{"schema":"https://assignee.net/schemas/benchmark-evidence-v1","schema_version":"1.0","contract_version":"benchmark-evidence-v1.0","contract_updated":"2026-06-01","schema_documentation":"https://assignee.net/schemas","changelog_url":"https://assignee.net/changelog","publisher":{"name":"Assignee Research","url":"https://assignee.net"},"html_url":"https://assignee.net/benchmarks/evidence?model=mBERT&bench=Xnli","json_url":"https://assignee.net/benchmarks/evidence.json?model=mBERT&bench=Xnli","model":"mBERT","benchmark":"Xnli","source_count":5,"source_coverage":{"record_count":5,"distinct_source_count":5,"coverage_level":"BROAD","basis":"distinct public paper URLs or titles in this evidence cluster"},"source_profile":{"source_url_count":5,"missing_source_url_count":0,"domains":["arxiv.org"],"year_min":2020,"year_max":2024,"basis":"public source URLs, source titles, and reported publication years in this evidence cluster"},"reported_range":{"min_score_pct":16.5,"max_score_pct":82.8},"spread_pp":66.3,"severity":"HIGH","entries":[{"model":"mBERT","benchmark":"Xnli","score_pct":82.8,"source_title":"From Zero to Hero: On the Limitations of Zero-Shot Cross-Lingual Transfer with Multilingual Transformers","source_url":"http://arxiv.org/abs/2005.00633v1","source_domain":"arxiv.org","year":2020},{"model":"mBERT","benchmark":"Xnli","score_pct":72.4,"source_title":"The Impact of Language Adapters in Cross-Lingual Transfer for NLU","source_url":"http://arxiv.org/abs/2402.00149v1","source_domain":"arxiv.org","year":2024},{"model":"mBERT","benchmark":"Xnli","score_pct":65.4,"source_title":"Feature Aggregation in Zero-Shot Cross-Lingual Transfer Using Multilingual BERT","source_url":"http://arxiv.org/abs/2205.08497v1","source_domain":"arxiv.org","year":2022},{"model":"mBERT","benchmark":"Xnli","score_pct":30.0,"source_title":"Bilingual Alignment Pre-Training for Zero-Shot Cross-Lingual Transfer","source_url":"http://arxiv.org/abs/2106.01732v2","source_domain":"arxiv.org","year":2021},{"model":"mBERT","benchmark":"Xnli","score_pct":16.5,"source_title":"XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization","source_url":"http://arxiv.org/abs/2003.11080v5","source_domain":"arxiv.org","year":2020}],"interpretation":"This record groups score claims extracted from papers for the same model and benchmark label. A nonzero spread means the public literature reports different values for this cluster.","limitations":["Differences are not automatically errors.","Reported values may differ because of prompts, dataset versions, evaluation protocols, scoring rules, preprocessing, fine-tuning, or reporting conventions.","Source papers remain authoritative for their own claims."]}