English
Related papers

Related papers: Pre-trained Models Perform the Best When Token Dis…

200 papers

Large language models (LLMs), typically designed as a function of next-word prediction, have excelled across extensive NLP tasks. Despite the generality, next-word prediction is often not an efficient formulation for many of the tasks,…

Computation and Language · Computer Science 2023-11-03 Yuheng Zha , Yichi Yang , Ruichen Li , Zhiting Hu

Many NLP applications, such as biomedical data and technical support, have 10-100 million tokens of in-domain data and limited computational resources for learning from it. How should we train a language model in this scenario? Most…

Computation and Language · Computer Science 2020-10-01 Charles Welch , Rada Mihalcea , Jonathan K. Kummerfeld

Pretraining large language models effectively requires strategic data selection, blending and ordering. However, key details about data mixtures especially their scalability to longer token horizons and larger model sizes remain…

Computation and Language · Computer Science 2024-12-23 Steven Feng , Shrimai Prabhumoye , Kezhi Kong , Dan Su , Mostofa Patwary , Mohammad Shoeybi , Bryan Catanzaro

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

While textual frequency has been validated as relevant to human cognition in reading speed, its relatedness to Large Language Models (LLMs) is seldom studied. We propose a novel research direction in terms of textual data frequency, which…

Computation and Language · Computer Science 2026-04-08 Hongyuan Adam Lu , Z. L. , Victor Wei , Zefan Zhang , Zhao Hong , Qiqi Xiang , Bowen Cao , Wai Lam

Scaling laws guide the development of large language models (LLMs) by offering estimates for the optimal balance of model size, tokens, and compute. More recently, loss-to-loss scaling laws that relate losses across pretraining datasets and…

Machine Learning · Computer Science 2026-05-21 Prasanna Mayilvahanan , Thaddäus Wiedemer , Sayak Mallick , Matthias Bethge , Wieland Brendel

The success of today's large language models (LLMs) depends on the observation that larger models perform better. However, the origin of this neural scaling law, that loss decreases as a power law with model size, remains unclear. We…

Machine Learning · Computer Science 2026-05-05 Yizhou Liu , Ziming Liu , Jeff Gore

While deep learning techniques have shown promising results in many natural language processing (NLP) tasks, it has not been widely applied to the clinical domain. The lack of large datasets and the pervasive use of domain-specific language…

Computation and Language · Computer Science 2019-06-20 Jiin Nam , Seunghyun Yoon , Kyomin Jung

An important body of quantitative linguistics is constituted by a series of statistical laws about language usage. Despite the importance of these linguistic laws, some of them are poorly formulated, and, more importantly, there is no…

Physics and Society · Physics 2020-11-09 Alvaro Corral , Isabel Serra

Training large language models (LLMs) from scratch can yield models with unique functionalities and strengths, but it is costly and often leads to redundant capabilities. A more cost-effective alternative is to fuse existing pre-trained…

Computation and Language · Computer Science 2025-09-23 Runjia Zeng , James Chenhao Liang , Cheng Han , Zhiwen Cao , Jiahao Liu , Xiaojun Quan , Yingjie Victor Chen , Lifu Huang , Tong Geng , Qifan Wang , Dongfang Liu

Tokenization is a foundational step in natural language processing (NLP) tasks, bridging raw text and language models. Existing tokenization approaches like Byte-Pair Encoding (BPE) originate from the field of data compression, and it has…

Computation and Language · Computer Science 2024-10-08 Craig W. Schmidt , Varshini Reddy , Haoran Zhang , Alec Alameddine , Omri Uzan , Yuval Pinter , Chris Tanner

We propose a novel scaling law for general-purpose decoder-only language models (LMs) trained on multilingual data, tackling the problem of balancing languages during multilingual pretraining. A primary challenge in studying multilingual…

Computation and Language · Computer Science 2024-12-05 Yifei He , Alon Benhaim , Barun Patra , Praneetha Vaddamanu , Sanchit Ahuja , Parul Chopra , Vishrav Chaudhary , Han Zhao , Xia Song

Knowing exactly how many data points need to be labeled to achieve a certain model performance is a hugely beneficial step towards reducing the overall budgets for annotation. It pertains to both active learning and traditional data…

Computation and Language · Computer Science 2023-07-04 Ernie Chang , Muhammad Hassan Rashid , Pin-Jie Lin , Changsheng Zhao , Vera Demberg , Yangyang Shi , Vikas Chandra

Tokenization is a hardcoded compression step which remains in the training pipeline of Large Language Models (LLMs), despite a general trend towards architectures becoming increasingly end-to-end. Prior work has shown promising results at…

Machine Learning · Computer Science 2026-02-17 Sam Dauncey , Roger Wattenhofer

We study empirical scaling laws for transfer learning between distributions in an unsupervised, fine-tuning setting. When we train increasingly large neural networks from-scratch on a fixed-size dataset, they eventually become data-limited…

Machine Learning · Computer Science 2021-02-03 Danny Hernandez , Jared Kaplan , Tom Henighan , Sam McCandlish

Current language models (LMs) use a fixed, static subword tokenizer. This default choice typically results in degraded efficiency and language capabilities, especially in languages other than English. To address this issue, we challenge the…

Computation and Language · Computer Science 2025-06-12 Darius Feher , Ivan Vulić , Benjamin Minixhofer

The mapping of lexical meanings to wordforms is a major feature of natural languages. While usage pressures might assign short words to frequent meanings (Zipf's law of abbreviation), the need for a productive and open-ended vocabulary,…

Computation and Language · Computer Science 2021-05-04 Tiago Pimentel , Irene Nikkarinen , Kyle Mahowald , Ryan Cotterell , Damián Blasi

How much data is required to learn the structure of a language via next-token prediction? We study this question for synthetic datasets generated via a Probabilistic Context-Free Grammar (PCFG) -- a tree-like generative model that captures…

Computation and Language · Computer Science 2024-10-30 Francesco Cagnetta , Matthieu Wyart

In the last years, researchers have realized the difficulties of fitting power-law distributions properly. These difficulties are higher in Zipf's systems, due to the discreteness of the variables and to the existence of two representations…

Data Analysis, Statistics and Probability · Physics 2022-11-29 Alvaro Corral , Isabel Serra , Ramon Ferrer-i-Cancho

Deep neural networks exhibit empirical neural scaling laws, with error decreasing as a power law with increasing model or data size, across a wide variety of architectures, tasks, and datasets. This universality suggests that scaling laws…

Machine Learning · Computer Science 2024-12-12 Ari Brill