01
Source
Curriculum-aligned Swahili language items (grammar, comprehension, vocabulary, literature), not maths or science taught in Swahili.
NILEAGI-SUB · Available
SUB-MCQ EDUCATION
2,569 expert-reviewed Swahili multiple-choice items from the Swahili-subject curriculum (Standard 3-7 and Form 2-4). First public release under NILEAGI-SUB; later sectors follow the same protocol.
2,569
Validated items
8
Grade bands
22.4%
Chance baseline
About this release
A model can sound fluent in Swahili and still fail on the vocabulary and reasoning in a Form 4 exam. This set gives a grade-stratified school signal for compact open models, scored with exact-match multiple choice under a controlled protocol. Featured in the Cohere Labs Open Science Community.
Items by grade
| Grade band | Items | Share |
|---|---|---|
| Standard 3 | 66 | 2.6% |
| Standard 4 | 116 | 4.5% |
| Standard 5 | 274 | 10.7% |
| Standard 6 | 651 | 25.3% |
| Standard 7 | 517 | 20.1% |
| Form 2 | 399 | 15.5% |
| Form 3 | 125 | 4.9% |
| Form 4 | 421 | 16.4% |
| Total | 2,569 | 100% |
How this data is built
02
Validate
Expert review, format checks, and grade-band assignment produce a clean multiple-choice set ready for controlled evaluation.
03
Evaluate
Open models run under identical prompts and deterministic decoding; accuracy is exact-match against the gold letter.
04
Release
Technical report, public story, dataset access paths, and benchmark pages so others can reproduce and extend the work.