{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ZJYYSL2BJ6FBGARBVTONBI6ELU","short_pith_number":"pith:ZJYYSL2B","canonical_record":{"source":{"id":"2403.09054","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-14T02:42:42Z","cross_cats_sorted":["cs.AI","cs.AR","cs.CL"],"title_canon_sha256":"ce98a54bb56906753337f3ee649fcf20639114da7f3080347db17751538f2ae4","abstract_canon_sha256":"9f851e67b5433c91e7dd61d734219680d12c1e226a5363ba0cd26410bc05e539"},"schema_version":"1.0"},"canonical_sha256":"ca71892f414f8a130221acdcd0a3c45d26e9fe9bebe07bda3028478463e0efd3","source":{"kind":"arxiv","id":"2403.09054","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.09054","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"arxiv_version","alias_value":"2403.09054v2","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09054","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_12","alias_value":"ZJYYSL2BJ6FB","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_16","alias_value":"ZJYYSL2BJ6FBGARB","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_8","alias_value":"ZJYYSL2B","created_at":"2026-07-05T08:05:11Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ZJYYSL2BJ6FBGARBVTONBI6ELU","target":"record","payload":{"canonical_record":{"source":{"id":"2403.09054","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-14T02:42:42Z","cross_cats_sorted":["cs.AI","cs.AR","cs.CL"],"title_canon_sha256":"ce98a54bb56906753337f3ee649fcf20639114da7f3080347db17751538f2ae4","abstract_canon_sha256":"9f851e67b5433c91e7dd61d734219680d12c1e226a5363ba0cd26410bc05e539"},"schema_version":"1.0"},"canonical_sha256":"ca71892f414f8a130221acdcd0a3c45d26e9fe9bebe07bda3028478463e0efd3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:05:11.941845Z","signature_b64":"rJdQaG1/tlrIrFdUXhQkUQrY70sTqPS0cpdEioxbYPKtJmsLF9eDqiG9aTLOu5sb13nFuYDoknSBGB8wS0F4AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ca71892f414f8a130221acdcd0a3c45d26e9fe9bebe07bda3028478463e0efd3","last_reissued_at":"2026-07-05T08:05:11.941348Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:05:11.941348Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.09054","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":"2026-07-05T08:05:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y2Y+Alpu72UJHlNjEIIzfKybZc95JX+IlxxbbBAZNbN8rwe2Fr4HjAY6IYJkQkJ5KxL3dUQvptZ/AqeTNns+Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T18:41:54.948971Z"},"content_sha256":"8eb0f3d96b3e134df049a42c2d9f17e8df5279b8eb4a0dbf18680102c97a6dd2","schema_version":"1.0","event_id":"sha256:8eb0f3d96b3e134df049a42c2d9f17e8df5279b8eb4a0dbf18680102c97a6dd2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ZJYYSL2BJ6FBGARBVTONBI6ELU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Keyformer: KV Cache Reduction through Key Tokens Selection for Efficient Generative Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.AR","cs.CL"],"primary_cat":"cs.LG","authors_text":"Akhil Arunkumar, Gaurav Jain, Ilya Soloveychik, Muhammad Adnan, Prashant J. Nair, Purushotham Kamath","submitted_at":"2024-03-14T02:42:42Z","abstract_excerpt":"Transformers have emerged as the underpinning architecture for Large Language Models (LLMs). In generative language models, the inference process involves two primary phases: prompt processing and token generation. Token generation, which constitutes the majority of the computational workload, primarily entails vector-matrix multiplications and interactions with the Key-Value (KV) Cache. This phase is constrained by memory bandwidth due to the overhead of transferring weights and KV cache values from the memory system to the computing units. This memory bottleneck becomes particularly pronounc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09054","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/2403.09054/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":"2026-07-05T08:05:11Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4cQauT/NbjX+cvuH3CnBVK7TZrSj+rTKPz8Zl3PTnz23/gMl6+N6fpOZDMx4CkFEU+He/VRrYZjibz/jrbBeBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T18:41:54.949363Z"},"content_sha256":"723d57bf452eb5f2805ce02f8bae308a10e300845f7d98df9842cfa258681b53","schema_version":"1.0","event_id":"sha256:723d57bf452eb5f2805ce02f8bae308a10e300845f7d98df9842cfa258681b53"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/bundle.json","state_url":"https://pith.