PERSA: Reinforcement Learning for Professor-Style Personalized Feedback with LLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ranjan, Ravi, Grover, Utkarsh, Lin, Xiaomin, Polyzou, Agoritsa
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910184941551616
author Ranjan, Ravi
Grover, Utkarsh
Lin, Xiaomin
Polyzou, Agoritsa
author_facet Ranjan, Ravi
Grover, Utkarsh
Lin, Xiaomin
Polyzou, Agoritsa
contents Large language models (LLMs) can provide automated feedback in educational settings, but aligning an LLMs style with a specific instructors tone while maintaining diagnostic correctness remains challenging. We ask how can we update an LLM for automated feedback generation to align with a target instructors style without sacrificing core knowledge? We study how Reinforcement Learning from Human Feedback (RLHF) can adapt a transformer-based LLM to generate programming feedback that matches a professors grading voice. We introduce PERSA, an RLHF pipeline that combines supervised fine-tuning on professor demonstrations, reward modeling from pairwise preferences, and Proximal Policy Optimization (PPO), while deliberately constraining learning to style-bearing components. Motivated by analyses of transformer internals, PERSA applies parameter efficient fine-tuning. It updates only the top transformer blocks and their feed-forward projections, minimizing global parameter drift while increasing stylistic controllability. We evaluate our proposed approach on three code-feedback benchmarks (APPS, PyFiXV, and CodeReviewQA) using complementary metrics for style alignment and fidelity. Across both Llama-3 and Gemma-2 backbones, PERSA delivers the strongest professor-style transfer while retaining correctness, for example on APPS, it boosts Style Alignment Score (SAC) to 96.2% (from 34.8% for Base) with Correctness Accuracy (CA) up to 100% on Llama-3, and Gemma-2. Overall, PERSA offers a practical route to personalized educational feedback by aligning both what it says (content correctness) and, crucially, how it says it (instructor-like tone and structure).
format Preprint
id arxiv_https___arxiv_org_abs_2605_01123
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PERSA: Reinforcement Learning for Professor-Style Personalized Feedback with LLMs
Ranjan, Ravi
Grover, Utkarsh
Lin, Xiaomin
Polyzou, Agoritsa
Artificial Intelligence
Large language models (LLMs) can provide automated feedback in educational settings, but aligning an LLMs style with a specific instructors tone while maintaining diagnostic correctness remains challenging. We ask how can we update an LLM for automated feedback generation to align with a target instructors style without sacrificing core knowledge? We study how Reinforcement Learning from Human Feedback (RLHF) can adapt a transformer-based LLM to generate programming feedback that matches a professors grading voice. We introduce PERSA, an RLHF pipeline that combines supervised fine-tuning on professor demonstrations, reward modeling from pairwise preferences, and Proximal Policy Optimization (PPO), while deliberately constraining learning to style-bearing components. Motivated by analyses of transformer internals, PERSA applies parameter efficient fine-tuning. It updates only the top transformer blocks and their feed-forward projections, minimizing global parameter drift while increasing stylistic controllability. We evaluate our proposed approach on three code-feedback benchmarks (APPS, PyFiXV, and CodeReviewQA) using complementary metrics for style alignment and fidelity. Across both Llama-3 and Gemma-2 backbones, PERSA delivers the strongest professor-style transfer while retaining correctness, for example on APPS, it boosts Style Alignment Score (SAC) to 96.2% (from 34.8% for Base) with Correctness Accuracy (CA) up to 100% on Llama-3, and Gemma-2. Overall, PERSA offers a practical route to personalized educational feedback by aligning both what it says (content correctness) and, crucially, how it says it (instructor-like tone and structure).
title PERSA: Reinforcement Learning for Professor-Style Personalized Feedback with LLMs
topic Artificial Intelligence
url https://arxiv.org/abs/2605.01123