{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:KFEIR6A4EPMDTKHEVP23BG4OMJ","short_pith_number":"pith:KFEIR6A4","canonical_record":{"source":{"id":"2404.15949","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T16:11:54Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"93af51c7e8e9536c9b0cb54f47e07d0466a547e76410cbfc0a9f9b2bef4214cb","abstract_canon_sha256":"bfa2d49d1330c5b8e9e9eae2a105c3dd3813d9738eb7b6f147430b005bffecb0"},"schema_version":"1.0"},"canonical_sha256":"514888f81c23d839a8e4abf5b09b8e625d82f034dd38923820a913d2705a2400","source":{"kind":"arxiv","id":"2404.15949","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.15949","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"arxiv_version","alias_value":"2404.15949v2","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.15949","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_12","alias_value":"KFEIR6A4EPMD","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_16","alias_value":"KFEIR6A4EPMDTKHE","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_8","alias_value":"KFEIR6A4","created_at":"2026-07-05T08:35:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:KFEIR6A4EPMDTKHEVP23BG4OMJ","target":"record","payload":{"canonical_record":{"source":{"id":"2404.15949","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T16:11:54Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"93af51c7e8e9536c9b0cb54f47e07d0466a547e76410cbfc0a9f9b2bef4214cb","abstract_canon_sha256":"bfa2d49d1330c5b8e9e9eae2a105c3dd3813d9738eb7b6f147430b005bffecb0"},"schema_version":"1.0"},"canonical_sha256":"514888f81c23d839a8e4abf5b09b8e625d82f034dd38923820a913d2705a2400","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:35:02.467782Z","signature_b64":"vnGSZSDMlZEQAvfEH3QlThs5xVieDlbCAGjjLNP0rUOYxkuht+79Ap2Zb016nfHFmzGfs/TB5zE7oeIMP5ZcDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"514888f81c23d839a8e4abf5b09b8e625d82f034dd38923820a913d2705a2400","last_reissued_at":"2026-07-05T08:35:02.467286Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:35:02.467286Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.15949","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":"2026-07-05T08:35:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gB6dyqTDeP4Q1xs0zVJyI65h8aXyy4mtIq82zvjyAq++RGq5uyOig0dp3auR9MRF3s4p0CZRzr3KcEufbp3iCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T07:15:36.051001Z"},"content_sha256":"f8ad387f5876092b453d1ef13b3f2fc2cb9b6281ee32ff28b5f787a48dc4b58c","schema_version":"1.0","event_id":"sha256:f8ad387f5876092b453d1ef13b3f2fc2cb9b6281ee32ff28b5f787a48dc4b58c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:KFEIR6A4EPMDTKHEVP23BG4OMJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CORM: Cache Optimization with Recent Message for Large Language Model Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chen Chen, Deng Cai, Haiyun Jiang, Jincheng Dai, Shuming Shi, Wei Bi, Zhuowei Huang","submitted_at":"2024-04-24T16:11:54Z","abstract_excerpt":"Large Language Models (LLMs), despite their remarkable performance across a wide range of tasks, necessitate substantial GPU memory and consume significant computational resources. Beyond the memory taken up by model weights, the memory used by the KV cache rises linearly with sequence length, becoming a primary bottleneck for inference. In this paper, we introduce an innovative method for optimizing the KV cache, which considerably minimizes its memory footprint. Upon thorough investigation, we discover that in most Transformer models, (i) there is a striking similarity between adjacent token"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.15949","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/2404.15949/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":"2026-07-05T08:35:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZUpCsn55hWQhOjCuSJnmfi1/MRd/6rCXJZaqcr3qOTNKS8rw692RbUDNhe7MK6DclPL/E/qRi/FqZqCdXfwUAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T07:15:36.051489Z"},"content_sha256":"dda73388ec327974d0e661c57fbf5118b901de73ab05d51a909685871800bc57","schema_version":"1.0","event_id":"sha256:dda73388ec327974d0e661c57fbf5118b901de73ab05d51a909685871800bc57"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/bundle.json","state