{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:D4FUX5SCFUXDHODEWDN3JAQUUR","short_pith_number":"pith:D4FUX5SC","canonical_record":{"source":{"id":"2010.06032","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T21:15:29Z","cross_cats_sorted":[],"title_canon_sha256":"1f08eadc734784e7ceda865f116a1dbeb31623d43bd5465d38b3dbd8ec297edd","abstract_canon_sha256":"64e922f40c5818b524f5882bc845569e5730b9684470c363b1cf108b122d22f6"},"schema_version":"1.0"},"canonical_sha256":"1f0b4bf6422d2e33b864b0dbb48214a446151414de46c90ddac34bfbe99f4510","source":{"kind":"arxiv","id":"2010.06032","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2010.06032","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"arxiv_version","alias_value":"2010.06032v2","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.06032","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_12","alias_value":"D4FUX5SCFUXD","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_16","alias_value":"D4FUX5SCFUXDHODE","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_8","alias_value":"D4FUX5SC","created_at":"2026-07-05T02:19:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:D4FUX5SCFUXDHODEWDN3JAQUUR","target":"record","payload":{"canonical_record":{"source":{"id":"2010.06032","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T21:15:29Z","cross_cats_sorted":[],"title_canon_sha256":"1f08eadc734784e7ceda865f116a1dbeb31623d43bd5465d38b3dbd8ec297edd","abstract_canon_sha256":"64e922f40c5818b524f5882bc845569e5730b9684470c363b1cf108b122d22f6"},"schema_version":"1.0"},"canonical_sha256":"1f0b4bf6422d2e33b864b0dbb48214a446151414de46c90ddac34bfbe99f4510","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:19:56.276628Z","signature_b64":"Era1fSoa/fgIwlD9+96BSKfbUGKXZ90CLQhu1sgF3aAfjhHIJLAwI28BSHNd7Ox61CXZQDBYxXOODyXCAKHVCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f0b4bf6422d2e33b864b0dbb48214a446151414de46c90ddac34bfbe99f4510","last_reissued_at":"2026-07-05T02:19:56.276139Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:19:56.276139Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2010.06032","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:19:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rlqj0QQow+hJNRmPZIJwInOByao3jIXXJRIestkZ/yxDakj0v8oGFXQpl6ycylf5Llhnk3LHIPVOsKeMVDxtDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T10:59:18.342830Z"},"content_sha256":"c8b1371cf45c296fb0a76aa9d2fe108e1361f4b3c392979a5b656780f03f6653","schema_version":"1.0","event_id":"sha256:c8b1371cf45c296fb0a76aa9d2fe108e1361f4b3c392979a5b656780f03f6653"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:D4FUX5SCFUXDHODEWDN3JAQUUR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Measuring and Reducing Gendered Correlations in Pre-trained Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alex Beutel, Ed Chi, Ellie Pavlick, Emily Pitler, Ian Tenney, Jilin Chen, Kellie Webster, Slav Petrov, Xuezhi Wang","submitted_at":"2020-10-12T21:15:29Z","abstract_excerpt":"Pre-trained models have revolutionized natural language understanding. However, researchers have found they can encode artifacts undesired in many applications, such as professions correlating with one gender more than another. We explore such gendered correlations as a case study for how to address unintended correlations in pre-trained models. We define metrics and reveal that it is possible for models with similar accuracy to encode correlations at very different rates. We show how measured correlations can be reduced with general-purpose techniques, and highlight the trade offs different s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.06032","kind":"arxiv","version":2},"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/2010.06032/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:19:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7mAJN5KoZYYTb15NwxYZcRIeaMBBZZOKcVT2FeJWHNxq87rNMzm/qMTgRLMu9VoBqFhKbb2tUXUs1HF6M72oCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T10:59:18.343206Z"},"content_sha256":"776f087207e0b19ce0e2b124f57616a7eb52e0bb9dc033f9b9e89d182af396a6","schema_version":"1.0","event_id":"sha256:776f087207e0b19ce0e2b124f57616a7eb52e0bb9dc033f9b9e89d182af396a6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