{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:I3X3D6E2G4MKLE5IMEH27R3FLF","short_pith_number":"pith:I3X3D6E2","schema_version":"1.0","canonical_sha256":"46efb1f89a3718a593a8610fafc765597a02112b17de2d3ac4dc74c5b21c2b00","source":{"kind":"arxiv","id":"2406.11194","version":4},"attestation_state":"computed","paper":{"title":"In-Context Editing: Learning Knowledge from Self-Induced Distributions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bangcheng Yang, Jiaqi Li, Kailin Jiang, Siyuan Qi, Xiaobo Wang, Yaodong Yang, Yifan Zhong, Zilong Zheng","submitted_at":"2024-06-17T04:00:04Z","abstract_excerpt":"In scenarios where language models must incorporate new information efficiently without extensive retraining, traditional fine-tuning methods are prone to overfitting, degraded generalization, and unnatural language generation. To address these limitations, we introduce Consistent In-Context Editing (ICE), a novel approach leveraging the model's in-context learning capability to optimize toward a contextual distribution rather than a one-hot target. ICE introduces a simple yet effective optimization framework for the model to internalize new knowledge by aligning its output distributions with "},"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":"2406.11194","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T04:00:04Z","cross_cats_sorted":[],"title_canon_sha256":"610aeab5bdcf3a934e579165f2e5af58580d0259d3f98cad4453b188f138ba0b","abstract_canon_sha256":"cdcc68cff24ca4046121fc598dca4eb434fbb32c2e7436f162bb081e1cdbfa32"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:41:47.018169Z","signature_b64":"gblCnKFkQogwo/FZVPZdlFG/4Gj4AEyMHIil6iW9cyO/h4MTZNiIr8evYn0g8sC0iqgkBsF8cvcdxP99qafMCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46efb1f89a3718a593a8610fafc765597a02112b17de2d3ac4dc74c5b21c2b00","last_reissued_at":"2026-07-05T10:41:47.017663Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:41:47.017663Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"In-Context Editing: Learning Knowledge from Self-Induced Distributions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bangcheng Yang, Jiaqi Li, Kailin Jiang, Siyuan Qi, Xiaobo Wang, Yaodong Yang, Yifan Zhong, Zilong Zheng","submitted_at":"2024-06-17T04:00:04Z","abstract_excerpt":"In scenarios where language models must incorporate new information efficiently without extensive retraining, traditional fine-tuning methods are prone to overfitting, degraded generalization, and unnatural language generation. To address these limitations, we introduce Consistent In-Context Editing (ICE), a novel approach leveraging the model's in-context learning capability to optimize toward a contextual distribution rather than a one-hot target. ICE introduces a simple yet effective optimization framework for the model to internalize new knowledge by aligning its output distributions with "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11194","kind":"arxiv","version":4},"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/2406.11194/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":"2406.11194","created_at":"2026-07-05T10:41:47.017722+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.11194v4","created_at":"2026-07-05T10:41:47.017722+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11194","created_at":"2026-07-05T10:41:47.017722+00:00"},{"alias_kind":"pith_short_12","alias_value":"I3X3D6E2G4MK","created_at":"2026-07-05T10:41:47.017722+00:00"},{"alias_kind":"pith_short_16","alias_value":"I3X3D6E2G4MKLE5I","created_at":"2026-07-05T10:41:47.017722+00:00"},{"alias_kind":"pith_short_8","alias_value":"I3X3D6E2","created_at":"2026-07-05T10:41:47.017722+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.18740","citing_title":"Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18740","citing_title":"Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19089","citing_title":"Towards Scalable Lifelong Knowledge Editing with Selective Knowledge Suppression","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05540","citing_title":"Learning to Edit Knowledge via Instruction-based Chain-of-Thought Prompting","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF","json":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF.json","graph_json":"https://pith.science/api/pith-number/I3X3D6E2G4MKLE5IMEH27R3FLF/graph.json","events_json":"https://pith.science/api/pith-number/I3X3D6E2G4MKLE5IMEH27R3FLF/events.json","paper":"https://pith.science/paper/I3X3D6E2"},"agent_actions":{"view_html":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF","download_json":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF.json","view_paper":"https://pith.science/paper/I3X3D6E2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.11194&json=true","fetch_graph":"https://pith.science/api/pith-number/I3X3D6E2G4MKLE5IMEH27R3FLF/graph.json","fetch_events":"https://pith.science/api/pith-number/I3X3D6E2G4MKLE5IMEH27R3FLF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF/action/storage_attestation","attest_author":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF/action/author_attestation","sign_citation":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF/action/citation_signature","submit_replication":"https://pith.science/pith/I3X3D6E2G4MKLE5IMEH27R3FLF/action/replication_record"}},"created_at":"2026-07-05T10:41:47.017722+00:00","updated_at":"2026-07-05T10:41:47.017722+00:00"}