--- license: cc-by-4.0 language: - sr multilinguality: monolingual pretty_name: BalkanBench SuperGLUE - Serbian size_categories: - 10K **Part of [BalkanBench](https://balkanbench.com)** - the open, reproducible > benchmark for language models across Serbian, Croatian, Montenegrin, and > Bosnian (BCMS). Live leaderboard at . > Background and motivation: [Release of BalkanBench - the vision behind it > (Medium, 2026-04-27)](https://medium.com/@permitt/release-of-balkanbench-vision-behind-it-fd1ba73be411). This is the **Serbian SuperGLUE** track of BalkanBench v0.1. Serbian is the official frozen track: the leaderboard's ranked average is computed over **6 ranked tasks**, with 2 diagnostic tasks reported separately. Croatian and Montenegrin previews live in sibling repos ([superglue-hr](https://huggingface.co/datasets/permitt/superglue-hr) / [superglue-mne](https://huggingface.co/datasets/permitt/superglue-mne)). ## What's inside 7 task configs are published here in their full train + validation + public test form (train + validation labelled here; the **test split lives only in the gated sibling repo**, see below): | Config | What it tests | Train | Validation | Test | |---------|----------------------------------------------------------|------:|-----------:|------:| | boolq | Yes/no question answering over passages | 9,427 | 3,270 | 3,245 | | cb | 3-way textual entailment (entail / contradict / neutral) | 250 | 56 | 250 | | copa | Causal reasoning between two alternatives | 400 | 100 | 500 | | multirc | Multi-sentence reading comprehension (multiple correct) | 27,243| 4,848 | 9,693 | | record | Commonsense cloze over news articles | 5,607 | 1,869 | 1,869 | | rte | Binary textual entailment | 2,490 | 277 | 3,000 | | wsc | Coreference resolution requiring world knowledge | 554 | 104 | 146 | | **TOTAL** | | **45,971** | **10,524** | **18,703** | For the **6 ranked v0.1 tasks** (BoolQ, CB, COPA, RTE, MultiRC, WSC) the total is **65,853 items** across all splits; ReCoRD ships in this repo for community use but is not part of the v0.1 ranked average. The **diagnostic tasks** AX-b (1,104) and AX-g (356) are Serbian-only and ship in the gated [superglue-sr-private](https://huggingface.co/datasets/permitt/superglue-sr-private) sibling repo with their test labels. ## Test split is gated The `test` split for every config lives **only in the gated sibling repo** (both inputs and labels): [`permitt/superglue-sr-private`](https://huggingface.co/datasets/permitt/superglue-sr-private), and are accessed only by the official scoring pipeline. This preserves leaderboard integrity: nobody (including model authors) can train against test labels, and `train` + `validation` remain a fully open and labeled playground. To get a number on the leaderboard you generate predictions on the public test inputs with `balkanbench predict`, then submit the `predictions.jsonl` through the [submission flow](https://balkanbench.com). ## Loading ```python from datasets import load_dataset # A specific task boolq = load_dataset("permitt/superglue-sr", "boolq") print(boolq["train"][0]) # All ranked tasks at once for cfg in ["boolq", "cb", "copa", "rte", "multirc", "wsc"]: ds = load_dataset("permitt/superglue-sr", cfg) print(cfg, {sp: len(ds[sp]) for sp in ds}) ``` Pin a revision for reproducible runs: ```python load_dataset("permitt/superglue-sr", "boolq", revision="v0.1.0-data") ``` ## Schema Each example carries an integer row id `idx` and the task-specific input fields plus a `label` column. For example, `cb`: ```python { "idx": 0, "premise": "Bio je to složen jezik. ...", "hypothesis": "Jezik je ogoljen.", "label": 0, # 0=entailment, 1=contradiction, 2=neutral } ``` Per-task field lists are documented in the BalkanBench source repo. ## Methodology - **Translation**: original SuperGLUE English -> Serbian via translation + human verification by native speakers. - **Frozen splits**: train/validation/test row counts and IDs are pinned at tag `v0.1.0-data`; reruns of an evaluation against this tag will see the exact same data. - **Test labels stay private**: see the gated sibling repo above. ## Citation If you use this dataset, please cite the Serbian SuperGLUE paper (Perović & Mihajlov, LoResLM 2026), the original SuperGLUE paper, and the BalkanBench release: ```bibtex @inproceedings{perovic-mihajlov-2026-serbian, title = "{S}erbian {S}uper{GLUE}: Towards an Evaluation Benchmark for {S}outh {S}lavic Language Models", author = "Perovic, Mitar and Mihajlov, Teodora", editor = "Hettiarachchi, Hansi and Ranasinghe, Tharindu and Plum, Alistair and Rayson, Paul and Mitkov, Ruslan and Gaber, Mohamed and Premasiri, Damith and Tan, Fiona Anting and Uyangodage, Lasitha", booktitle = "Proceedings of the Second Workshop on Language Models for Low-Resource Languages (LoResLM 2026)", month = mar, year = "2026", address = "Rabat, Morocco", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2026.loreslm-1.30/", doi = "10.18653/v1/2026.loreslm-1.30", pages = "347--361", ISBN = "979-8-89176-377-7" } @inproceedings{wang2019superglue, title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems}, author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R.}, booktitle={NeurIPS}, year={2019} } @misc{balkanbench2026superglue_sr, title={BalkanBench SuperGLUE-SR: Serbian SuperGLUE for the BCMS evaluation suite}, author={Perović, Mitar and contributors}, year={2026}, howpublished={\url{https://huggingface.co/datasets/permitt/superglue-sr}}, note={Part of BalkanBench v0.1, \url{https://balkanbench.com}} } ``` ## License CC-BY-4.0 (matches the original SuperGLUE source license). Translations and human verification are released under the same terms. ## Sponsor Compute for the official v0.1 evaluation is sponsored by **[Recrewty](https://recrewty.com)**. ## Links - Live leaderboard: - Project home: - Background: [Release of BalkanBench - the vision behind it (Medium)](https://medium.com/@permitt/release-of-balkanbench-vision-behind-it-fd1ba73be411) - Source code: - Sibling tracks: [SuperGLUE-HR (Croatian)](https://huggingface.co/datasets/permitt/superglue-hr) · [SuperGLUE-MNE (Montenegrin)](https://huggingface.co/datasets/permitt/superglue-mne) - Private (gated) labels: