{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MD6LH35VAF7XEOLQMSMDYO5ZSC","short_pith_number":"pith:MD6LH35V","schema_version":"1.0","canonical_sha256":"60fcb3efb5017f72397064983c3bb990b8310cf9c41d61202889f58a37565c53","source":{"kind":"arxiv","id":"2411.15242","version":1},"attestation_state":"computed","paper":{"title":"The Zamba2 Suite: Technical Report","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Anna Golubeva, Beren Millidge, James Whittington, Jonathan Pilault, Paolo Glorioso, Quentin Anthony, Vasudev Shyam, Yury Tokpanov","submitted_at":"2024-11-22T02:55:20Z","abstract_excerpt":"In this technical report, we present the Zamba2 series -- a suite of 1.2B, 2.7B, and 7.4B parameter hybrid Mamba2-transformer models that achieve state of the art performance against the leading open-weights models of their class, while achieving substantial gains in inference latency, throughput, and memory efficiency. The Zamba2 series builds upon our initial work with Zamba1-7B, optimizing its architecture, training and annealing datasets, and training for up to three trillion tokens. We provide open-source weights for all models of the Zamba2 series as well as instruction-tuned variants th"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2411.15242","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-22T02:55:20Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"4e15e61f6976f8f7afb033712d8fc41a44c781e8cb451503be5e018bbf679970","abstract_canon_sha256":"6088e4f4ed0235a2728de3d0aa60467b48cf434179c8037d48c89a0f7e2df5cc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:39:31.763249Z","signature_b64":"fW5jUPMz8oYp7QR4dYIfvsf/fTSmTsLkFFPq37GPOrN4GAHxXaPbSluTlFn7IZbK7Z8A3vqkhQkboy3mSw8sBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60fcb3efb5017f72397064983c3bb990b8310cf9c41d61202889f58a37565c53","last_reissued_at":"2026-07-05T09:39:31.762760Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:39:31.762760Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Zamba2 Suite: Technical Report","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Anna Golubeva, Beren Millidge, James Whittington, Jonathan Pilault, Paolo Glorioso, Quentin Anthony, Vasudev Shyam, Yury Tokpanov","submitted_at":"2024-11-22T02:55:20Z","abstract_excerpt":"In this technical report, we present the Zamba2 series -- a suite of 1.2B, 2.7B, and 7.4B parameter hybrid Mamba2-transformer models that achieve state of the art performance against the leading open-weights models of their class, while achieving substantial gains in inference latency, throughput, and memory efficiency. The Zamba2 series builds upon our initial work with Zamba1-7B, optimizing its architecture, training and annealing datasets, and training for up to three trillion tokens. We provide open-source weights for all models of the Zamba2 series as well as instruction-tuned variants th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.15242","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/2411.15242/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2411.15242","created_at":"2026-07-05T09:39:31.762816+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.15242v1","created_at":"2026-07-05T09:39:31.762816+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.15242","created_at":"2026-07-05T09:39:31.762816+00:00"},{"alias_kind":"pith_short_12","alias_value":"MD6LH35VAF7X","created_at":"2026-07-05T09:39:31.762816+00:00"},{"alias_kind":"pith_short_16","alias_value":"MD6LH35VAF7XEOLQ","created_at":"2026-07-05T09:39:31.762816+00:00"},{"alias_kind":"pith_short_8","alias_value":"MD6LH35V","created_at":"2026-07-05T09:39:31.762816+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24320","citing_title":"ZONOS2 Technical Report","ref_index":213,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03014","citing_title":"MOSAIC: Efficient Mixture-of-Agent Scheduling via Adaptive Aggregation and Inference Concurrency","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02332","citing_title":"Forget Attention: Importance-Aware Attention Is All You Need","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00390","citing_title":"Zamba2-VL Technical Report","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24320","citing_title":"ZONOS2 Technical Report","ref_index":213,"is_internal_anchor":false},{"citing_arxiv_id":"2503.08223","citing_title":"Will LLMs Scaling Hit the Wall? Breaking Barriers via Distributed Resources on Massive Edge Devices","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22416","citing_title":"Asymmetric Virtual Memory Paging for Hybrid Mamba-Transformer Inference","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2510.04800","citing_title":"Hybrid Architectures for Language Models: Systematic Analysis and Design Insights","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05066","citing_title":"The Impossibility Triangle of Long-Context Modeling","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05365","citing_title":"ZAYA1-8B Technical Report","ref_index":172,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22050","citing_title":"LayerBoost: Layer-Aware Attention Reduction for Efficient LLMs","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC","json":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC.json","graph_json":"https://pith.science/api/pith-number/MD6LH35VAF7XEOLQMSMDYO5ZSC/graph.json","events_json":"https://pith.science/api/pith-number/MD6LH35VAF7XEOLQMSMDYO5ZSC/events.json","paper":"https://pith.science/paper/MD6LH35V"},"agent_actions":{"view_html":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC","download_json":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC.json","view_paper":"https://pith.science/paper/MD6LH35V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.15242&json=true","fetch_graph":"https://pith.science/api/pith-number/MD6LH35VAF7XEOLQMSMDYO5ZSC/graph.json","fetch_events":"https://pith.science/api/pith-number/MD6LH35VAF7XEOLQMSMDYO5ZSC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC/action/storage_attestation","attest_author":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC/action/author_attestation","sign_citation":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC/action/citation_signature","submit_replication":"https://pith.science/pith/MD6LH35VAF7XEOLQMSMDYO5ZSC/action/replication_record"}},"created_at":"2026-07-05T09:39:31.762816+00:00","updated_at":"2026-07-05T09:39:31.762816+00:00"}