{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:3GWDRGDSWSUKMPROHMEA6JX73V","short_pith_number":"pith:3GWDRGDS","canonical_record":{"source":{"id":"2402.15220","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T09:29:19Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"18e0f512eee20ff73bacd1ab420b377f3a43c0e666d13cc57db0bb76be15c9d9","abstract_canon_sha256":"f8c7e03602fc4c92ec21fee827fbf7a7f3ec5bcc9f1d4c24785496076435b02a"},"schema_version":"1.0"},"canonical_sha256":"d9ac389872b4a8a63e2e3b080f26ffdd7115d37b3085f6d06986b70d06ce0f60","source":{"kind":"arxiv","id":"2402.15220","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.15220","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"arxiv_version","alias_value":"2402.15220v4","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15220","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_12","alias_value":"3GWDRGDSWSUK","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_16","alias_value":"3GWDRGDSWSUKMPRO","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_8","alias_value":"3GWDRGDS","created_at":"2026-07-05T08:50:54Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:3GWDRGDSWSUKMPROHMEA6JX73V","target":"record","payload":{"canonical_record":{"source":{"id":"2402.15220","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T09:29:19Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"18e0f512eee20ff73bacd1ab420b377f3a43c0e666d13cc57db0bb76be15c9d9","abstract_canon_sha256":"f8c7e03602fc4c92ec21fee827fbf7a7f3ec5bcc9f1d4c24785496076435b02a"},"schema_version":"1.0"},"canonical_sha256":"d9ac389872b4a8a63e2e3b080f26ffdd7115d37b3085f6d06986b70d06ce0f60","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:50:54.104308Z","signature_b64":"KhOpCmCG+fhl7+qONh3HiVget7NBOLS7bg9AN7A+tG3vxoy5cC0zuxdmaRgTJ1d/Zxon9OgUGB3vCb25UBwfAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d9ac389872b4a8a63e2e3b080f26ffdd7115d37b3085f6d06986b70d06ce0f60","last_reissued_at":"2026-07-05T08:50:54.103899Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:50:54.103899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.15220","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:50:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uaOcdCmix8nlQhibEBRZCi6nD/CApiO2R2EJGeRA75ruwwdnsZDSZDKTfrMa+PLNBRvlZBdoGKGTYN10EV/wDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T10:56:30.047702Z"},"content_sha256":"879eae20506750eba8dbf245479f16da48fcc46204d5278cf3b376e14a0de854","schema_version":"1.0","event_id":"sha256:879eae20506750eba8dbf245479f16da48fcc46204d5278cf3b376e14a0de854"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:3GWDRGDSWSUKMPROHMEA6JX73V","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"ChunkAttention: Efficient Self-Attention with Prefix-Aware KV Cache and Two-Phase Partition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Lu Ye, Yang Li, Yong Huang, Ze Tao","submitted_at":"2024-02-23T09:29:19Z","abstract_excerpt":"Self-attention is an essential component of large language models (LLM) but a significant source of inference latency for long sequences. In multi-tenant LLM serving scenarios, the compute and memory operation cost of self-attention can be optimized by using the probability that multiple LLM requests have shared system prompts in prefixes. In this paper, we introduce ChunkAttention, a prefix-aware self-attention module that can detect matching prompt prefixes across multiple requests and share their key/value tensors in memory at runtime to improve the memory utilization of KV cache. This is a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15220","kind":"arxiv","version":4},"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/2402.15220/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":"2026-07-05T08:50:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8CbLtiPidFEdLBy/SRRHCOkL80sSU4UN7AEuiNsXg6BUXhDmTgUsOyNmBtH8sIy7/SvyekGCyDDTk2avlMsjDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T10:56:30.048623Z"},"content_sha256":"93f90d2aa6260eb91a8f19214d98ad3797d809ce56aac85e79c2ef9049d9e33d","schema_version":"1.0","event_id":"sha256:93f90d2aa6260eb91a8f19214d98ad3797d809ce56aac85e79c2ef9049d9e33d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3GWDRGDSWSUKMPROHMEA6JX73V/bundle.json","state_url":"https://pith.science