# Asymmetric Conflict and Synergy in Post-training for LLM-based Multilingual Machine Translation

arXiv: https://arxiv.org/abs/2502.11223 · HTML: https://mlsanity.com/paper/2502.11223/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Feb 2025.

## Built on by (most cited first, 6 of 6)

- [UORA: Uniform Orthogonal Reinitialization Adaptation in Parameter-Efficient Fine-Tuning of Large Models](https://mlsanity.com/paper/2505.20154/) (2025)
- [Beyond Decoder-only: Large Language Models Can be Good Encoders for Machine Translation](https://mlsanity.com/paper/2503.06594/) (2025)
- [Bridging the Linguistic Divide: A Survey on Leveraging Large Language Models for Machine Translation](https://mlsanity.com/paper/2504.01919/) (2025)
- [Breach in the Shield: Unveiling the Vulnerabilities of Large Language Models](https://mlsanity.com/paper/2504.03714/) (2025)
- [NiuTrans.LMT: Toward Inclusive and Scalable Multilingual Machine Translation with LLMs](https://mlsanity.com/paper/2511.07003/) (2025)
- [Retrieval, Reward, and Training Protocols: What Matters in Training Search Agents?](https://mlsanity.com/paper/2605.27881/) (2026)

## Abstract (arXiv)

The emergence of Large Language Models (LLMs) has advanced the multilingual machine translation (MMT), yet the Curse of Multilinguality (CoM) remains a major challenge. Existing work in LLM-based MMT typically mitigates this issue via scaling up training and computation budget, which raises a critical question: Is scaling up the training and computation budget truly necessary for high-quality MMT, or can a deeper understanding of CoM provide a more efficient solution? To explore this problem, we analyze the linguistic conflicts and synergy, the underlying mechanism of CoM during post-training phase. We identify an asymmetric phenomenon in linguistic conflicts and synergy: the dominance of conflicts and synergy varies in different translation directions, leading to sub-optimal adaptation in existing post-training methods. We further find that a significant bottleneck in MMT appears to lie in post-training rather than multilingual pre-training, suggesting the need for more effective adaptation strategies. Building on these new insights, we propose a direction-aware training approach, combined with group-wise model merging, to address asymmetry in linguistic conflicts and synergy explicitly. Leveraging this strategy, our method fine-tunes X-ALMA-13B-Pretrain-trained only with multilingual pre-training-achieving comparable performance to XALMA-13B (only SFT) while using only 20B pretraining tokens and 17B parameters-5.5x fewer pretraining-tokens and 1.7x fewer model size-with just 0.85 COMET drop on Flores-200 testsets of 50 languages.