paper-with-me

Papers

Position Interpolation Improves ALiBi Extrapolation

2023-10-18 · Faisal Al-Khateeb, Nolan Dey, Daria Soboleva, Joel Hestness

Linear position interpolation helps pre-trained models using rotary position embeddings (RoPE) to extrapolate to longer sequence lengths. We propose using linear position interpolation to extend the extrapolation range of models using Attention with Linear Biases (ALiBi). We find position interpolation significantly improves extrapolation capability on upstream language modelling and downstream summarization and retrieval tasks.

📄 PDF Abstract BibTeX arXiv:2310.13017

Code (1)

ofirpress/attention_with_linear_biases pytorch

Tasks

Language ModellingPositionRetrieval

Similar Papers 제목 키워드 기반

Train Short, Test Long: Attention with Linear Biases Enables Input Length Extrapolation

2021-08-27 · ICLR 2022 4 · Ofir Press, Noah A. Smith, Mike Lewis

Since the introduction of the transformer model by Vaswani et al. (2017), a fundamental question has yet to be answered: how does a model achieve extrapolation at inference time for sequences that are longer than it saw …

Inductive BiasPlaying the Game of 2048PositionWord Embeddings

Dissecting Transformer Length Extrapolation via the Lens of Receptive Field Analysis

2022-12-20 · Ta-Chung Chi, Ting-Han Fan, Alexander I. Rudnicky, Peter J. Ramadge

Length extrapolation permits training a transformer language model on short sequences that preserves perplexities when tested on substantially longer sequences. A relative positional embedding design, ALiBi, has had the …

Language ModelingLanguage Modelling

Bayesian Attention Mechanism: A Probabilistic Framework for Positional Encoding and Context Length Extrapolation

2025-05-28 · Arthur S. Bianchessi, Rodrigo C. Barros, Lucas S. Kupssinskü

Transformer-based language models rely on positional encoding (PE) to handle token order and support context length extrapolation. However, existing PE methods lack theoretical clarity and rely on limited evaluation metr…

Information RetrievalRetrieval

Context-aware Biases for Length Extrapolation

2025-03-11 · Ali Veisi, Amir Mansourian

Transformers' ability to generalize to longer sequences than they have been trained on, known as length extrapolation, degrades as sequence length increases. Most of Relative Positional Encoding (RPE) methods address thi…

Decoder

Functional Interpolation for Relative Positions Improves Long Context Transformers

2023-10-06 · Shanda Li, Chong You, Guru Guruganesh, Joshua Ainslie 외

Preventing the performance decay of Transformers on inputs longer than those used for training has been an important challenge in extending the context length of these models. Though the Transformer architecture has fund…

Language ModelingLanguage ModellingPosition