{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5MNQVWWVPIG43I4A4PADNFLJJ2","short_pith_number":"pith:5MNQVWWV","schema_version":"1.0","canonical_sha256":"eb1b0adad57a0dcda380e3c03695694ea1e80934f6b03094d4abc158462bc2fc","source":{"kind":"arxiv","id":"2502.11646","version":3},"attestation_state":"computed","paper":{"title":"Hyper-SET: Designing Transformers via Hyperspherical Energy Minimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Difan Zou, Dong Xu, Yunzhe Hu","submitted_at":"2025-02-17T10:39:11Z","abstract_excerpt":"Transformer-based models have achieved remarkable success, but their core components, Transformer layers, are largely heuristics-driven and engineered from the bottom up, calling for a prototypical model with high interpretability and practical competence. To this end, we conceptualize a principled, top-down approach grounded in energy-based interpretation. Specifically, we formalize token dynamics as a joint maximum likelihood estimation on the hypersphere, featuring two properties: semantic alignment in the high-dimensional space and distributional uniformity in the low-dimensional space. By"},"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":"2502.11646","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-17T10:39:11Z","cross_cats_sorted":[],"title_canon_sha256":"b9093ca1baa2da6ef10887ee1e6363af7835ea6480bea61f42c25f58f5a28fed","abstract_canon_sha256":"b8cda0d99e581a2705cc3d1c582176d896f4496086af211d246439bfdfc12072"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:27.018381Z","signature_b64":"nd4hVGDt5lIPIYGGRWjedI9zGDCPbzTfsBGnYmsRhieofttaeYu7Gi/KhEtycBLDtakaDkSsL6asRoo7El+lBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb1b0adad57a0dcda380e3c03695694ea1e80934f6b03094d4abc158462bc2fc","last_reissued_at":"2026-07-05T11:12:27.017762Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:27.017762Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hyper-SET: Designing Transformers via Hyperspherical Energy Minimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Difan Zou, Dong Xu, Yunzhe Hu","submitted_at":"2025-02-17T10:39:11Z","abstract_excerpt":"Transformer-based models have achieved remarkable success, but their core components, Transformer layers, are largely heuristics-driven and engineered from the bottom up, calling for a prototypical model with high interpretability and practical competence. To this end, we conceptualize a principled, top-down approach grounded in energy-based interpretation. Specifically, we formalize token dynamics as a joint maximum likelihood estimation on the hypersphere, featuring two properties: semantic alignment in the high-dimensional space and distributional uniformity in the low-dimensional space. By"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.11646","kind":"arxiv","version":3},"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/2502.11646/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":"2502.11646","created_at":"2026-07-05T11:12:27.017832+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.11646v3","created_at":"2026-07-05T11:12:27.017832+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.11646","created_at":"2026-07-05T11:12:27.017832+00:00"},{"alias_kind":"pith_short_12","alias_value":"5MNQVWWVPIG4","created_at":"2026-07-05T11:12:27.017832+00:00"},{"alias_kind":"pith_short_16","alias_value":"5MNQVWWVPIG43I4A","created_at":"2026-07-05T11:12:27.017832+00:00"},{"alias_kind":"pith_short_8","alias_value":"5MNQVWWV","created_at":"2026-07-05T11:12:27.017832+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.13540","citing_title":"Provable Low-Frequency Bias of In-Context Learning of Representations","ref_index":2021,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2","json":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2.json","graph_json":"https://pith.science/api/pith-number/5MNQVWWVPIG43I4A4PADNFLJJ2/graph.json","events_json":"https://pith.science/api/pith-number/5MNQVWWVPIG43I4A4PADNFLJJ2/events.json","paper":"https://pith.science/paper/5MNQVWWV"},"agent_actions":{"view_html":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2","download_json":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2.json","view_paper":"https://pith.science/paper/5MNQVWWV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.11646&json=true","fetch_graph":"https://pith.science/api/pith-number/5MNQVWWVPIG43I4A4PADNFLJJ2/graph.json","fetch_events":"https://pith.science/api/pith-number/5MNQVWWVPIG43I4A4PADNFLJJ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2/action/storage_attestation","attest_author":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2/action/author_attestation","sign_citation":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2/action/citation_signature","submit_replication":"https://pith.science/pith/5MNQVWWVPIG43I4A4PADNFLJJ2/action/replication_record"}},"created_at":"2026-07-05T11:12:27.017832+00:00","updated_at":"2026-07-05T11:12:27.017832+00:00"}