HAERAE-HUB/KMMLU-HARD
KMMLU (Korean-MMLU) We propose KMMLU, a new Korean benchmark with 35,030 expert-level multiple-choice questions across 45 subjects ranging from humanities to STEM. Unlike previous Korean benchmarks that are translated from existing English benchmarks, KMMLU is collected from original Korean exams, capturing linguistic and cultural aspects of the Korean language. We test 26 publically available and proprietary LLMs, identifying significant room for improvement. The best… See the full description on the dataset page: https://huggingface.co/datasets/HAERAE-HUB/KMMLU-HARD.
| Author / organization | HAERAE-HUB |
|---|---|
| License (as declared) | cc-by-nd-4.0 |
| Downloads | 2.1K (as reported by Hugging Face) |
| Likes | 13 |
| Created | 12 January 2024 |
| Last updated at source | 9 March 2024 |
Popularity, last 3 days
2,090 now (-1 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Recent changes
- 16 Oct 2026 Description: KMMLU (Korean-MMLU) We propose KMMLU, a new Korean benchmark with 35,030 expert-level multiple-choice questions across 45 subjects ranging from humanities to STEM. Unlike previous Korean benchmarks that are translated from existing English → KMMLU (Korean-MMLU) We propose KMMLU, a new Korean benchmark with 35,030 expert-level multiple-choice questions across 45 subjects ranging from humanities to STEM. Unlike previous Korean benchmarks that are translated from existing English
Changes detected by Silytics between refreshes of the source.
More from HAERAE-HUB
- HAERAE-HUB/KMMLUDataset
- HAERAE-HUB/HAE_RAE_BENCH_1.1Dataset
- HAERAE-HUB/HRM8KDataset
- HAERAE-HUB/KUDGEDataset
- HAERAE-HUB/K2-EvalDataset
- HAERAE-HUB/KoSimpleEvalDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.