paper-with-me

Papers

Rethinking Tokenization: Crafting Better Tokenizers for Large Language Models

2024-03-01 · Jinbiao Yang

Tokenization significantly influences language models(LMs)' performance. This paper traces the evolution of tokenizers from word-level to subword-level, analyzing how they balance tokens and types to enhance model adaptability while controlling complexity. Despite subword tokenizers like Byte Pair Encoding (BPE) overcoming many word tokenizer limitations, they encounter difficulties in handling non-Latin languages and depend heavily on extensive training data and computational resources to grasp the nuances of multiword expressions (MWEs). This article argues that tokenizers, more than mere technical tools, should drawing inspiration from the cognitive science about human language processing. This study then introduces the "Principle of Least Effort" from cognitive science, that humans naturally seek to reduce cognitive effort, and discusses the benefits of this principle for tokenizer development. Based on this principle, the paper proposes that the Less-is-Better (LiB) model could be a new approach for LLM tokenizer. The LiB model can autonomously learn an integrated vocabulary consisting of subwords, words, and MWEs, which effectively reduces both the numbers of tokens and types. Comparative evaluations show that the LiB tokenizer outperforms existing word and BPE tokenizers, presenting an innovative method for tokenizer development, and hinting at the possibility of future cognitive science-based tokenizers being more efficient.

📄 PDF Abstract BibTeX arXiv:2403.00417

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…

Similar Papers 제목 키워드 기반

Parallel Tokenizers: Rethinking Vocabulary Design for Cross-Lingual Transfer

2025-10-07 · Muhammad Dehan Al Kautsar, Fajri Koto arxiv

Tokenization defines the foundation of multilingual language models by determining how words are represented and shared across languages. However, existing methods often fail to support effective cross-lingual transfer b…

Representation LearningEmotion ClassificationCross-Lingual TransferHate Speech Detection

Rethinking Tokenization for Rich Morphology: The Dominance of Unigram over BPE and Morphological Alignment

2025-08-11 · Saketh Reddy Vemula, Sandipan Dandapat, Dipti Misra Sharma, Parameswari Krishnamurthy arxiv

The relationship between tokenizer algorithm (e.g., Byte-Pair Encoding (BPE), Unigram), morphological alignment, tokenization quality (e.g., compression efficiency), and downstream performance remains largely unclear, pa…

Text Classification

Unpacking Tokenization: Evaluating Text Compression and its Correlation with Model Performance

2024-03-10 · Omer Goldman, Avi Caciularu, Matan Eyal, Kris Cao 외

Despite it being the cornerstone of BPE, the most common tokenization algorithm, the importance of compression in the tokenization process is still unclear. In this paper, we argue for the theoretical importance of compr…

Language ModelingLanguage ModellingText Compression

Impact of Tokenization on Language Models: An Analysis for Turkish

2021-11-16 · ACL ARR November 2021 11 · Anonymous

Tokenization is an important text preprocessing step to prepare input tokens for language models. WordPiece and BPE are de-facto methods employed by large language models, such as BERT and GPT. However, the impact of tok…

Comparative Analysis of Tokenization Algorithms for Low-Resource Language Dzongkha

2025-09-18 · Tandin Wangchuk, Tad Gonsalves arxiv

Large Language Models (LLMs) are gaining popularity and improving rapidly. Tokenizers are crucial components of natural language processing, especially for LLMs. Tokenizers break down input text into tokens that models c…

Sentiment AnalysisText Generation