{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:5FN2QWTUKBAPKOUJA3HK56MQVM","short_pith_number":"pith:5FN2QWTU","canonical_record":{"source":{"id":"2408.07092","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-11T18:40:36Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"1561c73eac3fb54b238b4c0ac5f7f6a38da1f64757242d2b0bba596991bf3e3f","abstract_canon_sha256":"61aa8748e2e77dac0ae940d127dd5866bc4ae0c110188003e9bd4d7d620902b6"},"schema_version":"1.0"},"canonical_sha256":"e95ba85a745040f53a8906ceaef990ab23906f85bfde58209122e52d0a30865d","source":{"kind":"arxiv","id":"2408.07092","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.07092","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"arxiv_version","alias_value":"2408.07092v2","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.07092","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_12","alias_value":"5FN2QWTUKBAP","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_16","alias_value":"5FN2QWTUKBAPKOUJ","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_8","alias_value":"5FN2QWTU","created_at":"2026-07-05T08:56:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:5FN2QWTUKBAPKOUJA3HK56MQVM","target":"record","payload":{"canonical_record":{"source":{"id":"2408.07092","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-11T18:40:36Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"1561c73eac3fb54b238b4c0ac5f7f6a38da1f64757242d2b0bba596991bf3e3f","abstract_canon_sha256":"61aa8748e2e77dac0ae940d127dd5866bc4ae0c110188003e9bd4d7d620902b6"},"schema_version":"1.0"},"canonical_sha256":"e95ba85a745040f53a8906ceaef990ab23906f85bfde58209122e52d0a30865d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:56:24.280069Z","signature_b64":"JYg08SJqh5nnqaHqpGOsA9/U35awsLXwNJCI3waPmsluB7ePrNHqF6/MrXn0rranWi/MajpP+bLAl+nv5FKKCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e95ba85a745040f53a8906ceaef990ab23906f85bfde58209122e52d0a30865d","last_reissued_at":"2026-07-05T08:56:24.279680Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:56:24.279680Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2408.07092","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:56:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Rma0ylNAcIrnLDeuHklWcKAyJYabeTREOtsXMQV/DJmHF/tJ83LCn8AusWJktQ2k8pQL9E27M8qOHlE3QfYqDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T07:05:27.212849Z"},"content_sha256":"d4415ba3f0f72821137598187672fd71381d53c9dbf85af0dff14af12a7459af","schema_version":"1.0","event_id":"sha256:d4415ba3f0f72821137598187672fd71381d53c9dbf85af0dff14af12a7459af"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:5FN2QWTUKBAPKOUJA3HK56MQVM","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Post-Training Sparse Attention with Double Sparsity","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Ion Stoica, Joseph E. Gonzalez, Lianmin Zheng, Shuo Yang, Ying Sheng","submitted_at":"2024-08-11T18:40:36Z","abstract_excerpt":"The inference process for large language models is slow and memory-intensive, with one of the most critical bottlenecks being excessive Key-Value (KV) cache accesses. This paper introduces \"Double Sparsity,\" a novel post-training sparse attention technique designed to alleviate this bottleneck by reducing KV cache access. Double Sparsity combines token sparsity, which focuses on utilizing only the important tokens for computing self-attention, with channel sparsity, an approach that uses important feature channels for identifying important tokens. Our key insight is that the pattern of channel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.07092","kind":"arxiv","version":2},"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/2408.07092/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:56:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Dli7hcb6+dbDTf+w+8Fo3UAMg05cKulevqcpzk4pbsrvNQD8kv0XvkUFn/AsvDno2zneVIgJS7Mz/b+OYSnTCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T07:05:27.213364Z"},"content_sha256":"a81e036ece5d621f4731b8e6196ebb32cfd1506bf39649596d40c0dc89204043","schema_version":"1.0","event_id":"sha256:a81e036ece5d621f4731b8e6196ebb32cfd1506bf39649596d40c0dc89204043"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