{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:WQ4Z65EXM6SQBBFGIZOZQOLS5K","short_pith_number":"pith:WQ4Z65EX","canonical_record":{"source":{"id":"2412.20677","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-30T03:05:45Z","cross_cats_sorted":[],"title_canon_sha256":"b87a688ebcc995dbb4622e66760063c0fd7329470e0a45b870702e001d8f35a4","abstract_canon_sha256":"dbe04b4c24319f5f9b77988f7c5916326ad433291fe7754e264d7dbf3819474d"},"schema_version":"1.0"},"canonical_sha256":"b4399f749767a50084a6465d983972ea802df571e997b915928eac274b8223ef","source":{"kind":"arxiv","id":"2412.20677","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.20677","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"arxiv_version","alias_value":"2412.20677v2","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.20677","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_12","alias_value":"WQ4Z65EXM6SQ","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_16","alias_value":"WQ4Z65EXM6SQBBFG","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_8","alias_value":"WQ4Z65EX","created_at":"2026-07-05T11:43:30Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:WQ4Z65EXM6SQBBFGIZOZQOLS5K","target":"record","payload":{"canonical_record":{"source":{"id":"2412.20677","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-30T03:05:45Z","cross_cats_sorted":[],"title_canon_sha256":"b87a688ebcc995dbb4622e66760063c0fd7329470e0a45b870702e001d8f35a4","abstract_canon_sha256":"dbe04b4c24319f5f9b77988f7c5916326ad433291fe7754e264d7dbf3819474d"},"schema_version":"1.0"},"canonical_sha256":"b4399f749767a50084a6465d983972ea802df571e997b915928eac274b8223ef","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:43:30.364664Z","signature_b64":"CVxFFzS74WvQtC9N1cOR4aUwYOS51b3waHpNQfm1DHKtPgnaQPmGQE9LOAqJsG6NjVbsft9PM6jYEVJRqqzQAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b4399f749767a50084a6465d983972ea802df571e997b915928eac274b8223ef","last_reissued_at":"2026-07-05T11:43:30.364013Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:43:30.364013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2412.20677","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":"2026-07-05T11:43:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xzeJH8qulYTQEdo8/MZNFC5ggkH3xRh4wMvWa7fsrrpf6DKxzp00vuEW0hVpMgP22gduiPcXGc7rqtwf/q6vCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T04:11:59.995282Z"},"content_sha256":"0e75b2e2a7f683438f5f323faee85eb743e779dc3a49463c1628094eb234e4cc","schema_version":"1.0","event_id":"sha256:0e75b2e2a7f683438f5f323faee85eb743e779dc3a49463c1628094eb234e4cc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:WQ4Z65EXM6SQBBFGIZOZQOLS5K","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Align Attention Heads Before Merging Them: An Effective Way for Converting MHA to GQA","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Feng Zhou, Qingyun Jin, Xiaohui Song, Zengchang Qin","submitted_at":"2024-12-30T03:05:45Z","abstract_excerpt":"Large language models (LLMs) have demonstrated exceptional performance across diverse natural language processing tasks. However, as the model size and the input sequence's length increase, the linearly increasing key-value (KV) cache significantly degrades inference throughput. Therefore, grouped-query attention (GQA), as an alternative to multi-head attention (MHA), has been widely introduced into LLMs. In this work, we propose a cost-effective method for converting MHA into GQA with any compression ratio of KV heads. The key point of our method lies in the application of Procrustes analysis"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.20677","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/2412.20677/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":"2026-07-05T11:43:30Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9m2XJMT078yJE09duxrN5PWnkACrhaMeyY58oXdHIRqnsCKa0r7WEfx0BgO+K8XvVJsLqZ752IAkiL+1LkKyBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T04:11:59.995962Z"},"content_sha256":"82b7cb70b50ef0c9f90332a7dce6ddc46428dd2b5c5524e333dd909782ce80e1","schema_version":"1.0","event_id":"sha256:82b7cb70b50ef0c9f90332a7dce6ddc46428dd2b5c5524e333dd909782ce80e1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/bundle.json","state