{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DENN4YJL7BHHZTGOVZTJRFAB4U","short_pith_number":"pith:DENN4YJL","schema_version":"1.0","canonical_sha256":"191ade612bf84e7cccceae66989401e538a7117282009f788d16beb3cc080c01","source":{"kind":"arxiv","id":"2606.02765","version":1},"attestation_state":"computed","paper":{"title":"Representational Capacity: Geometric Limits on Feature Representation in Transformer Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alexander Guha","submitted_at":"2026-06-01T18:28:56Z","abstract_excerpt":"Model dimension ($d_{model}$) is a fundamental hyperparameter in transformer language models, yet its role in setting the geometric limits of feature representation remains under-explored. Grounded in the Linear Representation and Superposition Hypotheses - which propose that models encode features as near-orthogonal directions in latent space - we develop a framework for estimating how many such directions a model can support. We first establish the embedding matrix as a measurable proxy for near-orthogonality constraints across the latent space: the boundary between meaningful token relation"},"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":"2606.02765","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-01T18:28:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f5531afbd721dcb9a903d55535c4acfcb0a3e7621a206514b54e81cf9201c57e","abstract_canon_sha256":"383572c455811403c6bd7a8678f6eff21c532648168c3fb4401456389c5866ea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:21.216863Z","signature_b64":"ADk22XJLv4khh9QwZSOoG7KvEgNR6daGhv23mT0vnqwy4cjFc4Vegw9OfcBTaqcD4HKELLorpcmccRua9lAXBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"191ade612bf84e7cccceae66989401e538a7117282009f788d16beb3cc080c01","last_reissued_at":"2026-06-03T01:05:21.216361Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:21.216361Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Representational Capacity: Geometric Limits on Feature Representation in Transformer Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Alexander Guha","submitted_at":"2026-06-01T18:28:56Z","abstract_excerpt":"Model dimension ($d_{model}$) is a fundamental hyperparameter in transformer language models, yet its role in setting the geometric limits of feature representation remains under-explored. Grounded in the Linear Representation and Superposition Hypotheses - which propose that models encode features as near-orthogonal directions in latent space - we develop a framework for estimating how many such directions a model can support. We first establish the embedding matrix as a measurable proxy for near-orthogonality constraints across the latent space: the boundary between meaningful token relation"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02765","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/2606.02765/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":"2606.02765","created_at":"2026-06-03T01:05:21.216424+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02765v1","created_at":"2026-06-03T01:05:21.216424+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02765","created_at":"2026-06-03T01:05:21.216424+00:00"},{"alias_kind":"pith_short_12","alias_value":"DENN4YJL7BHH","created_at":"2026-06-03T01:05:21.216424+00:00"},{"alias_kind":"pith_short_16","alias_value":"DENN4YJL7BHHZTGO","created_at":"2026-06-03T01:05:21.216424+00:00"},{"alias_kind":"pith_short_8","alias_value":"DENN4YJL","created_at":"2026-06-03T01:05:21.216424+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/DENN4YJL7BHHZTGOVZTJRFAB4U","json":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U.json","graph_json":"https://pith.science/api/pith-number/DENN4YJL7BHHZTGOVZTJRFAB4U/graph.json","events_json":"https://pith.science/api/pith-number/DENN4YJL7BHHZTGOVZTJRFAB4U/events.json","paper":"https://pith.science/paper/DENN4YJL"},"agent_actions":{"view_html":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U","download_json":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U.json","view_paper":"https://pith.science/paper/DENN4YJL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02765&json=true","fetch_graph":"https://pith.science/api/pith-number/DENN4YJL7BHHZTGOVZTJRFAB4U/graph.json","fetch_events":"https://pith.science/api/pith-number/DENN4YJL7BHHZTGOVZTJRFAB4U/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U/action/storage_attestation","attest_author":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U/action/author_attestation","sign_citation":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U/action/citation_signature","submit_replication":"https://pith.science/pith/DENN4YJL7BHHZTGOVZTJRFAB4U/action/replication_record"}},"created_at":"2026-06-03T01:05:21.216424+00:00","updated_at":"2026-06-03T01:05:21.216424+00:00"}