{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GX7OFEAGNMHRQLOIR44FGLG4MU","short_pith_number":"pith:GX7OFEAG","schema_version":"1.0","canonical_sha256":"35fee290066b0f182dc88f38532cdc6539a15489942ab8b13b1d69f5e3904081","source":{"kind":"arxiv","id":"2501.12997","version":2},"attestation_state":"computed","paper":{"title":"Ehrenfeucht-Haussler Rank and Chain of Thought","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alexander Kozachinskiy, Pablo Barcel\\'o, Tomasz Steifer","submitted_at":"2025-01-22T16:30:58Z","abstract_excerpt":"The notion of \\emph{rank} of a Boolean function has been a cornerstone in PAC learning theory, enabling quasipolynomial-time learning algorithms for polynomial-size decision trees. We present a novel characterization of rank, grounded in the well-known Transformer architecture. We show that the rank of a function $f$ corresponds to the minimum number of \\emph{Chain of Thought} (CoT) steps required by a single-layer Transformer with hard attention to compute $f$. Based on this characterization we establish tight bounds on the number of CoT steps required for specific problems, showing that \\(\\e"},"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":"2501.12997","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-01-22T16:30:58Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"72310fb510eacc576cba226e1cdbec392080d959e1e204e9b9f2371b0b3eda99","abstract_canon_sha256":"4134d5c8cdc9e0975e6722c30a9d3cce98ede5e4856139fd0986b37757b0459b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:01.083088Z","signature_b64":"u0kgnW0gcJ2EEcXgWBDgbbxqSPjmKv1GsYrwafQErbyi7TQH/Z09lh8Mx6O1QocL4HWDLNQamNu1lXHog7MJBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35fee290066b0f182dc88f38532cdc6539a15489942ab8b13b1d69f5e3904081","last_reissued_at":"2026-07-05T11:52:01.082605Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:01.082605Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ehrenfeucht-Haussler Rank and Chain of Thought","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alexander Kozachinskiy, Pablo Barcel\\'o, Tomasz Steifer","submitted_at":"2025-01-22T16:30:58Z","abstract_excerpt":"The notion of \\emph{rank} of a Boolean function has been a cornerstone in PAC learning theory, enabling quasipolynomial-time learning algorithms for polynomial-size decision trees. We present a novel characterization of rank, grounded in the well-known Transformer architecture. We show that the rank of a function $f$ corresponds to the minimum number of \\emph{Chain of Thought} (CoT) steps required by a single-layer Transformer with hard attention to compute $f$. Based on this characterization we establish tight bounds on the number of CoT steps required for specific problems, showing that \\(\\e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.12997","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/2501.12997/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":"2501.12997","created_at":"2026-07-05T11:52:01.082662+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.12997v2","created_at":"2026-07-05T11:52:01.082662+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.12997","created_at":"2026-07-05T11:52:01.082662+00:00"},{"alias_kind":"pith_short_12","alias_value":"GX7OFEAGNMHR","created_at":"2026-07-05T11:52:01.082662+00:00"},{"alias_kind":"pith_short_16","alias_value":"GX7OFEAGNMHRQLOI","created_at":"2026-07-05T11:52:01.082662+00:00"},{"alias_kind":"pith_short_8","alias_value":"GX7OFEAG","created_at":"2026-07-05T11:52:01.082662+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.00183","citing_title":"Agentic Transformers Provably Learn to Search via Reinforcement Learning","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2511.07372","citing_title":"Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-Training","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU","json":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU.json","graph_json":"https://pith.science/api/pith-number/GX7OFEAGNMHRQLOIR44FGLG4MU/graph.json","events_json":"https://pith.science/api/pith-number/GX7OFEAGNMHRQLOIR44FGLG4MU/events.json","paper":"https://pith.science/paper/GX7OFEAG"},"agent_actions":{"view_html":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU","download_json":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU.json","view_paper":"https://pith.science/paper/GX7OFEAG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.12997&json=true","fetch_graph":"https://pith.science/api/pith-number/GX7OFEAGNMHRQLOIR44FGLG4MU/graph.json","fetch_events":"https://pith.science/api/pith-number/GX7OFEAGNMHRQLOIR44FGLG4MU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU/action/storage_attestation","attest_author":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU/action/author_attestation","sign_citation":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU/action/citation_signature","submit_replication":"https://pith.science/pith/GX7OFEAGNMHRQLOIR44FGLG4MU/action/replication_record"}},"created_at":"2026-07-05T11:52:01.082662+00:00","updated_at":"2026-07-05T11:52:01.082662+00:00"}