{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ZBV2RIF3MBGQLBNMM5W7PPTZM4","short_pith_number":"pith:ZBV2RIF3","canonical_record":{"source":{"id":"2406.07056","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-11T08:37:33Z","cross_cats_sorted":[],"title_canon_sha256":"c27470cbadc8e7396557ae4724b3500bfc3f0ddf97ed9d5079d4d1658f2f18de","abstract_canon_sha256":"2296e336ae95d5b185f148ba6c75af6f9da25efe9757c398cdfb3d9f19312503"},"schema_version":"1.0"},"canonical_sha256":"c86ba8a0bb604d0585ac676df7be7967015793015f64bd2c92cc1648eb7e4d37","source":{"kind":"arxiv","id":"2406.07056","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.07056","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"arxiv_version","alias_value":"2406.07056v1","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.07056","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_12","alias_value":"ZBV2RIF3MBGQ","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_16","alias_value":"ZBV2RIF3MBGQLBNM","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_8","alias_value":"ZBV2RIF3","created_at":"2026-07-05T08:30:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ZBV2RIF3MBGQLBNMM5W7PPTZM4","target":"record","payload":{"canonical_record":{"source":{"id":"2406.07056","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-11T08:37:33Z","cross_cats_sorted":[],"title_canon_sha256":"c27470cbadc8e7396557ae4724b3500bfc3f0ddf97ed9d5079d4d1658f2f18de","abstract_canon_sha256":"2296e336ae95d5b185f148ba6c75af6f9da25efe9757c398cdfb3d9f19312503"},"schema_version":"1.0"},"canonical_sha256":"c86ba8a0bb604d0585ac676df7be7967015793015f64bd2c92cc1648eb7e4d37","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:30:17.663804Z","signature_b64":"aDaQ8t57wL8ixxGK9iq7OGS/merz9kLhCAM24S7utQIRBEBT/iSs3vv2mCgZssfS1jCAzHD+m417zCFd/mNfAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c86ba8a0bb604d0585ac676df7be7967015793015f64bd2c92cc1648eb7e4d37","last_reissued_at":"2026-07-05T08:30:17.663254Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:30:17.663254Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.07056","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":"2026-07-05T08:30:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yDHr8J0413pLfKmenp8n0MwfLy5I+69tw1PJlG8Z7KwDacyZ+Uvm7hEJVjRxSgciyFacKKmexkoxSXhKTtS4Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T20:08:56.000588Z"},"content_sha256":"cbd80c98a55455ef8df070453fc251289bf046c4d479a52f2e46a93bcb0e8edd","schema_version":"1.0","event_id":"sha256:cbd80c98a55455ef8df070453fc251289bf046c4d479a52f2e46a93bcb0e8edd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ZBV2RIF3MBGQLBNMM5W7PPTZM4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Effectively Compress KV Heads for LLM","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hao Yu, Jianxin Wu, Shen Li, Yong Li, Zelan Yang","submitted_at":"2024-06-11T08:37:33Z","abstract_excerpt":"The advent of pre-trained large language models (LLMs) has revolutionized various natural language processing tasks. These models predominantly employ an auto-regressive decoding mechanism that utilizes Key-Value (KV) caches to eliminate redundant calculations for previous tokens. Nevertheless, as context lengths and batch sizes increase, the linear expansion in memory footprint of KV caches becomes a key bottleneck of LLM deployment, which decreases generation speeds significantly. To mitigate this issue, previous techniques like multi-query attention (MQA) and grouped-query attention (GQA) h"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.07056","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/2406.07056/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":"2026-07-05T08:30:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"p6VvHn6kqBn4QswFyWcWAQ2kibuKfW3MRBMAWdy/rEin8cec8h8AZ7LbfcUCh9mrs0VqUzlfeyC+h2uel5TVCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T20:08:56.001200Z"},"content_sha256":"93d33029cd89db9b222ba37b5e5a894187f48943faceb1af7a302060cd93c414","schema_version":"1.0","event_id":"sha256:93d33029cd89db9b222ba37b5e5a894187f48943faceb1af7a302060cd93c414"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/bundle.json","state_url":"https://pith.science/