{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GCJCMFBDUKDEE7GLZSAWNWN2OS","short_pith_number":"pith:GCJCMFBD","schema_version":"1.0","canonical_sha256":"3092261423a286427ccbcc8166d9ba749a5d09ade6bb63488f8062cef1720522","source":{"kind":"arxiv","id":"2402.18865","version":1},"attestation_state":"computed","paper":{"title":"Analyzing and Reducing Catastrophic Forgetting in Parameter Efficient Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Lei Wang, Tianxiang Zhao, Weijieying Ren, Wei Qin, Xinlong Li","submitted_at":"2024-02-29T05:27:45Z","abstract_excerpt":"Existing research has shown that large language models (LLMs) exhibit remarkable performance in language understanding and generation. However, when LLMs are continuously fine-tuned on complex and diverse domain-specific downstream tasks, the inference performance on historical tasks decreases dramatically, which is known as a catastrophic forgetting problem. A trade-off needs to be kept between learning plasticity and memory stability. Plenty of existing works have explored strategies like memory replay, regularization and parameter isolation, but little is known about the geometric connectio"},"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":"2402.18865","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-29T05:27:45Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"7567802703904f1e1044b534b90e64d464f9a4304347c4e6e52a4aa53b43a3d9","abstract_canon_sha256":"a8030fae4e7c16a495101a9437668a1471b579f232ed583237b21a6a89bd0d97"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:50:28.910493Z","signature_b64":"tCOJdTv0yjgEg1Ks1LWKpRKJ3kXTa5ief6zuBjYkE+p1jz5PfZB7W2884H6/qKUcstyxsgeQMNvLpuSrIKmWBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3092261423a286427ccbcc8166d9ba749a5d09ade6bb63488f8062cef1720522","last_reissued_at":"2026-07-05T07:50:28.909976Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:50:28.909976Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Analyzing and Reducing Catastrophic Forgetting in Parameter Efficient Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Lei Wang, Tianxiang Zhao, Weijieying Ren, Wei Qin, Xinlong Li","submitted_at":"2024-02-29T05:27:45Z","abstract_excerpt":"Existing research has shown that large language models (LLMs) exhibit remarkable performance in language understanding and generation. However, when LLMs are continuously fine-tuned on complex and diverse domain-specific downstream tasks, the inference performance on historical tasks decreases dramatically, which is known as a catastrophic forgetting problem. A trade-off needs to be kept between learning plasticity and memory stability. Plenty of existing works have explored strategies like memory replay, regularization and parameter isolation, but little is known about the geometric connectio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.18865","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/2402.18865/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":"2402.18865","created_at":"2026-07-05T07:50:28.910036+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.18865v1","created_at":"2026-07-05T07:50:28.910036+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.18865","created_at":"2026-07-05T07:50:28.910036+00:00"},{"alias_kind":"pith_short_12","alias_value":"GCJCMFBDUKDE","created_at":"2026-07-05T07:50:28.910036+00:00"},{"alias_kind":"pith_short_16","alias_value":"GCJCMFBDUKDEE7GL","created_at":"2026-07-05T07:50:28.910036+00:00"},{"alias_kind":"pith_short_8","alias_value":"GCJCMFBD","created_at":"2026-07-05T07:50:28.910036+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28998","citing_title":"Reward-Free Code Alignment from Pretrained or Fine-Tuned LLM: Unpacking the Trade-offs for Code Generation","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2510.09885","citing_title":"Diffusion-Inspired Masked Fine-Tuning for Knowledge Injection in Autoregressive LLMs","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09664","citing_title":"FreeMOCA: Memory-Free Continual Learning for Malicious Code Analysis","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09664","citing_title":"FreeMOCA: Memory-Free Continual Learning for Malicious Code Analysis","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09315","citing_title":"Do Self-Evolving Agents Forget? Capability Degradation and Preservation in Lifelong LLM Agent Adaptation","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17051","citing_title":"Efficient Task Adaptation in Large Language Models via Selective Parameter Optimization","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS","json":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS.json","graph_json":"https://pith.science/api/pith-number/GCJCMFBDUKDEE7GLZSAWNWN2OS/graph.json","events_json":"https://pith.science/api/pith-number/GCJCMFBDUKDEE7GLZSAWNWN2OS/events.json","paper":"https://pith.science/paper/GCJCMFBD"},"agent_actions":{"view_html":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS","download_json":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS.json","view_paper":"https://pith.science/paper/GCJCMFBD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.18865&json=true","fetch_graph":"https://pith.science/api/pith-number/GCJCMFBDUKDEE7GLZSAWNWN2OS/graph.json","fetch_events":"https://pith.science/api/pith-number/GCJCMFBDUKDEE7GLZSAWNWN2OS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS/action/storage_attestation","attest_author":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS/action/author_attestation","sign_citation":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS/action/citation_signature","submit_replication":"https://pith.science/pith/GCJCMFBDUKDEE7GLZSAWNWN2OS/action/replication_record"}},"created_at":"2026-07-05T07:50:28.910036+00:00","updated_at":"2026-07-05T07:50:28.910036+00:00"}