# Efficient Training of Robust Traditional Chinese LLaMA-1B on a Single Consumer GPU: Continual Pre-training, SFT, and DPO

arXiv: https://arxiv.org/abs/2510.01616 · HTML: https://mlsanity.com/paper/2510.01616/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in LLM preference optimization methods, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 8 of its 20 references that are in the map.
- First posted on arXiv 2 Oct 2025.

## Placed in

- LLM alignment / RLHF / preference optimisation / [LLM preference optimization methods](https://mlsanity.com/alignment/llm-preference-optimization-methods/)

## Builds on (most cited first, 8 of 8)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Multilingual Large Language Models Are Not (Yet) Code-Switchers](https://mlsanity.com/paper/2305.14235/) (2023)
- [Taiwan LLM: Bridging the Linguistic Divide with a Culturally Aligned Language Model](https://mlsanity.com/paper/2311.17487/) (2023)
- [Code-Switching Curriculum Learning for Multilingual Transfer in LLMs](https://mlsanity.com/paper/2411.02460/) (2024)
- [Investigating and Scaling up Code-Switching for Multilingual Language Model Pre-Training](https://mlsanity.com/paper/2504.01801/) (2025)
- [Lost in the Mix: Evaluating LLM Understanding of Code-Switched Text](https://mlsanity.com/paper/2506.14012/) (2025)
- [Breaking Language Barriers: Equitable Performance in Multilingual Language Models](https://mlsanity.com/paper/2508.12662/) (2025)

## Abstract (arXiv)

Small Language Models (SLMs) enable cost-effective, on-device and latency-sensitive AI applications, yet their deployment in Traditional Chinese (TC) remains hindered by token-level instability - models unpredictably emit non-TC characters or code-switch into other languages. We address this practical reliability gap by creating PureTC-1B, a three-stage stabilization pipeline for Llama-3.2-1B-Instruct (an open-weight, instruction-tuned model released by Meta) using parameter-efficient LoRA adapters. Our method combines Continual Pre-Training (CPT) on TC-centric corpora, Supervised Fine-Tuning (SFT) with instruction data, and Direct Preference Optimization (DPO) using TC-adherence preferences to improve monolingual robustness without full-model retraining. On a benchmark designed to simulate real-world usage, PureTC-1B achieves a 51.3% relative reduction (micro-average) in non-TC output tokens versus the base model. On a Named Entity Translation (NET) task, PureTC-1B further reduces incorrect-language tokens by 77.2% relative to Llama-3B and 57.2% relative to Qwen-1.5B, indicating that robust TC adherence is attainable even at the 1B scale. The pipeline is reproducible, adapter-only, and hardware-friendly, offering practitioners a practical recipe to enhance language stability for TC and potentially other non-English languages.