{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:5DH535BUCFZN575GH7OWVXWZU4","short_pith_number":"pith:5DH535BU","canonical_record":{"source":{"id":"2506.07311","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-08T22:59:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9556f2f26a688e1e28a48a1b9e2668d88d44adfcff030372da9ff25b90303a73","abstract_canon_sha256":"293eca1608a8c014ec843f5ceec19ce2312f746e03d0c2f358394aec166b3d74"},"schema_version":"1.0"},"canonical_sha256":"e8cfddf4341172deffa63fdd6aded9a705c35cd77c97c16932a38240a5685cae","source":{"kind":"arxiv","id":"2506.07311","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.07311","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"arxiv_version","alias_value":"2506.07311v1","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.07311","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_12","alias_value":"5DH535BUCFZN","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_16","alias_value":"5DH535BUCFZN575G","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_8","alias_value":"5DH535BU","created_at":"2026-07-05T11:18:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:5DH535BUCFZN575GH7OWVXWZU4","target":"record","payload":{"canonical_record":{"source":{"id":"2506.07311","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-08T22:59:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9556f2f26a688e1e28a48a1b9e2668d88d44adfcff030372da9ff25b90303a73","abstract_canon_sha256":"293eca1608a8c014ec843f5ceec19ce2312f746e03d0c2f358394aec166b3d74"},"schema_version":"1.0"},"canonical_sha256":"e8cfddf4341172deffa63fdd6aded9a705c35cd77c97c16932a38240a5685cae","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:24.851289Z","signature_b64":"aqIQ0aplx0vmuME4qQsKmkCDGAsXa4KK/MGxWPlw2QX4UYbzV8lrxeatupyWuXZR7hgeO241Mmaml/FG1s8YCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e8cfddf4341172deffa63fdd6aded9a705c35cd77c97c16932a38240a5685cae","last_reissued_at":"2026-07-05T11:18:24.850824Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:24.850824Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.07311","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:18:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4KYpgh7f35c1MFEItw5By/JPiaL2ueq80CbI6GWKoBwPow//ng1joWw3EPHzfz4OxvvxvojMmVyhp3g9ZBlGBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T01:54:14.185543Z"},"content_sha256":"b4711c9d8a91e8f4a07c347a6893a1e38acc0297430c0b7f2c69f0042e8d8ea9","schema_version":"1.0","event_id":"sha256:b4711c9d8a91e8f4a07c347a6893a1e38acc0297430c0b7f2c69f0042e8d8ea9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:5DH535BUCFZN575GH7OWVXWZU4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Paged Attention Meets FlexAttention: Unlocking Long-Context Efficiency in Deployed Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Antoni Viros i Martin, Herman Saini, Kaoutar El Maghraoui, Neil Dhillon, Thomas Joshi","submitted_at":"2025-06-08T22:59:20Z","abstract_excerpt":"Large Language Models (LLMs) encounter severe memory inefficiencies during long-context inference due to conventional handling of key-value (KV) caches. In this work, we introduce a novel integration of PagedAttention with PyTorch's FlexAttention, addressing internal fragmentation and inefficiencies associated with monolithic KV cache allocations. Implemented within IBM's Foundation Model Stack (FMS), our fused attention kernel efficiently gathers scattered KV data. Our benchmarks on an NVIDIA L4 GPU (24GB) demonstrate significantly reduced inference latency, growing only linearly (~2x) with s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.07311","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/2506.07311/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":"2026-07-05T11:18:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"C/pOXaMpSQpKQo2zPK0C223ZSrCI0fTFksONfnhj3w6kCeDQxHNCHItO7qMIi6bzeXdzChXCx/0dd8OEDLY7Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T01:54:14.186254Z"},"content_sha256":"d98382ebc4933f16eb1628fe92510418869dfe76a489df05a894b991a9e4d2ac","schema_version":"1.0","event_id":"sha256:d98382ebc4933f16eb1628fe92510418869dfe76a489df05a894b991a9e4d2ac"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5DH535BUCFZN575GH7OWVXWZU4/bundle.json","state_url":"https://pith.science/pith