{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4SBKT3TQGWEDVJMVGKGBGWYOZD","short_pith_number":"pith:4SBKT3TQ","schema_version":"1.0","canonical_sha256":"e482a9ee7035883aa595328c135b0ec8dccdb0d2b50b462906facc4e3ca2ae40","source":{"kind":"arxiv","id":"2605.19893","version":1},"attestation_state":"computed","paper":{"title":"SpecSA: Bridging Speculative Decoding and Sparse Attention for Efficient LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.OS","authors_text":"Nuo Shen, Rong Gu, Sheng Zhong, Yuhang Zhou, Zhibin Wang, Ziyu Zhong","submitted_at":"2026-05-19T14:24:27Z","abstract_excerpt":"Speculative decoding and dynamic sparse attention are two complementary approaches for accelerating long-context LLM inference: the former amortizes target-model execution across multiple verifier queries, while the latter reduces each query's KV-cache working set. Directly combining them, however, exposes a structural mismatch: speculative verification relies on cross-query commonality, whereas dynamic sparse attention assigns query-specific sparse layouts. This mismatch limits KV-block reuse, amplifies NSA's branch-wise overheads, and makes verification strategy selection input- and regime-d"},"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":"2605.19893","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.OS","submitted_at":"2026-05-19T14:24:27Z","cross_cats_sorted":[],"title_canon_sha256":"b3c256edeb9bd3d11f22ac82fc12f58aeb6c3a47e6577d2a38d64fcc1e4d31b0","abstract_canon_sha256":"b42144ba0d8b1a36b6db1cc3ddedec7489506c0f1c00f3978d1f96c02aed2a15"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:06:19.725473Z","signature_b64":"pZtNHhkeH9yP5s6IsJg8qln6+9pKqmhWk2bUA9EYFnPveyZsNAX909XzVYQAkoXVOasJv5O2RN3bzNIi5EkgBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e482a9ee7035883aa595328c135b0ec8dccdb0d2b50b462906facc4e3ca2ae40","last_reissued_at":"2026-05-20T01:06:19.724920Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:06:19.724920Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SpecSA: Bridging Speculative Decoding and Sparse Attention for Efficient LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.OS","authors_text":"Nuo Shen, Rong Gu, Sheng Zhong, Yuhang Zhou, Zhibin Wang, Ziyu Zhong","submitted_at":"2026-05-19T14:24:27Z","abstract_excerpt":"Speculative decoding and dynamic sparse attention are two complementary approaches for accelerating long-context LLM inference: the former amortizes target-model execution across multiple verifier queries, while the latter reduces each query's KV-cache working set. Directly combining them, however, exposes a structural mismatch: speculative verification relies on cross-query commonality, whereas dynamic sparse attention assigns query-specific sparse layouts. This mismatch limits KV-block reuse, amplifies NSA's branch-wise overheads, and makes verification strategy selection input- and regime-d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19893","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/2605.19893/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":"2605.19893","created_at":"2026-05-20T01:06:19.725012+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19893v1","created_at":"2026-05-20T01:06:19.725012+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19893","created_at":"2026-05-20T01:06:19.725012+00:00"},{"alias_kind":"pith_short_12","alias_value":"4SBKT3TQGWED","created_at":"2026-05-20T01:06:19.725012+00:00"},{"alias_kind":"pith_short_16","alias_value":"4SBKT3TQGWEDVJMV","created_at":"2026-05-20T01:06:19.725012+00:00"},{"alias_kind":"pith_short_8","alias_value":"4SBKT3TQ","created_at":"2026-05-20T01:06:19.725012+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD","json":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD.json","graph_json":"https://pith.science/api/pith-number/4SBKT3TQGWEDVJMVGKGBGWYOZD/graph.json","events_json":"https://pith.science/api/pith-number/4SBKT3TQGWEDVJMVGKGBGWYOZD/events.json","paper":"https://pith.science/paper/4SBKT3TQ"},"agent_actions":{"view_html":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD","download_json":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD.json","view_paper":"https://pith.science/paper/4SBKT3TQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19893&json=true","fetch_graph":"https://pith.science/api/pith-number/4SBKT3TQGWEDVJMVGKGBGWYOZD/graph.json","fetch_events":"https://pith.science/api/pith-number/4SBKT3TQGWEDVJMVGKGBGWYOZD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD/action/storage_attestation","attest_author":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD/action/author_attestation","sign_citation":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD/action/citation_signature","submit_replication":"https://pith.science/pith/4SBKT3TQGWEDVJMVGKGBGWYOZD/action/replication_record"}},"created_at":"2026-05-20T01:06:19.725012+00:00","updated_at":"2026-05-20T01:06:19.725012+00:00"}