Salvato in:
Dettagli Bibliografici
Autori principali: Shamrai, Maksym, Hamolia, Vladyslav
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2508.11676
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913993630679040
author Shamrai, Maksym
Hamolia, Vladyslav
author_facet Shamrai, Maksym
Hamolia, Vladyslav
contents We introduce a novel framework that utilizes the internal weight activations of modern Large Language Models (LLMs) to construct a metric space of languages. Unlike traditional approaches based on hand-crafted linguistic features, our method automatically derives high-dimensional vector representations by computing weight importance scores via an adapted pruning algorithm. Our approach captures intrinsic language characteristics that reflect linguistic phenomena. We validate our approach across diverse datasets and multilingual LLMs, covering 106 languages. The results align well with established linguistic families while also revealing unexpected inter-language connections that may indicate historical contact or language evolution. The source code, computed language latent vectors, and visualization tool are made publicly available at https://github.com/mshamrai/deep-language-geometry.
format Preprint
id arxiv_https___arxiv_org_abs_2508_11676
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Deep Language Geometry: Constructing a Metric Space from LLM Weights
Shamrai, Maksym
Hamolia, Vladyslav
Computation and Language
Artificial Intelligence
Machine Learning
We introduce a novel framework that utilizes the internal weight activations of modern Large Language Models (LLMs) to construct a metric space of languages. Unlike traditional approaches based on hand-crafted linguistic features, our method automatically derives high-dimensional vector representations by computing weight importance scores via an adapted pruning algorithm. Our approach captures intrinsic language characteristics that reflect linguistic phenomena. We validate our approach across diverse datasets and multilingual LLMs, covering 106 languages. The results align well with established linguistic families while also revealing unexpected inter-language connections that may indicate historical contact or language evolution. The source code, computed language latent vectors, and visualization tool are made publicly available at https://github.com/mshamrai/deep-language-geometry.
title Deep Language Geometry: Constructing a Metric Space from LLM Weights
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2508.11676