{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:2ZOLRUPR6RLA3MNJGY2TVXZPTO","short_pith_number":"pith:2ZOLRUPR","schema_version":"1.0","canonical_sha256":"d65cb8d1f1f4560db1a936353adf2f9ba5323cf3f8758e4fe750778f5621fc85","source":{"kind":"arxiv","id":"2112.15121","version":2},"attestation_state":"computed","paper":{"title":"On the Role of Neural Collapse in Transfer Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andr\\'as Gy\\\"orgy, Marcus Hutter, Tomer Galanti","submitted_at":"2021-12-30T16:36:26Z","abstract_excerpt":"We study the ability of foundation models to learn representations for classification that are transferable to new, unseen classes. Recent results in the literature show that representations learned by a single classifier over many classes are competitive on few-shot learning problems with representations learned by special-purpose algorithms designed for such problems. In this paper we provide an explanation for this behavior based on the recently observed phenomenon that the features learned by overparameterized classification networks show an interesting clustering property, called neural 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":"2112.15121","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-12-30T16:36:26Z","cross_cats_sorted":[],"title_canon_sha256":"6c79c426115486e76c5e9372d25d12c2d01f998118cb687eaf36c2c17a882c7f","abstract_canon_sha256":"bbf7f31eba4d39b6d1ab5880956ee0dcd737f6d7eecc29314651c0b5d56df315"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:45:13.136591Z","signature_b64":"BFtTisabUtMnPSCu1H+wOoAUVWd4Dx20UTOUTtKC0n5FsQBllH7GoGSFg5RbT59bvk54CwaVcD7WT6WSp7Z/DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d65cb8d1f1f4560db1a936353adf2f9ba5323cf3f8758e4fe750778f5621fc85","last_reissued_at":"2026-07-05T03:45:13.136080Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:45:13.136080Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Role of Neural Collapse in Transfer Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andr\\'as Gy\\\"orgy, Marcus Hutter, Tomer Galanti","submitted_at":"2021-12-30T16:36:26Z","abstract_excerpt":"We study the ability of foundation models to learn representations for classification that are transferable to new, unseen classes. Recent results in the literature show that representations learned by a single classifier over many classes are competitive on few-shot learning problems with representations learned by special-purpose algorithms designed for such problems. In this paper we provide an explanation for this behavior based on the recently observed phenomenon that the features learned by overparameterized classification networks show an interesting clustering property, called neural c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.15121","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/2112.15121/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":"2112.15121","created_at":"2026-07-05T03:45:13.136148+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.15121v2","created_at":"2026-07-05T03:45:13.136148+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.15121","created_at":"2026-07-05T03:45:13.136148+00:00"},{"alias_kind":"pith_short_12","alias_value":"2ZOLRUPR6RLA","created_at":"2026-07-05T03:45:13.136148+00:00"},{"alias_kind":"pith_short_16","alias_value":"2ZOLRUPR6RLA3MNJ","created_at":"2026-07-05T03:45:13.136148+00:00"},{"alias_kind":"pith_short_8","alias_value":"2ZOLRUPR","created_at":"2026-07-05T03:45:13.136148+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":86,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":86,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23087","citing_title":"The Implicit Bias of Depth: From Neural Collapse to Softmax Codes","ref_index":115,"is_internal_anchor":false},{"citing_arxiv_id":"2510.01105","citing_title":"Geometric Analysis of Neural Regression Collapse via Intrinsic Dimension","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO","json":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO.json","graph_json":"https://pith.science/api/pith-number/2ZOLRUPR6RLA3MNJGY2TVXZPTO/graph.json","events_json":"https://pith.science/api/pith-number/2ZOLRUPR6RLA3MNJGY2TVXZPTO/events.json","paper":"https://pith.science/paper/2ZOLRUPR"},"agent_actions":{"view_html":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO","download_json":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO.json","view_paper":"https://pith.science/paper/2ZOLRUPR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.15121&json=true","fetch_graph":"https://pith.science/api/pith-number/2ZOLRUPR6RLA3MNJGY2TVXZPTO/graph.json","fetch_events":"https://pith.science/api/pith-number/2ZOLRUPR6RLA3MNJGY2TVXZPTO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO/action/storage_attestation","attest_author":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO/action/author_attestation","sign_citation":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO/action/citation_signature","submit_replication":"https://pith.science/pith/2ZOLRUPR6RLA3MNJGY2TVXZPTO/action/replication_record"}},"created_at":"2026-07-05T03:45:13.136148+00:00","updated_at":"2026-07-05T03:45:13.136148+00:00"}