AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages

• URL: http://arxiv.org/abs/2601.06395v1
• Date: Sat, 10 Jan 2026 02:39:31 GMT
• Title: AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages
• Authors: Hao Yu, Tianyi Xu, Michael A. Hedderich, Wassim Hamidouche, Syed Waqas Zamir, David Ifeoluwa Adelani,
• Abstract summary: Large language models (LLMs) are increasingly multilingual, yet open models continue to underperform relative to proprietary systems.<n>We present textttAfriqueLLM, a suite of open LLMs adapted to 20 African languages through CPT on 26B tokens.
• Score: 30.309928265469427
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) are increasingly multilingual, yet open models continue to underperform relative to proprietary systems, with the gap most pronounced for African languages. Continued pre-training (CPT) offers a practical route to language adaptation, but improvements on demanding capabilities such as mathematical reasoning often remain limited. This limitation is driven in part by the uneven domain coverage and missing task-relevant knowledge that characterize many low-resource language corpora. We present \texttt{AfriqueLLM}, a suite of open LLMs adapted to 20 African languages through CPT on 26B tokens. We perform a comprehensive empirical study across five base models spanning sizes and architectures, including Llama 3.1, Gemma 3, and Qwen 3, and systematically analyze how CPT data composition shapes downstream performance. In particular, we vary mixtures that include math, code, and synthetic translated data, and evaluate the resulting models on a range of multilingual benchmarks. Our results identify data composition as the primary driver of CPT gains. Adding math, code, and synthetic translated data yields consistent improvements, including on reasoning-oriented evaluations. Within a fixed architecture, larger models typically improve performance, but architectural choices dominate scale when comparing across model families. Moreover, strong multilingual performance in the base model does not reliably predict post-CPT outcomes; robust architectures coupled with task-aligned data provide a more dependable recipe. Finally, our best models improve long-context performance, including document-level translation. Models have been released on [Huggingface](https://huggingface.co/collections/McGill-NLP/afriquellm).

Related papers

• What Matters When Building Universal Multilingual Named Entity Recognition Models? [12.125413756152833]
  We introduce Otter, a universal multilingual NER model supporting over 100 languages.<n>Otter achieves consistent improvements over strong multilingual NER baselines, outperforming GLiNER-x-base by 5.3pp in F1.<n>We release model checkpoints, training and evaluation code to facilitate and future research.
  arXiv  Detail & Related papers  (2026-01-09T23:02:37Z)
• Pretraining Language Models to Ponder in Continuous Space [50.52734567589996]
  We introduce this pondering process into language models by repeatedly invoking the forward process within a single token generation step.<n>We show that the model can learn to ponder in this way through self-supervised learning, without any human annotations.
  arXiv  Detail & Related papers  (2025-05-27T03:47:33Z)
• Improving Multilingual Math Reasoning for African Languages [49.27985213689457]
  We conduct experiments to evaluate different combinations of data types (translated versus synthetically generated), training stages (pre-training versus post-training), and other model adaptation configurations.<n>Our experiments focuses on mathematical reasoning tasks, using the Llama 3.1 model family as our base model.
  arXiv  Detail & Related papers  (2025-05-26T11:35:01Z)
• The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs [45.08958917457921]
  Large language models (LLMs) still struggle across tasks outside of high-resource languages.<n>In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce.
  arXiv  Detail & Related papers  (2025-05-23T20:28:31Z)
• Lugha-Llama: Adapting Large Language Models for African Languages [48.97516583523523]
  Large language models (LLMs) have achieved impressive results in a wide range of natural language applications.<n>We consider how to adapt LLMs to low-resource African languages.<n>We find that combining curated data from African languages with high-quality English educational texts results in a training mix that substantially improves the model's performance on these languages.
  arXiv  Detail & Related papers  (2025-04-09T02:25:53Z)
• P-MMEval: A Parallel Multilingual Multitask Benchmark for Consistent Evaluation of LLMs [84.24644520272835]
  We introduce P-MMEval, a large-scale benchmark covering effective fundamental and capability-specialized datasets.<n>P-MMEval delivers consistent language coverage across various datasets and provides parallel samples.<n>We conduct extensive experiments on representative multilingual model series to compare performances across models and tasks.
  arXiv  Detail & Related papers  (2024-11-14T01:29:36Z)
• Accelerating Large Language Model Pretraining via LFR Pedagogy: Learn, Focus, and Review [50.78587571704713]
  Learn-Focus-Review (LFR) is a dynamic training approach that adapts to the model's learning progress.<n>LFR tracks the model's learning performance across data blocks (sequences of tokens) and prioritizes revisiting challenging regions of the dataset.<n>Compared to baseline models trained on the full datasets, LFR consistently achieved lower perplexity and higher accuracy.
  arXiv  Detail & Related papers  (2024-09-10T00:59:18Z)
• InkubaLM: A small language model for low-resource African languages [9.426968756845389]
  InkubaLM is a small language model with 0.4 billion parameters. It achieves performance comparable to models with significantly larger parameter counts. It demonstrates remarkable consistency across multiple languages.
  arXiv  Detail & Related papers  (2024-08-30T05:42:31Z)
• The Languini Kitchen: Enabling Language Modelling Research at Different Scales of Compute [66.84421705029624]
  We introduce an experimental protocol that enables model comparisons based on equivalent compute, measured in accelerator hours. We pre-process an existing large, diverse, and high-quality dataset of books that surpasses existing academic benchmarks in quality, diversity, and document length. This work also provides two baseline models: a feed-forward model derived from the GPT-2 architecture and a recurrent model in the form of a novel LSTM with ten-fold throughput.
  arXiv  Detail & Related papers  (2023-09-20T10:31:17Z)
• Improving Massively Multilingual ASR With Auxiliary CTC Objectives [40.10307386370194]
  We introduce our work on improving performance on FLEURS, a 102-language open ASR benchmark. We investigate techniques inspired from recent Connectionist Temporal Classification ( CTC) studies to help the model handle the large number of languages. Our state-of-the-art systems using self-supervised models with the Conformer architecture improve over the results of prior work on FLEURS by a relative 28.4% CER.
  arXiv  Detail & Related papers  (2023-02-24T18:59:51Z)
• Mixed-Lingual Pre-training for Cross-lingual Summarization [54.4823498438831]
  Cross-lingual Summarization aims at producing a summary in the target language for an article in the source language. We propose a solution based on mixed-lingual pre-training that leverages both cross-lingual tasks like translation and monolingual tasks like masked language models. Our model achieves an improvement of 2.82 (English to Chinese) and 1.15 (Chinese to English) ROUGE-1 scores over state-of-the-art results.
  arXiv  Detail & Related papers  (2020-10-18T00:21:53Z)
• Grounded Compositional Outputs for Adaptive Language Modeling [59.02706635250856]
  A language model's vocabulary$-$typically selected before training and permanently fixed later$-$affects its size. We propose a fully compositional output embedding layer for language models. To our knowledge, the result is the first word-level language model with a size that does not depend on the training vocabulary.
  arXiv  Detail & Related papers  (2020-09-24T07:21:14Z)
• A Data Efficient End-To-End Spoken Language Understanding Architecture [22.823732899634518]
  We introduce a data efficient system which is trained end-to-end, with no additional, pre-trained external module. The proposed model achieves a reasonable size and competitive results with respect to state-of-the-art while using a small training dataset.
  arXiv  Detail & Related papers  (2020-02-14T10:24:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.