# Application-Driven Pedagogical Knowledge Optimization of Open-Source LLMs via Reinforcement Learning and Supervised Fine-Tuning

Source: [Application-Driven Pedagogical Knowledge Optimization of Open-Source LLMs via Reinforcement Learning and Supervised Fine-Tuning](<https://www.sota2.com/research/paper/69d732717d0172140ec76a15>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-10T05:00:04.898Z

Authors: Navan Preet Singh, Xiaokun Wang, Anurag Garikipati, Madalina Ciobanu, Qingqing Mao, Ritankar Das

Published: 2026-04-07T19:16:50.000Z

## Abstract

We present an innovative multi-stage optimization strategy combining reinforcement learning (RL) and supervised fine-tuning (SFT) to enhance the pedagogical knowledge of large language models (LLMs), as illustrated by EduQwen 32B-RL1, EduQwen 32B-SFT, and an optional third-stage model EduQwen 32B-SFT-RL2: (1) RL optimization that implements progressive difficulty training, focuses on challenging examples, and employs extended reasoning rollouts; (2) a subsequent SFT phase that leverages the RL-trained model to synthesize high-quality training data with difficulty-weighted sampling; and (3) an optional second round of RL optimization. EduQwen 32B-RL1, EduQwen 32B-SFT, and EduQwen 32B-SFT-RL2 are an application-driven family of open-source pedagogical LLMs built on a dense Qwen3-32B backbone. These models remarkably achieve high enough accuracy on the Cross-Domain Pedagogical Knowledge (CDPK) Benchmark to establish new state-of-the-art (SOTA) results across the interactive Pedagogy Benchmark Leaderboard and surpass significantly larger proprietary systems such as the previous benchmark leader Gemini-3 Pro. These dense 32-billion-parameter models demonstrate that domain-specialized optimization can transform mid-sized open-source LLMs into true pedagogical domain experts that outperform much larger general-purpose systems, while preserving the transparency, customizability, and cost-efficiency required for responsible educational AI deployment.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Pedagogical Knowledge Evaluation on CDPK (Cross-Domain Pedagogical Knowledge) Benchmark](<https://www.sota2.com/research/sota/pedagogical-knowledge-evaluation-on-cdpk-cross-domain-pedagogical-knowledge-benchmark.md>): Pedagogical Knowledge Evaluation; CDPK (Cross-Domain Pedagogical Knowledge) Benchmark; Replicated Accuracy: 96.52.

- [LLM Tutoring Evaluation on TutorBench](<https://www.sota2.com/research/sota/llm-tutoring-evaluation-on-tutorbench.md>): LLM Tutoring Evaluation; TutorBench; Text-Only Score: 64.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.06385>)

[PDF](<https://arxiv.org/pdf/2604.06385.pdf>)