/pith/3GWDRGDSWSUKMPROHMEA6JX73V/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3GWDRGDSWSUKMPROHMEA6JX73V/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":"2026-08-06T10:56:30Z","links":{"resolver":"https://pith.science/pith/3GWDRGDSWSUKMPROHMEA6JX73V","bundle":"https://pith.science/pith/3GWDRGDSWSUKMPROHMEA6JX73V/bundle.json","state":"https://pith.science/pith/3GWDRGDSWSUKMPROHMEA6JX73V/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3GWDRGDSWSUKMPROHMEA6JX73V/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:3GWDRGDSWSUKMPROHMEA6JX73V","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":"f8c7e03602fc4c92ec21fee827fbf7a7f3ec5bcc9f1d4c24785496076435b02a","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T09:29:19Z","title_canon_sha256":"18e0f512eee20ff73bacd1ab420b377f3a43c0e666d13cc57db0bb76be15c9d9"},"schema_version":"1.0","source":{"id":"2402.15220","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.15220","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"arxiv_version","alias_value":"2402.15220v4","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15220","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_12","alias_value":"3GWDRGDSWSUK","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_16","alias_value":"3GWDRGDSWSUKMPRO","created_at":"2026-07-05T08:50:54Z"},{"alias_kind":"pith_short_8","alias_value":"3GWDRGDS","created_at":"2026-07-05T08:50:54Z"}],"graph_snapshots":[{"event_id":"sha256:93f90d2aa6260eb91a8f19214d98ad3797d809ce56aac85e79c2ef9049d9e33d","target":"graph","created_at":"2026-07-05T08:50:54Z","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/2402.15220/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Self-attention is an essential component of large language models (LLM) but a significant source of inference latency for long sequences. In multi-tenant LLM serving scenarios, the compute and memory operation cost of self-attention can be optimized by using the probability that multiple LLM requests have shared system prompts in prefixes. In this paper, we introduce ChunkAttention, a prefix-aware self-attention module that can detect matching prompt prefixes across multiple requests and share their key/value tensors in memory at runtime to improve the memory utilization of KV cache. This is a","authors_text":"Lu Ye, Yang Li, Yong Huang, Ze Tao","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T09:29:19Z","title":"ChunkAttention: Efficient Self-Attention with Prefix-Aware KV Cache and Two-Phase Partition"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15220","kind":"arxiv","version":4},"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":"sha256:879eae20506750eba8dbf245479f16da48fcc46204d5278cf3b376e14a0de854","target":"record","created_at":"2026-07-05T08:50:54Z","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":"f8c7e03602fc4c92ec21fee827fbf7a7f3ec5bcc9f1d4c24785496076435b02a","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T09:29:19Z","title_canon_sha256":"18e0f512eee20ff73bacd1ab420b377f3a43c0e666d13cc57db0bb76be15c9d9"},"schema_version":"1.0","source":{"id":"2402.15220","kind":"arxiv","version":4}},"canonical_sha256":"d9ac389872b4a8a63e2e3b080f26ffdd7115d37b3085f6d06986b70d06ce0f60","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d9ac389872b4a8a63e2e3b080f26ffdd7115d37b3085f6d06986b70d06ce0f60","first_computed_at":"2026-07-05T08:50:54.103899Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:50:54.103899Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KhOpCmCG+fhl7+qONh3HiVget7NBOLS7bg9AN7A+tG3vxoy5cC0zuxdmaRgTJ1d/Zxon9OgUGB3vCb25UBwfAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:50:54.104308Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.15220","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:879eae20506750eba8dbf245479f16da48fcc46204d5278cf3b376e14a0de854","sha256:93f90d2aa6260eb91a8f19214d98ad3797d809ce56aac85e79c2ef9049d9e33d"],"state_sha256":"0e0c6c96e7c666a034da7a4c2da61e340e4fa61dd74bdc43fec5b1f7d75f8720"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"W3vqmIvlpN5BcwPQMOkrC9oXdzJ/Mw3Esso9XvYUlrvT4m0gwVGeYbi/ZizWpa42M56NNuddtS5oy84WzAOGDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T10:56:30.055913Z","bundle_sha256":"6f5bfe8b97c263a592bf3b50394c0aa41715fca37be5bbc06949204cbfd558db"}}