SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.20732 · 2026

MzansiText and MzansiLM: An Open Corpus and Decoder-Only Language Model for South African Languages

Francois Meyer, Jan Buys, Anri Lombard, Simbarashe Mawere, Temi Aina, Ethan Wolff, Sbonelo Gumede, Elan Novick

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2603.20732.json · badge: /api/ran/2603.20732.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
Anri-Lombard/sallm canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
FunctionStatusWhere it lives
count_trainable_parameters Ran
ran, honoured the contract we drafted
Anri-Lombard/sallm/src/main/sallm/utils.py
code served (permissive licence) · get_code("d1bf7ae924336923")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import nn

torch.manual_seed(0)

# Create a simple model with trainable parameters
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

# Ensure all parameters require gradients (default behavior)
for param in model.parameters():
    param.requires_grad = True

ARGS = {
    "model": model
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
to_resolved_dict Ran
ran; our contract draft was wrong, not the code
Anri-Lombard/sallm/src/main/sallm/configs/common.py
code served (permissive licence) · get_code("8c8854f8c2a9984f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from omegaconf import DictConfig, OmegaConf

# Test case 1: DictConfig input
config_dict = DictConfig({'model': 'transformer', 'lr': 0.001, 'batch_size': 32})

# Test case 2: Plain dict input
plain_dict = {'optimizer': 'adam', 'epochs': 10, 'layers': 3}

# Test case 3: DictConfig with interpolations/references that need resolving
config_with_refs = OmegaConf.create({
    'base_lr': 0.01,
    'warmup_lr': '${base_lr}',
    'model_name': 'bert',
    'nested': {
        'hidden_size': 768,
        'num_heads': 12
    }
})

ARGS = {
    'value': config_dict,
    'name': 'config'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
omegaconf — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Decoder-only language models can be adapted to diverse tasks through instruction finetuning, but the extent to which this generalizes at small scale for low-resource languages remains unclear. We focus on the languages of South Africa, where we are not aware of a publicly available decoder-only model that explicitly targets all eleven official written languages, nine of which are low-resource. We introduce MzansiText, a curated multilingual pretraining corpus with a reproducible filtering pipeline, and MzansiLM, a 125M-parameter language model trained from scratch. We evaluate MzansiLM on natural language understanding and generation using three adaptation regimes: monolingual task-specific finetuning, multilingual task-specific finetuning, and general multi-task instruction finetuning. Monolingual task-specific finetuning achieves strong performance on data-to-text generation, reaching 20.65 BLEU on isiXhosa and competing with encoder-decoder baselines over ten times larger. Multilingual task-specific finetuning benefits closely related languages on topic classification, achieving 78.5% macro-F1 on isiXhosa news classification. While MzansiLM adapts effectively to supervised NLU and NLG tasks, few-shot reasoning remains challenging at this model size, with performance near chance even for much larger decoder-only models. We release MzansiText and MzansiLM to provide a reproducible decoder-only baseline and clear guidance on adaptation strategies for South African languages at small scale.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2603.20732")
get_code_for_paper("2603.20732")
have("2603.20732")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2603.20732.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2603.20732.svg)](https://syntology.ai/paper/2603.20732)

Connect an agent — have() is free.