_url":"https://pith.science/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/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":"2026-08-16T07:15:36Z","links":{"resolver":"https://pith.science/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ","bundle":"https://pith.science/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/bundle.json","state":"https://pith.science/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KFEIR6A4EPMDTKHEVP23BG4OMJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:KFEIR6A4EPMDTKHEVP23BG4OMJ","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":"bfa2d49d1330c5b8e9e9eae2a105c3dd3813d9738eb7b6f147430b005bffecb0","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T16:11:54Z","title_canon_sha256":"93af51c7e8e9536c9b0cb54f47e07d0466a547e76410cbfc0a9f9b2bef4214cb"},"schema_version":"1.0","source":{"id":"2404.15949","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.15949","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"arxiv_version","alias_value":"2404.15949v2","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.15949","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_12","alias_value":"KFEIR6A4EPMD","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_16","alias_value":"KFEIR6A4EPMDTKHE","created_at":"2026-07-05T08:35:02Z"},{"alias_kind":"pith_short_8","alias_value":"KFEIR6A4","created_at":"2026-07-05T08:35:02Z"}],"graph_snapshots":[{"event_id":"sha256:dda73388ec327974d0e661c57fbf5118b901de73ab05d51a909685871800bc57","target":"graph","created_at":"2026-07-05T08:35:02Z","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/2404.15949/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs), despite their remarkable performance across a wide range of tasks, necessitate substantial GPU memory and consume significant computational resources. Beyond the memory taken up by model weights, the memory used by the KV cache rises linearly with sequence length, becoming a primary bottleneck for inference. In this paper, we introduce an innovative method for optimizing the KV cache, which considerably minimizes its memory footprint. Upon thorough investigation, we discover that in most Transformer models, (i) there is a striking similarity between adjacent token","authors_text":"Chen Chen, Deng Cai, Haiyun Jiang, Jincheng Dai, Shuming Shi, Wei Bi, Zhuowei Huang","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T16:11:54Z","title":"CORM: Cache Optimization with Recent Message for Large Language Model Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.15949","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:f8ad387f5876092b453d1ef13b3f2fc2cb9b6281ee32ff28b5f787a48dc4b58c","target":"record","created_at":"2026-07-05T08:35:02Z","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":"bfa2d49d1330c5b8e9e9eae2a105c3dd3813d9738eb7b6f147430b005bffecb0","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-24T16:11:54Z","title_canon_sha256":"93af51c7e8e9536c9b0cb54f47e07d0466a547e76410cbfc0a9f9b2bef4214cb"},"schema_version":"1.0","source":{"id":"2404.15949","kind":"arxiv","version":2}},"canonical_sha256":"514888f81c23d839a8e4abf5b09b8e625d82f034dd38923820a913d2705a2400","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"514888f81c23d839a8e4abf5b09b8e625d82f034dd38923820a913d2705a2400","first_computed_at":"2026-07-05T08:35:02.467286Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:35:02.467286Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"vnGSZSDMlZEQAvfEH3QlThs5xVieDlbCAGjjLNP0rUOYxkuht+79Ap2Zb016nfHFmzGfs/TB5zE7oeIMP5ZcDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:35:02.467782Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.15949","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f8ad387f5876092b453d1ef13b3f2fc2cb9b6281ee32ff28b5f787a48dc4b58c","sha256:dda73388ec327974d0e661c57fbf5118b901de73ab05d51a909685871800bc57"],"state_sha256":"36b22ba6185fc9869847d45ff5e8fff7ab35565f124ab23f8dc4c64f44eb1a0d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YO4A5HXA4n14/3G3QpDPfOPCEYi6cEeYITUD1AcTSZIi+CL7+u48H2x9TdAb7I7ozJMjvh23pSTWliFQeJw6CQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-16T07:15:36.056317Z","bundle_sha256":"a1038c009b8945f48e0642714c0b23379018657425ccb148dff11638cfe00165"}}