Teaching Old Tokenizers New Words: Efficient Tokenizer Adaptation for Pre-trained Models

• URL: http://arxiv.org/abs/2512.03989v1
• Date: Wed, 03 Dec 2025 17:20:16 GMT
• Title: Teaching Old Tokenizers New Words: Efficient Tokenizer Adaptation for Pre-trained Models
• Authors: Taido Purason, Pavel Chizhov, Ivan P. Yamshchikov, Mark Fishel,
• Abstract summary: Tokenizer adaptation plays an important role in transferring pre-trained language models to new domains or languages.<n>The common approach to extension trains a new tokenizer on domain-specific text and appends the tokens that do not overlap with the existing vocabulary.<n>We propose continued BPE training, which adapts a pre-trained tokenizer by continuing the BPE merge learning process on new data.
• Score: 12.218365713546214
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Tokenizer adaptation plays an important role in transferring pre-trained language models to new domains or languages. In this work, we address two complementary aspects of this process: vocabulary extension and pruning. The common approach to extension trains a new tokenizer on domain-specific text and appends the tokens that do not overlap with the existing vocabulary, which often results in many tokens that are unreachable or never used. We propose continued BPE training, which adapts a pre-trained tokenizer by continuing the BPE merge learning process on new data. Experiments across multiple languages and model families show that this approach improves tokenization efficiency and leads to better utilization of added vocabulary. We also introduce leaf-based vocabulary pruning, which removes redundant tokens while preserving model quality. Together, these methods provide practical tools for controlled vocabulary modification, which we release as an open-source package.

Related papers

• AdaptBPE: From General Purpose to Specialized Tokenizers [18.70903226766322]
  We propose a post-training adaptation strategy that selectively replaces low-utility tokens with more relevant ones based on their frequency in an adaptation corpus.<n>Our algorithm identifies the token inventory that most effectively encodes the adaptation corpus for a given target vocabulary size.<n>This method serves as a lightweight adaptation mechanism, akin to a vocabulary fine-tuning process, enabling optimized tokenization for specific domains or tasks.
  arXiv  Detail & Related papers  (2026-01-29T12:59:40Z)
• HYPEROFA: Expanding LLM Vocabulary to New Languages via Hypernetwork-Based Embedding Initialization [50.27950279695363]
  Many pre-trained language models (PLMs) exhibit suboptimal performance on mid- and low-resource languages.<n>A common strategy to address this is to introduce new tokens specific to the target languages, initialize their embeddings, and apply continual pre-training on target-language data.<n>We propose HYPEROFA, a hypernetwork-based approach for more adaptive token embedding.
  arXiv  Detail & Related papers  (2025-04-21T19:40:32Z)
• Scaling LLM Pre-training with Vocabulary Curriculum [0.0]
  We introduce vocabulary curriculum learning, an approach that improves pretraining efficiency with log-linear scaling gains relative to vocabulary size.<n>Our method alternates between entropy-guided vocabulary expansion and model optimization, enabling models to learn transferable representations across diverse tokenization granularities.<n>Experiments on small-scale GPT models demonstrate improved scaling efficiency, reinforcing the effectiveness of dynamic tokenization.
  arXiv  Detail & Related papers  (2025-02-25T07:18:29Z)
• Batching BPE Tokenization Merges [55.2480439325792]
  BatchBPE is an open-source pure Python implementation of the Byte Pair algorithm. It is used to train a high quality tokenizer on a basic laptop.
  arXiv  Detail & Related papers  (2024-08-05T09:37:21Z)
• Continuously Learning New Words in Automatic Speech Recognition [56.972851337263755]
  We propose a self-supervised continual learning approach for Automatic Speech Recognition.<n>We use a memory-enhanced ASR model from the literature to decode new words from the slides.<n>We show that with this approach, we obtain increasing performance on the new words when they occur more frequently.
  arXiv  Detail & Related papers  (2024-01-09T10:39:17Z)
• Embedding structure matters: Comparing methods to adapt multilingual vocabularies to new languages [20.17308477850864]
  Pre-trained multilingual language models underpin a large portion of modern NLP tools outside of English. We propose several simple techniques to replace a cross-lingual vocabulary with a compact, language-specific one.
  arXiv  Detail & Related papers  (2023-09-09T04:27:18Z)
• Evolving Dictionary Representation for Few-shot Class-incremental Learning [34.887690018011675]
  We tackle a challenging and practical continual learning scenario named few-shot class-incremental learning (FSCIL) InFSCIL, labeled data are given for classes in a base session but very limited labeled instances are available for new incremental classes. We propose deep dictionary learning which is a hybrid learning architecture that combines dictionary learning and visual representation learning.
  arXiv  Detail & Related papers  (2023-05-03T04:30:34Z)
• Semantic Tokenizer for Enhanced Natural Language Processing [32.605667552915854]
  We present a novel tokenizer that uses semantics to drive vocabulary construction. The tokenizer more than doubles the number of wordforms represented in the vocabulary.
  arXiv  Detail & Related papers  (2023-04-24T19:33:41Z)
• VECO 2.0: Cross-lingual Language Model Pre-training with Multi-granularity Contrastive Learning [56.47303426167584]
  We propose a cross-lingual pre-trained model VECO2.0 based on contrastive learning with multi-granularity alignments. Specifically, the sequence-to-sequence alignment is induced to maximize the similarity of the parallel pairs and minimize the non-parallel pairs. token-to-token alignment is integrated to bridge the gap between synonymous tokens excavated via the thesaurus dictionary from the other unpaired tokens in a bilingual instance.
  arXiv  Detail & Related papers  (2023-04-17T12:23:41Z)
• Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval [109.62363167257664]
  We propose a generative model for learning multilingual text embeddings. Our model operates on parallel data in $N$ languages. We evaluate this method on a suite of tasks including semantic similarity, bitext mining, and cross-lingual question retrieval.
  arXiv  Detail & Related papers  (2022-12-21T02:41:40Z)
• Continual Learning in Multilingual NMT via Language-Specific Embeddings [92.91823064720232]
  It consists in replacing the shared vocabulary with a small language-specific vocabulary and fine-tuning the new embeddings on the new language's parallel data. Because the parameters of the original model are not modified, its performance on the initial languages does not degrade.
  arXiv  Detail & Related papers  (2021-10-20T10:38:57Z)

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.