paper-with-me

Papers

Transformer Language Models without Positional Encodings Still Learn Positional Information

2022-03-30 · Adi Haviv, Ori Ram, Ofir Press, Peter Izsak, Omer Levy

Causal transformer language models (LMs), such as GPT-3, typically require some form of positional encoding, such as positional embeddings. However, we show that LMs without any explicit positional encoding are still competitive with standard models, and that this phenomenon is robust across different datasets, model sizes, and sequence lengths. Probing experiments reveal that such models acquire an implicit notion of absolute positions throughout the network, effectively compensating for the missing information. We conjecture that causal attention enables the model to infer the number of predecessors that each token can attend to, thereby approximating its absolute position. Our findings indicate that causal LMs might derive positional awareness not only from the explicit positioning mechanism, but also from the effects of the causal mask.

📄 PDF Abstract BibTeX arXiv:2203.16634

Code (1)

adihaviv/nopos 공식 구현 pytorch

Tasks

Position

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Cosine Annealing Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before…
Linear Warmup With Cosine Annealing Linear Warmup With Cosine Annealing is a learning rate schedule where we increase the learning rate linearly for $n$ updates and then anneal according to a cosine schedule…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
{Dispute@FaQ-s}How to file a dispute with Expedia? How to file a dispute with Expedia? To file a complaint against Expedia, first try contacting their customer service directly. You can reach them by phone at…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

Position Information Emerges in Causal Transformers Without Positional Encodings via Similarity of Nearby Embeddings

2024-12-30 · Chunsheng Zuo, Pavel Guerzhoy, Michael Guerzhoy

Transformers with causal attention can solve tasks that require positional information without using positional encodings. In this work, we propose and investigate a new hypothesis about how positional information can be…

Position

HyPE-GT: where Graph Transformers meet Hyperbolic Positional Encodings

2023-12-11 · Kushal Bose, Swagatam Das

Graph Transformers (GTs) facilitate the comprehension of graph-structured data by calculating the self-attention of node pairs without considering node position information. To address this limitation, we introduce an in…

Position

The Impact of Positional Encodings on Multilingual Compression

2021-09-11 · EMNLP 2021 11 · Vinit Ravishankar, Anders Søgaard

In order to preserve word-order information in a non-autoregressive setting, transformer architectures tend to include positional knowledge, by (for instance) adding positional encodings to token embeddings. Several modi…

Inductive Bias

Positional Encoding Field

2025-10-23 · Yunpeng Bai, Haoxiang Li, Qixing Huang arxiv

Diffusion Transformers (DiTs) have emerged as the dominant architecture for visual generation, powering state-of-the-art image and video models. By representing images as patch tokens with positional encodings (PEs), DiT…

Novel View SynthesisImage Editing

Explicitly Encoding Structural Symmetry is Key to Length Generalization in Arithmetic Tasks

2024-06-04 · Mahdi Sabbaghi, George Pappas, Hamed Hassani, Surbhi Goel

Despite the success of Transformers on language understanding, code generation, and logical reasoning, they still fail to generalize over length on basic arithmetic tasks such as addition and multiplication. A major reas…

Code GenerationLogical ReasoningOut-of-Distribution Generalization