/5DH535BUCFZN575GH7OWVXWZU4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5DH535BUCFZN575GH7OWVXWZU4/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":"2026-08-11T01:54:14Z","links":{"resolver":"https://pith.science/pith/5DH535BUCFZN575GH7OWVXWZU4","bundle":"https://pith.science/pith/5DH535BUCFZN575GH7OWVXWZU4/bundle.json","state":"https://pith.science/pith/5DH535BUCFZN575GH7OWVXWZU4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5DH535BUCFZN575GH7OWVXWZU4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:5DH535BUCFZN575GH7OWVXWZU4","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":"293eca1608a8c014ec843f5ceec19ce2312f746e03d0c2f358394aec166b3d74","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-08T22:59:20Z","title_canon_sha256":"9556f2f26a688e1e28a48a1b9e2668d88d44adfcff030372da9ff25b90303a73"},"schema_version":"1.0","source":{"id":"2506.07311","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.07311","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"arxiv_version","alias_value":"2506.07311v1","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.07311","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_12","alias_value":"5DH535BUCFZN","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_16","alias_value":"5DH535BUCFZN575G","created_at":"2026-07-05T11:18:24Z"},{"alias_kind":"pith_short_8","alias_value":"5DH535BU","created_at":"2026-07-05T11:18:24Z"}],"graph_snapshots":[{"event_id":"sha256:d98382ebc4933f16eb1628fe92510418869dfe76a489df05a894b991a9e4d2ac","target":"graph","created_at":"2026-07-05T11:18:24Z","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/2506.07311/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) encounter severe memory inefficiencies during long-context inference due to conventional handling of key-value (KV) caches. In this work, we introduce a novel integration of PagedAttention with PyTorch's FlexAttention, addressing internal fragmentation and inefficiencies associated with monolithic KV cache allocations. Implemented within IBM's Foundation Model Stack (FMS), our fused attention kernel efficiently gathers scattered KV data. Our benchmarks on an NVIDIA L4 GPU (24GB) demonstrate significantly reduced inference latency, growing only linearly (~2x) with s","authors_text":"Antoni Viros i Martin, Herman Saini, Kaoutar El Maghraoui, Neil Dhillon, Thomas Joshi","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-08T22:59:20Z","title":"Paged Attention Meets FlexAttention: Unlocking Long-Context Efficiency in Deployed Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.07311","kind":"arxiv","version":1},"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:b4711c9d8a91e8f4a07c347a6893a1e38acc0297430c0b7f2c69f0042e8d8ea9","target":"record","created_at":"2026-07-05T11:18:24Z","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":"293eca1608a8c014ec843f5ceec19ce2312f746e03d0c2f358394aec166b3d74","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-08T22:59:20Z","title_canon_sha256":"9556f2f26a688e1e28a48a1b9e2668d88d44adfcff030372da9ff25b90303a73"},"schema_version":"1.0","source":{"id":"2506.07311","kind":"arxiv","version":1}},"canonical_sha256":"e8cfddf4341172deffa63fdd6aded9a705c35cd77c97c16932a38240a5685cae","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e8cfddf4341172deffa63fdd6aded9a705c35cd77c97c16932a38240a5685cae","first_computed_at":"2026-07-05T11:18:24.850824Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:18:24.850824Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"aqIQ0aplx0vmuME4qQsKmkCDGAsXa4KK/MGxWPlw2QX4UYbzV8lrxeatupyWuXZR7hgeO241Mmaml/FG1s8YCA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:18:24.851289Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.07311","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b4711c9d8a91e8f4a07c347a6893a1e38acc0297430c0b7f2c69f0042e8d8ea9","sha256:d98382ebc4933f16eb1628fe92510418869dfe76a489df05a894b991a9e4d2ac"],"state_sha256":"f898ac92815acf43b4833c1ed16539eb879c1bab76ffdbb1890b379b35829979"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eXWTal6VTGk28arGO2I0Li6x7N/G6q93rvL5VF6WWhdIj6wHwYGlGE9LXMLIsCvJBggjVFyXOc0VqSk9fqRrBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T01:54:14.191371Z","bundle_sha256":"84896ac200698e5db9f6173da69f427f5cf94a1b079123e3690cd5c6fe1ccc33"}}