English
Related papers

Related papers: Token Homogenization under Positional Bias

200 papers

Large Language Models (LLMs) have demonstrated remarkable generalization across diverse tasks, leading individuals to increasingly use them as personal assistants and universal computing engines. Nevertheless, a notable obstacle emerges…

Machine Learning · Computer Science 2023-09-13 Dimitris Spathis , Fahim Kawsar

Tokenisation - "the process of splitting text into atomic parts" (Brezina & Timperley, 2017: 1) - is a crucial step for corpus linguistics, as it provides the basis for any applicable quantitative method (e.g. collocations) while ensuring…

Computation and Language · Computer Science 2025-07-03 Matteo Di Cristofaro

This paper investigates contextual word representation models from the lens of similarity analysis. Given a collection of trained models, we measure the similarity of their internal representations and attention. Critically, these models…

Computation and Language · Computer Science 2020-05-05 John M. Wu , Yonatan Belinkov , Hassan Sajjad , Nadir Durrani , Fahim Dalvi , James Glass

Vision Transformers have emerged as powerful, scalable and versatile representation learners. To capture both global and local features, a learnable [CLS] class token is typically prepended to the input sequence of patch tokens. Despite…

Computer Vision and Pattern Recognition · Computer Science 2026-02-10 Alexis Marouani , Oriane Siméoni , Hervé Jégou , Piotr Bojanowski , Huy V. Vo

Transformers have impressive generalization capabilities on tasks with a fixed context length. However, they fail to generalize to sequences of arbitrary length, even for seemingly simple tasks such as duplicating a string. Moreover, simply…

Augmenting large language models (LLMs) with auxiliary tokens has emerged as a promising strategy for enhancing model performance. In this work, we introduce a lightweight method termed latent tokens; these are dummy tokens that may be…

Machine Learning · Computer Science 2025-05-20 Yuchang Sun , Yanxi Chen , Yaliang Li , Bolin Ding

Vector quantization is a technique in machine learning that discretizes continuous representations into a set of discrete vectors. It is widely employed in tokenizing data representations for large language models, diffusion models, and…

Machine Learning · Computer Science 2024-11-26 Wenhao Zhao , Qiran Zou , Rushi Shah , Dianbo Liu

Tokenization is the first step in every language model (LM), yet it never takes the sounds of words into account. We investigate how tokenization influences text-only LMs' ability to represent phonological knowledge. Through a series of…

Computation and Language · Computer Science 2026-04-21 Disen Liao , Freda Shi

Representing token embeddings as probability distributions over learned manifolds allows for more flexible contextual inference, reducing representational rigidity while enhancing semantic granularity. Comparative evaluations demonstrate…

Computation and Language · Computer Science 2025-04-25 Christopher Nightingale , Dominic Lavington , Jonathan Thistlethwaite , Sebastian Penhaligon , Thomas Belinski , David Boldo

The Transformer architecture has become prominent in developing large causal language models. However, mechanisms to explain its capabilities are not well understood. Focused on the training process, here we establish a meta-learning view…

Machine Learning · Computer Science 2024-03-26 Xinbo Wu , Lav R. Varshney

Large language models encode the correlational structure present in natural language by fitting segments of utterances (tokens) into a high dimensional ambient latent space upon which the models then operate. We assert that in order to…

Differential Geometry · Mathematics 2024-10-14 Michael Robinson , Sourya Dey , Shauna Sweet

Good quality explanations strengthen the understanding of language models and data. Feature attribution methods, such as Integrated Gradient, are a type of post-hoc explainer that can provide token-level insights. However, explanations on…

Computation and Language · Computer Science 2026-04-21 Jonathan Kamp , Roos Bakker , Dominique Blok

Pretraining large language models (LLMs) with next-token prediction has led to remarkable advances, yet the context-dependent nature of token embeddings in such models results in high intra-class variance and inter-class similarity, thus…

Computation and Language · Computer Science 2026-05-12 Yan Sun , Guoxia Wang , Jinle Zeng , JiaBin Yang , Shuai Li , Li Shen , Dacheng Tao , DianHai Yu , Haifeng Wang

The choice of tokenizer can profoundly impact language model performance, yet accessible and reliable evaluations of tokenizer quality remain an open challenge. Inspired by scaling consistency, we show that smaller models can accurately…

Computation and Language · Computer Science 2025-06-04 Jonas F. Lotz , António V. Lopes , Stephan Peitz , Hendra Setiawan , Leonardo Emili

In this work, we provide a systematic and comprehensive empirical comparison of pretrained multilingual language models versus their monolingual counterparts with regard to their monolingual task performance. We study a set of nine…

Computation and Language · Computer Science 2021-06-03 Phillip Rust , Jonas Pfeiffer , Ivan Vulić , Sebastian Ruder , Iryna Gurevych

Representing continuous time is a critical and under-explored challenge in modeling temporal event sequences with large language models (LLMs). Various strategies like byte-level representations or calendar tokens have been proposed.…

Computation and Language · Computer Science 2026-05-12 Zefang Liu , Nam H. Nguyen , Yinzhu Quan , Shi-Xiong Zhang

Transformer-based pre-trained models with millions of parameters require large storage. Recent approaches tackle this shortcoming by training adapters, but these approaches still require a relatively large number of parameters. In this…

Computation and Language · Computer Science 2023-01-31 Chin-Lun Fu , Zih-Ching Chen , Yun-Ru Lee , Hung-yi Lee

Language models lack the notion of interchangeable tokens: symbols that are semantically equivalent yet distinct, such as bound variables in formal logic. This limitation prevents generalization to larger vocabularies and hinders the…

Computation and Language · Computer Science 2025-06-19 İlker Işık , Ramazan Gokberk Cinbis , Ebru Aydin Gol

In order to preserve word-order information in a non-autoregressive setting, transformer architectures tend to include positional knowledge, by (for instance) adding positional encodings to token embeddings. Several modifications have been…

Computation and Language · Computer Science 2021-09-14 Vinit Ravishankar , Anders Søgaard

It is a widely known issue that Transformers, when trained on shorter sequences, fail to generalize robustly to longer ones at test time. This raises the question of whether Transformer models are real reasoning engines, despite their…

Machine Learning · Computer Science 2025-04-04 Ruining Li , Gabrijel Boduljak , Jensen , Zhou