{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:6NIHJWXXHOHV7DWN36PTOH4K5O","short_pith_number":"pith:6NIHJWXX","schema_version":"1.0","canonical_sha256":"f35074daf73b8f5f8ecddf9f371f8aeba576c01eec87cb16abc149b6637446c6","source":{"kind":"arxiv","id":"2411.06068","version":1},"attestation_state":"computed","paper":{"title":"Zyda-2: a 5 Trillion Token High-Quality Dataset","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Beren Millidge, Paolo Glorioso, Quentin Anthony, Yury Tokpanov","submitted_at":"2024-11-09T04:57:41Z","abstract_excerpt":"In this technical report, we present Zyda-2: a five trillion token dataset for language model pretraining. Zyda-2 was used to train our Zamba2 series of models which are state-of-the-art for their weight class. We build Zyda-2 by collating high-quality open-source tokens such as FineWeb and DCLM, then distilling them to the highest-quality subset via cross-deduplication and model-based quality filtering. Zyda-2 is released under a permissive open license, and is available at https://huggingface.co/datasets/Zyphra/Zyda-2"},"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":"2411.06068","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-09T04:57:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f5a5fc3f0a627db59cf02c728564d1665d6aaa1bab127ca53297e333575d5da9","abstract_canon_sha256":"1c65e8667a682357a3a60abec86dfb5db98bc6bde695dbcac93e16ec38f80a5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:33:16.518853Z","signature_b64":"3elKUQDdluNUW1QnDXnRPWntMVW3Gpq3eh35Kp1GVcvGv/qUpgK33r99AEY0Q3ITqWt5Fw5mJyWQFPVxA+IxAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f35074daf73b8f5f8ecddf9f371f8aeba576c01eec87cb16abc149b6637446c6","last_reissued_at":"2026-07-05T09:33:16.518164Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:33:16.518164Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Zyda-2: a 5 Trillion Token High-Quality Dataset","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Beren Millidge, Paolo Glorioso, Quentin Anthony, Yury Tokpanov","submitted_at":"2024-11-09T04:57:41Z","abstract_excerpt":"In this technical report, we present Zyda-2: a five trillion token dataset for language model pretraining. Zyda-2 was used to train our Zamba2 series of models which are state-of-the-art for their weight class. We build Zyda-2 by collating high-quality open-source tokens such as FineWeb and DCLM, then distilling them to the highest-quality subset via cross-deduplication and model-based quality filtering. Zyda-2 is released under a permissive open license, and is available at https://huggingface.co/datasets/Zyphra/Zyda-2"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.06068","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/2411.06068/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":"2411.06068","created_at":"2026-07-05T09:33:16.518306+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.06068v1","created_at":"2026-07-05T09:33:16.518306+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.06068","created_at":"2026-07-05T09:33:16.518306+00:00"},{"alias_kind":"pith_short_12","alias_value":"6NIHJWXXHOHV","created_at":"2026-07-05T09:33:16.518306+00:00"},{"alias_kind":"pith_short_16","alias_value":"6NIHJWXXHOHV7DWN","created_at":"2026-07-05T09:33:16.518306+00:00"},{"alias_kind":"pith_short_8","alias_value":"6NIHJWXX","created_at":"2026-07-05T09:33:16.518306+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27866","citing_title":"FlexMoE: One-for-All Nested Intra-Expert Pruning for MoE Language Models","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O","json":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O.json","graph_json":"https://pith.science/api/pith-number/6NIHJWXXHOHV7DWN36PTOH4K5O/graph.json","events_json":"https://pith.science/api/pith-number/6NIHJWXXHOHV7DWN36PTOH4K5O/events.json","paper":"https://pith.science/paper/6NIHJWXX"},"agent_actions":{"view_html":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O","download_json":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O.json","view_paper":"https://pith.science/paper/6NIHJWXX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.06068&json=true","fetch_graph":"https://pith.science/api/pith-number/6NIHJWXXHOHV7DWN36PTOH4K5O/graph.json","fetch_events":"https://pith.science/api/pith-number/6NIHJWXXHOHV7DWN36PTOH4K5O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O/action/storage_attestation","attest_author":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O/action/author_attestation","sign_citation":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O/action/citation_signature","submit_replication":"https://pith.science/pith/6NIHJWXXHOHV7DWN36PTOH4K5O/action/replication_record"}},"created_at":"2026-07-05T09:33:16.518306+00:00","updated_at":"2026-07-05T09:33:16.518306+00:00"}