In recent years, Large Language Models (LLMs) have demonstrated exceptional proficiency across a broad spectrum of Natural Language Processing (NLP) tasks, including Machine Translation. However, previous methods predominantly relied on iterative processes such as instruction fine-tuning or continual pre-training, leaving unexplored the challenges of training LLMs solely on parallel data. In this work, we introduce Plume (Parallel Language Model), a collection of three 2B LLMs featuring varying vocabulary sizes (32k, 128k, and 256k) trained exclusively on Catalan-centric parallel examples. These models perform comparably to previous encoder-decoder architectures on 16 supervised translation directions and 56 zero-shot ones. Utilizing this set of models, we conduct a thorough investigation into the translation capabilities of LLMs, probing their performance, the role of vocabulary size, the impact of the different elements of the prompt, and their cross-lingual representation space. We find that larger vocabulary sizes improve zero-shot performance and that different layers specialize in distinct aspects of the prompt, such as language-specific tags. We further show that as the vocabulary size grows, a larger number of attention heads can be pruned with minimal loss in translation quality, achieving a reduction of over 64.7% in attention heads.

High-quality machine translation requires datasets that not only ensure linguistic accuracy but also capture regional and cultural nuances. While many existing benchmarks, such as FLORES-200, rely on English as a pivot language, this approach can overlook the specificity of direct language pairs, particularly for underrepresented combinations like Catalan-Chinese. In this study, we demonstrate that even with a relatively small dataset of approximately 1,000 sentences, we can significantly improve MT localization. To this end, we introduce a dataset specifically designed to enhance Catalan-to-Chinese translation by prioritizing regionally and culturally specific topics. Unlike pivot-based datasets, our data source ensures a more faithful representation of Catalan linguistic and cultural elements, leading to more accurate translations of local terms and expressions. Using this dataset, we demonstrate better performance over the English-pivot FLORES-200 dev set and achieve competitive results on the FLORES-200 devtest set when evaluated with neural-based metrics. We release this dataset as both a human-preference resource and a benchmark for Catalan-Chinese translation. Additionally, we include Spanish translations for each sentence, facilitating extensions to Spanish-Chinese translation tasks.

We introduce MT-Lens, a framework designed to evaluate Machine Translation (MT) systems across a variety of tasks, including translation quality, gender bias detection, added toxicity, and robustness to misspellings. While several toolkits have become very popular for benchmarking the capabilities of Large Language Models (LLMs), existing evaluation tools often lack the ability to thoroughly assess the diverse aspects of MT performance. MT-Lens addresses these limitations by extending the capabilities of LM-eval-harness for MT, supporting state-of-the-art datasets and a wide range of evaluation metrics. It also offers a user-friendly platform to compare systems and analyze translations with interactive visualizations. MT-Lens aims to broaden access to evaluation strategies that go beyond traditional translation quality evaluation, enabling researchers and engineers to better understand the performance of a NMT model and also easily measure system’s biases.

In this paper, we present the SalamandraTA family of models, an improved iteration of Salamandra LLMs (Gonzalez-Agirre et al., 2025) specifically trained to achieve strong performance in translation-related tasks for 38 European languages. SalamandraTA comes in two scales: 2B and 7B parameters. For both versions, we applied the same training recipe with a first step of continual pre-training on parallel data, and a second step of supervised fine-tuning on high-quality instructions.The BSC submission to the WMT25 General Machine Translation shared task is based on the 7B variant of SalamandraTA. We first extended the model vocabulary to support the additional non-European languages included in the task. This was followed by a second phase of continual pretraining and supervised fine-tuning, carefully designed to optimize performance across all translation directions for this year’s shared task. For decoding, we employed two quality-aware strategies: Minimum Bayes Risk Decoding and Translation Reranking using Comet and Comet-kiwi.We publicly release both the 2B and 7B versions of SalamandraTA, along with the newer SalamandraTA-v2 model, on Hugging Face.

This paper describes the BSC’s submission to the AmericasNLP 2024 Shared Task. We participated in the Spanish to Quechua and Spanish to Guarani tasks. In this paper we show that by using LoRA adapters we can achieve similar performance as a full parameter fine-tuning by only training 14.2% of the total number of parameters. Our systems achieved the highest ChrF++ scores and ranked first for both directions in the final results outperforming strong baseline systems in the provided development and test datasets.

This paper studies gender bias in machine translation through the lens of Large Language Models (LLMs). Four widely-used test sets are employed to benchmark various base LLMs, comparing their translation quality and gender bias against state-of-the-art Neural Machine Translation (NMT) models for English to Catalan (En → Ca) and English to Spanish (En → Es) translation directions. Our findings reveal pervasive gender bias across all models, with base LLMs exhibiting a higher degree of bias compared to NMT models. To combat this bias, we explore prompting engineering techniques applied to an instruction-tuned LLM. We identify a prompt structure that significantly reduces gender bias by up to 12% on the WinoMT evaluation dataset compared to more straightforward prompts. These results significantly reduce the gender bias accuracy gap between LLMs and traditional NMT systems.

This paper presents a comprehensive evaluation of gender bias in English-Catalan machine translation, encompassing the creation of a novel language resource and an analysis of translation quality across four different tokenization models. The study introduces a new dataset derived from the MuST-SHE corpus, focusing on gender-neutral terms that necessitate gendered translations in Catalan. The results reveal noteworthy gender bias across all translation models, with a consistent preference for masculine forms. Notably, the study finds that when context is available, BPE and Sentencepiece Unigram tokenization methods outperform others, achieving higher accuracy in gender translation. However, when no context is provided, Morfessor outputs more feminine forms than other tokenization methods, albeit still a small percentage. The study also reflects that stereotypes present in the data are amplified in the translation output. Ultimately, this work serves as a valuable resource for addressing and mitigating gender bias in machine translation, emphasizing the need for improved awareness and sensitivity to gender issues in natural language processing applications.

In this paper, we present the two strategies employed for the WMT24 Shared Task on Translation into Low-Resource Languages of Spain. We participated in the language pairs of Spanish-to-Aragonese, Spanish-to-Aranese, and Spanish-to-Asturian, developing neural-based translation systems and moving away from rule-based approaches for these language directions. To create these models, two distinct strategies were employed. The first strategy involved a thorough cleaning process and curation of the limited provided data, followed by fine-tuning the multilingual NLLB-200-600M model (Constrained Submission). The other strategy involved training a transformer from scratch using a vast amount of synthetic data (Open Submission). Both approaches relied on generated synthetic data and resulted in high ChrF and BLEU scores. However, given the characteristics of the task, the strategy used in the Constrained Submission resulted in higher scores that surpassed the baselines across the three translation directions, whereas the strategy employed in the Open Submission yielded slightly lower scores than the highest baseline.