{"ID":2864013,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.25543","arxiv_id":"2509.25543","title":"Aligning Multilingual Reasoning with Verifiable Semantics from a High-Resource Expert Model","abstract":"While reinforcement learning has advanced the reasoning abilities of Large Language Models (LLMs), these gains are largely confined to English, creating a significant performance disparity across languages. To address this, we introduce Pivot-Based Reinforcement Learning with Semantically Verifiable Rewards (PB-RLSVR), a novel framework that enhances multilingual reasoning by circumventing the need for human-annotated data in target languages. Our approach employs a high-performing English LLM as a \"pivot\" model to generate reference responses for reasoning tasks. A multilingual model is then rewarded based on the semantic equivalence of its responses to the English reference, effectively transferring the pivot model's reasoning capabilities across languages. We investigate several cross-lingual semantic reward functions, including those based on embeddings and machine translation. Extensive experiments on a suite of multilingual reasoning benchmarks show that our method significantly narrows the performance gap between English and other languages, substantially outperforming traditional PPO baselines. Specifically, our PB-RLSVR framework improves the average multilingual performance of Llama-3.1-8B-Instruct and Qwen3-32B by 16.41% and 10.17%, respectively, demonstrating a powerful and data-efficient approach to building truly multilingual reasoning agents.","short_abstract":"While reinforcement learning has advanced the reasoning abilities of Large Language Models (LLMs), these gains are largely confined to English, creating a significant performance disparity across languages. To address this, we introduce Pivot-Based Reinforcement Learning with Semantically Verifiable Rewards (PB-RLSVR),...","url_abs":"https://arxiv.org/abs/2509.25543","url_pdf":"https://arxiv.org/pdf/2509.25543v1","authors":"[\"Fahim Faisal\",\"Kaiqiang Song\",\"Song Wang\",\"Simin Ma\",\"Shujian Liu\",\"Haoyun Deng\",\"Sathish Reddy Indurthi\"]","published":"2025-09-29T22:03:11Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"Reinforcement Learning\",\"Large Language Model\",\"Language Model\"]","has_code":false}