mlsanityData to 14 Oct 2026

AdaMame: A Training Recipe for Adaptive Multilingual Reasoning

arXiv 2606.15080 ↗First posted 13 Jun 2026cs.CL cs.AIcited by 0

On the map: LLM reasoning → Reasoning benchmarks and evaluation

Abstract (arXiv)

While Large Reasoning Models (LRMs) show strong performance in English, they often fail to reason in the language of the query, a phenomenon known as language collapse. Existing RL-based fixes typically add a binary language fidelity reward to the accuracy objective, yet still incur trade-off in accuracy, mid-trace code-switching, and excessive token usage. In this work, we propose AdaMame, a two-stage training recipe for multilingual mathematical reasoning that addresses these limitations by adaptively aligning the reasoning language to the query language without compromising accuracy. The first SFT stage fine-tunes on non-MT reasoning traces across five languages to establish multilingual reasoning capability. In the subsequent RL stage, we introduce AdaMame-GRPO, an adaptation of Group Relative Policy Optimization (GRPO) in which a query-conditioned alignment factor grows progressively during training, guiding the model to first explore diverse reasoning languages before exploiting reasoning in the query language. Evaluated across two benchmarks, two LRMs, and 12 languages, AdaMame-GRPO achieves Pareto-optimal performance across reasoning accuracy, language fidelity, and token efficiency over all baselines, with the strongest gains on out-of-domain, lower-resource languages.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 24 of its 65 references that are in the map. First posted on arXiv 13 Jun 2026.

Builds on

24 of 65 references.

Closest ancestors · 8

Same direction, newest first

  1. 2026ReasonXL: Shifting LLM Reasoning Language Without Sacrificing Performance14 Apr
  2. 2026R3S: Refining and Recovering Reinforcement Signals for Multilingual Understanding and Reasoning5 Feb
  3. 2025Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning7 Jul
  4. 2025Cross-lingual Collapse: How Language-Centric Foundation Models Shape Reasoning in Large Language Models6 Jun
  5. 2025Language Matters: How Do Multilingual Input and Reasoning Paths Affect Large Reasoning Models?23 May
  6. 2025Qwen3 Technical Report14 May
  7. 2025Crosslingual Reasoning through Test-Time Scaling8 May
  8. 2025Could Thinking Multilingually Empower LLM Reasoning?16 Apr

Other references · 16

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  3. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  4. 2024Qwen2.5 Technical Reportcited by 4,942
  5. 2025Understanding R1-Zero-Like Training: A Critical Perspectivecited by 1,345
  6. 2025OpenAI GPT-5 System Cardcited by 989
  7. 2022Language Models are Multilingual Chain-of-Thought Reasonerscited by 530
  8. 2025Stop Overthinking: A Survey on Efficient Reasoning for Large Language Modelscited by 464
  9. 2025A Survey on Test-Time Scaling in Large Language Models: What, How, Where, and How Well?cited by 223
  10. 2025Magistralcited by 60
  11. 2025Long Chain-of-Thought Reasoning Across Languagescited by 19
  12. 2025Language Models' Factuality Depends on the Language of Inquirycited by 14
4 more
  1. 2026Ares: Adaptive Reasoning Effort Selection for Efficient LLM Agentscited by 8
  2. 2025A Comparative Analysis of LLM Adaptation: SFT, LoRA, and ICL in Data-Scarce Scenarioscited by 4
  3. 2026ExpLang: Improved Exploration and Exploitation in LLM Reasoning with On-Policy Thinking Language Selectioncited by 3
  4. 2026Think Multilingual, Not Harder: A Data-Efficient Framework for Teaching Reasoning Models to Code-Switchcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025AdaMCoT: Rethinking Cross-Lingual Factual Reasoning through Adaptive Multilingual Chain-of-Thought
  2. 2026LANG: Reinforcement Learning for Multilingual Reasoning with Language-Adaptive Hint Guidance
  3. 2026Cross-lingual Self-Consistency for Multilingual Reasoning with Language Models
  4. 2026Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning
  5. 2026What Makes Good Multilingual Reasoning? Disentangling Traces with Measurable Features
  6. 2026Rethinking the Multilingual Reasoning Gap with Layer Swap
  7. 2025MMATH: A Multilingual Benchmark for Mathematical Reasoning
  8. 2025EfficientXLang: Towards Improving Token Efficiency Through Cross-Lingual Reasoning