Logo Lanfrica
  • Home
  • Atlas
  • Insights
  • Docs
  • Sign in

© 2026 Lanfrica. All rights reserved. All copyrights of the resources shown on the Lanfrica website belong to the original copyright holders, unless explicitly stated otherwise.

taresco/COHERE-GLOBAL-MMLU-FILTERED-MATH

Domain:

natural language processing

Record type:

dataset
Creator:
tar
Host:
This is a filtered subset of the CohereLabs/Global-MMLU dataset. This dataset only included mathematics (abstract algebra, high school mathematics & elementary mathematics) and it contains 10 languages. We created this subset to use for evaluation of math capabilities of LLMs on African Languages.

Visit

huggingface.co

Tasks

question answering

Languages

AmharicHausaIgboShonaSomaliSwahili

Tags

Math

Licenses

apache-2.0

Similar

Global MMLU-Lite (Chichewa, Māori, Inuktitut)taresco/xP3x_african_subsettaresco/open_math_instruct_v2_translated_african_languagesaritra1804/Cohere-LinguaLinksmutuvi/cohere-swahili-ndiziAfrican Languages - OpenAI MMLU

Global MMLU-Lite (Chichewa, Māori, Inuktitut)

Global MMLU-Lite is a multilingual evaluation benchmark for LLMs covering 18 languages. This dataset

taresco/xP3x_african_subset

This version of xP3x includes only African Languages + Arabic, English, French, & Portuguese. See Co

taresco/open_math_instruct_v2_translated_african_languages

This is a set of 41k nvidia/OpenMathInstruct-2 questions translated into 9 African languages using A

aritra1804/Cohere-LinguaLink

This project uses a multilingual embedding model to align sentences in one language ( preferably a l

smutuvi/cohere-swahili-ndizi

African Languages - OpenAI MMLU

This is a filtered subset of the openai/MMMLU dataset. This dataset only included mathematics (abstr