Parallel Scaling Law: Unveiling Reasoning Generalization through A Cross-Linguistic Perspective

2025/10/02 by Wen Yang, Junhong Wu, Yang, Wen +7
Computer Science · #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques

paper · pdf · doi:10.48550/arxiv.2510.02272

openalex publication_date 2025/10/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Recent advancements in Reinforcement Post-Training (RPT) have significantly enhanced the capabilities of Large Reasoning Models (LRMs), sparking increased interest in the generalization of RL-based reasoning. While existing work has primarily focused on investigating its generalization across tasks or modalities, this study proposes a novel cross-linguistic perspective to investigate reasoning generalization. This raises a crucial question: Does the reasoning capability achieved from English RPT effectively transfer to other languages? We address this by systematically evaluating English-centric LRMs on multilingual reasoning benchmarks and introducing a metric to quantify cross-lingual transferability. Our findings reveal that cross-lingual transferability varies significantly across initial model, target language, and training paradigm. Through interventional studies, we find that models with stronger initial English capabilities tend to over-rely on English-specific patterns, leading to diminished cross-lingual generalization. To address this, we conduct a thorough parallel training study. Experimental results yield three key findings: First-Parallel Leap, a substantial leap in performance when transitioning from monolingual to just a single parallel language, and a predictable Parallel Scaling Law, revealing that cross-lingual reasoning transfer follows a power-law with the number of training parallel languages. Moreover, we identify the discrepancy between actual monolingual performance and the power-law prediction as Monolingual Generalization Gap, indicating that English-centric LRMs fail to fully generalize across languages. Our study challenges the assumption that LRM reasoning mirrors human cognition, providing critical insights for the development of more language-agnostic LRMs.

Citations

• Look Again, Think Slowly: Enhancing Visual Reflection in Vision-Language Models
• ACE-RL: Adaptive Constraint-Enhanced Reward for Long-form Generation Reinforcement Learning
• Does Learning Mathematical Problem-Solving Generalize to Broader Reasoning?
• Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning
• Just Go Parallel: Improving the Multilingual Capabilities of Large Language Models
• OpenThoughts: Data Recipes for Reasoning Models
• KTAE: A Model-Free Algorithm to Key-Tokens Advantage Estimation in Mathematical Reasoning
• Language Mixing in Reasoning Language Models: Patterns, Impact, and Internal Causes
• Crosslingual Reasoning through Test-Time Scaling
• X-Reasoner: Towards Generalizable Reasoning Across Modalities and Domains
• Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Model
• Understanding R1-Zero-Like Training: A Critical Perspective
• SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
• DAPO: An Open-Source LLM Reinforcement Learning System at Scale
• Implicit Cross-Lingual Rewarding for Efficient Multilingual Preference Alignment
• LIMO: Less is More for Reasoning
• s1: Simple test-time scaling
• SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training
• Kimi k1.5: Scaling Reinforcement Learning with LLMs
• DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning
• Tulu 3: Pushing Frontiers in Open Language Model Post-Training
• Cross-lingual Transfer of Reward Models in Multilingual Alignment
• Language Imbalance Driven Rewarding for Multilingual Self-improving
• Layer Swapping for Zero-Shot Cross-Lingual Transfer in Large Language Models
• Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
• X-Instruction: Aligning Language Model in Low-resource Languages with Self-curated Cross-lingual Instructions
• Reuse Your Rewards: Reward Model Transfer for Zero-Shot Cross-Lingual Alignment
• Revealing the Parallel Multilingual Learning within Large Language Models
• Zero-shot cross-lingual transfer in instruction tuning of large language models
• BioXP-0.5B: Explainable Medical-AI via RL-GRPO
• GPQA: A Graduate-Level Google-Proof Q&A Benchmark
• Efficient Memory Management for Large Language Model Serving with PagedAttention
• BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
• Measuring Mathematical Problem Solving With the MATH Dataset
• Trust Region Policy Optimization
• When Models Reason in Your Language: Controlling Thinking Language Comes at the Cost of Accuracy
• OpenAI o1 System Card

Related

• Beyond English-Centric Training: How Reinforcement Learning Improves Cross-Lingual Reasoning in LLMs
• Cross-lingual Collapse: How Language-Centric Foundation Models Shape Reasoning in Large Language Models
• Efficient Multilingual Reasoning Transfer via Progressive Code-Switching
• Crosslingual Reasoning through Test-Time Scaling
• Think Natively: Unlocking Multilingual Reasoning with Consistency-Enhanced Reinforcement Learning