# Enhancing Translation Accuracy of Large Language Models through Continual Pre-Training on Parallel Data

arXiv: https://arxiv.org/abs/2407.03145 · HTML: https://mlsanity.com/paper/2407.03145/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 29 of its 53 references that are in the map.
- First posted on arXiv 3 Jul 2024.

## Builds on (most cited first, 29 of 29)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [No Language Left Behind: Scaling Human-Centered Machine Translation](https://mlsanity.com/paper/2207.04672/) (2022)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Multilingual Denoising Pre-training for Neural Machine Translation](https://mlsanity.com/paper/2001.08210/) (2020)
- [Training Deep Nets with Sublinear Memory Cost](https://mlsanity.com/paper/1604.06174/) (2016)
- [Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation](https://mlsanity.com/paper/2401.08417/) (2024)
- [How Good Are GPT Models at Machine Translation? A Comprehensive Evaluation](https://mlsanity.com/paper/2302.09210/) (2023)
- [Efficient and Effective Text Encoding for Chinese LLaMA and Alpaca](https://mlsanity.com/paper/2304.08177/) (2023)
- [Tower: An Open Multilingual Large Language Model for Translation-Related Tasks](https://mlsanity.com/paper/2402.17733/) (2024)
- [Fine-tuned Language Models are Continual Learners](https://mlsanity.com/paper/2205.12393/) (2022)
- [Lifelong Pretraining: Continually Adapting Language Models to Emerging Corpora](https://mlsanity.com/paper/2110.08534/) (2021)
- [Searching for Needles in a Haystack: On the Role of Incidental Bilingualism in PaLM's Translation Capability](https://mlsanity.com/paper/2305.10266/) (2023)
- [A Shocking Amount of the Web is Machine Translated: Insights from Multi-Way Parallelism](https://mlsanity.com/paper/2401.05749/) (2024)
- [nmT5 -- Is parallel data still relevant for pre-training massively multilingual language models?](https://mlsanity.com/paper/2106.02171/) (2021)
- [Findings of the First Shared Task on Machine Translation Robustness](https://mlsanity.com/paper/1906.11943/) (2019)
- [JParaCrawl v3.0: A Large-scale English-Japanese Parallel Corpus](https://mlsanity.com/paper/2202.12607/) (2022)
- [JESC: Japanese-English Subtitle Corpus](https://mlsanity.com/paper/1710.10639/) (2017)
- [Designing the Business Conversation Corpus](https://mlsanity.com/paper/2008.01940/) (2020)
- [LEALLA: Learning Lightweight Language-agnostic Sentence Embeddings with Knowledge Distillation](https://mlsanity.com/paper/2302.08387/) (2023)
- [Cross-Lingual Supervision improves Large Language Models Pre-training](https://mlsanity.com/paper/2305.11778/) (2023)

## Built on by (most cited first, 7 of 7)

- [X-ALMA: Plug & Play Modules and Adaptive Rejection for Quality Translation at Scale](https://mlsanity.com/paper/2410.03115/) (2024)
- [EMMA-500: Enhancing Massively Multilingual Adaptation of Large Language Models](https://mlsanity.com/paper/2409.17892/) (2024)
- [Trans-Zero: Self-Play Incentivizes Large Language Models for Multilingual Translation Without Parallel Data](https://mlsanity.com/paper/2504.14669/) (2025)
- [Massively Multilingual Adaptation of Large Language Models Using Bilingual Translation Data](https://mlsanity.com/paper/2506.00469/) (2025)
- [Bridging Linguistic Gaps: Cross-Lingual Mapping in Pre-Training and Dataset for Enhanced Multilingual LLM Performance](https://mlsanity.com/paper/2604.10590/) (2026)
- [Copy First, Translate Later: Interpreting Translation Dynamics in Multilingual Pretraining](https://mlsanity.com/paper/2604.17633/) (2026)
- [Do Large Language Models Grasp The Grammar? Evidence from Grammar-Book-Guided Probing in Luxembourgish](https://mlsanity.com/paper/2510.24856/) (2025)

## Abstract (arXiv)

In this paper, we propose a two-phase training approach where pre-trained large language models are continually pre-trained on parallel data and then supervised fine-tuned with a small amount of high-quality parallel data. To investigate the effectiveness of our proposed approach, we conducted continual pre-training with a 3.8B-parameter model and parallel data across eight different formats. We evaluate these methods on thirteen test sets for Japanese-to-English and English-to-Japanese translation. The results demonstrate that when utilizing parallel data in continual pre-training, it is essential to alternate between source and target sentences. Additionally, we demonstrated that the translation accuracy improves only for translation directions where the order of source and target sentences aligns between continual pre-training data and inference. In addition, we demonstrate that the LLM-based translation model is more robust in translating spoken language and achieves higher accuracy with less training data compared to supervised encoder-decoder models. We also show that the highest accuracy is achieved when the data for continual pre-training consists of interleaved source and target sentences and when tags are added to the source sentences.