{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:GRYDKMS3454UPKJ5WM6RWQ4E75","short_pith_number":"pith:GRYDKMS3","schema_version":"1.0","canonical_sha256":"347035325be77947a93db33d1b4384ff6795053d88076a5735c85aa95a349cb2","source":{"kind":"arxiv","id":"1906.01502","version":1},"attestation_state":"computed","paper":{"title":"How multilingual is Multilingual BERT?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dan Garrette, Eva Schlinger, Telmo Pires","submitted_at":"2019-06-04T15:12:47Z","abstract_excerpt":"In this paper, we show that Multilingual BERT (M-BERT), released by Devlin et al. (2018) as a single language model pre-trained from monolingual corpora in 104 languages, is surprisingly good at zero-shot cross-lingual model transfer, in which task-specific annotations in one language are used to fine-tune the model for evaluation in another language. To understand why, we present a large number of probing experiments, showing that transfer is possible even to languages in different scripts, that transfer works best between typologically similar languages, that monolingual corpora can train mo"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"1906.01502","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-06-04T15:12:47Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"eca98acadc2d6e7de87e4caca471d8c823e7e8337d376e3c4907ad68da77cdf0","abstract_canon_sha256":"be4de2bb58702cfe7b3dbddbb4687efa0d5095f8339d9b2f22bdedaa71e595c0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:44:16.360487Z","signature_b64":"EBBxGsukn0OdjmjAHW6ou4BWbRDUIK/ETTKnT88b16uXL5JKa++GeDzLbg7keYM6K/Nn12RdpovNygkm0IvcDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"347035325be77947a93db33d1b4384ff6795053d88076a5735c85aa95a349cb2","last_reissued_at":"2026-05-17T23:44:16.359883Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:44:16.359883Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How multilingual is Multilingual BERT?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Dan Garrette, Eva Schlinger, Telmo Pires","submitted_at":"2019-06-04T15:12:47Z","abstract_excerpt":"In this paper, we show that Multilingual BERT (M-BERT), released by Devlin et al. (2018) as a single language model pre-trained from monolingual corpora in 104 languages, is surprisingly good at zero-shot cross-lingual model transfer, in which task-specific annotations in one language are used to fine-tune the model for evaluation in another language. To understand why, we present a large number of probing experiments, showing that transfer is possible even to languages in different scripts, that transfer works best between typologically similar languages, that monolingual corpora can train mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1906.01502","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"1906.01502","created_at":"2026-05-17T23:44:16.359974+00:00"},{"alias_kind":"arxiv_version","alias_value":"1906.01502v1","created_at":"2026-05-17T23:44:16.359974+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1906.01502","created_at":"2026-05-17T23:44:16.359974+00:00"},{"alias_kind":"pith_short_12","alias_value":"GRYDKMS3454U","created_at":"2026-05-18T12:33:18.533446+00:00"},{"alias_kind":"pith_short_16","alias_value":"GRYDKMS3454UPKJ5","created_at":"2026-05-18T12:33:18.533446+00:00"},{"alias_kind":"pith_short_8","alias_value":"GRYDKMS3","created_at":"2026-05-18T12:33:18.533446+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"1907.02052","citing_title":"Patent Claim Generation by Fine-Tuning OpenAI GPT-2","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14404","citing_title":"Knowledge Beyond Language: Bridging the Gap in Multilingual Machine Unlearning Evaluation","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2002.08155","citing_title":"CodeBERT: A Pre-Trained Model for Programming and Natural Languages","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2402.03216","citing_title":"M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10627","citing_title":"Computational Lesions in Multilingual Language Models Separate Shared and Language-specific Brain Alignment","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75","json":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75.json","graph_json":"https://pith.science/api/pith-number/GRYDKMS3454UPKJ5WM6RWQ4E75/graph.json","events_json":"https://pith.science/api/pith-number/GRYDKMS3454UPKJ5WM6RWQ4E75/events.json","paper":"https://pith.science/paper/GRYDKMS3"},"agent_actions":{"view_html":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75","download_json":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75.json","view_paper":"https://pith.science/paper/GRYDKMS3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1906.01502&json=true","fetch_graph":"https://pith.science/api/pith-number/GRYDKMS3454UPKJ5WM6RWQ4E75/graph.json","fetch_events":"https://pith.science/api/pith-number/GRYDKMS3454UPKJ5WM6RWQ4E75/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75/action/storage_attestation","attest_author":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75/action/author_attestation","sign_citation":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75/action/citation_signature","submit_replication":"https://pith.science/pith/GRYDKMS3454UPKJ5WM6RWQ4E75/action/replication_record"}},"created_at":"2026-05-17T23:44:16.359974+00:00","updated_at":"2026-05-17T23:44:16.359974+00:00"}