{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2WVXFXZ2EOHTWPIR2HPKVYITW2","short_pith_number":"pith:2WVXFXZ2","schema_version":"1.0","canonical_sha256":"d5ab72df3a238f3b3d11d1deaae113b6ba1c2914b4081ef310b2c621c1d12f92","source":{"kind":"arxiv","id":"2411.17116","version":3},"attestation_state":"computed","paper":{"title":"Star Attention: Efficient LLM Inference over Long Sequences","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Fei Jia, Shantanu Acharya","submitted_at":"2024-11-26T05:10:04Z","abstract_excerpt":"Inference with Transformer-based Large Language Models (LLMs) on long sequences is both costly and slow due to the quadratic complexity of the self-attention mechanism. We introduce Star Attention, a two-phase block-sparse approximation that improves computational efficiency by sharding attention across multiple hosts while minimizing communication overhead. In the first phase, the context is processed using blockwise-local attention across hosts, in parallel. In the second phase, query and response tokens attend to all prior cached tokens through sequence-global attention. Star Attention inte"},"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":"2411.17116","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-26T05:10:04Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"e1ca483d269c98cb29239deaa32bf351b2e711de5db98108606aaad2520ec8a0","abstract_canon_sha256":"d6bd18514a74cb9492580197d43554f103ef929db5231a09871d40070c3d9873"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:22.692903Z","signature_b64":"cH0c0GeniENzB1Y399Pa9jljWE2tvoKT9iMnbp17+LOA6gJRW2D1OmbsYCaloLfbbDHiyIzm/x4aiZwpzcYgDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d5ab72df3a238f3b3d11d1deaae113b6ba1c2914b4081ef310b2c621c1d12f92","last_reissued_at":"2026-07-05T11:12:22.692300Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:22.692300Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Star Attention: Efficient LLM Inference over Long Sequences","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Boris Ginsburg, Fei Jia, Shantanu Acharya","submitted_at":"2024-11-26T05:10:04Z","abstract_excerpt":"Inference with Transformer-based Large Language Models (LLMs) on long sequences is both costly and slow due to the quadratic complexity of the self-attention mechanism. We introduce Star Attention, a two-phase block-sparse approximation that improves computational efficiency by sharding attention across multiple hosts while minimizing communication overhead. In the first phase, the context is processed using blockwise-local attention across hosts, in parallel. In the second phase, query and response tokens attend to all prior cached tokens through sequence-global attention. Star Attention inte"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.17116","kind":"arxiv","version":3},"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/2411.17116/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":"2411.17116","created_at":"2026-07-05T11:12:22.692372+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.17116v3","created_at":"2026-07-05T11:12:22.692372+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.17116","created_at":"2026-07-05T11:12:22.692372+00:00"},{"alias_kind":"pith_short_12","alias_value":"2WVXFXZ2EOHT","created_at":"2026-07-05T11:12:22.692372+00:00"},{"alias_kind":"pith_short_16","alias_value":"2WVXFXZ2EOHTWPIR","created_at":"2026-07-05T11:12:22.692372+00:00"},{"alias_kind":"pith_short_8","alias_value":"2WVXFXZ2","created_at":"2026-07-05T11:12:22.692372+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.13334","citing_title":"A Survey of Context Engineering for Large Language Models","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12491","citing_title":"Elastic Attention Cores for Scalable Vision Transformers","ref_index":64,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2","json":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2.json","graph_json":"https://pith.science/api/pith-number/2WVXFXZ2EOHTWPIR2HPKVYITW2/graph.json","events_json":"https://pith.science/api/pith-number/2WVXFXZ2EOHTWPIR2HPKVYITW2/events.json","paper":"https://pith.science/paper/2WVXFXZ2"},"agent_actions":{"view_html":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2","download_json":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2.json","view_paper":"https://pith.science/paper/2WVXFXZ2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.17116&json=true","fetch_graph":"https://pith.science/api/pith-number/2WVXFXZ2EOHTWPIR2HPKVYITW2/graph.json","fetch_events":"https://pith.science/api/pith-number/2WVXFXZ2EOHTWPIR2HPKVYITW2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2/action/storage_attestation","attest_author":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2/action/author_attestation","sign_citation":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2/action/citation_signature","submit_replication":"https://pith.science/pith/2WVXFXZ2EOHTWPIR2HPKVYITW2/action/replication_record"}},"created_at":"2026-07-05T11:12:22.692372+00:00","updated_at":"2026-07-05T11:12:22.692372+00:00"}