{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:B5PKKP3KF5BPPBDTHIZ5O2CCYR","short_pith_number":"pith:B5PKKP3K","schema_version":"1.0","canonical_sha256":"0f5ea53f6a2f42f784733a33d76842c46fa2e85f95e9fcd4a802fd86e29355cf","source":{"kind":"arxiv","id":"2104.01767","version":3},"attestation_state":"computed","paper":{"title":"WhiteningBERT: An Easy Unsupervised Sentence Embedding Approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daxin Jiang, Duyu Tang, Junjie Huang, Linjun Shou, Ming Gong, Nan Duan, Shuai Lu, Wanjun Zhong","submitted_at":"2021-04-05T04:30:28Z","abstract_excerpt":"Producing the embedding of a sentence in an unsupervised way is valuable to natural language matching and retrieval problems in practice. In this work, we conduct a thorough examination of pretrained model based unsupervised sentence embeddings. We study on four pretrained models and conduct massive experiments on seven datasets regarding sentence semantics. We have there main findings. First, averaging all tokens is better than only using [CLS] vector. Second, combining both top andbottom layers is better than only using top layers. Lastly, an easy whitening-based vector normalization strateg"},"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":"2104.01767","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-04-05T04:30:28Z","cross_cats_sorted":[],"title_canon_sha256":"777feed218cdf03042cc26070d07ef7db9fa975919a11ee76f3bef85038dc96c","abstract_canon_sha256":"8505255dbf37fe37992f0dadcb36d1f74f5fd4e9a9d87056c2fe6d463d614dc2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:30:32.789855Z","signature_b64":"05FsDdn4Ly47wx9jbPwsvo8gnXWVzzf1OeekabUmfAEV6phqazbcFqjsztuxffaoD213DfDqzKPMLYI6wondBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0f5ea53f6a2f42f784733a33d76842c46fa2e85f95e9fcd4a802fd86e29355cf","last_reissued_at":"2026-07-05T02:30:32.789323Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:30:32.789323Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"WhiteningBERT: An Easy Unsupervised Sentence Embedding Approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daxin Jiang, Duyu Tang, Junjie Huang, Linjun Shou, Ming Gong, Nan Duan, Shuai Lu, Wanjun Zhong","submitted_at":"2021-04-05T04:30:28Z","abstract_excerpt":"Producing the embedding of a sentence in an unsupervised way is valuable to natural language matching and retrieval problems in practice. In this work, we conduct a thorough examination of pretrained model based unsupervised sentence embeddings. We study on four pretrained models and conduct massive experiments on seven datasets regarding sentence semantics. We have there main findings. First, averaging all tokens is better than only using [CLS] vector. Second, combining both top andbottom layers is better than only using top layers. Lastly, an easy whitening-based vector normalization strateg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.01767","kind":"arxiv","version":3},"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/2104.01767/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":"2104.01767","created_at":"2026-07-05T02:30:32.789393+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.01767v3","created_at":"2026-07-05T02:30:32.789393+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.01767","created_at":"2026-07-05T02:30:32.789393+00:00"},{"alias_kind":"pith_short_12","alias_value":"B5PKKP3KF5BP","created_at":"2026-07-05T02:30:32.789393+00:00"},{"alias_kind":"pith_short_16","alias_value":"B5PKKP3KF5BPPBDT","created_at":"2026-07-05T02:30:32.789393+00:00"},{"alias_kind":"pith_short_8","alias_value":"B5PKKP3K","created_at":"2026-07-05T02:30:32.789393+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2104.08821","citing_title":"SimCSE: Simple Contrastive Learning of Sentence Embeddings","ref_index":119,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR","json":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR.json","graph_json":"https://pith.science/api/pith-number/B5PKKP3KF5BPPBDTHIZ5O2CCYR/graph.json","events_json":"https://pith.science/api/pith-number/B5PKKP3KF5BPPBDTHIZ5O2CCYR/events.json","paper":"https://pith.science/paper/B5PKKP3K"},"agent_actions":{"view_html":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR","download_json":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR.json","view_paper":"https://pith.science/paper/B5PKKP3K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.01767&json=true","fetch_graph":"https://pith.science/api/pith-number/B5PKKP3KF5BPPBDTHIZ5O2CCYR/graph.json","fetch_events":"https://pith.science/api/pith-number/B5PKKP3KF5BPPBDTHIZ5O2CCYR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR/action/storage_attestation","attest_author":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR/action/author_attestation","sign_citation":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR/action/citation_signature","submit_replication":"https://pith.science/pith/B5PKKP3KF5BPPBDTHIZ5O2CCYR/action/replication_record"}},"created_at":"2026-07-05T02:30:32.789393+00:00","updated_at":"2026-07-05T02:30:32.789393+00:00"}