_url":"https://pith.science/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/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":"2026-08-13T04:11:59Z","links":{"resolver":"https://pith.science/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K","bundle":"https://pith.science/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/bundle.json","state":"https://pith.science/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WQ4Z65EXM6SQBBFGIZOZQOLS5K/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:WQ4Z65EXM6SQBBFGIZOZQOLS5K","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":"dbe04b4c24319f5f9b77988f7c5916326ad433291fe7754e264d7dbf3819474d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-30T03:05:45Z","title_canon_sha256":"b87a688ebcc995dbb4622e66760063c0fd7329470e0a45b870702e001d8f35a4"},"schema_version":"1.0","source":{"id":"2412.20677","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.20677","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"arxiv_version","alias_value":"2412.20677v2","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.20677","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_12","alias_value":"WQ4Z65EXM6SQ","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_16","alias_value":"WQ4Z65EXM6SQBBFG","created_at":"2026-07-05T11:43:30Z"},{"alias_kind":"pith_short_8","alias_value":"WQ4Z65EX","created_at":"2026-07-05T11:43:30Z"}],"graph_snapshots":[{"event_id":"sha256:82b7cb70b50ef0c9f90332a7dce6ddc46428dd2b5c5524e333dd909782ce80e1","target":"graph","created_at":"2026-07-05T11:43:30Z","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/2412.20677/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) have demonstrated exceptional performance across diverse natural language processing tasks. However, as the model size and the input sequence's length increase, the linearly increasing key-value (KV) cache significantly degrades inference throughput. Therefore, grouped-query attention (GQA), as an alternative to multi-head attention (MHA), has been widely introduced into LLMs. In this work, we propose a cost-effective method for converting MHA into GQA with any compression ratio of KV heads. The key point of our method lies in the application of Procrustes analysis","authors_text":"Feng Zhou, Qingyun Jin, Xiaohui Song, Zengchang Qin","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-30T03:05:45Z","title":"Align Attention Heads Before Merging Them: An Effective Way for Converting MHA to GQA"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.20677","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":"sha256:0e75b2e2a7f683438f5f323faee85eb743e779dc3a49463c1628094eb234e4cc","target":"record","created_at":"2026-07-05T11:43:30Z","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":"dbe04b4c24319f5f9b77988f7c5916326ad433291fe7754e264d7dbf3819474d","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-30T03:05:45Z","title_canon_sha256":"b87a688ebcc995dbb4622e66760063c0fd7329470e0a45b870702e001d8f35a4"},"schema_version":"1.0","source":{"id":"2412.20677","kind":"arxiv","version":2}},"canonical_sha256":"b4399f749767a50084a6465d983972ea802df571e997b915928eac274b8223ef","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b4399f749767a50084a6465d983972ea802df571e997b915928eac274b8223ef","first_computed_at":"2026-07-05T11:43:30.364013Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:43:30.364013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CVxFFzS74WvQtC9N1cOR4aUwYOS51b3waHpNQfm1DHKtPgnaQPmGQE9LOAqJsG6NjVbsft9PM6jYEVJRqqzQAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:43:30.364664Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.20677","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0e75b2e2a7f683438f5f323faee85eb743e779dc3a49463c1628094eb234e4cc","sha256:82b7cb70b50ef0c9f90332a7dce6ddc46428dd2b5c5524e333dd909782ce80e1"],"state_sha256":"522ff87b6f2261357872129e69fd5dab6f2fa4085bfaba84a34fbcfd0bea6c3e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QRXNRqibJJvfQ4Ey92w2gkWU4RzVWIUUkok58EFUH4+r+gaWwVcYzCZrnnfashE8J6FnaINt70139uHLiPIjCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-13T04:11:59.999927Z","bundle_sha256":"6b5fb2e444cbc680cf422e402f9c486fbf8421f47b2887f3c1f1e9c8a192a8a7"}}