Datasets:
Update MultilingualPoetryRetrieval v1.2.0: merge hard negatives into qrels (score 0–3); drop hard_negatives_* configs; refresh Configs card
7b88967 verified | cff-version: 1.2.0 | |
| message: If you use this dataset, please cite it as below. | |
| title: "Multilingual Poetry Retrieval: Multi-Aspect Graded Retrieval for Multilingual Poetry" | |
| version: "1.2.0" | |
| type: dataset | |
| license: CC-BY-NC-SA-4.0 | |
| authors: | |
| - name: "PoetryMTEB Contributors" | |
| url: "https://huggingface.co/datasets/PoetryMTEB/MultilingualPoetryRetrieval" | |
| repository-code: "https://huggingface.co/datasets/PoetryMTEB/MultilingualPoetryRetrieval" | |
| keywords: | |
| - poetry | |
| - multilingual | |
| - retrieval | |
| - mteb | |
| - poetrymteb | |
| - embedding-evaluation | |
| abstract: >- | |
| PoetryMTEB BEIR-style multi-aspect multilingual poetry retrieval benchmark | |
| covering emotion, intent, theme, and thought under MONO/CROSS settings, | |
| with graded relevance scores (0-3) and train / silver / gold splits. | |