Optimizing LLMs for Italian: Reducing Token Fertility and Enhancing Efficiency Through Vocabulary Adaptation

• URL: http://arxiv.org/abs/2504.17025v1
• Date: Wed, 23 Apr 2025 18:12:27 GMT
• Title: Optimizing LLMs for Italian: Reducing Token Fertility and Enhancing Efficiency Through Vocabulary Adaptation
• Authors: Luca Moroni, Giovanni Puccetti, Pere-Lluis Huguet Cabot, Andrei Stefan Bejgu, Edoardo Barba, Alessio Miaschi, Felice Dell'Orletta, Andrea Esuli, Roberto Navigli,
• Abstract summary: State-of-the-art Large Language Models (LLMs) can handle other languages, due to language contamination or some degree of multilingual pretraining data, but not optimized for non-English languages.<n>In this work, we thoroughly compare a variety of vocabulary adaptation techniques for optimizing English LLMs for the Italian language.<n>We adapt two LLMs: Mistral-7b-v0.1, reducing token fertility by 25%, and Llama-3.1-8B, optimizing the vocabulary and reducing the number of parameters by 1 billion.
• Score: 39.60572668223083
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The number of pretrained Large Language Models (LLMs) is increasing steadily, though the majority are designed predominantly for the English language. While state-of-the-art LLMs can handle other languages, due to language contamination or some degree of multilingual pretraining data, they are not optimized for non-English languages, leading to inefficient encoding (high token "fertility") and slower inference speed. In this work, we thoroughly compare a variety of vocabulary adaptation techniques for optimizing English LLMs for the Italian language, and put forward Semantic Alignment Vocabulary Adaptation (SAVA), a novel method that leverages neural mapping for vocabulary substitution. SAVA achieves competitive performance across multiple downstream tasks, enhancing grounded alignment strategies. We adapt two LLMs: Mistral-7b-v0.1, reducing token fertility by 25\%, and Llama-3.1-8B, optimizing the vocabulary and reducing the number of parameters by 1 billion. We show that, following the adaptation of the vocabulary, these models can recover their performance with a relatively limited stage of continual training on the target language. Finally, we test the capabilities of the adapted models on various multi-choice and generative tasks.

Related papers

• Franken-Adapter: Cross-Lingual Adaptation of LLMs by Embedding Surgery [31.516243610548635]
  We present $textitFranken-Adapter$, a modular language adaptation approach for decoder-only Large Language Models.<n>Our method begins by creating customized vocabularies for target languages and performing language adaptation through embedding tuning on multilingual data.<n>Experiments on $ttGemma2$ models with up to 27B parameters demonstrate improvements of up to 20% across 96 languages, spanning both discriminative and generative tasks.
  arXiv  Detail & Related papers  (2025-02-12T00:38:11Z)
• Adapters for Altering LLM Vocabularies: What Languages Benefit the Most? [23.83290627671739]
  VocADT is a novel method for vocabulary adaptation using adapter modules that are trained to learn the optimal linear combination of existing embeddings. We find that Latin-script languages and highly fragmented languages benefit the most from vocabulary adaptation.
  arXiv  Detail & Related papers  (2024-10-12T20:45:24Z)
• Lens: Rethinking Multilingual Enhancement for Large Language Models [70.85065197789639]
  Lens is a novel approach to enhance multilingual capabilities of large language models (LLMs) It operates by manipulating the hidden representations within the language-agnostic and language-specific subspaces from top layers of LLMs. It achieves superior results with much fewer computational resources compared to existing post-training approaches.
  arXiv  Detail & Related papers  (2024-10-06T08:51:30Z)
• Exploring Design Choices for Building Language-Specific LLMs [36.32622880071991]
  We study building language-specific language models by adapting monolingual and multilingual models. We find that the initial performance of LLM does not always correlate with the final performance after the adaptation.
  arXiv  Detail & Related papers  (2024-06-20T18:47:43Z)
• Bridging the Gap: Dynamic Learning Strategies for Improving Multilingual Performance in LLMs [15.911445732909849]
  Large language models (LLMs) are at the forefront of transforming numerous domains globally. However, their inclusivity and effectiveness remain limited for non-Latin scripts and low-resource languages. This paper tackles the imperative challenge of enhancing the multilingual performance of LLMs without extensive training or fine-tuning.
  arXiv  Detail & Related papers  (2024-05-28T16:56:42Z)
• Getting More from Less: Large Language Models are Good Spontaneous Multilingual Learners [67.85635044939836]
  Large Language Models (LLMs) have shown impressive language capabilities. In this work, we investigate the spontaneous multilingual alignment improvement of LLMs. We find that LLMs instruction-tuned on the question translation data (i.e. without annotated answers) are able to encourage the alignment between English and a wide range of languages.
  arXiv  Detail & Related papers  (2024-05-22T16:46:19Z)
• An Empirical Study on Cross-lingual Vocabulary Adaptation for Efficient Language Model Inference [38.1823640848362]
  State-of-the-art generative large language models (LLMs) disproportionately relies on English-centric tokenizers, vocabulary and pre-training data. Recent studies have shown that their inference efficiency deteriorates when generating text in languages other than English.
  arXiv  Detail & Related papers  (2024-02-16T14:15:15Z)
• The Ups and Downs of Large Language Model Inference with Vocabulary Trimming by Language Heuristics [74.99898531299148]
  This research examines vocabulary trimming (VT) inspired by restricting embedding entries to the language of interest to bolster time and memory efficiency. We apply two languages to trim the full vocabulary - Unicode-based script filtering and corpus-based selection - to different language families and sizes. It is found that VT reduces the memory usage of small models by nearly 50% and has an upper bound of 25% improvement in generation speed.
  arXiv  Detail & Related papers  (2023-11-16T09:35:50Z)
• UNKs Everywhere: Adapting Multilingual Language Models to New Scripts [103.79021395138423]
  Massively multilingual language models such as multilingual BERT (mBERT) and XLM-R offer state-of-the-art cross-lingual transfer performance on a range of NLP tasks. Due to their limited capacity and large differences in pretraining data, there is a profound performance gap between resource-rich and resource-poor target languages. We propose novel data-efficient methods that enable quick and effective adaptation of pretrained multilingual models to such low-resource languages and unseen scripts.
  arXiv  Detail & Related papers  (2020-12-31T11:37:28Z)
• Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models [63.92643612630657]
  This paper attempts to peek into the black-box of multilingual optimization through the lens of loss function geometry. We find that gradient similarity measured along the optimization trajectory is an important signal, which correlates well with language proximity. We derive a simple and scalable optimization procedure, named Gradient Vaccine, which encourages more geometrically aligned parameter updates for close tasks.
  arXiv  Detail & Related papers  (2020-10-12T17:26:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.