{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BFXK4QWPH3UHICTVPME4ZHTQMU","short_pith_number":"pith:BFXK4QWP","schema_version":"1.0","canonical_sha256":"096eae42cf3ee8740a757b09cc9e7065153814d1462c48f99060f1945cfd34f6","source":{"kind":"arxiv","id":"2607.01299","version":1},"attestation_state":"computed","paper":{"title":"HYPIC: Accelerating Hybrid-Attention LLM Serving with Position-Independent Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Junhao Hu, Juntong Wu, Minghao Li, Weihang Chen, Xiaoxu Chen, Yang Liu, Yifei Liu","submitted_at":"2026-07-01T14:03:56Z","abstract_excerpt":"In retrieval augmented generation (RAG) and agentic LLM serving, prompts are assembled from independent segments into long contexts, making the prefill stage dominate the per-request computation cost. To this cost, two directions have emerged in parallel: position-independent caching (PIC) admits KV reuse for non-contiguous segments shared across different requests, while hybrid-attention models reduce computation complexity by replacing most full-attention layers with linear attention. However, they cannot coexist: applying PIC to hybrid-attention models breaks down because per-token KV-cache"},"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":"2607.01299","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DC","submitted_at":"2026-07-01T14:03:56Z","cross_cats_sorted":[],"title_canon_sha256":"c166c16181297b9eb01337e9decde585b8321242e5340656ce93dbd2c9bb9d29","abstract_canon_sha256":"56b526c66601fedfd34cb7da6bc861665ee59503c74fc1b1decd7aba98861669"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T00:16:56.655451Z","signature_b64":"JvuOCmHDB3JtqGseWXZxs0aDpiR/jeV4fCo/nY+ZhyOmfI2dQkZPPc1yExq+RB1f0QA2z6UqGUNol6ZD4xLaDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"096eae42cf3ee8740a757b09cc9e7065153814d1462c48f99060f1945cfd34f6","last_reissued_at":"2026-07-03T00:16:56.655066Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T00:16:56.655066Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HYPIC: Accelerating Hybrid-Attention LLM Serving with Position-Independent Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.DC","authors_text":"Junhao Hu, Juntong Wu, Minghao Li, Weihang Chen, Xiaoxu Chen, Yang Liu, Yifei Liu","submitted_at":"2026-07-01T14:03:56Z","abstract_excerpt":"In retrieval augmented generation (RAG) and agentic LLM serving, prompts are assembled from independent segments into long contexts, making the prefill stage dominate the per-request computation cost. To this cost, two directions have emerged in parallel: position-independent caching (PIC) admits KV reuse for non-contiguous segments shared across different requests, while hybrid-attention models reduce computation complexity by replacing most full-attention layers with linear attention. However, they cannot coexist: applying PIC to hybrid-attention models breaks down because per-token KV-cache"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.01299","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/2607.01299/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":"2607.01299","created_at":"2026-07-03T00:16:56.655130+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.01299v1","created_at":"2026-07-03T00:16:56.655130+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.01299","created_at":"2026-07-03T00:16:56.655130+00:00"},{"alias_kind":"pith_short_12","alias_value":"BFXK4QWPH3UH","created_at":"2026-07-03T00:16:56.655130+00:00"},{"alias_kind":"pith_short_16","alias_value":"BFXK4QWPH3UHICTV","created_at":"2026-07-03T00:16:56.655130+00:00"},{"alias_kind":"pith_short_8","alias_value":"BFXK4QWP","created_at":"2026-07-03T00:16:56.655130+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/BFXK4QWPH3UHICTVPME4ZHTQMU","json":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU.json","graph_json":"https://pith.science/api/pith-number/BFXK4QWPH3UHICTVPME4ZHTQMU/graph.json","events_json":"https://pith.science/api/pith-number/BFXK4QWPH3UHICTVPME4ZHTQMU/events.json","paper":"https://pith.science/paper/BFXK4QWP"},"agent_actions":{"view_html":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU","download_json":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU.json","view_paper":"https://pith.science/paper/BFXK4QWP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.01299&json=true","fetch_graph":"https://pith.science/api/pith-number/BFXK4QWPH3UHICTVPME4ZHTQMU/graph.json","fetch_events":"https://pith.science/api/pith-number/BFXK4QWPH3UHICTVPME4ZHTQMU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU/action/storage_attestation","attest_author":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU/action/author_attestation","sign_citation":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU/action/citation_signature","submit_replication":"https://pith.science/pith/BFXK4QWPH3UHICTVPME4ZHTQMU/action/replication_record"}},"created_at":"2026-07-03T00:16:56.655130+00:00","updated_at":"2026-07-03T00:16:56.655130+00:00"}