{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:4E5DK7Z22RSW3XAQIHKC2DO37E","short_pith_number":"pith:4E5DK7Z2","schema_version":"1.0","canonical_sha256":"e13a357f3ad4656ddc1041d42d0ddbf917867d0586850e7a2bc9580ca1eade97","source":{"kind":"arxiv","id":"2210.14975","version":1},"attestation_state":"computed","paper":{"title":"MABEL: Attenuating Gender Bias using Textual Entailment Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Christiane Fellbaum, Danqi Chen, Jacqueline He, Mengzhou Xia","submitted_at":"2022-10-26T18:36:58Z","abstract_excerpt":"Pre-trained language models encode undesirable social biases, which are further exacerbated in downstream use. To this end, we propose MABEL (a Method for Attenuating Gender Bias using Entailment Labels), an intermediate pre-training approach for mitigating gender bias in contextualized representations. Key to our approach is the use of a contrastive learning objective on counterfactually augmented, gender-balanced entailment pairs from natural language inference (NLI) datasets. We also introduce an alignment regularizer that pulls identical entailment pairs along opposite gender directions cl"},"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":"2210.14975","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-26T18:36:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f25c8a190ec14b40bc0e9adff8be81a54f546f6639be19d554af6ae3fac25591","abstract_canon_sha256":"d7aeb16675aeb21096ae3054a9515b18f11fae0418d15851761f392664547818"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:11:00.620384Z","signature_b64":"O2AkGzqjrj7CcJtd+cpoov3gGJqsZSUDRw95/Q8D+DzQzRFbuSPnM5hyo9XRsDbZK/MXnUomI7t0JThuNljyBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e13a357f3ad4656ddc1041d42d0ddbf917867d0586850e7a2bc9580ca1eade97","last_reissued_at":"2026-07-05T05:11:00.619889Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:11:00.619889Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MABEL: Attenuating Gender Bias using Textual Entailment Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Christiane Fellbaum, Danqi Chen, Jacqueline He, Mengzhou Xia","submitted_at":"2022-10-26T18:36:58Z","abstract_excerpt":"Pre-trained language models encode undesirable social biases, which are further exacerbated in downstream use. To this end, we propose MABEL (a Method for Attenuating Gender Bias using Entailment Labels), an intermediate pre-training approach for mitigating gender bias in contextualized representations. Key to our approach is the use of a contrastive learning objective on counterfactually augmented, gender-balanced entailment pairs from natural language inference (NLI) datasets. We also introduce an alignment regularizer that pulls identical entailment pairs along opposite gender directions cl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.14975","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/2210.14975/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":"2210.14975","created_at":"2026-07-05T05:11:00.619943+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.14975v1","created_at":"2026-07-05T05:11:00.619943+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.14975","created_at":"2026-07-05T05:11:00.619943+00:00"},{"alias_kind":"pith_short_12","alias_value":"4E5DK7Z22RSW","created_at":"2026-07-05T05:11:00.619943+00:00"},{"alias_kind":"pith_short_16","alias_value":"4E5DK7Z22RSW3XAQ","created_at":"2026-07-05T05:11:00.619943+00:00"},{"alias_kind":"pith_short_8","alias_value":"4E5DK7Z2","created_at":"2026-07-05T05:11:00.619943+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.10407","citing_title":"Addressing Bias in Generative AI: Challenges and Research Opportunities in Information Management","ref_index":183,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E","json":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E.json","graph_json":"https://pith.science/api/pith-number/4E5DK7Z22RSW3XAQIHKC2DO37E/graph.json","events_json":"https://pith.science/api/pith-number/4E5DK7Z22RSW3XAQIHKC2DO37E/events.json","paper":"https://pith.science/paper/4E5DK7Z2"},"agent_actions":{"view_html":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E","download_json":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E.json","view_paper":"https://pith.science/paper/4E5DK7Z2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.14975&json=true","fetch_graph":"https://pith.science/api/pith-number/4E5DK7Z22RSW3XAQIHKC2DO37E/graph.json","fetch_events":"https://pith.science/api/pith-number/4E5DK7Z22RSW3XAQIHKC2DO37E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E/action/storage_attestation","attest_author":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E/action/author_attestation","sign_citation":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E/action/citation_signature","submit_replication":"https://pith.science/pith/4E5DK7Z22RSW3XAQIHKC2DO37E/action/replication_record"}},"created_at":"2026-07-05T05:11:00.619943+00:00","updated_at":"2026-07-05T05:11:00.619943+00:00"}