{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:DA34JOLBERMITB3YUAQ63E7TEQ","short_pith_number":"pith:DA34JOLB","schema_version":"1.0","canonical_sha256":"1837c4b9612458898778a021ed93f32417cb1bd51ceb24099ccfdb8141d302d4","source":{"kind":"arxiv","id":"2010.14534","version":1},"attestation_state":"computed","paper":{"title":"Unmasking Contextual Stereotypes: Measuring and Mitigating BERT's Gender Bias","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Albert Gatt, Malvina Nissim, Marion Bartl","submitted_at":"2020-10-27T18:06:09Z","abstract_excerpt":"Contextualized word embeddings have been replacing standard embeddings as the representational knowledge source of choice in NLP systems. Since a variety of biases have previously been found in standard word embeddings, it is crucial to assess biases encoded in their replacements as well. Focusing on BERT (Devlin et al., 2018), we measure gender bias by studying associations between gender-denoting target words and names of professions in English and German, comparing the findings with real-world workforce statistics. We mitigate bias by fine-tuning BERT on the GAP corpus (Webster et al., 2018"},"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":"2010.14534","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2020-10-27T18:06:09Z","cross_cats_sorted":[],"title_canon_sha256":"cdee90f9b75b2173139e4136014b4494b9b16e1070d0804095a7f76db6985167","abstract_canon_sha256":"8889052fa7af2fab82280d6218a9ea4d9450931da44e37484f8d3220a2efcb54"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:47:16.193729Z","signature_b64":"qs1AwnBaKLOqk3Gc7jawpxsyMn3IwQbs7517+zthKMatIqNC+6crYyTHXGmzFSrucyEbSuTb9X+kXwZ4KNozCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1837c4b9612458898778a021ed93f32417cb1bd51ceb24099ccfdb8141d302d4","last_reissued_at":"2026-07-05T01:47:16.193363Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:47:16.193363Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unmasking Contextual Stereotypes: Measuring and Mitigating BERT's Gender Bias","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Albert Gatt, Malvina Nissim, Marion Bartl","submitted_at":"2020-10-27T18:06:09Z","abstract_excerpt":"Contextualized word embeddings have been replacing standard embeddings as the representational knowledge source of choice in NLP systems. Since a variety of biases have previously been found in standard word embeddings, it is crucial to assess biases encoded in their replacements as well. Focusing on BERT (Devlin et al., 2018), we measure gender bias by studying associations between gender-denoting target words and names of professions in English and German, comparing the findings with real-world workforce statistics. We mitigate bias by fine-tuning BERT on the GAP corpus (Webster et al., 2018"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.14534","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2010.14534/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"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":"2010.14534","created_at":"2026-07-05T01:47:16.193417+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.14534v1","created_at":"2026-07-05T01:47:16.193417+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.14534","created_at":"2026-07-05T01:47:16.193417+00:00"},{"alias_kind":"pith_short_12","alias_value":"DA34JOLBERMI","created_at":"2026-07-05T01:47:16.193417+00:00"},{"alias_kind":"pith_short_16","alias_value":"DA34JOLBERMITB3Y","created_at":"2026-07-05T01:47:16.193417+00:00"},{"alias_kind":"pith_short_8","alias_value":"DA34JOLB","created_at":"2026-07-05T01:47:16.193417+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2408.12935","citing_title":"AI Safety Landscape for Large Language Models: Taxonomy, State-of-the-art, and Future Directions","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ","json":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ.json","graph_json":"https://pith.science/api/pith-number/DA34JOLBERMITB3YUAQ63E7TEQ/graph.json","events_json":"https://pith.science/api/pith-number/DA34JOLBERMITB3YUAQ63E7TEQ/events.json","paper":"https://pith.science/paper/DA34JOLB"},"agent_actions":{"view_html":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ","download_json":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ.json","view_paper":"https://pith.science/paper/DA34JOLB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.14534&json=true","fetch_graph":"https://pith.science/api/pith-number/DA34JOLBERMITB3YUAQ63E7TEQ/graph.json","fetch_events":"https://pith.science/api/pith-number/DA34JOLBERMITB3YUAQ63E7TEQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ/action/storage_attestation","attest_author":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ/action/author_attestation","sign_citation":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ/action/citation_signature","submit_replication":"https://pith.science/pith/DA34JOLBERMITB3YUAQ63E7TEQ/action/replication_record"}},"created_at":"2026-07-05T01:47:16.193417+00:00","updated_at":"2026-07-05T01:47:16.193417+00:00"}