{"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"}