{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:6KHQGLUQE3UT7Z7FB6WCZWUO64","short_pith_number":"pith:6KHQGLUQ","schema_version":"1.0","canonical_sha256":"f28f032e9026e93fe7e50fac2cda8ef700105363db97b1bfc85e722f9bed52ce","source":{"kind":"arxiv","id":"2411.06171","version":1},"attestation_state":"computed","paper":{"title":"SEEKR: Selective Attention-Guided Knowledge Retention for Continual Learning of Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Haiyun Guo, Jinghan He, Jinqiao Wang, Kuan Zhu, Ming Tang, Zihan Zhao","submitted_at":"2024-11-09T13:02:36Z","abstract_excerpt":"Continual learning (CL) is crucial for language models to dynamically adapt to the evolving real-world demands. To mitigate the catastrophic forgetting problem in CL, data replay has been proven a simple and effective strategy, and the subsequent data-replay-based distillation can further enhance the performance. However, existing methods fail to fully exploit the knowledge embedded in models from previous tasks, resulting in the need for a relatively large number of replay samples to achieve good results. In this work, we first explore and emphasize the importance of attention weights in know"},"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":"2411.06171","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-09T13:02:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"c4e06d91e6795c47740449522dbc1f431c36bb36884ae27fdb0f137142037981","abstract_canon_sha256":"9466f420340a524c7502cfff52b1bf43042497522e9f472bb15e9d9bc1a467ae"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:33:42.409406Z","signature_b64":"1vhY7CZ2eXEyE6P2wHfJVZ3NKwaRE5bkPaLje56b6T2PT2rSyhnyABJvEdJG9e6TbZq/hwk77NuBMOZo6teHAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f28f032e9026e93fe7e50fac2cda8ef700105363db97b1bfc85e722f9bed52ce","last_reissued_at":"2026-07-05T09:33:42.409010Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:33:42.409010Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SEEKR: Selective Attention-Guided Knowledge Retention for Continual Learning of Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Haiyun Guo, Jinghan He, Jinqiao Wang, Kuan Zhu, Ming Tang, Zihan Zhao","submitted_at":"2024-11-09T13:02:36Z","abstract_excerpt":"Continual learning (CL) is crucial for language models to dynamically adapt to the evolving real-world demands. To mitigate the catastrophic forgetting problem in CL, data replay has been proven a simple and effective strategy, and the subsequent data-replay-based distillation can further enhance the performance. However, existing methods fail to fully exploit the knowledge embedded in models from previous tasks, resulting in the need for a relatively large number of replay samples to achieve good results. In this work, we first explore and emphasize the importance of attention weights in know"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.06171","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/2411.06171/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":"2411.06171","created_at":"2026-07-05T09:33:42.409066+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.06171v1","created_at":"2026-07-05T09:33:42.409066+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.06171","created_at":"2026-07-05T09:33:42.409066+00:00"},{"alias_kind":"pith_short_12","alias_value":"6KHQGLUQE3UT","created_at":"2026-07-05T09:33:42.409066+00:00"},{"alias_kind":"pith_short_16","alias_value":"6KHQGLUQE3UT7Z7F","created_at":"2026-07-05T09:33:42.409066+00:00"},{"alias_kind":"pith_short_8","alias_value":"6KHQGLUQ","created_at":"2026-07-05T09:33:42.409066+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.09623","citing_title":"Analytic Task Scheduler: Recursive Least Squares Based Method for Continual Learning in Embodied Foundation Models","ref_index":6,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64","json":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64.json","graph_json":"https://pith.science/api/pith-number/6KHQGLUQE3UT7Z7FB6WCZWUO64/graph.json","events_json":"https://pith.science/api/pith-number/6KHQGLUQE3UT7Z7FB6WCZWUO64/events.json","paper":"https://pith.science/paper/6KHQGLUQ"},"agent_actions":{"view_html":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64","download_json":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64.json","view_paper":"https://pith.science/paper/6KHQGLUQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.06171&json=true","fetch_graph":"https://pith.science/api/pith-number/6KHQGLUQE3UT7Z7FB6WCZWUO64/graph.json","fetch_events":"https://pith.science/api/pith-number/6KHQGLUQE3UT7Z7FB6WCZWUO64/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64/action/storage_attestation","attest_author":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64/action/author_attestation","sign_citation":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64/action/citation_signature","submit_replication":"https://pith.science/pith/6KHQGLUQE3UT7Z7FB6WCZWUO64/action/replication_record"}},"created_at":"2026-07-05T09:33:42.409066+00:00","updated_at":"2026-07-05T09:33:42.409066+00:00"}