{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WV5JWCN67KVCXYNHRC4242BRSI","short_pith_number":"pith:WV5JWCN6","schema_version":"1.0","canonical_sha256":"b57a9b09befaaa2be1a788b9ae6831922b33116ffffa86afe26469c4fde69d12","source":{"kind":"arxiv","id":"2402.03902","version":2},"attestation_state":"computed","paper":{"title":"A phase transition between positional and semantic learning in a solvable model of dot-product attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Florent Krzakala, Freya Behrens, Hugo Cui, Lenka Zdeborov\\'a","submitted_at":"2024-02-06T11:13:54Z","abstract_excerpt":"Many empirical studies have provided evidence for the emergence of algorithmic mechanisms (abilities) in the learning of language models, that lead to qualitative improvements of the model capabilities. Yet, a theoretical characterization of how such mechanisms emerge remains elusive. In this paper, we take a step in this direction by providing a tight theoretical analysis of the emergence of semantic attention in a solvable model of dot-product attention. More precisely, we consider a non-linear self-attention layer with trainable tied and low-rank query and key matrices. In the asymptotic li"},"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":"2402.03902","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-06T11:13:54Z","cross_cats_sorted":[],"title_canon_sha256":"dd4e0c1a599a339388732ba11304b3a6f55142b23630c7d58a5b71e2a137928b","abstract_canon_sha256":"27f5ab9e5f9979f2aeaea643aac346da5163a9898f7affc896fdbc4096003685"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:10:39.565108Z","signature_b64":"i6JCMi6/TaNs6kTq8leieEMDMGvrj6k88SEL1F4RtO/oUTpmUpLWtgBjW5c6HryTIfQsk/tEaDs+lqiu/GLSBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b57a9b09befaaa2be1a788b9ae6831922b33116ffffa86afe26469c4fde69d12","last_reissued_at":"2026-07-05T10:10:39.564725Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:10:39.564725Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A phase transition between positional and semantic learning in a solvable model of dot-product attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Florent Krzakala, Freya Behrens, Hugo Cui, Lenka Zdeborov\\'a","submitted_at":"2024-02-06T11:13:54Z","abstract_excerpt":"Many empirical studies have provided evidence for the emergence of algorithmic mechanisms (abilities) in the learning of language models, that lead to qualitative improvements of the model capabilities. Yet, a theoretical characterization of how such mechanisms emerge remains elusive. In this paper, we take a step in this direction by providing a tight theoretical analysis of the emergence of semantic attention in a solvable model of dot-product attention. More precisely, we consider a non-linear self-attention layer with trainable tied and low-rank query and key matrices. In the asymptotic li"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.03902","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/2402.03902/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":"2402.03902","created_at":"2026-07-05T10:10:39.564777+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.03902v2","created_at":"2026-07-05T10:10:39.564777+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.03902","created_at":"2026-07-05T10:10:39.564777+00:00"},{"alias_kind":"pith_short_12","alias_value":"WV5JWCN67KVC","created_at":"2026-07-05T10:10:39.564777+00:00"},{"alias_kind":"pith_short_16","alias_value":"WV5JWCN67KVCXYNH","created_at":"2026-07-05T10:10:39.564777+00:00"},{"alias_kind":"pith_short_8","alias_value":"WV5JWCN6","created_at":"2026-07-05T10:10:39.564777+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.12391","citing_title":"Physics of Skill Learning","ref_index":43,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI","json":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI.json","graph_json":"https://pith.science/api/pith-number/WV5JWCN67KVCXYNHRC4242BRSI/graph.json","events_json":"https://pith.science/api/pith-number/WV5JWCN67KVCXYNHRC4242BRSI/events.json","paper":"https://pith.science/paper/WV5JWCN6"},"agent_actions":{"view_html":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI","download_json":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI.json","view_paper":"https://pith.science/paper/WV5JWCN6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.03902&json=true","fetch_graph":"https://pith.science/api/pith-number/WV5JWCN67KVCXYNHRC4242BRSI/graph.json","fetch_events":"https://pith.science/api/pith-number/WV5JWCN67KVCXYNHRC4242BRSI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI/action/storage_attestation","attest_author":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI/action/author_attestation","sign_citation":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI/action/citation_signature","submit_replication":"https://pith.science/pith/WV5JWCN67KVCXYNHRC4242BRSI/action/replication_record"}},"created_at":"2026-07-05T10:10:39.564777+00:00","updated_at":"2026-07-05T10:10:39.564777+00:00"}