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/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":"2026-08-07T20:08:56Z","links":{"resolver":"https://pith.science/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4","bundle":"https://pith.science/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/bundle.json","state":"https://pith.science/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZBV2RIF3MBGQLBNMM5W7PPTZM4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ZBV2RIF3MBGQLBNMM5W7PPTZM4","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":"2296e336ae95d5b185f148ba6c75af6f9da25efe9757c398cdfb3d9f19312503","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-11T08:37:33Z","title_canon_sha256":"c27470cbadc8e7396557ae4724b3500bfc3f0ddf97ed9d5079d4d1658f2f18de"},"schema_version":"1.0","source":{"id":"2406.07056","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.07056","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"arxiv_version","alias_value":"2406.07056v1","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.07056","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_12","alias_value":"ZBV2RIF3MBGQ","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_16","alias_value":"ZBV2RIF3MBGQLBNM","created_at":"2026-07-05T08:30:17Z"},{"alias_kind":"pith_short_8","alias_value":"ZBV2RIF3","created_at":"2026-07-05T08:30:17Z"}],"graph_snapshots":[{"event_id":"sha256:93d33029cd89db9b222ba37b5e5a894187f48943faceb1af7a302060cd93c414","target":"graph","created_at":"2026-07-05T08:30:17Z","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/2406.07056/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The advent of pre-trained large language models (LLMs) has revolutionized various natural language processing tasks. These models predominantly employ an auto-regressive decoding mechanism that utilizes Key-Value (KV) caches to eliminate redundant calculations for previous tokens. Nevertheless, as context lengths and batch sizes increase, the linear expansion in memory footprint of KV caches becomes a key bottleneck of LLM deployment, which decreases generation speeds significantly. To mitigate this issue, previous techniques like multi-query attention (MQA) and grouped-query attention (GQA) h","authors_text":"Hao Yu, Jianxin Wu, Shen Li, Yong Li, Zelan Yang","cross_cats":[],"headline":"","license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-11T08:37:33Z","title":"Effectively Compress KV Heads for LLM"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.07056","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:cbd80c98a55455ef8df070453fc251289bf046c4d479a52f2e46a93bcb0e8edd","target":"record","created_at":"2026-07-05T08:30:17Z","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":"2296e336ae95d5b185f148ba6c75af6f9da25efe9757c398cdfb3d9f19312503","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-11T08:37:33Z","title_canon_sha256":"c27470cbadc8e7396557ae4724b3500bfc3f0ddf97ed9d5079d4d1658f2f18de"},"schema_version":"1.0","source":{"id":"2406.07056","kind":"arxiv","version":1}},"canonical_sha256":"c86ba8a0bb604d0585ac676df7be7967015793015f64bd2c92cc1648eb7e4d37","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c86ba8a0bb604d0585ac676df7be7967015793015f64bd2c92cc1648eb7e4d37","first_computed_at":"2026-07-05T08:30:17.663254Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:30:17.663254Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"aDaQ8t57wL8ixxGK9iq7OGS/merz9kLhCAM24S7utQIRBEBT/iSs3vv2mCgZssfS1jCAzHD+m417zCFd/mNfAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:30:17.663804Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.07056","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cbd80c98a55455ef8df070453fc251289bf046c4d479a52f2e46a93bcb0e8edd","sha256:93d33029cd89db9b222ba37b5e5a894187f48943faceb1af7a302060cd93c414"],"state_sha256":"d35fcc3088eb839595bc1600f36696ece76317648cbbe4460a191ead6ee20170"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dE/16CSKNELmeZt86O4DU+Nv2BcVasumlHhDdgqIs+7kzoozJG36nEAifdLy2Tm8IYljpD03DtfDySejJm9RCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T20:08:56.006205Z","bundle_sha256":"3aec06b6f014accb1e64c706b634fe72bf491d05bc4611faef4989e8f925a354"}}