Benchmarking Sociolinguistic Diversity in Swahili NLP: A Taxonomy-Guided Approach

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Oketch, Kezia, Lalor, John P., Abbasi, Ahmed
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918127384657920
author Oketch, Kezia
Lalor, John P.
Abbasi, Ahmed
author_facet Oketch, Kezia
Lalor, John P.
Abbasi, Ahmed
contents We introduce the first taxonomy-guided evaluation of Swahili NLP, addressing gaps in sociolinguistic diversity. Drawing on health-related psychometric tasks, we collect a dataset of 2,170 free-text responses from Kenyan speakers. The data exhibits tribal influences, urban vernacular, code-mixing, and loanwords. We develop a structured taxonomy and use it as a lens for examining model prediction errors across pre-trained and instruction-tuned language models. Our findings advance culturally grounded evaluation frameworks and highlight the role of sociolinguistic variation in shaping model performance.
format Preprint
id arxiv_https___arxiv_org_abs_2508_14051
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Benchmarking Sociolinguistic Diversity in Swahili NLP: A Taxonomy-Guided Approach
Oketch, Kezia
Lalor, John P.
Abbasi, Ahmed
Computation and Language
Computers and Society
We introduce the first taxonomy-guided evaluation of Swahili NLP, addressing gaps in sociolinguistic diversity. Drawing on health-related psychometric tasks, we collect a dataset of 2,170 free-text responses from Kenyan speakers. The data exhibits tribal influences, urban vernacular, code-mixing, and loanwords. We develop a structured taxonomy and use it as a lens for examining model prediction errors across pre-trained and instruction-tuned language models. Our findings advance culturally grounded evaluation frameworks and highlight the role of sociolinguistic variation in shaping model performance.
title Benchmarking Sociolinguistic Diversity in Swahili NLP: A Taxonomy-Guided Approach
topic Computation and Language
Computers and Society
url https://arxiv.org/abs/2508.14051