{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BGT73LMPV2BIP6EP5EQ247JERO","short_pith_number":"pith:BGT73LMP","schema_version":"1.0","canonical_sha256":"09a7fdad8fae8287f88fe921ae7d248baf738d86943286fb4c449847e7778903","source":{"kind":"arxiv","id":"2607.19771","version":1},"attestation_state":"computed","paper":{"title":"An Isotropy-Preserving Spectral Cap for Muon: Theory and Three Case Studies","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiachun Li","submitted_at":"2026-07-22T05:39:00Z","abstract_excerpt":"Muon and related matrix-sign optimizers are increasingly used to pre-train large language models, but their effect on the internal geometry of individual weight matrices is not well understood. This preliminary report proposes a unified framework built on a single idealizing assumption -- exact scale invariance of the loss under weight rescaling, which holds approximately in normalization-heavy networks. Under this assumption, plain SGD carries a built-in 1/||W|| brake on its update size, whereas Muon's matrix-sign step removes that brake, so both the Frobenius and spectral norms drift outward"},"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":"2607.19771","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-22T05:39:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7bcab4ebecab087c4661bfa83fb4408b75cf44645e1c274e35d22380463767c2","abstract_canon_sha256":"ca94a256633e1ed5820ad5bc6ffc44ec30df96dfe4ce3284026b1ed1be367d4e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-23T01:23:58.318450Z","signature_b64":"61L3uZCUM4q2cCr7do8ZnLTpNUbRuac8CR9S2J0gh/Rj6TNqQnlL3qY5Fzi3hwYB9wzuPBoUU/hVUW0WnOcPBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"09a7fdad8fae8287f88fe921ae7d248baf738d86943286fb4c449847e7778903","last_reissued_at":"2026-07-23T01:23:58.317540Z","signature_status":"signed_v1","first_computed_at":"2026-07-23T01:23:58.317540Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Isotropy-Preserving Spectral Cap for Muon: Theory and Three Case Studies","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiachun Li","submitted_at":"2026-07-22T05:39:00Z","abstract_excerpt":"Muon and related matrix-sign optimizers are increasingly used to pre-train large language models, but their effect on the internal geometry of individual weight matrices is not well understood. This preliminary report proposes a unified framework built on a single idealizing assumption -- exact scale invariance of the loss under weight rescaling, which holds approximately in normalization-heavy networks. Under this assumption, plain SGD carries a built-in 1/||W|| brake on its update size, whereas Muon's matrix-sign step removes that brake, so both the Frobenius and spectral norms drift outward"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.19771","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/2607.19771/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":"2607.19771","created_at":"2026-07-23T01:23:58.317959+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.19771v1","created_at":"2026-07-23T01:23:58.317959+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.19771","created_at":"2026-07-23T01:23:58.317959+00:00"},{"alias_kind":"pith_short_12","alias_value":"BGT73LMPV2BI","created_at":"2026-07-23T01:23:58.317959+00:00"},{"alias_kind":"pith_short_16","alias_value":"BGT73LMPV2BIP6EP","created_at":"2026-07-23T01:23:58.317959+00:00"},{"alias_kind":"pith_short_8","alias_value":"BGT73LMP","created_at":"2026-07-23T01:23:58.317959+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/BGT73LMPV2BIP6EP5EQ247JERO","json":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO.json","graph_json":"https://pith.science/api/pith-number/BGT73LMPV2BIP6EP5EQ247JERO/graph.json","events_json":"https://pith.science/api/pith-number/BGT73LMPV2BIP6EP5EQ247JERO/events.json","paper":"https://pith.science/paper/BGT73LMP"},"agent_actions":{"view_html":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO","download_json":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO.json","view_paper":"https://pith.science/paper/BGT73LMP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.19771&json=true","fetch_graph":"https://pith.science/api/pith-number/BGT73LMPV2BIP6EP5EQ247JERO/graph.json","fetch_events":"https://pith.science/api/pith-number/BGT73LMPV2BIP6EP5EQ247JERO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO/action/storage_attestation","attest_author":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO/action/author_attestation","sign_citation":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO/action/citation_signature","submit_replication":"https://pith.science/pith/BGT73LMPV2BIP6EP5EQ247JERO/action/replication_record"}},"created_at":"2026-07-23T01:23:58.317959+00:00","updated_at":"2026-07-23T01:23:58.317959+00:00"}