{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:K4VSHN72IFQ65ZUVQA7O5G4I2O","short_pith_number":"pith:K4VSHN72","schema_version":"1.0","canonical_sha256":"572b23b7fa4161eee695803eee9b88d38119007a2439a560cc5cc835a3bc4bee","source":{"kind":"arxiv","id":"2404.07413","version":1},"attestation_state":"computed","paper":{"title":"JetMoE: Reaching Llama2 Performance with 0.1M Dollars","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Tianle Cai, Yikang Shen, Zengyi Qin, Zhen Guo","submitted_at":"2024-04-11T00:52:39Z","abstract_excerpt":"Large Language Models (LLMs) have achieved remarkable results, but their increasing resource demand has become a major obstacle to the development of powerful and accessible super-human intelligence. This report introduces JetMoE-8B, a new LLM trained with less than $0.1 million, using 1.25T tokens from carefully mixed open-source corpora and 30,000 H100 GPU hours. Despite its low cost, the JetMoE-8B demonstrates impressive performance, with JetMoE-8B outperforming the Llama2-7B model and JetMoE-8B-Chat surpassing the Llama2-13B-Chat model. These results suggest that LLM training can be much m"},"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":"2404.07413","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-11T00:52:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c3dbd2ab6dd35e3983bbe72e0702fcab9266c5b3850be7fd8b11920cb542d9d9","abstract_canon_sha256":"12119e5da4208dafc35a3012cf5ab93f20e31b6ba0e7e500f9db2a3e241946bb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:06:52.211023Z","signature_b64":"PN93Eis9K3jmCNki1uTWLwbc9Nu5RWIr+c7DFWkEhEOqjfKZwMAIjb+F7dPsbVfH64CPkocVjOGLSUjm0LNBAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"572b23b7fa4161eee695803eee9b88d38119007a2439a560cc5cc835a3bc4bee","last_reissued_at":"2026-07-05T08:06:52.210589Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:06:52.210589Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"JetMoE: Reaching Llama2 Performance with 0.1M Dollars","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Tianle Cai, Yikang Shen, Zengyi Qin, Zhen Guo","submitted_at":"2024-04-11T00:52:39Z","abstract_excerpt":"Large Language Models (LLMs) have achieved remarkable results, but their increasing resource demand has become a major obstacle to the development of powerful and accessible super-human intelligence. This report introduces JetMoE-8B, a new LLM trained with less than $0.1 million, using 1.25T tokens from carefully mixed open-source corpora and 30,000 H100 GPU hours. Despite its low cost, the JetMoE-8B demonstrates impressive performance, with JetMoE-8B outperforming the Llama2-7B model and JetMoE-8B-Chat surpassing the Llama2-13B-Chat model. These results suggest that LLM training can be much m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.07413","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/2404.07413/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":"2404.07413","created_at":"2026-07-05T08:06:52.210655+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.07413v1","created_at":"2026-07-05T08:06:52.210655+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.07413","created_at":"2026-07-05T08:06:52.210655+00:00"},{"alias_kind":"pith_short_12","alias_value":"K4VSHN72IFQ6","created_at":"2026-07-05T08:06:52.210655+00:00"},{"alias_kind":"pith_short_16","alias_value":"K4VSHN72IFQ65ZUV","created_at":"2026-07-05T08:06:52.210655+00:00"},{"alias_kind":"pith_short_8","alias_value":"K4VSHN72","created_at":"2026-07-05T08:06:52.210655+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25821","citing_title":"SARA: Unlocking Multilingual Knowledge in Mixture-of-Experts via Semantically Anchored Routing Alignment","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24320","citing_title":"ZONOS2 Technical Report","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24320","citing_title":"ZONOS2 Technical Report","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06501","citing_title":"Cubit: Token Mixer with Kernel Ridge Regression","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2411.04996","citing_title":"Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11537","citing_title":"Fast MoE Inference via Predictive Prefetching and Expert Replication","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08292","citing_title":"Hierarchical Mixture-of-Experts with Two-Stage Optimization","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06501","citing_title":"Cubit: Token Mixer with Kernel Ridge Regression","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05365","citing_title":"ZAYA1-8B Technical Report","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O","json":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O.json","graph_json":"https://pith.science/api/pith-number/K4VSHN72IFQ65ZUVQA7O5G4I2O/graph.json","events_json":"https://pith.science/api/pith-number/K4VSHN72IFQ65ZUVQA7O5G4I2O/events.json","paper":"https://pith.science/paper/K4VSHN72"},"agent_actions":{"view_html":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O","download_json":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O.json","view_paper":"https://pith.science/paper/K4VSHN72","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.07413&json=true","fetch_graph":"https://pith.science/api/pith-number/K4VSHN72IFQ65ZUVQA7O5G4I2O/graph.json","fetch_events":"https://pith.science/api/pith-number/K4VSHN72IFQ65ZUVQA7O5G4I2O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O/action/storage_attestation","attest_author":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O/action/author_attestation","sign_citation":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O/action/citation_signature","submit_replication":"https://pith.science/pith/K4VSHN72IFQ65ZUVQA7O5G4I2O/action/replication_record"}},"created_at":"2026-07-05T08:06:52.210655+00:00","updated_at":"2026-07-05T08:06:52.210655+00:00"}