{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5L4LIMTGMIVJC53MWCZ2UIVR7M","short_pith_number":"pith:5L4LIMTG","schema_version":"1.0","canonical_sha256":"eaf8b43266622a91776cb0b3aa22b1fb399a105ec6bfcc7e7f8a6cc9f5a747f8","source":{"kind":"arxiv","id":"2603.12658","version":2},"attestation_state":"computed","paper":{"title":"Beyond Static Models: An Evolving Framework for Continual Learning in Large Language Models across Training Stages","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hongfei Ye, Hongyang Chen, Kunchi Li, Xuemin Lin, Zhongwu Sun","submitted_at":"2026-03-13T05:01:25Z","abstract_excerpt":"Continual learning (CL) has emerged as a pivotal paradigm to enable large language models (LLMs) to dynamically adapt to evolving knowledge and sequential tasks while mitigating catastrophic forgetting, a critical limitation of the static pre-training paradigm inherent to modern LLMs. This paper presents a comprehensive overview of CL methodologies tailored for LLMs, structured around three core training stages: continual pre-training, continual fine-tuning, and continual alignment. Beyond the canonical taxonomy of rehearsal-, regularization-, and architecture-based methods, we further subdivi"},"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":"2603.12658","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-13T05:01:25Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d3f4e48492c0c64fb7db0a674209ff89476c6f4f014dcc701aedb273ef05dd64","abstract_canon_sha256":"5d04b3cfcec064d4c3c5d84edf6f9bd4d37afef4a0c1a6d6fc0e671b5897ee6a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-11T01:24:40.612160Z","signature_b64":"ZpnVKaQYvxeDi5nwkO++6eyavgR8NDlgzM7eQIib5mdTyz/DmfoclcLBvbSp3ozfSlIbgQ62QDx6iv8VaotEDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eaf8b43266622a91776cb0b3aa22b1fb399a105ec6bfcc7e7f8a6cc9f5a747f8","last_reissued_at":"2026-08-11T01:24:40.609785Z","signature_status":"signed_v1","first_computed_at":"2026-08-11T01:24:40.609785Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Static Models: An Evolving Framework for Continual Learning in Large Language Models across Training Stages","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hongfei Ye, Hongyang Chen, Kunchi Li, Xuemin Lin, Zhongwu Sun","submitted_at":"2026-03-13T05:01:25Z","abstract_excerpt":"Continual learning (CL) has emerged as a pivotal paradigm to enable large language models (LLMs) to dynamically adapt to evolving knowledge and sequential tasks while mitigating catastrophic forgetting, a critical limitation of the static pre-training paradigm inherent to modern LLMs. This paper presents a comprehensive overview of CL methodologies tailored for LLMs, structured around three core training stages: continual pre-training, continual fine-tuning, and continual alignment. Beyond the canonical taxonomy of rehearsal-, regularization-, and architecture-based methods, we further subdivi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.12658","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/2603.12658/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":"2603.12658","created_at":"2026-08-11T01:24:40.610507+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.12658v2","created_at":"2026-08-11T01:24:40.610507+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.12658","created_at":"2026-08-11T01:24:40.610507+00:00"},{"alias_kind":"pith_short_12","alias_value":"5L4LIMTGMIVJ","created_at":"2026-08-11T01:24:40.610507+00:00"},{"alias_kind":"pith_short_16","alias_value":"5L4LIMTGMIVJC53M","created_at":"2026-08-11T01:24:40.610507+00:00"},{"alias_kind":"pith_short_8","alias_value":"5L4LIMTG","created_at":"2026-08-11T01:24:40.610507+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2606.27634","citing_title":"Continual Learning for Sequential Personalization of Small Language Models: A Stability Monitoring Analysis","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07076","citing_title":"Self-Consolidating Language Models: Continual Knowledge Incorporation from Context","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.22407","citing_title":"Hidden Failure Modes of Gradient Modification under Adam in Continual Learning, and Adaptive Decoupled Moment Routing as a Repair","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2604.08224","citing_title":"Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07076","citing_title":"Self-Consolidating Language Models: Continual Knowledge Incorporation from Context","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M","json":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M.json","graph_json":"https://pith.science/api/pith-number/5L4LIMTGMIVJC53MWCZ2UIVR7M/graph.json","events_json":"https://pith.science/api/pith-number/5L4LIMTGMIVJC53MWCZ2UIVR7M/events.json","paper":"https://pith.science/paper/5L4LIMTG"},"agent_actions":{"view_html":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M","download_json":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M.json","view_paper":"https://pith.science/paper/5L4LIMTG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.12658&json=true","fetch_graph":"https://pith.science/api/pith-number/5L4LIMTGMIVJC53MWCZ2UIVR7M/graph.json","fetch_events":"https://pith.science/api/pith-number/5L4LIMTGMIVJC53MWCZ2UIVR7M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M/action/storage_attestation","attest_author":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M/action/author_attestation","sign_citation":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M/action/citation_signature","submit_replication":"https://pith.science/pith/5L4LIMTGMIVJC53MWCZ2UIVR7M/action/replication_record"}},"created_at":"2026-08-11T01:24:40.610507+00:00","updated_at":"2026-08-11T01:24:40.610507+00:00"}