{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XFH44COMPMI2AUXC5UUGGPXMSI","short_pith_number":"pith:XFH44COM","schema_version":"1.0","canonical_sha256":"b94fce09cc7b11a052e2ed28633eec92140737f9f5c8312a050fbe837e78bf8e","source":{"kind":"arxiv","id":"2411.07175","version":3},"attestation_state":"computed","paper":{"title":"Continual Memorization of Factoids in Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adithya Bhaskar, Dan Friedman, Danqi Chen, Howard Chen, Jiayi Geng","submitted_at":"2024-11-11T17:56:15Z","abstract_excerpt":"As new knowledge rapidly accumulates, language models (LMs) with pretrained knowledge quickly become obsolete. A common approach to updating LMs is fine-tuning them directly on new knowledge. However, recent studies have shown that fine-tuning for memorization may be ineffective in storing knowledge or may exacerbate hallucinations. In this work, we introduce a setting we call continual memorization, where a model must memorize and retain a set of factoids through multiple stages of fine-tuning on subsequent datasets. We characterized the forgetting patterns through extensive experiments and s"},"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":"2411.07175","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-11T17:56:15Z","cross_cats_sorted":[],"title_canon_sha256":"e1a0e422c7394d69e15520a4565258db394f0c38d5339ee057c71083b3829292","abstract_canon_sha256":"8330656d5a5ffcf3840f15017f0a63a0f7161144e911e7fb14de5bdd74e4bca5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-29T01:14:00.988486Z","signature_b64":"qECdRP0il4wZ0/6DmsJAgtWQeWcX/Acrr40QY3i5LUN5zR6DuvCSbmuEDlQOsrq2MgTLhmzgeRJN2YHjlMuBDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b94fce09cc7b11a052e2ed28633eec92140737f9f5c8312a050fbe837e78bf8e","last_reissued_at":"2026-06-29T01:14:00.987965Z","signature_status":"signed_v1","first_computed_at":"2026-06-29T01:14:00.987965Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Continual Memorization of Factoids in Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adithya Bhaskar, Dan Friedman, Danqi Chen, Howard Chen, Jiayi Geng","submitted_at":"2024-11-11T17:56:15Z","abstract_excerpt":"As new knowledge rapidly accumulates, language models (LMs) with pretrained knowledge quickly become obsolete. A common approach to updating LMs is fine-tuning them directly on new knowledge. However, recent studies have shown that fine-tuning for memorization may be ineffective in storing knowledge or may exacerbate hallucinations. In this work, we introduce a setting we call continual memorization, where a model must memorize and retain a set of factoids through multiple stages of fine-tuning on subsequent datasets. We characterized the forgetting patterns through extensive experiments and s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.07175","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/2411.07175/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":"2411.07175","created_at":"2026-06-29T01:14:00.988025+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.07175v3","created_at":"2026-06-29T01:14:00.988025+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.07175","created_at":"2026-06-29T01:14:00.988025+00:00"},{"alias_kind":"pith_short_12","alias_value":"XFH44COMPMI2","created_at":"2026-06-29T01:14:00.988025+00:00"},{"alias_kind":"pith_short_16","alias_value":"XFH44COMPMI2AUXC","created_at":"2026-06-29T01:14:00.988025+00:00"},{"alias_kind":"pith_short_8","alias_value":"XFH44COM","created_at":"2026-06-29T01:14:00.988025+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.20005","citing_title":"Fine-Tuning Without Forgetting via Loss-Adaptive Learning Rates","ref_index":4,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI","json":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI.json","graph_json":"https://pith.science/api/pith-number/XFH44COMPMI2AUXC5UUGGPXMSI/graph.json","events_json":"https://pith.science/api/pith-number/XFH44COMPMI2AUXC5UUGGPXMSI/events.json","paper":"https://pith.science/paper/XFH44COM"},"agent_actions":{"view_html":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI","download_json":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI.json","view_paper":"https://pith.science/paper/XFH44COM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.07175&json=true","fetch_graph":"https://pith.science/api/pith-number/XFH44COMPMI2AUXC5UUGGPXMSI/graph.json","fetch_events":"https://pith.science/api/pith-number/XFH44COMPMI2AUXC5UUGGPXMSI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI/action/storage_attestation","attest_author":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI/action/author_attestation","sign_citation":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI/action/citation_signature","submit_replication":"https://pith.science/pith/XFH44COMPMI2AUXC5UUGGPXMSI/action/replication_record"}},"created_at":"2026-06-29T01:14:00.988025+00:00","updated_at":"2026-06-29T01:14:00.988025+00:00"}