XL-Sum
by Hasan et al. / University of Edinburgh · free · Last verified 2026-03-17
XL-Sum is a large-scale benchmark dataset for multilingual abstractive summarization. It contains 1.35 million article-summary pairs from BBC News across 44 languages, designed to evaluate a model's ability to generate concise summaries across diverse linguistic families and writing systems.
https://github.com/csebuetnlp/xl-sum ↗B
B—Above Average
Adoption: BQuality: AFreshness: B+Citations: B+Engagement: F
Specifications
- License
- CC BY-NC-SA 4.0
- Pricing
- free
- Capabilities
- abstractive-summarization-evaluation, multilingual-nlp-benchmarking, cross-lingual-transfer-learning-assessment, low-resource-language-summarization, news-article-summarization, text-generation-evaluation, rouge-score-computation
- Integrations
- Use Cases
- [object Object], [object Object], [object Object], [object Object], [object Object]
- API Available
- No
- Evaluated Models
- mbart-large-cc25, mt5-xl, gpt-4o, llama-3-70b
- Metrics
- rouge-1, rouge-2, rouge-l
- Methodology
- Test splits per language evaluated with standard ROUGE metrics (R-1, R-2, R-L). Macro-average across all 44 languages is the primary comparison metric. Models generate abstractive summaries; no extractive oracle is used.
- Last Run
- 2026-01-05
- Tags
- summarization, multilingual, news, bbc, benchmark, dataset, nlp, text-generation, cross-lingual, abstractive-summarization, evaluation
- Added
- 2026-03-17
- Completeness
- 0.7%
Index Score
62.2Adoption
68
Quality
85
Freshness
72
Citations
72
Engagement
0