science/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/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":"2026-07-06T18:41:54Z","links":{"resolver":"https://pith.science/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU","bundle":"https://pith.science/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/bundle.json","state":"https://pith.science/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZJYYSL2BJ6FBGARBVTONBI6ELU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ZJYYSL2BJ6FBGARBVTONBI6ELU","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":"9f851e67b5433c91e7dd61d734219680d12c1e226a5363ba0cd26410bc05e539","cross_cats_sorted":["cs.AI","cs.AR","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-14T02:42:42Z","title_canon_sha256":"ce98a54bb56906753337f3ee649fcf20639114da7f3080347db17751538f2ae4"},"schema_version":"1.0","source":{"id":"2403.09054","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.09054","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"arxiv_version","alias_value":"2403.09054v2","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.09054","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_12","alias_value":"ZJYYSL2BJ6FB","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_16","alias_value":"ZJYYSL2BJ6FBGARB","created_at":"2026-07-05T08:05:11Z"},{"alias_kind":"pith_short_8","alias_value":"ZJYYSL2B","created_at":"2026-07-05T08:05:11Z"}],"graph_snapshots":[{"event_id":"sha256:723d57bf452eb5f2805ce02f8bae308a10e300845f7d98df9842cfa258681b53","target":"graph","created_at":"2026-07-05T08:05:11Z","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/2403.09054/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformers have emerged as the underpinning architecture for Large Language Models (LLMs). In generative language models, the inference process involves two primary phases: prompt processing and token generation. Token generation, which constitutes the majority of the computational workload, primarily entails vector-matrix multiplications and interactions with the Key-Value (KV) Cache. This phase is constrained by memory bandwidth due to the overhead of transferring weights and KV cache values from the memory system to the computing units. This memory bottleneck becomes particularly pronounc","authors_text":"Akhil Arunkumar, Gaurav Jain, Ilya Soloveychik, Muhammad Adnan, Prashant J. Nair, Purushotham Kamath","cross_cats":["cs.AI","cs.AR","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-14T02:42:42Z","title":"Keyformer: KV Cache Reduction through Key Tokens Selection for Efficient Generative Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.09054","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":"sha256:8eb0f3d96b3e134df049a42c2d9f17e8df5279b8eb4a0dbf18680102c97a6dd2","target":"record","created_at":"2026-07-05T08:05:11Z","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":"9f851e67b5433c91e7dd61d734219680d12c1e226a5363ba0cd26410bc05e539","cross_cats_sorted":["cs.AI","cs.AR","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-03-14T02:42:42Z","title_canon_sha256":"ce98a54bb56906753337f3ee649fcf20639114da7f3080347db17751538f2ae4"},"schema_version":"1.0","source":{"id":"2403.09054","kind":"arxiv","version":2}},"canonical_sha256":"ca71892f414f8a130221acdcd0a3c45d26e9fe9bebe07bda3028478463e0efd3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ca71892f414f8a130221acdcd0a3c45d26e9fe9bebe07bda3028478463e0efd3","first_computed_at":"2026-07-05T08:05:11.941348Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:05:11.941348Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rJdQaG1/tlrIrFdUXhQkUQrY70sTqPS0cpdEioxbYPKtJmsLF9eDqiG9aTLOu5sb13nFuYDoknSBGB8wS0F4AA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:05:11.941845Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.09054","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8eb0f3d96b3e134df049a42c2d9f17e8df5279b8eb4a0dbf18680102c97a6dd2","sha256:723d57bf452eb5f2805ce02f8bae308a10e300845f7d98df9842cfa258681b53"],"state_sha256":"9693a83fd2e02f61e546c7ab47823f8bf417da308491eb23159e87f5c8d2528f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1pF5UglrariDA2s0huJQ3Y51hzWY1ca/AoitUzw20PBSi7S9EXVY8nnqITUlgYjxqZm9Pl7XdVS1BGSvrKKBDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T18:41:54.951393Z","bundle_sha256":"9d0e1bc041cb6c38ebc0b98a0fbba3254f2ae5b7a71ffd28957da094d36e9779"}}