English
Related papers

Related papers: Training-Free Looped Transformers

200 papers

Transformers evaluated in a single, fixed-depth pass are provably limited in expressive power to the constant-depth circuit class TC0. Running a Transformer autoregressively removes that ceiling -- first in next-token prediction and, more…

Machine Learning · Computer Science 2025-07-21 Mrinal Mathur , Mike Doan , Barak Pearlmutter , Sergey Plis

While looped transformers (termed as Looped-Attn) often outperform standard transformers (termed as Single-Attn) on complex reasoning tasks, the mechanism for this advantage remains underexplored. In this paper, we explain this phenomenon…

Machine Learning · Computer Science 2026-01-07 Zixuan Gong , Yong Liu , Jiaye Teng

Multiplications are responsible for most of the computational cost involved in neural network training and inference. Recent research has thus looked for ways to reduce the cost associated with them. Inspired by Mogami (2020), we replace…

Machine Learning · Computer Science 2023-10-26 Atli Kosson , Martin Jaggi

Transformers lack an explicit architectural mechanism for storing and organizing knowledge acquired during training. We introduce learnable sparse memory banks: a set of latent tokens, randomly initialized and trained end-to-end, that…

Machine Learning · Computer Science 2026-03-24 Tasmay Pankaj Tibrewal , Pritish Saha , Ankit Meda , Kunal Singh , Pradeep Moturi

Looped Transformers have emerged as an efficient and powerful class of models for reasoning in the language domain. Recent studies show that these models achieve strong performance on algorithmic and reasoning tasks, suggesting that looped…

Computation and Language · Computer Science 2026-02-13 Ahmadreza Jeddi , Marco Ciccone , Babak Taati

Transformer-based QA models use input-wide self-attention -- i.e. across both the question and the input passage -- at all layers, causing them to be slow and memory-intensive. It turns out that we can get by without input-wide…

Computation and Language · Computer Science 2020-05-05 Qingqing Cao , Harsh Trivedi , Aruna Balasubramanian , Niranjan Balasubramanian

As large language models (LLMs) grow in scale and capability, understanding their internal mechanisms becomes increasingly critical. Sparse autoencoders (SAEs) have emerged as a key tool in mechanistic interpretability, enabling the…

Computation and Language · Computer Science 2025-06-10 Jiaming Li , Haoran Ye , Yukun Chen , Xinyue Li , Lei Zhang , Hamid Alinejad-Rokny , Jimmy Chih-Hsien Peng , Min Yang

In deep learning, models typically reuse the same parameters for all inputs. Mixture of Experts (MoE) defies this and instead selects different parameters for each incoming example. The result is a sparsely-activated model -- with…

Machine Learning · Computer Science 2022-06-20 William Fedus , Barret Zoph , Noam Shazeer

In order to reduce the computational complexity of large language models, great efforts have been made to to improve the efficiency of transformer models such as linear attention and flash-attention. However, the model size and…

Computation and Language · Computer Science 2026-02-04 Ning Ding , Yehui Tang , Haochen Qin , Zhenli Zhou , Chao Xu , Lin Li , Kai Han , Heng Liao , Yunhe Wang

Deploying transformer models in practice is challenging due to their inference cost, which scales quadratically with input sequence length. To address this, we present a novel Learned Token Pruning (LTP) method which adaptively removes…

Computation and Language · Computer Science 2022-06-06 Sehoon Kim , Sheng Shen , David Thorsley , Amir Gholami , Woosuk Kwon , Joseph Hassoun , Kurt Keutzer

This paper reveals that large language models (LLMs), despite being trained solely on textual data, are surprisingly strong encoders for purely visual tasks in the absence of language. Even more intriguingly, this can be achieved by a…

Computer Vision and Pattern Recognition · Computer Science 2024-05-07 Ziqi Pang , Ziyang Xie , Yunze Man , Yu-Xiong Wang

Trainable layers such as convolutional building blocks are the standard network design choices by learning parameters to capture the global context through successive spatial operations. When designing an efficient network, trainable layers…

Computer Vision and Pattern Recognition · Computer Science 2022-03-22 Dongyoon Han , YoungJoon Yoo , Beomyoung Kim , Byeongho Heo

Transformer-based models generally allocate the same amount of computation for each token in a given sequence. We develop a simple but effective "token dropping" method to accelerate the pretraining of transformer models, such as BERT,…

Computation and Language · Computer Science 2022-03-25 Le Hou , Richard Yuanzhe Pang , Tianyi Zhou , Yuexin Wu , Xinying Song , Xiaodan Song , Denny Zhou

Continual Learning (CL) involves training a machine learning model in a sequential manner to learn new information while retaining previously learned tasks without the presence of previous training data. Although there has been significant…

Computer Vision and Pattern Recognition · Computer Science 2023-08-23 Anurag Roy , Vinay Kumar Verma , Sravan Voonna , Kripabandhu Ghosh , Saptarshi Ghosh , Abir Das

In contrast to RNNs, which compress their history into a single hidden state, Transformers can attend to all past tokens directly. However, standard Transformers rely solely on the hidden state from the previous layer to represent the…

Machine Learning · Computer Science 2025-05-29 Gleb Gerasimov , Yaroslav Aksenov , Nikita Balagansky , Viacheslav Sinii , Daniil Gavrilov

We study model-free Q-learning in finite-horizon episodic Markov Decision Processes (MDPs) with stationary dynamics across episodes. We identify a central issue in nascent model-free posterior-sampling works: the reliance on delayed…

Machine Learning · Computer Science 2026-05-21 Sofia R. Miskala-Dinc , Aviva Prins

Understanding why Transformers perform so well remains challenging due to their non-convex optimization landscape. In this work, we analyze a shallow Transformer with $m$ independent heads trained by projected gradient descent in the kernel…

Machine Learning · Computer Science 2026-04-03 Enes Arda , Semih Cayci , Atilla Eryilmaz

Meta-training, which fine-tunes the language model (LM) on various downstream tasks by maximizing the likelihood of the target label given the task instruction and input instance, has improved the zero-shot task generalization performance.…

Computation and Language · Computer Science 2023-06-07 Seonghyeon Ye , Doyoung Kim , Joel Jang , Joongbo Shin , Minjoon Seo

Transformers have the capacity to act as supervised learning algorithms: by properly encoding a set of labeled training ("in-context") examples and an unlabeled test example into an input sequence of vectors of the same dimension, the…

Machine Learning · Computer Science 2024-12-16 Spencer Frei , Gal Vardi

Recent advances have reformulated diffusion models as deterministic ordinary differential equations (ODEs) through the framework of flow matching, providing a unified formulation for the noise-to-data generative process. Various…

Computer Vision and Pattern Recognition · Computer Science 2026-02-11 Yan Luo , Henry Huang , Todd Y. Zhou , Mengyu Wang