English
Related papers

Related papers: Critical attention scaling in long-context transfo…

200 papers

The success of Transformer language models is widely credited to their dot-product attention mechanism, which interweaves a set of key design principles: mixing information across positions (enabling multi-token interactions),…

Computation and Language · Computer Science 2025-10-14 Huiyin Xue , Nafise Sadat Moosavi , Nikolaos Aletras

Mechanistic interpretability assumes that circuit analysis becomes harder as models scale. We challenge this assumption by showing that the attention architecture matters more than parameter count. Studying three circuit types across Pythia…

Computation and Language · Computer Science 2026-05-12 Sohan Venkatesh

The transformer architecture has become the foundation of modern Large Language Models (LLMs), yet its theoretical properties are still not well understood. As with classic neural networks, a common approach to improve these models is to…

Artificial Intelligence · Computer Science 2025-09-26 Álvaro Rodríguez Abella , João Pedro Silvestre , Paulo Tabuada

Efficient parallelization of Large Language Models (LLMs) with long sequences is essential but challenging due to their significant computational and memory demands, particularly stemming from communication bottlenecks in attention…

Distributed, Parallel, and Cluster Computing · Computer Science 2024-12-31 Zongwu Wang , Fangxin Liu , Mingshuai Li , Li Jiang

Large language models based on the transformer architecture can solve highly complex tasks, yet their fundamental limitations on simple algorithmic problems remain poorly understood. In this work, we focus on basic counting tasks and…

Computation and Language · Computer Science 2026-02-26 Gilad Yehudai , Haim Kaplan , Guy Dar , Royi Rassin , Asma Ghandeharioun , Mor Geva , Amir Globerson

Length-dependent logit rescaling is widely used to stabilize long-context self-attention, but existing analyses and methods suggest conflicting inverse-temperature laws for the context length $n$, ranging from $(\log n)^{1/2}$ to $\log n$…

Machine Learning · Statistics 2026-05-14 Tomohiro Hayase , Ryo Karakida

Efficient inference in Large Vision-Language Models is constrained by the high cost of processing thousands of visual tokens, yet it remains unclear which tokens and computations can be safely removed. While attention scores are commonly…

Computer Vision and Pattern Recognition · Computer Science 2026-02-03 Samyak Jha , Junho Kim

Transformers have driven remarkable breakthroughs in natural language processing and computer vision, yet their standard attention mechanism still imposes O(N^2) complexity, hindering scalability to longer sequences. We introduce…

Machine Learning · Computer Science 2026-01-06 Yoshihiro Yamada

Large language models (LLMs) excel at natural language understanding and generation but remain vulnerable to factual errors, limiting their reliability in knowledge-intensive tasks. While decoding-time strategies provide a promising…

Artificial Intelligence · Computer Science 2025-10-06 Jingze Zhu , Yongliang Wu , Wenbo Zhu , Jiawang Cao , Yanqiang Zheng , Jiawei Chen , Xu Yang , Bernt Schiele , Jonas Fischer , Xinting Hu

We present a novel non attention based architecture for large language models (LLMs) that efficiently handles very long context windows, on the order of hundreds of thousands to potentially millions of tokens. Unlike traditional Transformer…

Machine Learning · Computer Science 2025-06-04 Andrew Kiruluta , Preethi Raju , Priscilla Burity

Recent works show we can linearize large language models (LLMs) -- swapping the quadratic attentions of popular Transformer-based LLMs with subquadratic analogs, such as linear attention -- avoiding the expensive pretraining costs. However,…

Machine Learning · Computer Science 2025-03-07 Michael Zhang , Simran Arora , Rahul Chalamala , Alan Wu , Benjamin Spector , Aaryan Singhal , Krithik Ramesh , Christopher Ré

Despite the progress made in sentence-level NMT, current systems still fall short at achieving fluent, good quality translation for a full document. Recent works in context-aware NMT consider only a few previous sentences as context and may…

Computation and Language · Computer Science 2019-05-27 Sameen Maruf , André F. T. Martins , Gholamreza Haffari

This study addresses the problem of hallucinated span detection in the outputs of large language models. It has received less attention than output-level hallucination detection despite its practical importance. Prior work has shown that…

Computation and Language · Computer Science 2025-09-16 Yuya Ogasa , Yuki Arase

Transformer models have achieved state-of-the-art results in a wide range of NLP tasks including summarization. Training and inference using large transformer models can be computationally expensive. Previous work has focused on one…

Computation and Language · Computer Science 2021-09-10 Potsawee Manakul , Mark J. F. Gales

We explore the internal mechanisms of how bias emerges in large language models (LLMs) when provided with ambiguous comparative prompts: inputs that compare or enforce choosing between two or more entities without providing clear context…

Computation and Language · Computer Science 2024-10-31 Rishabh Adiga , Besmira Nushi , Varun Chandrasekaran

In large-scale recommender systems, ultra-long user behavior sequences encode rich signals of evolving interests. Extending sequence length generally improves accuracy, but directly modeling such sequences in production is infeasible due to…

Information Retrieval · Computer Science 2025-08-26 Kaiyuan Li , Yongxiang Tang , Yanhua Cheng , Yong Bai , Yanxiang Zeng , Chao Wang , Xialong Liu , Peng Jiang

Extended sequence generation often leads to degradation in contextual consistency due to the inability of conventional self-attention mechanisms to effectively retain long-range dependencies. Existing approaches, including memory…

Computation and Language · Computer Science 2025-01-30 Jonathan Teel , Jocasta Cumberbatch , Raphael Benington , Quentin Baskerville

Scaling the context size of large language models (LLMs) enables them to perform various new tasks, e.g., book summarization. However, the memory cost of the Key and Value (KV) cache in attention significantly limits the practical…

Computation and Language · Computer Science 2024-10-03 Zhiyu Guo , Hidetaka Kamigaito , Taro Watanabe

The self-attention mechanism has significantly advanced the field of natural language processing, facilitating the development of advanced language-learning machines. Although its utility is widely acknowledged, the precise mechanisms of…

Computation and Language · Computer Science 2026-02-04 Tal Halevi , Yarden Tzach , Ronit D. Gross , Shalom Rosner , Ido Kanter

Transformer-based models have achieved dominant performance in numerous NLP tasks. Despite their remarkable successes, pre-trained transformers such as BERT suffer from a computationally expensive self-attention mechanism that interacts…

Computation and Language · Computer Science 2024-06-04 Jungmin Yun , Mihyeon Kim , Youngbin Kim