{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SA2KG4OMU5RPEMJUA6M3CUBLGR","short_pith_number":"pith:SA2KG4OM","schema_version":"1.0","canonical_sha256":"9034a371cca762f231340799b1502b344f59e06335b005bce2b5dcd8f9e4c682","source":{"kind":"arxiv","id":"2602.05305","version":3},"attestation_state":"computed","paper":{"title":"FlashBlock: Attention Caching for Efficient Long-Context Block Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Bohan Zhuang, Jianfei Cai, Zhuokun Chen","submitted_at":"2026-02-05T04:57:21Z","abstract_excerpt":"Generating long-form content, such as minute-long videos and extended texts, is increasingly important for modern generative models. Block diffusion improves inference efficiency via KV caching and block-wise causal inference and has been widely adopted in diffusion language models and video generation. However, in long-context settings, block diffusion still incurs substantial overhead from repeatedly computing attention over a growing KV cache. We identify an underexplored property of block diffusion: cross-step redundancy of attention within a block. Our analysis shows that attention output"},"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":"2602.05305","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-05T04:57:21Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"5911320277ef401a105b6991423f9089d02b2112e8bb1149e656a29b4f95a39d","abstract_canon_sha256":"4344c84700855967a365bf33eb6d7542150c78cb58f1ba079b976823cad10da9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:18.174836Z","signature_b64":"vhTbkBxx42jGoPP7/oBTQoH8jqT5PNJcPCkU08CJm15cS7qbWG3LwAnvlWeBdV/27sQeQrTKEnzDsYWMbMSdBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9034a371cca762f231340799b1502b344f59e06335b005bce2b5dcd8f9e4c682","last_reissued_at":"2026-07-07T02:17:18.173957Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:18.173957Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FlashBlock: Attention Caching for Efficient Long-Context Block Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Bohan Zhuang, Jianfei Cai, Zhuokun Chen","submitted_at":"2026-02-05T04:57:21Z","abstract_excerpt":"Generating long-form content, such as minute-long videos and extended texts, is increasingly important for modern generative models. Block diffusion improves inference efficiency via KV caching and block-wise causal inference and has been widely adopted in diffusion language models and video generation. However, in long-context settings, block diffusion still incurs substantial overhead from repeatedly computing attention over a growing KV cache. We identify an underexplored property of block diffusion: cross-step redundancy of attention within a block. Our analysis shows that attention output"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.05305","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/2602.05305/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":"2602.05305","created_at":"2026-07-07T02:17:18.174084+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.05305v3","created_at":"2026-07-07T02:17:18.174084+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.05305","created_at":"2026-07-07T02:17:18.174084+00:00"},{"alias_kind":"pith_short_12","alias_value":"SA2KG4OMU5RP","created_at":"2026-07-07T02:17:18.174084+00:00"},{"alias_kind":"pith_short_16","alias_value":"SA2KG4OMU5RPEMJU","created_at":"2026-07-07T02:17:18.174084+00:00"},{"alias_kind":"pith_short_8","alias_value":"SA2KG4OM","created_at":"2026-07-07T02:17:18.174084+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.18165","citing_title":"Elastic-dLLM: Position Preserving Context Compression and Augmentation of Diffusion LLMs","ref_index":7,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR","json":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR.json","graph_json":"https://pith.science/api/pith-number/SA2KG4OMU5RPEMJUA6M3CUBLGR/graph.json","events_json":"https://pith.science/api/pith-number/SA2KG4OMU5RPEMJUA6M3CUBLGR/events.json","paper":"https://pith.science/paper/SA2KG4OM"},"agent_actions":{"view_html":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR","download_json":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR.json","view_paper":"https://pith.science/paper/SA2KG4OM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.05305&json=true","fetch_graph":"https://pith.science/api/pith-number/SA2KG4OMU5RPEMJUA6M3CUBLGR/graph.json","fetch_events":"https://pith.science/api/pith-number/SA2KG4OMU5RPEMJUA6M3CUBLGR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR/action/storage_attestation","attest_author":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR/action/author_attestation","sign_citation":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR/action/citation_signature","submit_replication":"https://pith.science/pith/SA2KG4OMU5RPEMJUA6M3CUBLGR/action/replication_record"}},"created_at":"2026-07-07T02:17:18.174084+00:00","updated_at":"2026-07-07T02:17:18.174084+00:00"}