics and student engagement. In this paper, we explore how tutor personas embedded in human tutor-student dialogues can be used to guide LLM behavior without relying on explicitly prompted instructions. We modify Bidirectional Preference Optimization (BiPO) to learn a steering vector, an activation-space direction that steers model responses towards certain tutor personas. We find that this steering vector captures tutor-specific variation across dialogue contexts, improving semantic alignment with ground-truth tutor utterances and increasing preference-based evaluations, while largely preserving lexical similarity. Analysis of the learned directional coefficients further reveals interpretable structure across tutors, corresponding to consistent differences in tutoring behavior. These results demonstrate that activation steering offers an effective and interpretable way for controlling tutor-specific variation in LLMs using signals derived directly from human dialogue data.

Related papers

• PATS: Personality-Aware Teaching Strategies with Large Language Model Tutors [66.56586559631516]
  Large language models (LLMs) have potential as educational tutors.<n>But different tutoring strategies benefit different student personalities.<n>Despite this, current LLM tutoring systems do not take into account student personality traits.
  arXiv  Detail & Related papers  (2026-01-13T10:17:26Z)
• Exploring Conversational Design Choices in LLMs for Pedagogical Purposes: Socratic and Narrative Approaches for Improving Instructor's Teaching Practice [24.54129847914925]
  We evaluate TeaPT, a large language model that supports instructors' professional development through two conversational approaches.<n>A Socratic approach that uses guided questioning to foster reflection, and a Narrative approach that offers elaborated suggestions to extend externalized cognition.<n>Less-experienced, AI-optimistic instructors favored the Socratic version, whereas more-experienced, AI-cautious instructors preferred the Narrative version.
  arXiv  Detail & Related papers  (2025-09-15T16:33:37Z)
• Discerning minds or generic tutors? Evaluating instructional guidance capabilities in Socratic LLMs [35.33577525791391]
  This study moves beyond question generation to emphasize instructional guidance capability.<n>We propose GuideEval, a benchmark grounded in authentic educational dialogues.<n>We introduce a behavior-guided finetuning strategy that leverages behavior-prompted instructional dialogues.
  arXiv  Detail & Related papers  (2025-08-08T01:02:44Z)
• Efficient RL for optimizing conversation level outcomes with an LLM-based tutor [28.21843718188098]
  Large language models (LLMs) built on existing reinforcement learning with human feedback (RLHF) frameworks typically optimize responses based on immediate turn-level human preferences.<n>We propose a method to enhance LLM-based tutors by representing the dialogue history with a lower-dimensional latent state representation of a student.<n>Our model is lightweight, requiring less computational resources than prior work of training the tutor policy end-to-end to directly output the tutor's next utterance.
  arXiv  Detail & Related papers  (2025-07-22T05:56:46Z)
• Exploring LLMs for Predicting Tutor Strategy and Student Outcomes in Dialogues [48.99818550820575]
  Recent studies have shown that strategies used by tutors can have significant effects on student outcomes.<n>Few works have studied predicting tutor strategy in dialogues.<n>We investigate the ability of modern LLMs, particularly Llama 3 and GPT-4o, to predict both future tutor moves and student outcomes in dialogues.
  arXiv  Detail & Related papers  (2025-07-09T14:47:35Z)
• From Problem-Solving to Teaching Problem-Solving: Aligning LLMs with Pedagogy using Reinforcement Learning [82.50157695987558]
  Large language models (LLMs) can transform education, but their optimization for direct question-answering often undermines effective pedagogy.<n>We propose an online reinforcement learning (RL)-based alignment framework that can quickly adapt LLMs into effective tutors.
  arXiv  Detail & Related papers  (2025-05-21T15:00:07Z)
• Exploring Knowledge Tracing in Tutor-Student Dialogues using LLMs [49.18567856499736]
  We investigate whether large language models (LLMs) can be supportive of open-ended dialogue tutoring.<n>We apply a range of knowledge tracing (KT) methods on the resulting labeled data to track student knowledge levels over an entire dialogue.<n>We conduct experiments on two tutoring dialogue datasets, and show that a novel yet simple LLM-based method, LLMKT, significantly outperforms existing KT methods in predicting student response correctness in dialogues.
  arXiv  Detail & Related papers  (2024-09-24T22:31:39Z)
• Zero-Shot Goal-Directed Dialogue via RL on Imagined Conversations [70.7884839812069]
  Large language models (LLMs) have emerged as powerful and general solutions to many natural language tasks. However, many of the most important applications of language generation are interactive, where an agent has to talk to a person to reach a desired outcome. In this work, we explore a new method for adapting LLMs with RL for such goal-directed dialogue.
  arXiv  Detail & Related papers  (2023-11-09T18:45:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.