Studying the Korean Word-Chain Game with RLVR: Mitigating Reward Conflicts via Curriculum Learning

2025/10/03 by Donghwan Rho, Rho, Donghwan
Computer Science · Arts and Humanities · #Education and Learning Interventions #Educational Systems and Policies #Translation Studies and Practices

paper · pdf · doi:10.48550/arxiv.2510.03394

Abstract

Reinforcement learning with verifiable rewards (RLVR) is a promising approach for training large language models (LLMs) with stronger reasoning abilities. It has also been applied to a variety of logic puzzles. In this work, we study the Korean word-chain game using RLVR. We show that rule-derived rewards can naturally conflict, and demonstrate through experiments that a curriculum-learning scheme mitigates these conflicts. Our findings motivate further studies of puzzle tasks in diverse languages.

Citations

• VCRL: Variance-based Curriculum Reinforcement Learning for Large Language Models
• REASONING GYM: Reasoning Environments for Reinforcement Learning with Verifiable Rewards
• Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles
• Qwen3 Technical Report
• Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learning
• DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning
• Tulu 3: Pushing Frontiers in Open Language Model Post-Training
• Curriculum Reinforcement Learning for Complex Reward Functions
• Curriculum Learning: A Survey
• Curriculum Learning: A Survey
• Reverse Curriculum Generation for Reinforcement Learning
• Reinforcement Learning: An Introduction

Related

• Can One Domain Help Others? A Data-Centric Study on Multi-Domain Reasoning via Reinforcement Learning
• Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
• Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecycle
• VCRL: Variance-based Curriculum Reinforcement Learning for Large Language Models
• The Reasoning Boundary Paradox: How Reinforcement Learning Constrains Language Models