{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:DVBXQN3HTOC66ZYODVSNLI36WI","short_pith_number":"pith:DVBXQN3H","schema_version":"1.0","canonical_sha256":"1d437837679b85ef670e1d64d5a37eb202d9258ff8be32281d50158fa84f419e","source":{"kind":"arxiv","id":"2210.10951","version":2},"attestation_state":"computed","paper":{"title":"Automatic Document Selection for Efficient Encoder Pretraining","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Benjamin Van Durme, Jo\\~ao Sedoc, Patrick Xia, Yukun Feng","submitted_at":"2022-10-20T01:45:02Z","abstract_excerpt":"Building pretrained language models is considered expensive and data-intensive, but must we increase dataset size to achieve better performance? We propose an alternative to larger training sets by automatically identifying smaller yet domain-representative subsets. We extend Cynical Data Selection, a statistical sentence scoring method that conditions on a representative target domain corpus. As an example, we treat the OntoNotes corpus as a target domain and pretrain a RoBERTa-like encoder from a cynically selected subset of the Pile. On both perplexity and across several downstream tasks in"},"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":"2210.10951","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-10-20T01:45:02Z","cross_cats_sorted":[],"title_canon_sha256":"819a79de0a73d5c92586960cb8a52d6e69b770320c13d1e169199e612b34dffd","abstract_canon_sha256":"d35d75bbd4128a08c0e6764670b36551beb50b423f86f1510ccc0913e25d189c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:10:41.193716Z","signature_b64":"304R/ewzuqtHJKiqIwAI3oYiyUs8ctueap81+CuO08itIKNGbrkG1lN2p1JTN3eyrEq9GIma23izfbFxdypCAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d437837679b85ef670e1d64d5a37eb202d9258ff8be32281d50158fa84f419e","last_reissued_at":"2026-07-05T05:10:41.193260Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:10:41.193260Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Automatic Document Selection for Efficient Encoder Pretraining","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Benjamin Van Durme, Jo\\~ao Sedoc, Patrick Xia, Yukun Feng","submitted_at":"2022-10-20T01:45:02Z","abstract_excerpt":"Building pretrained language models is considered expensive and data-intensive, but must we increase dataset size to achieve better performance? We propose an alternative to larger training sets by automatically identifying smaller yet domain-representative subsets. We extend Cynical Data Selection, a statistical sentence scoring method that conditions on a representative target domain corpus. As an example, we treat the OntoNotes corpus as a target domain and pretrain a RoBERTa-like encoder from a cynically selected subset of the Pile. On both perplexity and across several downstream tasks in"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.10951","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/2210.10951/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":"2210.10951","created_at":"2026-07-05T05:10:41.193322+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.10951v2","created_at":"2026-07-05T05:10:41.193322+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.10951","created_at":"2026-07-05T05:10:41.193322+00:00"},{"alias_kind":"pith_short_12","alias_value":"DVBXQN3HTOC6","created_at":"2026-07-05T05:10:41.193322+00:00"},{"alias_kind":"pith_short_16","alias_value":"DVBXQN3HTOC66ZYO","created_at":"2026-07-05T05:10:41.193322+00:00"},{"alias_kind":"pith_short_8","alias_value":"DVBXQN3H","created_at":"2026-07-05T05:10:41.193322+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.15553","citing_title":"Approximating Language Model Training Data from Weights","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI","json":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI.json","graph_json":"https://pith.science/api/pith-number/DVBXQN3HTOC66ZYODVSNLI36WI/graph.json","events_json":"https://pith.science/api/pith-number/DVBXQN3HTOC66ZYODVSNLI36WI/events.json","paper":"https://pith.science/paper/DVBXQN3H"},"agent_actions":{"view_html":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI","download_json":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI.json","view_paper":"https://pith.science/paper/DVBXQN3H","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.10951&json=true","fetch_graph":"https://pith.science/api/pith-number/DVBXQN3HTOC66ZYODVSNLI36WI/graph.json","fetch_events":"https://pith.science/api/pith-number/DVBXQN3HTOC66ZYODVSNLI36WI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI/action/storage_attestation","attest_author":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI/action/author_attestation","sign_citation":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI/action/citation_signature","submit_replication":"https://pith.science/pith/DVBXQN3HTOC66ZYODVSNLI36WI/action/replication_record"}},"created_at":"2026-07-05T05:10:41.193322+00:00","updated_at":"2026-07-05T05:10:41.193322+00:00"}