Saved in:
Bibliographic Details
Main Authors: Nam, Hyunji, Gottesman, Omer, Zhang, Amy, Foster, Dean, Brunskill, Emma, Ungar, Lyle
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.16252
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918101391507456
author Nam, Hyunji
Gottesman, Omer
Zhang, Amy
Foster, Dean
Brunskill, Emma
Ungar, Lyle
author_facet Nam, Hyunji
Gottesman, Omer
Zhang, Amy
Foster, Dean
Brunskill, Emma
Ungar, Lyle
contents Large language models (LLMs) built on existing reinforcement learning with human feedback (RLHF) frameworks typically optimize responses based on immediate turn-level human preferences. However, this approach falls short in multi-turn dialogue settings, such as online math tutoring. We propose a method to enhance LLM-based tutors by representing the dialogue history with a lower-dimensional latent state representation of a student and optimizing a long-term policy to determine high-level actions based on the latent state. The goal is to better align the tutor's behavior with the long-term objective of guiding the student towards solving a target math problem on their own. 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. Our experiment results demonstrate that these modifications lead to improved long-term outcomes compared to prompting in LLM-simulated tutoring tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2507_16252
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Efficient RL for optimizing conversation level outcomes with an LLM-based tutor
Nam, Hyunji
Gottesman, Omer
Zhang, Amy
Foster, Dean
Brunskill, Emma
Ungar, Lyle
Computation and Language
Artificial Intelligence
Large language models (LLMs) built on existing reinforcement learning with human feedback (RLHF) frameworks typically optimize responses based on immediate turn-level human preferences. However, this approach falls short in multi-turn dialogue settings, such as online math tutoring. We propose a method to enhance LLM-based tutors by representing the dialogue history with a lower-dimensional latent state representation of a student and optimizing a long-term policy to determine high-level actions based on the latent state. The goal is to better align the tutor's behavior with the long-term objective of guiding the student towards solving a target math problem on their own. 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. Our experiment results demonstrate that these modifications lead to improved long-term outcomes compared to prompting in LLM-simulated tutoring tasks.
title Efficient RL for optimizing conversation level outcomes with an LLM-based tutor
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2507.16252