Native Multilingual Chain-of-Thought Reasoning in Low-Resource Southeast Asian Languages
On the map: LLM reasoning → Reasoning benchmarks and evaluation
Abstract (arXiv)
Large Language Models have achieved substantial progress in reasoning capabilities. Yet in low-resource native settings, many suffer from cross-lingual collapse, reverting to English during intermediate steps that require complex logical reasoning. This presents a cold-start bottleneck for policy optimization, whereas standard fine-tuning risks catastrophic forgetting due to cross-lingual representation drift. To address these challenges, we introduce the Onramp-Sequence Cross-Distillation (OSCD), a post-training algorithm that projects high-resource reasoning trajectories into low-resource vocabulary subspaces during generative training rollouts via an integrated translator agentic loop, ensuring the stable and efficient translation of dynamically generated reference samples for fine-tuning. This is coupled with joint-embedding semantic alignment of both reference and target-language reasoning traces, thereby bridging the pairwise cross-lingual representational gaps. Comprehensive evaluations using the AIME25 and HMMT25 benchmarks demonstrate that OSCD yields up to 3.2 times overall improvements in native Southeast Asian languages for mathematical reasoning, of which the joint-embedding semantic alignment component contributes up to 6.4% improvements in linguistic debiasing over translation-only baselines.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning. Builds on 12 of its 35 references that are in the map. First posted on arXiv 1 Aug 2026.
Builds on
12 of 35 references.
Closest ancestors · 3
Same direction, newest first
- 2025Why Do Multilingual Reasoning Gaps Emerge in Reasoning Language Models?31 Oct
- 2025Learn Globally, Speak Locally: Bridging the Gaps in Multilingual Reasoning7 Jul
- 2025Cross-lingual Collapse: How Language-Centric Foundation Models Shape Reasoning in Large Language Models6 Jun
Other references · 9
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
- 2026LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixelscited by 237
- 2026Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMscited by 115
- 2024Learning and Leveraging World Models in Visual Representation Learningcited by 71
- 2025Language-Specific Latent Process Hinders Cross-Lingual Performancecited by 20
- 2025Long Chain-of-Thought Reasoning Across Languagescited by 18
- 2025JEPA-Reasoner: Decoupling Latent Reasoning from Token Generationcited by 4
- 2025Sparse Subnetwork Enhancement for Underrepresented Languages in Large Language Modelscited by 1
- 2026Semantic Step Prediction: Multi-Step Latent Forecasting in LLM Reasoning Trajectories via Step Samplingcited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025Cross-Lingual Consistency: A Novel Inference Framework for Advancing Reasoning in Large Language Models
- 2025AdaMCoT: Rethinking Cross-Lingual Factual Reasoning through Adaptive Multilingual Chain-of-Thought
- 2024LinguaLIFT: An Effective Two-stage Instruction Tuning Framework for Low-Resource Language Reasoning
- 2026Less Languages, Less Tokens: An Efficient Unified Logic Cross-lingual Chain-of-Thought Reasoning Framework
- 2024xCoT: Cross-lingual Instruction Tuning for Cross-lingual Chain-of-Thought Reasoning
- 2026Crosslingual On-Policy Self-Distillation for Multilingual Reasoning
- 2026Rethinking the Multilingual Reasoning Gap with Layer Swap
- 2025Reasoning Transfer for an Extremely Low-Resource and Endangered Language: Bridging Languages Through Sample-Efficient Language Understanding