{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:IUSWN3DLTOHWLWIHWHHDDMASBU","short_pith_number":"pith:IUSWN3DL","schema_version":"1.0","canonical_sha256":"452566ec6b9b8f65d907b1ce31b0120d199813c2adf1506984edee04c00ba46e","source":{"kind":"arxiv","id":"2606.09951","version":1},"attestation_state":"computed","paper":{"title":"Hasse Diagrams for Attention: A Partial Order Framework for Designing Transformer Masks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chentao Li, Han Guo","submitted_at":"2026-06-08T09:27:47Z","abstract_excerpt":"During the training of large Transformer models, attention masks regulate the scope and direction of information flow across a sequence. Numerous mask variants exist, and operators such as FlexAttention already support arbitrary attention masks. Nevertheless, a systematic formal analysis of the information-flow structure induced by arbitrary masks has been missing. This paper develops a complete theoretical framework. We prove that, with sufficient depth, the information flow of a multi-layer Transformer converges to a Hasse diagram -- a directed acyclic graph representing a partial order. Bui"},"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":"2606.09951","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-08T09:27:47Z","cross_cats_sorted":[],"title_canon_sha256":"524077fcadf109fbe26980185bc9f41dacf3b273a6b14e1fdc491d1ae0461c70","abstract_canon_sha256":"464650b0af85bdf584761c6168f3be56f56e12da8ffcf71d5370e1cc57482939"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-10T00:08:34.539959Z","signature_b64":"oOMZah1jzkMm9zTQsFsdHhj4Kyw1iwQje0zBb0wV3kqbmJpTHOf3J0zc38fLscrc/BgSPHole41mphvYW/9dAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"452566ec6b9b8f65d907b1ce31b0120d199813c2adf1506984edee04c00ba46e","last_reissued_at":"2026-06-10T00:08:34.539043Z","signature_status":"signed_v1","first_computed_at":"2026-06-10T00:08:34.539043Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hasse Diagrams for Attention: A Partial Order Framework for Designing Transformer Masks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chentao Li, Han Guo","submitted_at":"2026-06-08T09:27:47Z","abstract_excerpt":"During the training of large Transformer models, attention masks regulate the scope and direction of information flow across a sequence. Numerous mask variants exist, and operators such as FlexAttention already support arbitrary attention masks. Nevertheless, a systematic formal analysis of the information-flow structure induced by arbitrary masks has been missing. This paper develops a complete theoretical framework. We prove that, with sufficient depth, the information flow of a multi-layer Transformer converges to a Hasse diagram -- a directed acyclic graph representing a partial order. Bui"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09951","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/2606.09951/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":"2606.09951","created_at":"2026-06-10T00:08:34.539200+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09951v1","created_at":"2026-06-10T00:08:34.539200+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09951","created_at":"2026-06-10T00:08:34.539200+00:00"},{"alias_kind":"pith_short_12","alias_value":"IUSWN3DLTOHW","created_at":"2026-06-10T00:08:34.539200+00:00"},{"alias_kind":"pith_short_16","alias_value":"IUSWN3DLTOHWLWIH","created_at":"2026-06-10T00:08:34.539200+00:00"},{"alias_kind":"pith_short_8","alias_value":"IUSWN3DL","created_at":"2026-06-10T00:08:34.539200+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/IUSWN3DLTOHWLWIHWHHDDMASBU","json":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU.json","graph_json":"https://pith.science/api/pith-number/IUSWN3DLTOHWLWIHWHHDDMASBU/graph.json","events_json":"https://pith.science/api/pith-number/IUSWN3DLTOHWLWIHWHHDDMASBU/events.json","paper":"https://pith.science/paper/IUSWN3DL"},"agent_actions":{"view_html":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU","download_json":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU.json","view_paper":"https://pith.science/paper/IUSWN3DL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09951&json=true","fetch_graph":"https://pith.science/api/pith-number/IUSWN3DLTOHWLWIHWHHDDMASBU/graph.json","fetch_events":"https://pith.science/api/pith-number/IUSWN3DLTOHWLWIHWHHDDMASBU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU/action/storage_attestation","attest_author":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU/action/author_attestation","sign_citation":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU/action/citation_signature","submit_replication":"https://pith.science/pith/IUSWN3DLTOHWLWIHWHHDDMASBU/action/replication_record"}},"created_at":"2026-06-10T00:08:34.539200+00:00","updated_at":"2026-06-10T00:08:34.539200+00:00"}