# CantoneseLLM v2: Reasoning in a Low-Resource Language

arXiv: https://arxiv.org/abs/2609.06970 · HTML: https://mlsanity.com/paper/2609.06970/

## 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 19 of its 46 references that are in the map.
- First posted on arXiv 7 Sep 2026.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 19 of 19)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [Scaling Language Models: Methods, Analysis & Insights from Training Gopher](https://mlsanity.com/paper/2112.11446/) (2021)
- [Gemma 4 Technical Report](https://mlsanity.com/paper/2607.02770/) (2026)
- [Kimi K3: Open Frontier Intelligence](https://mlsanity.com/paper/2607.24653/) (2026)
- [Continual Pre-Training for Cross-Lingual LLM Adaptation: Enhancing Japanese Language Capabilities](https://mlsanity.com/paper/2404.17790/) (2024)
- [Taiwan LLM: Bridging the Linguistic Divide with a Culturally Aligned Language Model](https://mlsanity.com/paper/2311.17487/) (2023)
- [Nemotron-Math: Efficient Long-Context Distillation of Mathematical Reasoning from Multi-Mode Supervision](https://mlsanity.com/paper/2512.15489/) (2025)
- [Efficient and Effective Vocabulary Expansion Towards Multilingual Large Language Models](https://mlsanity.com/paper/2402.14714/) (2024)
- [VinaLLaMA: LLaMA-based Vietnamese Foundation Model](https://mlsanity.com/paper/2312.11011/) (2023)
- [EuroLLM-22B: Technical Report](https://mlsanity.com/paper/2602.05879/) (2026)
- [Typhoon-S: Minimal Open Post-Training for Sovereign Large Language Models](https://mlsanity.com/paper/2601.18129/) (2026)
- [CARTE: A Benchmark for Mapping Language Model Knowledge Across France](https://mlsanity.com/paper/2606.01995/) (2026)
- [A Sovereign, Open-Source Foundation Model for German and English](https://mlsanity.com/paper/2607.09424/) (2026)

## Abstract (arXiv)

Cantonese is widely spoken but remains low-resource in written data, with no large corpus of native Cantonese reasoning traces available for model training. We develop and release CantoneseLLM v2, comprising models based on Qwen3 8B and 30B-A3B. The models are trained through CPT on 784 million Cantonese and Hong Kong-related tokens, chat-vector merging, SFT, DPO, and RLVR. Evaluation across the training stages shows that chat-vector merging transfers instruction following but preserves the donor model's reasoning language, while SFT with limited Cantonese reasoning data substantially shortens or removes reasoning traces and reduces benchmark performance. DPO restores the reasoning-block format, particularly for the 8B model, but recovers only part of the lost performance. The RLVR training with Cantonese language and Traditional Chinese scripts as multiplicative constraints introduced Cantonese language alignment and restored the lost performance. The 30B-A3B model reaches 73.16 on HKCanto-Eval, within 1.20 points of its merged checkpoint, while retaining the Cantonese reasoning behaviour absent from that checkpoint. We release the model checkpoints, the training environments, and a thirteen-year Traditional Chinese Common Crawl dataset. The models can be accessed at https://huggingface.co/collections/hon9kon9ize/cantonesellm-v20