English
Related papers

Related papers: Token embeddings violate the manifold hypothesis

200 papers

Hubness, the tendency for a few points to be among the nearest neighbours of a disproportionate number of other points, commonly arises when applying standard distance measures to high-dimensional data, often negatively impacting…

Computation and Language · Computer Science 2025-10-20 Beatrix M. G. Nielsen , Iuri Macocco , Marco Baroni

Large Language Models (LLMs) are widely deployed in real-world applications, yet little is known about their training dynamics at the token level. Evaluation typically relies on aggregated training loss, measured at the batch level, which…

Computation and Language · Computer Science 2024-10-17 Andrea Pinto , Tomer Galanti , Randall Balestriero

We evaluate LLMs' language understanding capacities on simple inference tasks that most humans find trivial. Specifically, we target (i) grammatically-specified entailments, (ii) premises with evidential adverbs of uncertainty, and (iii)…

Computation and Language · Computer Science 2024-04-12 Victoria Basmov , Yoav Goldberg , Reut Tsarfaty

Typical deep clustering methods, while achieving notable progress, can only provide one clustering result per dataset. This limitation arises from their assumption of a fixed underlying data distribution, which may fail to meet user needs…

Machine Learning · Computer Science 2025-12-02 Xinyue Wang , Yuheng Jia , Hui Liu , Junhui Hou

Extracting sentence embeddings from large language models (LLMs) is a promising direction, as LLMs have demonstrated stronger semantic understanding capabilities. Previous studies typically focus on prompt engineering to elicit sentence…

Computation and Language · Computer Science 2025-07-04 Yuchen Fu , Zifeng Cheng , Zhiwei Jiang , Zhonghui Wang , Yafeng Yin , Zhengliang Li , Qing Gu

The manifold hypothesis, which assumes that data lies on or close to an unknown manifold of low intrinsic dimension, is a staple of modern machine learning research. However, recent work has shown that real-world data exhibits distinct…

Machine Learning · Computer Science 2023-06-16 Julius von Rohrscheidt , Bastian Rieck

The manifold hypothesis presumes that high-dimensional data lies on or near a low-dimensional manifold. While the utility of encoding geometric structure has been demonstrated empirically, rigorous analysis of its impact on the learnability…

Machine Learning · Computer Science 2024-06-04 Bobak T. Kiani , Jason Wang , Melanie Weber

Tokenization is an important preprocessing step in the training and inference of large language models (LLMs). While there has been extensive research on the expressive power of the neural achitectures used in LLMs, the impact of…

Computation and Language · Computer Science 2024-12-05 Saibo Geng , Sankalp Gambhir , Chris Wendler , Robert West

Accurately quantifying uncertainty in large language models (LLMs) is crucial for their reliable deployment, especially in high-stakes applications. Current state-of-the-art methods for measuring semantic uncertainty in LLMs rely on strict…

Machine Learning · Computer Science 2024-10-31 Yashvir S. Grewal , Edwin V. Bonilla , Thang D. Bui

We study machine learning of phenomenologically relevant properties of string compactifications, which arise in the context of heterotic line bundle models. Both supervised and unsupervised learning are considered. We find that, for a fixed…

High Energy Physics - Theory · Physics 2020-03-31 Rehan Deen , Yang-Hui He , Seung-Joo Lee , Andre Lukas

Understanding the latent space geometry of large language models (LLMs) is key to interpreting their behavior and improving alignment. Yet it remains unclear to what extent LLMs linearly organize representations related to semantic…

Computation and Language · Computer Science 2026-01-22 Baturay Saglam , Paul Kassianik , Blaine Nelson , Sajana Weerawardhena , Yaron Singer , Amin Karbasi

Word embeddings represent language vocabularies as clouds of $d$-dimensional points. We investigate how information is conveyed by the general shape of these clouds, instead of representing the semantic meaning of each token. Specifically,…

Computation and Language · Computer Science 2025-01-15 Ondřej Draganov , Steven Skiena

In this work, we observe an interesting phenomenon: it is possible to generate reversible sentence embeddings that allow an LLM to reconstruct the original text exactly, without modifying the model's weights. This is achieved by introducing…

Computation and Language · Computer Science 2026-01-09 Ignacio Sastre , Aiala Rosá

There is a large ongoing scientific effort in mechanistic interpretability to map embeddings and internal representations of AI systems into human-understandable concepts. A key element of this effort is the linear representation…

Machine Learning · Computer Science 2025-05-27 Alexander Modell , Patrick Rubin-Delanchy , Nick Whiteley

Large Language Models (LLMs) are known to produce very high-quality tests and responses to our queries. But how much can we trust this generated text? In this paper, we study the problem of uncertainty quantification in LLMs. We propose a…

Computation and Language · Computer Science 2025-04-28 Muhammad Mubashar , Shireen Kudukkil Manchingal , Fabio Cuzzolin

Psychological research consistently finds that human ratings of words across diverse semantic scales can be reduced to a low-dimensional form with relatively little information loss. We find that the semantic associations encoded in the…

Computation and Language · Computer Science 2025-08-15 Austin C. Kozlowski , Callin Dai , Andrei Boutyline

The manifold hypothesis suggests that word vectors live on a submanifold within their ambient vector space. We argue that we should, more accurately, expect them to live on a pinched manifold: a singular quotient of a manifold obtained by…

Computation and Language · Computer Science 2020-11-19 Alexander Jakubowski , Milica Gašić , Marcus Zibrowius

The embedding space of language models is widely believed to capture the semantic relationships; for instance, embeddings of digits often exhibit an ordered structure that corresponds to their natural sequence. However, the mechanisms…

Machine Learning · Computer Science 2025-09-25 Junjie Yao , Zhi-Qin John Xu

Multi-modal Large Language Models (MLLMs) have demonstrated remarkable capabilities in understanding and generating content across various modalities, such as images and text. However, their interpretability remains a challenge, hindering…

Computer Vision and Pattern Recognition · Computer Science 2024-05-29 Loris Giulivi , Giacomo Boracchi

Despite their widespread use, the mechanisms by which large language models (LLMs) represent and regulate uncertainty in next-token predictions remain largely unexplored. This study investigates two critical components believed to influence…

Machine Learning · Computer Science 2024-11-11 Alessandro Stolfo , Ben Wu , Wes Gurnee , Yonatan Belinkov , Xingyi Song , Mrinmaya Sachan , Neel Nanda