English
Related papers

Related papers: Residual Connections and the Causal Shift: Uncover…

200 papers

We seek to understand how the representations of individual tokens and the structure of the learned feature space evolve between layers in deep neural networks under different learning objectives. We focus on the Transformers for our…

Computation and Language · Computer Science 2019-09-05 Elena Voita , Rico Sennrich , Ivan Titov

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

The geometric evolution of token representations in large language models (LLMs) presents a fundamental paradox: while human language inherently organizes semantic information in low-dimensional spaces ($\sim 10^1$ dimensions), modern LLMs…

Computation and Language · Computer Science 2025-03-31 Zhuo-Yang Song , Zeyu Li , Qing-Hong Cao , Ming-xing Luo , Hua Xing Zhu

Token prediction stability remains a challenge in autoregressive generative models, where minor variations in early inference steps often lead to significant semantic drift over extended sequences. A structured modulation mechanism was…

Modern large language models (LLMs) excel at tasks that require storing and retrieving knowledge, such as factual recall and question answering. Transformers are central to this capability because they can encode information during training…

Machine Learning · Statistics 2026-03-18 Nuri Mert Vural , Alberto Bietti , Mahdi Soltanolkotabi , Denny Wu

We study Latent Recurrent Transformer (LRT), a lightweight augmentation of autoregressive transformers that reuses a high-level source-layer hidden state from the previous token as recurrent memory for the next token. Because this source…

Masked diffusion language models (MDLMs) have emerged as a promising alternative to dominant autoregressive approaches. Although they achieve competitive performance on several tasks, a substantial gap remains in open-ended text generation.…

Computation and Language · Computer Science 2026-02-02 Mengyu Ye , Ryosuke Takahashi , Keito Kudo , Jun Suzuki

Transformer networks have become the preferred architecture for many tasks due to their state-of-the-art performance. However, the optimal way to implement residual connections in Transformer, which are essential for effective training, is…

Computation and Language · Computer Science 2023-05-01 Shufang Xie , Huishuai Zhang , Junliang Guo , Xu Tan , Jiang Bian , Hany Hassan Awadalla , Arul Menezes , Tao Qin , Rui Yan

While large language models demonstrate remarkable capabilities at task-specific applications through fine-tuning, extending these benefits across diverse languages is essential for broad accessibility. However, effective cross-lingual…

Computation and Language · Computer Science 2025-06-03 Danni Liu , Jan Niehues

Transformer models systematically favor certain token positions, yet the architectural origins of this position bias remain poorly understood. This bias is closely connected to the Lost-in-the-Middle phenomenon, where models underutilize…

Machine Learning · Computer Science 2026-05-28 Hanna Herasimchyk , Robin Labryga , Tomislav Prusina , Sören Laue

This research aims to unravel how large language models (LLMs) iteratively refine token predictions through internal processing. We utilized a logit lens technique to analyze the model's token predictions derived from intermediate…

Computation and Language · Computer Science 2025-06-10 Jaturong Kongmanee

In this work, we argue that large language models (LLMs), though trained to predict only the next token, exhibit emergent planning behaviors: $\textbf{their hidden representations encode future outputs beyond the next token}$. Through…

Computation and Language · Computer Science 2025-08-05 Zhichen Dong , Zhanhui Zhou , Zhixuan Liu , Chao Yang , Chaochao Lu

This paper investigates how Large Language Models (LLMs) represent non-English tokens -- a question that remains underexplored despite recent progress. We propose a lightweight intervention method using representation steering, where a…

Computation and Language · Computer Science 2025-08-27 Omar Mahmoud , Buddhika Laknath Semage , Thommen George Karimpanal , Santu Rana

The mechanisms by which reasoning training reshapes LLMs' internal computations remain unclear. We study lightweight steering vectors inserted into the base model's residual stream and trained with a reinforcement-learning objective. These…

Large Multimodal Models (LMMs) have achieved remarkable success in vision-language tasks, yet their vast parameter counts are often underutilized during both training and inference. In this work, we embrace the idea of looping back to move…

Machine Learning · Computer Science 2026-02-11 Ruihan Xu , Yuting Gao , Lan Wang , Jianing Li , Weihao Chen , Qingpei Guo , Ming Yang , Shiliang Zhang

While there is much recent interest in studying why Transformer-based large language models make predictions the way they do, the complex computations performed within each layer have made their behavior somewhat opaque. To mitigate this…

Computation and Language · Computer Science 2023-06-06 Byung-Doh Oh , William Schuler

Transformers have achieved state-of-the-art performance in language modeling tasks. However, the reasons behind their tremendous success are still unclear. In this paper, towards a better understanding, we train a Transformer model on a…

Machine Learning · Statistics 2024-06-06 Michael E. Sander , Raja Giryes , Taiji Suzuki , Mathieu Blondel , Gabriel Peyré

Representation Misdirection for Unlearning (RMU), which steers model representation in the intermediate layer to a target random representation, is an effective method for large language model (LLM) unlearning. Despite its high performance,…

Computation and Language · Computer Science 2025-02-07 Dang Huu-Tien , Trung-Tin Pham , Hoang Thanh-Tung , Naoya Inoue

Transformers underpin modern large language models (LLMs) and are commonly assumed to be behaviorally unstructured at random initialization, with all meaningful preferences emerging only through large-scale training. We challenge this…

Machine Learning · Statistics 2026-02-06 Siquan Li , Yao Tong , Haonan Wang , Tianyang Hu

Recent studies have revealed various manifestations of position bias in transformer architectures, from the "lost-in-the-middle" phenomenon to attention sinks, yet a comprehensive theoretical understanding of how attention masks and…

Machine Learning · Computer Science 2025-08-12 Xinyi Wu , Yifei Wang , Stefanie Jegelka , Ali Jadbabaie
‹ Prev 1 2 3 10 Next ›