{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:2GXI3UQRUBI2KDAZA3Y7PRHQ3T","short_pith_number":"pith:2GXI3UQR","schema_version":"1.0","canonical_sha256":"d1ae8dd211a051a50c1906f1f7c4f0dcca5e53ccc248cfd17825281a8c0a9bb5","source":{"kind":"arxiv","id":"2304.02207","version":1},"attestation_state":"computed","paper":{"title":"Algorithm and Hardness for Dynamic Attention Maintenance in Large Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CC"],"primary_cat":"cs.DS","authors_text":"Jan van den Brand, Tianyi Zhou, Zhao Song","submitted_at":"2023-04-05T03:43:38Z","abstract_excerpt":"Large language models (LLMs) have made fundamental changes in human life. The attention scheme is one of the key components over all the LLMs, such as BERT, GPT-1, Transformers, GPT-2, 3, 3.5 and 4. Inspired by previous theoretical study of static version of the attention multiplication problem [Zandieh, Han, Daliri, and Karbasi arXiv 2023, Alman and Song arXiv 2023]. In this work, we formally define a dynamic version of attention matrix multiplication problem. There are matrices $Q,K, V \\in \\mathbb{R}^{n \\times d}$, they represent query, key and value in LLMs. In each iteration we update one "},"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":"2304.02207","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.DS","submitted_at":"2023-04-05T03:43:38Z","cross_cats_sorted":["cs.CC"],"title_canon_sha256":"9238dbc6b7ce679e6fd593c590053dde7f4212b15144445f543b492c5c67fe0e","abstract_canon_sha256":"857066d8e7940fe54ec70a70cc1efaa602acba6104cc9779b8d90aa54d436885"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:58:23.229871Z","signature_b64":"RbAwoIif9iYU2yF/x1e0KSaji7cUHFjcmzQt/P0/KrT0IozB6mxngjKKPfYGZVTkPFy0cGNO/T0AtvepEERQAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d1ae8dd211a051a50c1906f1f7c4f0dcca5e53ccc248cfd17825281a8c0a9bb5","last_reissued_at":"2026-07-05T05:58:23.229430Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:58:23.229430Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Algorithm and Hardness for Dynamic Attention Maintenance in Large Language Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CC"],"primary_cat":"cs.DS","authors_text":"Jan van den Brand, Tianyi Zhou, Zhao Song","submitted_at":"2023-04-05T03:43:38Z","abstract_excerpt":"Large language models (LLMs) have made fundamental changes in human life. The attention scheme is one of the key components over all the LLMs, such as BERT, GPT-1, Transformers, GPT-2, 3, 3.5 and 4. Inspired by previous theoretical study of static version of the attention multiplication problem [Zandieh, Han, Daliri, and Karbasi arXiv 2023, Alman and Song arXiv 2023]. In this work, we formally define a dynamic version of attention matrix multiplication problem. There are matrices $Q,K, V \\in \\mathbb{R}^{n \\times d}$, they represent query, key and value in LLMs. In each iteration we update one "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.02207","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/2304.02207/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":"2304.02207","created_at":"2026-07-05T05:58:23.229491+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.02207v1","created_at":"2026-07-05T05:58:23.229491+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.02207","created_at":"2026-07-05T05:58:23.229491+00:00"},{"alias_kind":"pith_short_12","alias_value":"2GXI3UQRUBI2","created_at":"2026-07-05T05:58:23.229491+00:00"},{"alias_kind":"pith_short_16","alias_value":"2GXI3UQRUBI2KDAZ","created_at":"2026-07-05T05:58:23.229491+00:00"},{"alias_kind":"pith_short_8","alias_value":"2GXI3UQR","created_at":"2026-07-05T05:58:23.229491+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2306.14048","citing_title":"H$_2$O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models","ref_index":100,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T","json":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T.json","graph_json":"https://pith.science/api/pith-number/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/graph.json","events_json":"https://pith.science/api/pith-number/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/events.json","paper":"https://pith.science/paper/2GXI3UQR"},"agent_actions":{"view_html":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T","download_json":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T.json","view_paper":"https://pith.science/paper/2GXI3UQR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.02207&json=true","fetch_graph":"https://pith.science/api/pith-number/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/graph.json","fetch_events":"https://pith.science/api/pith-number/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/action/storage_attestation","attest_author":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/action/author_attestation","sign_citation":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/action/citation_signature","submit_replication":"https://pith.science/pith/2GXI3UQRUBI2KDAZA3Y7PRHQ3T/action/replication_record"}},"created_at":"2026-07-05T05:58:23.229491+00:00","updated_at":"2026-07-05T05:58:23.229491+00:00"}