/bundle.json","state_url":"https://pith.science/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-05T10:59:18Z","links":{"resolver":"https://pith.science/pith/D4FUX5SCFUXDHODEWDN3JAQUUR","bundle":"https://pith.science/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/bundle.json","state":"https://pith.science/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/D4FUX5SCFUXDHODEWDN3JAQUUR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:D4FUX5SCFUXDHODEWDN3JAQUUR","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"64e922f40c5818b524f5882bc845569e5730b9684470c363b1cf108b122d22f6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T21:15:29Z","title_canon_sha256":"1f08eadc734784e7ceda865f116a1dbeb31623d43bd5465d38b3dbd8ec297edd"},"schema_version":"1.0","source":{"id":"2010.06032","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2010.06032","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"arxiv_version","alias_value":"2010.06032v2","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.06032","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_12","alias_value":"D4FUX5SCFUXD","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_16","alias_value":"D4FUX5SCFUXDHODE","created_at":"2026-07-05T02:19:56Z"},{"alias_kind":"pith_short_8","alias_value":"D4FUX5SC","created_at":"2026-07-05T02:19:56Z"}],"graph_snapshots":[{"event_id":"sha256:776f087207e0b19ce0e2b124f57616a7eb52e0bb9dc033f9b9e89d182af396a6","target":"graph","created_at":"2026-07-05T02:19:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2010.06032/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pre-trained models have revolutionized natural language understanding. However, researchers have found they can encode artifacts undesired in many applications, such as professions correlating with one gender more than another. We explore such gendered correlations as a case study for how to address unintended correlations in pre-trained models. We define metrics and reveal that it is possible for models with similar accuracy to encode correlations at very different rates. We show how measured correlations can be reduced with general-purpose techniques, and highlight the trade offs different s","authors_text":"Alex Beutel, Ed Chi, Ellie Pavlick, Emily Pitler, Ian Tenney, Jilin Chen, Kellie Webster, Slav Petrov, Xuezhi Wang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T21:15:29Z","title":"Measuring and Reducing Gendered Correlations in Pre-trained Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.06032","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c8b1371cf45c296fb0a76aa9d2fe108e1361f4b3c392979a5b656780f03f6653","target":"record","created_at":"2026-07-05T02:19:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"64e922f40c5818b524f5882bc845569e5730b9684470c363b1cf108b122d22f6","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-10-12T21:15:29Z","title_canon_sha256":"1f08eadc734784e7ceda865f116a1dbeb31623d43bd5465d38b3dbd8ec297edd"},"schema_version":"1.0","source":{"id":"2010.06032","kind":"arxiv","version":2}},"canonical_sha256":"1f0b4bf6422d2e33b864b0dbb48214a446151414de46c90ddac34bfbe99f4510","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1f0b4bf6422d2e33b864b0dbb48214a446151414de46c90ddac34bfbe99f4510","first_computed_at":"2026-07-05T02:19:56.276139Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:19:56.276139Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Era1fSoa/fgIwlD9+96BSKfbUGKXZ90CLQhu1sgF3aAfjhHIJLAwI28BSHNd7Ox61CXZQDBYxXOODyXCAKHVCA==","signature_status":"signed_v1","signed_at":"2026-07-05T02:19:56.276628Z","signed_message":"canonical_sha256_bytes"},"source_id":"2010.06032","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c8b1371cf45c296fb0a76aa9d2fe108e1361f4b3c392979a5b656780f03f6653","sha256:776f087207e0b19ce0e2b124f57616a7eb52e0bb9dc033f9b9e89d182af396a6"],"state_sha256":"7683b9c109b84fc21ec2237e1e69158956c3aadcdd8d7c46bf06271af548e3b4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FoJ+d+ZVrGt5qbL4r2zbVgV4vSTZX+u3vv4OOkwDOrPY6fnBVhpuaDcXCSGvlxzX1jOqxs9dYgLJmlnulUGiBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-05T10:59:18.345438Z","bundle_sha256":"95ce7bbb927eee73fc0ac0e18f51b1c8967ff60a555656764ddff174780d607b"}}