{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UPD67XBGDNQA2ICAJBVGI74TRE","short_pith_number":"pith:UPD67XBG","schema_version":"1.0","canonical_sha256":"a3c7efdc261b600d2040486a647f9389330558f5d20c1bfd2066c18dc018935d","source":{"kind":"arxiv","id":"2606.07404","version":1},"attestation_state":"computed","paper":{"title":"Reversible Foundations: Training a 120B Sparse MoE through State-Preserving Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Rohan Shravan","submitted_at":"2026-06-05T15:48:42Z","abstract_excerpt":"This paper reports on training a hundred-billion-parameter sparse mixture of experts on a single eight-GPU node, end to end. LightningLM 0.1V is a recurrence-backbone language model family grown in four stages from a small dense seed, through a 5B and a 9B mixture of experts, to a 120B model with 460 routed experts under top-12 routing. Each larger model is grown from the trained weights of the smaller one; active parameters rise monotonically from 1.78B at the dense seed to 5.93B at 120B (about 5% of the 118.67B stored). The full lineage runs on single nodes, the larger stages at 8K context, "},"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":"2606.07404","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-05T15:48:42Z","cross_cats_sorted":[],"title_canon_sha256":"233726e2f888f7142a513e5e525459fbe3b544f329e6738f940febbab7391414","abstract_canon_sha256":"41c3c2aa942ae7be8ae0cc183c4414f2ed7a17712d2325c860f5bbd309335c5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:05:25.149209Z","signature_b64":"gsJyl+FPKccWv/Z9VNrAhQxkkTBnKLhRyhoARZZ6Jd3+o0qIQKmmDuI12/LCaTLjIrbpw+wTTpO+xKpUZkB5DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a3c7efdc261b600d2040486a647f9389330558f5d20c1bfd2066c18dc018935d","last_reissued_at":"2026-06-08T01:05:25.148513Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:05:25.148513Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reversible Foundations: Training a 120B Sparse MoE through State-Preserving Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Rohan Shravan","submitted_at":"2026-06-05T15:48:42Z","abstract_excerpt":"This paper reports on training a hundred-billion-parameter sparse mixture of experts on a single eight-GPU node, end to end. LightningLM 0.1V is a recurrence-backbone language model family grown in four stages from a small dense seed, through a 5B and a 9B mixture of experts, to a 120B model with 460 routed experts under top-12 routing. Each larger model is grown from the trained weights of the smaller one; active parameters rise monotonically from 1.78B at the dense seed to 5.93B at 120B (about 5% of the 118.67B stored). The full lineage runs on single nodes, the larger stages at 8K context, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07404","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/2606.07404/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":"2606.07404","created_at":"2026-06-08T01:05:25.148621+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07404v1","created_at":"2026-06-08T01:05:25.148621+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07404","created_at":"2026-06-08T01:05:25.148621+00:00"},{"alias_kind":"pith_short_12","alias_value":"UPD67XBGDNQA","created_at":"2026-06-08T01:05:25.148621+00:00"},{"alias_kind":"pith_short_16","alias_value":"UPD67XBGDNQA2ICA","created_at":"2026-06-08T01:05:25.148621+00:00"},{"alias_kind":"pith_short_8","alias_value":"UPD67XBG","created_at":"2026-06-08T01:05:25.148621+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE","json":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE.json","graph_json":"https://pith.science/api/pith-number/UPD67XBGDNQA2ICAJBVGI74TRE/graph.json","events_json":"https://pith.science/api/pith-number/UPD67XBGDNQA2ICAJBVGI74TRE/events.json","paper":"https://pith.science/paper/UPD67XBG"},"agent_actions":{"view_html":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE","download_json":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE.json","view_paper":"https://pith.science/paper/UPD67XBG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07404&json=true","fetch_graph":"https://pith.science/api/pith-number/UPD67XBGDNQA2ICAJBVGI74TRE/graph.json","fetch_events":"https://pith.science/api/pith-number/UPD67XBGDNQA2ICAJBVGI74TRE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE/action/storage_attestation","attest_author":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE/action/author_attestation","sign_citation":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE/action/citation_signature","submit_replication":"https://pith.science/pith/UPD67XBGDNQA2ICAJBVGI74TRE/action/replication_record"}},"created_at":"2026-06-08T01:05:25.148621+00:00","updated_at":"2026-06-08T01:05:25.148621+00:00"}