{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KR4H55MISIAXZOCLH2WLEEVWZC","short_pith_number":"pith:KR4H55MI","schema_version":"1.0","canonical_sha256":"54787ef58892017cb84b3eacb212b6c8b1ff220e18c08176bf073a251a6be7ce","source":{"kind":"arxiv","id":"2605.21070","version":1},"attestation_state":"computed","paper":{"title":"Towards Understanding Self-Pretraining for Sequence Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Antonio Orvieto, Loredana Zollo, Omar Coser, Paolo Soda","submitted_at":"2026-05-20T11:56:15Z","abstract_excerpt":"Amos et al. (2024) showed that the accuracy of Transformer models in sequence classification can be significantly improved by first pretraining with a masked token prediction objective without external data or augmentation, a procedure referred to as self-pretraining (SPT). While the primary objective of Amos et al. (2024) was to showcase that Transformers can achieve strong performance on the Long-Range Arena (LRA), their pipeline raises more fundamental questions: How does SPT drive optimization to better solutions? Why can standard supervised training fail in Transformers? To better underst"},"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":"2605.21070","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T11:56:15Z","cross_cats_sorted":[],"title_canon_sha256":"00a7b9858dd812e94bc1e906508cd168da88fcb890ff5bf6afb80017b9d5bde4","abstract_canon_sha256":"85f1a4dacb5c8fe274dd389d8560bcc2acecd8ffd2a0f0c475fa8fe4662db61e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:35.123433Z","signature_b64":"qDqL3dYfzWQ+nOJnWzKWTVRtMacsilPcdPPqdK+U45oGF2VvAhFtYdH3orYWgOFRFIf+hgztydWW1ChP+SeGCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"54787ef58892017cb84b3eacb212b6c8b1ff220e18c08176bf073a251a6be7ce","last_reissued_at":"2026-05-21T01:05:35.122682Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:35.122682Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Understanding Self-Pretraining for Sequence Classification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Antonio Orvieto, Loredana Zollo, Omar Coser, Paolo Soda","submitted_at":"2026-05-20T11:56:15Z","abstract_excerpt":"Amos et al. (2024) showed that the accuracy of Transformer models in sequence classification can be significantly improved by first pretraining with a masked token prediction objective without external data or augmentation, a procedure referred to as self-pretraining (SPT). While the primary objective of Amos et al. (2024) was to showcase that Transformers can achieve strong performance on the Long-Range Arena (LRA), their pipeline raises more fundamental questions: How does SPT drive optimization to better solutions? Why can standard supervised training fail in Transformers? To better underst"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21070","kind":"arxiv","version":1},"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/2605.21070/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":"2605.21070","created_at":"2026-05-21T01:05:35.122797+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21070v1","created_at":"2026-05-21T01:05:35.122797+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21070","created_at":"2026-05-21T01:05:35.122797+00:00"},{"alias_kind":"pith_short_12","alias_value":"KR4H55MISIAX","created_at":"2026-05-21T01:05:35.122797+00:00"},{"alias_kind":"pith_short_16","alias_value":"KR4H55MISIAXZOCL","created_at":"2026-05-21T01:05:35.122797+00:00"},{"alias_kind":"pith_short_8","alias_value":"KR4H55MI","created_at":"2026-05-21T01:05:35.122797+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC","json":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC.json","graph_json":"https://pith.science/api/pith-number/KR4H55MISIAXZOCLH2WLEEVWZC/graph.json","events_json":"https://pith.science/api/pith-number/KR4H55MISIAXZOCLH2WLEEVWZC/events.json","paper":"https://pith.science/paper/KR4H55MI"},"agent_actions":{"view_html":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC","download_json":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC.json","view_paper":"https://pith.science/paper/KR4H55MI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21070&json=true","fetch_graph":"https://pith.science/api/pith-number/KR4H55MISIAXZOCLH2WLEEVWZC/graph.json","fetch_events":"https://pith.science/api/pith-number/KR4H55MISIAXZOCLH2WLEEVWZC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC/action/storage_attestation","attest_author":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC/action/author_attestation","sign_citation":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC/action/citation_signature","submit_replication":"https://pith.science/pith/KR4H55MISIAXZOCLH2WLEEVWZC/action/replication_record"}},"created_at":"2026-05-21T01:05:35.122797+00:00","updated_at":"2026-05-21T01:05:35.122797+00:00"}