/bundle.json","state_url":"https://pith.science/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-05T07:05:27Z","links":{"resolver":"https://pith.science/pith/5FN2QWTUKBAPKOUJA3HK56MQVM","bundle":"https://pith.science/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/bundle.json","state":"https://pith.science/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5FN2QWTUKBAPKOUJA3HK56MQVM/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:5FN2QWTUKBAPKOUJA3HK56MQVM","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"61aa8748e2e77dac0ae940d127dd5866bc4ae0c110188003e9bd4d7d620902b6","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-11T18:40:36Z","title_canon_sha256":"1561c73eac3fb54b238b4c0ac5f7f6a38da1f64757242d2b0bba596991bf3e3f"},"schema_version":"1.0","source":{"id":"2408.07092","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.07092","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"arxiv_version","alias_value":"2408.07092v2","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.07092","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_12","alias_value":"5FN2QWTUKBAP","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_16","alias_value":"5FN2QWTUKBAPKOUJ","created_at":"2026-07-05T08:56:24Z"},{"alias_kind":"pith_short_8","alias_value":"5FN2QWTU","created_at":"2026-07-05T08:56:24Z"}],"graph_snapshots":[{"event_id":"sha256:a81e036ece5d621f4731b8e6196ebb32cfd1506bf39649596d40c0dc89204043","target":"graph","created_at":"2026-07-05T08:56:24Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2408.07092/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The inference process for large language models is slow and memory-intensive, with one of the most critical bottlenecks being excessive Key-Value (KV) cache accesses. This paper introduces \"Double Sparsity,\" a novel post-training sparse attention technique designed to alleviate this bottleneck by reducing KV cache access. Double Sparsity combines token sparsity, which focuses on utilizing only the important tokens for computing self-attention, with channel sparsity, an approach that uses important feature channels for identifying important tokens. Our key insight is that the pattern of channel","authors_text":"Ion Stoica, Joseph E. Gonzalez, Lianmin Zheng, Shuo Yang, Ying Sheng","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-11T18:40:36Z","title":"Post-Training Sparse Attention with Double Sparsity"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.07092","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:d4415ba3f0f72821137598187672fd71381d53c9dbf85af0dff14af12a7459af","target":"record","created_at":"2026-07-05T08:56:24Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"61aa8748e2e77dac0ae940d127dd5866bc4ae0c110188003e9bd4d7d620902b6","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-08-11T18:40:36Z","title_canon_sha256":"1561c73eac3fb54b238b4c0ac5f7f6a38da1f64757242d2b0bba596991bf3e3f"},"schema_version":"1.0","source":{"id":"2408.07092","kind":"arxiv","version":2}},"canonical_sha256":"e95ba85a745040f53a8906ceaef990ab23906f85bfde58209122e52d0a30865d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e95ba85a745040f53a8906ceaef990ab23906f85bfde58209122e52d0a30865d","first_computed_at":"2026-07-05T08:56:24.279680Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:56:24.279680Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JYg08SJqh5nnqaHqpGOsA9/U35awsLXwNJCI3waPmsluB7ePrNHqF6/MrXn0rranWi/MajpP+bLAl+nv5FKKCg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:56:24.280069Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.07092","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d4415ba3f0f72821137598187672fd71381d53c9dbf85af0dff14af12a7459af","sha256:a81e036ece5d621f4731b8e6196ebb32cfd1506bf39649596d40c0dc89204043"],"state_sha256":"fad4c3e80ffd230177ee5d9fcd2bffc945cd8a1e0f49093a0c468a5c0126fe73"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R8RYJt8/zfFjYxDNBdOzmtH9uJKEqa8bnk9g1L38Pj5136Snh462R2K6HPsZrbDkmqcVYQZXiAG8yQK5q+LQDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T07:05:27.216789Z","bundle_sha256":"4b95800970dcd9f998108e64648a4b06adeec6d439227079cdf83d19d64ba23d"}}