English
Related papers

Related papers: Learning Less Is More: Premature Upper-Layer Atten…

200 papers

Transformers are mostly relying on softmax attention, which introduces quadratic complexity with respect to sequence length and remains a major bottleneck for efficient inference. Prior work on linear or hybrid attention typically replaces…

Pre-trained language models (PLM) have demonstrated their effectiveness for a broad range of information retrieval and natural language processing tasks. As the core part of PLM, multi-head self-attention is appealing for its ability to…

Computation and Language · Computer Science 2022-04-07 Shanshan Wang , Zhumin Chen , Zhaochun Ren , Huasheng Liang , Qiang Yan , Pengjie Ren

Natural language processing (NLP) enables the understanding and generation of meaningful human language, typically using a pre-trained complex architecture on a large dataset to learn the language and next fine-tune its weights to implement…

Computation and Language · Computer Science 2025-09-04 Yarden Tzach , Ronit D. Gross , Ella Koresh , Shalom Rosner , Or Shpringer , Tal Halevi , Ido Kanter

Transformer-based language models (LMs) are at the core of modern NLP, but their internal prediction construction process is opaque and largely not understood. In this work, we make a substantial step towards unveiling this underlying…

Computation and Language · Computer Science 2022-10-14 Mor Geva , Avi Caciularu , Kevin Ro Wang , Yoav Goldberg

We study length generalization in sequence models on a composite problem involving both state tracking and associative recall. Prior work finds that recurrent networks handle state tracking well but struggle with recall, whereas…

Machine Learning · Computer Science 2025-10-02 Buu Phan , Reza Ebrahimi , Sanjay Haresh , Roland Memisevic

Large Language Models (LLMs) often assign disproportionate attention to the first token, a phenomenon known as the attention sink. Several recent approaches aim to address this issue, including Sink Attention in GPT-OSS and Gated Attention…

Computation and Language · Computer Science 2026-05-28 Zizhuo Fu , Wenxuan Zeng , Runsheng Wang , Meng Li

As Large Language Models (LLMs) achieve remarkable empirical success through scaling model and data size, pretraining has become increasingly critical yet computationally prohibitive, hindering rapid development. Despite the availability of…

Computation and Language · Computer Science 2026-02-06 Ji Zhao , Yufei Gu , Shitong Shao , Xun Zhou , Liang Xiang , Zeke Xie

Video Large Language Models (VLLMs) incur substantial prefilling cost due to the large number of visual tokens. While attention-based token pruning offers a promising acceleration strategy, applying it at shallow decoder layers often causes…

Computer Vision and Pattern Recognition · Computer Science 2026-03-17 Yingjie Xia , Tao Liu , Jinglei Shi , Qingsong Xie , Heng Guo , Jian Yang , Xi Wang

Large vision-language models (LVLMs) often hallucinate content that is fluent yet unsupported by the image, limiting their reliability in real-world deployment. We show that a key failure mode arises from route competition: even when visual…

Computer Vision and Pattern Recognition · Computer Science 2026-05-26 Zhe Cheng , Wenyu Chen , Fode Zhang , Dehuan Shen

Central to the success of Transformers is the attention block, which effectively models global dependencies among input tokens associated to a dataset. However, we theoretically demonstrate that standard attention mechanisms in transformers…

Machine Learning · Computer Science 2026-03-31 Hemanth Saratchandran

Transformer architectures have emerged as promising deep learning (DL) tools for modeling complex sequence-to-sequence interactions in channel decoding. However, current transformer-based decoders for error correction codes (ECCs)…

Signal Processing · Electrical Eng. & Systems 2025-07-22 Hongzhi Zhu , Wei Xu , Xiaohu You

We propose a general framework for decoding quantum error-correcting codes with generative modeling. The model utilizes autoregressive neural networks, specifically Transformers, to learn the joint probability of logical operators and…

Quantum Physics · Physics 2023-07-19 Hanyan Cao , Feng Pan , Yijia Wang , Pan Zhang

The Transformer is a fully attention-based alternative to recurrent networks that has achieved state-of-the-art results across a range of NLP tasks. In this paper, we analyze the structure of attention in a Transformer language model, the…

Computation and Language · Computer Science 2019-06-20 Jesse Vig , Yonatan Belinkov

Variational autoencoders have been widely applied for natural language generation, however, there are two long-standing problems: information under-representation and posterior collapse. The former arises from the fact that only the last…

Machine Learning · Computer Science 2021-06-17 Xianghong Fang , Haoli Bai , Zenglin Xu , Michael Lyu , Irwin King

Decoder-only transformer networks have become incredibly popular for language modeling tasks. State-of-the-art models can have over a hundred transformer blocks, containing billions of trainable parameters, and are trained on trillions of…

Computation and Language · Computer Science 2025-05-13 Isaac Gerber

Deep pretrained language models have achieved great success in the way of pretraining first and then fine-tuning. But such a sequential transfer learning paradigm often confronts the catastrophic forgetting problem and leads to sub-optimal…

Computation and Language · Computer Science 2020-04-28 Sanyuan Chen , Yutai Hou , Yiming Cui , Wanxiang Che , Ting Liu , Xiangzhan Yu

Linear attention Transformers and their gated variants, celebrated for enabling parallel training and efficient recurrent inference, still fall short in recall-intensive tasks compared to traditional Transformers and demand significant…

Computation and Language · Computer Science 2024-11-01 Yu Zhang , Songlin Yang , Ruijie Zhu , Yue Zhang , Leyang Cui , Yiqiao Wang , Bolun Wang , Freda Shi , Bailin Wang , Wei Bi , Peng Zhou , Guohong Fu

Decoder-only language models, such as GPT and LLaMA, generally decode on the last layer. Motivated by human's hierarchical thinking capability, we propose that a hierarchical decoder architecture could be built with different layers…

Computation and Language · Computer Science 2025-09-30 Yihong Wang , Zhonglin Jiang , Ningyuan Xi , Yue Zhao , Qingqing Gu , Xiyuan Chen , Hao Wu , Sheng Xu , Hange Zhou , Yong Chen , Luo Ji

This paper explores the improvement of post-training quantization (PTQ) after knowledge distillation in the Whisper speech foundation model family. We address the challenge of outliers in weights and activation tensors, known to impede…

Sound · Computer Science 2024-06-18 Dominik Wagner , Ilja Baumann , Korbinian Riedhammer , Tobias Bocklet

Since 2017, the Transformer-based models play critical roles in various downstream Natural Language Processing tasks. However, a common limitation of the attention mechanism utilized in Transformer Encoder is that it cannot automatically…

Computation and Language · Computer Science 2022-04-20 Ziyang Luo , Yadong Xi , Jing Ma , Zhiwei Yang , Xiaoxi Mao , Changjie Fan , Rongsheng Zhang