{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:7DI3UHSAISIVF7DAXIBNCI7DOP","short_pith_number":"pith:7DI3UHSA","schema_version":"1.0","canonical_sha256":"f8d1ba1e40449152fc60ba02d123e373d56ecd8503cda5e92a97955160d874c1","source":{"kind":"arxiv","id":"1908.07898","version":2},"attestation_state":"computed","paper":{"title":"Are We Modeling the Task or the Annotator? An Investigation of Annotator Bias in Natural Language Understanding Datasets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jonathan Berant, Mor Geva, Yoav Goldberg","submitted_at":"2019-08-21T14:48:01Z","abstract_excerpt":"Crowdsourcing has been the prevalent paradigm for creating natural language understanding datasets in recent years. A common crowdsourcing practice is to recruit a small number of high-quality workers, and have them massively generate examples. Having only a few workers generate the majority of examples raises concerns about data diversity, especially when workers freely generate sentences. In this paper, we perform a series of experiments showing these concerns are evident in three recent NLP datasets. We show that model performance improves when training with annotator identifiers as feature"},"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":"1908.07898","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-08-21T14:48:01Z","cross_cats_sorted":[],"title_canon_sha256":"3b14f0a9509550b53ed728a25ae1ffd325c48482f8351e8d4cfec018f51769c8","abstract_canon_sha256":"b86083d5c0354f9348d956a26deea5105997666890fa7dd55a36766e7dbae94b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:00:16.216450Z","signature_b64":"tHFd0MOsfNb1inxzURGy0E9A6ra+kinpLiahcxBLx8OF1gubqxnOYESoflb5OADwRyEgjR9jtoIyT5sTwUo1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f8d1ba1e40449152fc60ba02d123e373d56ecd8503cda5e92a97955160d874c1","last_reissued_at":"2026-07-05T00:00:16.216007Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:00:16.216007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Are We Modeling the Task or the Annotator? An Investigation of Annotator Bias in Natural Language Understanding Datasets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jonathan Berant, Mor Geva, Yoav Goldberg","submitted_at":"2019-08-21T14:48:01Z","abstract_excerpt":"Crowdsourcing has been the prevalent paradigm for creating natural language understanding datasets in recent years. A common crowdsourcing practice is to recruit a small number of high-quality workers, and have them massively generate examples. Having only a few workers generate the majority of examples raises concerns about data diversity, especially when workers freely generate sentences. In this paper, we perform a series of experiments showing these concerns are evident in three recent NLP datasets. We show that model performance improves when training with annotator identifiers as feature"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1908.07898","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/1908.07898/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":"1908.07898","created_at":"2026-07-05T00:00:16.216064+00:00"},{"alias_kind":"arxiv_version","alias_value":"1908.07898v2","created_at":"2026-07-05T00:00:16.216064+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1908.07898","created_at":"2026-07-05T00:00:16.216064+00:00"},{"alias_kind":"pith_short_12","alias_value":"7DI3UHSAISIV","created_at":"2026-07-05T00:00:16.216064+00:00"},{"alias_kind":"pith_short_16","alias_value":"7DI3UHSAISIVF7DA","created_at":"2026-07-05T00:00:16.216064+00:00"},{"alias_kind":"pith_short_8","alias_value":"7DI3UHSA","created_at":"2026-07-05T00:00:16.216064+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.03259","citing_title":"From Pre-trained Models to Large Language Models: A Comprehensive Survey of AI-Driven Psychological Computing","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP","json":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP.json","graph_json":"https://pith.science/api/pith-number/7DI3UHSAISIVF7DAXIBNCI7DOP/graph.json","events_json":"https://pith.science/api/pith-number/7DI3UHSAISIVF7DAXIBNCI7DOP/events.json","paper":"https://pith.science/paper/7DI3UHSA"},"agent_actions":{"view_html":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP","download_json":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP.json","view_paper":"https://pith.science/paper/7DI3UHSA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1908.07898&json=true","fetch_graph":"https://pith.science/api/pith-number/7DI3UHSAISIVF7DAXIBNCI7DOP/graph.json","fetch_events":"https://pith.science/api/pith-number/7DI3UHSAISIVF7DAXIBNCI7DOP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP/action/storage_attestation","attest_author":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP/action/author_attestation","sign_citation":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP/action/citation_signature","submit_replication":"https://pith.science/pith/7DI3UHSAISIVF7DAXIBNCI7DOP/action/replication_record"}},"created_at":"2026-07-05T00:00:16.216064+00:00","updated_at":"2026-07-05T00:00:16.216064+00:00"}