paper-with-me

홈 › Papers

A path to natural language through tokenisation and transformers

2026-01-06 · David S. Berman, Alexander G. Stapleton arxiv

Natural languages exhibit striking regularities in their statistical structure, including notably the emergence of Zipf's and Heaps' laws. Despite this, it remains broadly unclear how these properties relate to the modern tokenisation schemes used in contemporary transformer models. In this note, we analyse the information content (as measured by the Shannon entropy) of various corpora under the assumption of a Zipfian frequency distribution, and derive a closed-form expression for the slot entropy expectation value. We then empirically investigate how byte--pair encoding (BPE) transforms corpus statistics, showing that recursive applications of BPE drive token frequencies toward a Zipfian power law while inducing a characteristic growth pattern in empirical entropy. Utilizing the ability of transformers to learn context dependent token probability distributions, we train language models on corpora tokenised at varying BPE depths, revealing that the model predictive entropies increasingly agree with Zipf-derived predictions as the BPE depth increases. Attention-based diagnostics further indicate that deeper tokenisation reduces local token dependencies, bringing the empirical distribution closer to the weakly dependent (near IID) regime. Together, these results clarify how BPE acts not only as a compression mechanism but also as a statistical transform that reconstructs key informational properties of natural language.

📄 PDF Abstract BibTeX arXiv:2601.03368

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Tokenisation over Bounded Alphabets is Hard

2025-11-19 · Violeta Kastreva, Philip Whittington, Dennis Komm, Tiago Pimentel arxiv

Recent works have shown that tokenisation is NP-complete. However, these works assume tokenisation is applied to inputs with unboundedly large alphabets -- an unrealistic assumption, given that in practice tokenisers ope…

Improving Tokenisation by Alternative Treatment of Spaces

2022-04-08 · Edward Gow-Smith, Harish Tayyar Madabushi, Carolina Scarton, Aline Villavicencio

Tokenisation is the first step in almost all NLP tasks, and state-of-the-art transformer-based language models all use subword tokenisation algorithms to process input text. Existing algorithms have problems, often produ…

Natural Language Understanding

Improving Tokenisation by Alternative Treatment of Spaces

2021-12-17 · ACL ARR December 2022 12 · Anonymous

Tokenisation is the first step in almost all NLP tasks, and state-of-the-art transformer-based language models all use subword tokenisation algorithms to process input text. Existing algorithms have problems, often produ…

Natural Language Understanding

Innovative tokenisation of structured data for LLM training

2025-08-03 · Kayvan Karim, Hani Ragab Hassen. Hadj Batatia arxiv

Data representation remains a fundamental challenge in machine learning, particularly when adapting sequence-based architectures like Transformers and Large Language Models (LLMs) for structured tabular data. Existing me…

Network Intrusion Detection

Multidimensional Byte Pair Encoding: Shortened Sequences for Improved Visual Data Generation

2024-11-15 · Tim Elsner, Paula Usinger, Julius Nehring-Wirxel, Gregor Kobsik 외

In language processing, transformers benefit greatly from text being condensed. This is achieved through a larger vocabulary that captures word fragments instead of plain characters. This is often done with Byte Pair Enc…