Guardado en:
Detalles Bibliográficos
Autor principal: Tabak, Akbay
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2401.11626
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913243453194240
author Tabak, Akbay
author_facet Tabak, Akbay
contents Freely Long-Thinking Transformer (FraiLT) is an improved transformer model designed to enhance processing capabilities without scaling up size. It utilizes a recursive approach, iterating over a subset of layers multiple times, and introduces iteration encodings to maintain awareness across these cycles. Iteration encoding allows FraiLT to achieve the interpretive depth of larger models in a compact form. When evaluated on a synthetic story dataset, FraiLT outperformed larger models, showcasing its ability to deliver high-quality performance while reducing memory demands. This model represents a step forward towards more efficient and accessible language models.
format Preprint
id arxiv_https___arxiv_org_abs_2401_11626
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Freely Long-Thinking Transformer (FraiLT)
Tabak, Akbay
Machine Learning
Computation and Language
Freely Long-Thinking Transformer (FraiLT) is an improved transformer model designed to enhance processing capabilities without scaling up size. It utilizes a recursive approach, iterating over a subset of layers multiple times, and introduces iteration encodings to maintain awareness across these cycles. Iteration encoding allows FraiLT to achieve the interpretive depth of larger models in a compact form. When evaluated on a synthetic story dataset, FraiLT outperformed larger models, showcasing its ability to deliver high-quality performance while reducing memory demands. This model represents a step forward towards more efficient and accessible language models.
title Freely Long-Thinking Transformer (FraiLT)
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2401.11626