{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SVIKAMOO7W4DCIQPLFXFHE7BGM","short_pith_number":"pith:SVIKAMOO","schema_version":"1.0","canonical_sha256":"9550a031cefdb831220f596e5393e1332195fe8277a7210e477053ac6186cbb9","source":{"kind":"arxiv","id":"2410.03529","version":2},"attestation_state":"computed","paper":{"title":"No Need to Talk: Asynchronous Mixture of Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Anastasiia Filippova, Angelos Katharopoulos, David Grangier, Ronan Collobert","submitted_at":"2024-10-04T15:50:10Z","abstract_excerpt":"We introduce SMALLTALK LM, an innovative method for training a mixture of language models in an almost asynchronous manner. Each model of the mixture specializes in distinct parts of the data distribution, without the need for high-bandwidth communication between the nodes training each model. At inference, a lightweight router directs a given sequence to a single expert, according to a short prefix. This inference scheme naturally uses a fraction of the parameters from the overall mixture model. Unlike prior works on asynchronous LLM training, our routing method does not rely on full corpus c"},"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":"2410.03529","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-04T15:50:10Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"8a671275919d2b9bc64f5e7f0b26aab55bc2d25782f86cac2a3408f3b3026630","abstract_canon_sha256":"7c2b065f259dbca80590394e627b82175286e2443e7a0a7a53a24a6a6fcef66a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:47.106167Z","signature_b64":"tu0bi7gMQog/wTFMmaY19d/yrycW9qUeDY2H7hImI9CNIFwIM3Q+hbNx61/NFHvtgD82nlzoNpzndQAf8OzbAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9550a031cefdb831220f596e5393e1332195fe8277a7210e477053ac6186cbb9","last_reissued_at":"2026-07-05T10:49:47.105545Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:47.105545Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"No Need to Talk: Asynchronous Mixture of Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Anastasiia Filippova, Angelos Katharopoulos, David Grangier, Ronan Collobert","submitted_at":"2024-10-04T15:50:10Z","abstract_excerpt":"We introduce SMALLTALK LM, an innovative method for training a mixture of language models in an almost asynchronous manner. Each model of the mixture specializes in distinct parts of the data distribution, without the need for high-bandwidth communication between the nodes training each model. At inference, a lightweight router directs a given sequence to a single expert, according to a short prefix. This inference scheme naturally uses a fraction of the parameters from the overall mixture model. Unlike prior works on asynchronous LLM training, our routing method does not rely on full corpus c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.03529","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/2410.03529/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":"2410.03529","created_at":"2026-07-05T10:49:47.105616+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.03529v2","created_at":"2026-07-05T10:49:47.105616+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.03529","created_at":"2026-07-05T10:49:47.105616+00:00"},{"alias_kind":"pith_short_12","alias_value":"SVIKAMOO7W4D","created_at":"2026-07-05T10:49:47.105616+00:00"},{"alias_kind":"pith_short_16","alias_value":"SVIKAMOO7W4DCIQP","created_at":"2026-07-05T10:49:47.105616+00:00"},{"alias_kind":"pith_short_8","alias_value":"SVIKAMOO","created_at":"2026-07-05T10:49:47.105616+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.00075","citing_title":"BTS: Harmonizing Specialized Experts into a Generalist LLM","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM","json":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM.json","graph_json":"https://pith.science/api/pith-number/SVIKAMOO7W4DCIQPLFXFHE7BGM/graph.json","events_json":"https://pith.science/api/pith-number/SVIKAMOO7W4DCIQPLFXFHE7BGM/events.json","paper":"https://pith.science/paper/SVIKAMOO"},"agent_actions":{"view_html":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM","download_json":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM.json","view_paper":"https://pith.science/paper/SVIKAMOO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.03529&json=true","fetch_graph":"https://pith.science/api/pith-number/SVIKAMOO7W4DCIQPLFXFHE7BGM/graph.json","fetch_events":"https://pith.science/api/pith-number/SVIKAMOO7W4DCIQPLFXFHE7BGM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM/action/storage_attestation","attest_author":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM/action/author_attestation","sign_citation":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM/action/citation_signature","submit_replication":"https://pith.science/pith/SVIKAMOO7W4DCIQPLFXFHE7BGM/action/replication_record"}},"created_at":"2026-07-05T10:49:47.105616+00:00","updated_at":"2026-07-05T10:49:47.105616+00:00"}