Studying the Korean Word-Chain Game with RLVR: Mitigating Reward Conflicts via Curriculum Learning
This work addresses reward conflicts in RLVR for a specific puzzle task, which is incremental in nature.
The paper tackled reward conflicts in reinforcement learning with verifiable rewards (RLVR) for the Korean word-chain game, demonstrating that a curriculum-learning scheme mitigates these conflicts, as shown through experimental results.
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.