{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:HAEOAJW65TX2KG6LAFHAYMUNXO","short_pith_number":"pith:HAEOAJW6","schema_version":"1.0","canonical_sha256":"3808e026deecefa51bcb014e0c328dbb88d3c946928971dbc5f94fe4a6d02a11","source":{"kind":"arxiv","id":"2109.06827","version":2},"attestation_state":"computed","paper":{"title":"Types of Out-of-Distribution Texts and How to Detect Them","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"He He, Udit Arora, William Huang","submitted_at":"2021-09-14T17:12:38Z","abstract_excerpt":"Despite agreement on the importance of detecting out-of-distribution (OOD) examples, there is little consensus on the formal definition of OOD examples and how to best detect them. We categorize these examples by whether they exhibit a background shift or a semantic shift, and find that the two major approaches to OOD detection, model calibration and density estimation (language modeling for text), have distinct behavior on these types of OOD data. Across 14 pairs of in-distribution and OOD English natural language understanding datasets, we find that density estimation methods consistently be"},"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":"2109.06827","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-09-14T17:12:38Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"859ba4c40544c18339159ed5b0a14e540011093ad2eb4d46f31444f6b810af14","abstract_canon_sha256":"05ca5ae77cd2381361d2532c2dd23f9de240c4636e82e5cfe87aa234a2f62a2f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:41:04.023492Z","signature_b64":"4nUxP9v37VPw7sjX++tTUayyXn+VTA9gC4CfS5IfnH/RChSW8ScadFXl62cwUYOiMC/80157frgjAQSE2umVBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3808e026deecefa51bcb014e0c328dbb88d3c946928971dbc5f94fe4a6d02a11","last_reissued_at":"2026-07-05T03:41:04.023020Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:41:04.023020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Types of Out-of-Distribution Texts and How to Detect Them","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"He He, Udit Arora, William Huang","submitted_at":"2021-09-14T17:12:38Z","abstract_excerpt":"Despite agreement on the importance of detecting out-of-distribution (OOD) examples, there is little consensus on the formal definition of OOD examples and how to best detect them. We categorize these examples by whether they exhibit a background shift or a semantic shift, and find that the two major approaches to OOD detection, model calibration and density estimation (language modeling for text), have distinct behavior on these types of OOD data. Across 14 pairs of in-distribution and OOD English natural language understanding datasets, we find that density estimation methods consistently be"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.06827","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/2109.06827/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":"2109.06827","created_at":"2026-07-05T03:41:04.023078+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.06827v2","created_at":"2026-07-05T03:41:04.023078+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.06827","created_at":"2026-07-05T03:41:04.023078+00:00"},{"alias_kind":"pith_short_12","alias_value":"HAEOAJW65TX2","created_at":"2026-07-05T03:41:04.023078+00:00"},{"alias_kind":"pith_short_16","alias_value":"HAEOAJW65TX2KG6L","created_at":"2026-07-05T03:41:04.023078+00:00"},{"alias_kind":"pith_short_8","alias_value":"HAEOAJW6","created_at":"2026-07-05T03:41:04.023078+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.00418","citing_title":"Dual Debiasing for Noisy In-Context Learning for Text Generation","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO","json":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO.json","graph_json":"https://pith.science/api/pith-number/HAEOAJW65TX2KG6LAFHAYMUNXO/graph.json","events_json":"https://pith.science/api/pith-number/HAEOAJW65TX2KG6LAFHAYMUNXO/events.json","paper":"https://pith.science/paper/HAEOAJW6"},"agent_actions":{"view_html":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO","download_json":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO.json","view_paper":"https://pith.science/paper/HAEOAJW6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.06827&json=true","fetch_graph":"https://pith.science/api/pith-number/HAEOAJW65TX2KG6LAFHAYMUNXO/graph.json","fetch_events":"https://pith.science/api/pith-number/HAEOAJW65TX2KG6LAFHAYMUNXO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO/action/storage_attestation","attest_author":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO/action/author_attestation","sign_citation":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO/action/citation_signature","submit_replication":"https://pith.science/pith/HAEOAJW65TX2KG6LAFHAYMUNXO/action/replication_record"}},"created_at":"2026-07-05T03:41:04.023078+00:00","updated_at":"2026-07-05T03:41:04.023078+00:00"}