中文
相关论文

相关论文: Latent Positional Information is in the Self-Atten…

200 篇论文

One of the most striking features of Large Language Models (LLMs) is their ability to learn in-context. Namely at inference time an LLM is able to learn new patterns without any additional weight update when these patterns are presented in…

计算与语言 · 计算机科学 2025-12-24 Benoit Dherin , Michael Munn , Hanna Mazzawi , Michael Wunder , Javier Gonzalvo

Embedding models are crucial for tasks in Information Retrieval (IR) and semantic similarity measurement, yet their handling of longer texts and associated positional biases remains underexplored. In this study, we investigate the impact of…

计算与语言 · 计算机科学 2026-01-01 Reagan J. Lee , Samarth Goel , Kannan Ramchandran

Despite being the current de-facto models in most NLP tasks, transformers are often limited to short sequences due to their quadratic attention complexity on the number of tokens. Several attempts to address this issue were studied, either…

计算与语言 · 计算机科学 2023-07-19 Amine Abdaoui , Sourav Dutta

Multilingual pretrained language models (MPLMs) exhibit multilinguality and are well suited for transfer across languages. Most MPLMs are trained in an unsupervised fashion and the relationship between their objective and multilinguality is…

计算与语言 · 计算机科学 2021-09-17 Sheng Liang , Philipp Dufter , Hinrich Schütze

Vector embeddings derived from large language models (LLMs) show promise in capturing latent information from the literature. Interestingly, these can be integrated into material embeddings, potentially useful for data-driven predictions of…

计算与语言 · 计算机科学 2024-09-19 Luke P. J. Gilligan , Matteo Cobelli , Hasan M. Sayeed , Taylor D. Sparks , Stefano Sanvito

Representation learning for text via pretraining a language model on a large corpus has become a standard starting point for building NLP systems. This approach stands in contrast to autoencoders, also trained on raw text, but with the…

计算与语言 · 计算机科学 2021-09-14 Ivan Montero , Nikolaos Pappas , Noah A. Smith

Discrete latent spaces in variational autoencoders have been shown to effectively capture the data distribution for many real-world problems such as natural language understanding, human intent prediction, and visual scene representation.…

机器学习 · 计算机科学 2021-01-19 Masha Itkina , Boris Ivanovic , Ransalu Senanayake , Mykel J. Kochenderfer , Marco Pavone

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…

计算机视觉与模式识别 · 计算机科学 2024-05-07 Ziqi Pang , Ziyang Xie , Yunze Man , Yu-Xiong Wang

Sentence embeddings from transformer models encode in a fixed length vector much linguistic information. We explore the hypothesis that these embeddings consist of overlapping layers of information that can be separated, and on which…

计算与语言 · 计算机科学 2024-07-03 Vivi Nastase , Paola Merlo

So far, expensive finetuning beyond the pretraining sequence length has been a requirement for effectively extending the context of language models (LM). In this work, we break this key bottleneck by Dropping the Positional Embeddings of…

计算与语言 · 计算机科学 2025-12-16 Yoav Gelberg , Koshi Eguchi , Takuya Akiba , Edoardo Cetin

Neural language models learn word representations, or embeddings, that capture rich linguistic and conceptual information. Here we investigate the embeddings learned by neural machine translation models, a recently-developed class of neural…

计算与语言 · 计算机科学 2015-04-06 Felix Hill , Kyunghyun Cho , Sebastien Jean , Coline Devin , Yoshua Bengio

Do pretrained language models have knowledge regarding the surface information of tokens? We examined the surface information stored in word or subword embeddings acquired by pretrained language models from the perspectives of token length,…

计算与语言 · 计算机科学 2024-02-23 Tatsuya Hiraoka , Naoaki Okazaki

A central question in multilingual language modeling is whether large language models (LLMs) develop a universal concept representation, disentangled from specific languages. In this paper, we address this question by analyzing latent…

计算与语言 · 计算机科学 2025-06-26 Clément Dumas , Chris Wendler , Veniamin Veselovsky , Giovanni Monea , Robert West

Vision Transformers (ViTs) lack the hierarchical inductive biases inherent to Convolutional Neural Networks (CNNs), theoretically allowing them to maintain high-dimensional representations throughout all layers. However, recent observations…

计算机视觉与模式识别 · 计算机科学 2025-12-09 Kanishk Awadhiya

Vision transformers (ViTs) have demonstrated remarkable performance in a variety of vision tasks. Despite their promising capabilities, training a ViT requires a large amount of diverse data. Several studies empirically found that using…

计算机视觉与模式识别 · 计算机科学 2024-05-24 Bum Jun Kim , Sang Woo Kim

Token representations in high-dimensional latent spaces often exhibit redundancy, limiting computational efficiency and reducing structural coherence across model layers. Hierarchical latent space folding introduces a structured…

计算与语言 · 计算机科学 2025-08-11 Fenella Harcourt , Naderdel Piero , Gilbert Sutherland , Daphne Holloway , Harriet Bracknell , Julian Ormsby

What happens when a language model thinks without words? Standard reasoning LLMs verbalize intermediate steps as chain-of-thought; latent reasoning transformers (LRTs) instead perform deliberation entirely in continuous hidden space. We…

计算与语言 · 计算机科学 2026-02-10 Jasmine Cui , Charles Ye

We conducted empirical experiments to assess the transferability of a light curve transformer to datasets with different cadences and magnitude distributions using various positional encodings (PEs). We proposed a new approach to…

Causal self-attention provides positional information to Transformer decoders. Prior work has shown that stacks of causal self-attention layers alone induce a positional bias in attention scores toward earlier tokens. However, this differs…

计算与语言 · 计算机科学 2026-04-21 Junu Kim , Xiao Liu , Zhenghao Lin , Lei Ji , Yeyun Gong , Edward Choi

The Transformer model has achieved state-of-the-art performance in many sequence modeling tasks. However, how to leverage model capacity with large or variable depths is still an open challenge. We present a probabilistic framework to…

计算与语言 · 计算机科学 2020-10-19 Xian Li , Asa Cooper Stickland , Yuqing Tang , Xiang Kong