English
Related papers

Related papers: How BPE Affects Memorization in Transformers

200 papers

Language models typically need to be trained or finetuned in order to acquire new knowledge, which involves updating their weights. We instead envision language models that can simply read and memorize new data at inference time, thus…

Machine Learning · Computer Science 2022-03-18 Yuhuai Wu , Markus N. Rabe , DeLesley Hutchins , Christian Szegedy

Large language models (LMs) have been shown to memorize parts of their training data, and when prompted appropriately, they will emit the memorized training data verbatim. This is undesirable because memorization violates privacy (exposing…

Machine Learning · Computer Science 2023-03-07 Nicholas Carlini , Daphne Ippolito , Matthew Jagielski , Katherine Lee , Florian Tramer , Chiyuan Zhang

Since hardware resources are limited, the objective of training deep learning models is typically to maximize accuracy subject to the time and memory constraints of training and inference. We study the impact of model size in this setting,…

Computation and Language · Computer Science 2020-06-24 Zhuohan Li , Eric Wallace , Sheng Shen , Kevin Lin , Kurt Keutzer , Dan Klein , Joseph E. Gonzalez

The effectiveness of Neural Machine Translation (NMT) models largely depends on the vocabulary used at training; small vocabularies can lead to out-of-vocabulary problems -- large ones, to memory issues. Subword (SW) tokenization has been…

Computation and Language · Computer Science 2023-03-02 J. Pourmostafa Roshan Sharami , D. Shterionov , P. Spronck

Recent research demonstrated that training large language models involves memorization of a significant fraction of training data. Such memorization can lead to privacy violations when training on sensitive user data and thus motivates the…

Machine Learning · Computer Science 2025-10-29 Vitaly Feldman , Guy Kornowski , Xin Lyu

Large protein language models are adept at capturing the underlying evolutionary information in primary structures, offering significant practical value for protein engineering. Compared to natural language models, protein amino acid…

Computation and Language · Computer Science 2023-10-27 Yang Tan , Mingchen Li , Pan Tan , Ziyi Zhou , Huiqun Yu , Guisheng Fan , Liang Hong

Traditional greedy tokenization methods have been a critical step in Natural Language Processing (NLP), influencing how text is converted into tokens and directly impacting model performance. While subword tokenizers like Byte-Pair Encoding…

Computation and Language · Computer Science 2025-05-05 Bharath Raj , Garvit Suri , Vikrant Dewangan , Raghav Sonavane

While NLP models significantly impact our lives, there are rising concerns about privacy invasion. Although federated learning enhances privacy, attackers may recover private training data by exploiting model parameters and gradients.…

Artificial Intelligence · Computer Science 2024-10-23 Mengjiao Zhang , Jia Xu

Pretraining large language models is a costly process. To make this process more efficient, several methods have been proposed to optimize model architecture/parametrization and hardware use. On the parametrization side, $\mu P$ (Maximal…

Machine Learning · Computer Science 2025-06-19 Soufiane Hayou , Liyuan Liu

Large Language Models are affected by the phenomena of memorizing and forgetting their training data. But how do these vary by model size? We work towards this question by investigating how the model size affects the model's ability to…

Computation and Language · Computer Science 2022-12-06 Ayrton San Joaquin , Ardy Haroen

This study investigates the impact of morphological typology on tokenization and language modeling performance. We focus on languages with synthetic and analytical morphological structures and examine their productivity when tokenized using…

Computation and Language · Computer Science 2024-11-01 Iñigo Parra

Language modeling is a fundamental task in natural language processing, which has been thoroughly explored with various architectures and hyperparameters. However, few studies focus on the effect of sub-word segmentation on the performance…

Computation and Language · Computer Science 2023-10-30 Jue Hou , Anisia Katinskaia , Anh-Duc Vu , Roman Yangarber

The assumption across nearly all language model (LM) tokenization schemes is that tokens should be subwords, i.e., contained within word boundaries. While providing a seemingly reasonable inductive bias, is this common practice limiting the…

Computation and Language · Computer Science 2025-08-28 Alisa Liu , Jonathan Hayase , Valentin Hofmann , Sewoong Oh , Noah A. Smith , Yejin Choi

Subword tokenization is an essential part of modern large language models (LLMs), yet its specific contributions to training efficiency and model performance remain poorly understood. In this work, we decouple the effects of subword…

Computation and Language · Computer Science 2026-05-15 Théo Gigant , Bowen Peng , Jeffrey Quesnelle

Large Transformer models have achieved impressive performance in many natural language tasks. In particular, Transformer based language models have been shown to have great capabilities in encoding factual knowledge in their vast amount of…

Computation and Language · Computer Science 2020-12-02 Chen Zhu , Ankit Singh Rawat , Manzil Zaheer , Srinadh Bhojanapalli , Daliang Li , Felix Yu , Sanjiv Kumar

Byte Pair Encoding (BPE) is a widely used tokenization algorithm, whose tokens cannot extend across pre-tokenization boundaries, functionally limiting it to representing at most full words. The BoundlessBPE and SuperBPE algorithms extend…

Computation and Language · Computer Science 2026-04-08 Craig W. Schmidt , Chris Tanner , Yuval Pinter

Byte-pair encoding (BPE) is a ubiquitous algorithm in the subword tokenization process of language models as it provides multiple benefits. However, this process is solely based on pre-training data statistics, making it hard for the…

Computation and Language · Computer Science 2021-09-27 Gustavo Aguilar , Bryan McCann , Tong Niu , Nazneen Rajani , Nitish Keskar , Thamar Solorio

This paper studies how the model architecture and data configurations influence the empirical memorization capacity of generative transformers. The models are trained using synthetic text datasets derived from the Systematized Nomenclature…

Computation and Language · Computer Science 2025-06-18 Anton Changalidis , Aki Härmä

In this work, we show a fundamental limitation in vocabulary adaptation approaches that use Byte-Pair Encoding (BPE) tokenization scheme for fine-tuning pretrained language models (PLMs) to expert domains. Current approaches trivially…

Computation and Language · Computer Science 2025-04-29 Gunjan Balde , Soumyadeep Roy , Mainack Mondal , Niloy Ganguly

Human memory is fleeting. As words are processed, the exact wordforms that make up incoming sentences are rapidly lost. Cognitive scientists have long believed that this limitation of memory may, paradoxically, help in learning language -…

Computation and Language · Computer Science 2026-05-11 Abishek Thamma , Micha Heilbron