{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MJZLP5RZV6NPWPBQ3VKEIZTSIT","short_pith_number":"pith:MJZLP5RZ","schema_version":"1.0","canonical_sha256":"6272b7f639af9afb3c30dd5444667244f004b44eb1637484857cfa1840f19155","source":{"kind":"arxiv","id":"2501.10661","version":1},"attestation_state":"computed","paper":{"title":"Unveiling the Mystery of Weight in Large Foundation Models: Gaussian Distribution Never Fades","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Chongjie Si, Jingjing Jiang, Wei Shen","submitted_at":"2025-01-18T05:43:17Z","abstract_excerpt":"This paper presents a pioneering exploration of the mechanisms underlying large foundation models' (LFMs) weights, aiming to simplify AI research. Through extensive observation and analysis on prevailing LFMs, we find that regardless of initialization strategies, their weights predominantly follow a Gaussian distribution, with occasional sharp, inverted T-shaped, or linear patterns. We further discover that the weights share the i.i.d. properties of Gaussian noise, and explore their direct relationship. We find that transformation weights can be derived from Gaussian noise, and they primarily "},"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":"2501.10661","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-18T05:43:17Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"87c1b2b5f3a6f1e8a9d957615b3c30d85b3a4676ce026d0e143e224fb52c730b","abstract_canon_sha256":"669ff3d995b1521fed5ddd60342920a1db3955b08326a6f1bf917e13d6b7d292"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:02:31.569194Z","signature_b64":"EY+1b82Q/VBF9IwgF2ghI35YXRXjm2Q3BudvMeqZdYvRdBju5Ymcizm0j92bN5enBJtLlMe7B4ydqvLCjv6wAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6272b7f639af9afb3c30dd5444667244f004b44eb1637484857cfa1840f19155","last_reissued_at":"2026-07-05T10:02:31.568672Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:02:31.568672Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unveiling the Mystery of Weight in Large Foundation Models: Gaussian Distribution Never Fades","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Chongjie Si, Jingjing Jiang, Wei Shen","submitted_at":"2025-01-18T05:43:17Z","abstract_excerpt":"This paper presents a pioneering exploration of the mechanisms underlying large foundation models' (LFMs) weights, aiming to simplify AI research. Through extensive observation and analysis on prevailing LFMs, we find that regardless of initialization strategies, their weights predominantly follow a Gaussian distribution, with occasional sharp, inverted T-shaped, or linear patterns. We further discover that the weights share the i.i.d. properties of Gaussian noise, and explore their direct relationship. We find that transformation weights can be derived from Gaussian noise, and they primarily "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.10661","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/2501.10661/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":"2501.10661","created_at":"2026-07-05T10:02:31.568747+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.10661v1","created_at":"2026-07-05T10:02:31.568747+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.10661","created_at":"2026-07-05T10:02:31.568747+00:00"},{"alias_kind":"pith_short_12","alias_value":"MJZLP5RZV6NP","created_at":"2026-07-05T10:02:31.568747+00:00"},{"alias_kind":"pith_short_16","alias_value":"MJZLP5RZV6NPWPBQ","created_at":"2026-07-05T10:02:31.568747+00:00"},{"alias_kind":"pith_short_8","alias_value":"MJZLP5RZ","created_at":"2026-07-05T10:02:31.568747+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.27844","citing_title":"ZipCCL: Efficient Lossless Data Compression of Communication Collectives for Accelerating LLM Training","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT","json":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT.json","graph_json":"https://pith.science/api/pith-number/MJZLP5RZV6NPWPBQ3VKEIZTSIT/graph.json","events_json":"https://pith.science/api/pith-number/MJZLP5RZV6NPWPBQ3VKEIZTSIT/events.json","paper":"https://pith.science/paper/MJZLP5RZ"},"agent_actions":{"view_html":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT","download_json":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT.json","view_paper":"https://pith.science/paper/MJZLP5RZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.10661&json=true","fetch_graph":"https://pith.science/api/pith-number/MJZLP5RZV6NPWPBQ3VKEIZTSIT/graph.json","fetch_events":"https://pith.science/api/pith-number/MJZLP5RZV6NPWPBQ3VKEIZTSIT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT/action/storage_attestation","attest_author":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT/action/author_attestation","sign_citation":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT/action/citation_signature","submit_replication":"https://pith.science/pith/MJZLP5RZV6NPWPBQ3VKEIZTSIT/action/replication_record"}},"created_at":"2026-07-05T10:02:31.568747+00:00","updated_at":"2026-07-05T10:02:31.568747+00:00"}