{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CRA5KNN6HE7N5WVM63Q5MZF4CT","short_pith_number":"pith:CRA5KNN6","schema_version":"1.0","canonical_sha256":"1441d535be393ededaacf6e1d664bc14df763bf41363c335a5eeae2da9688c37","source":{"kind":"arxiv","id":"2410.01243","version":2},"attestation_state":"computed","paper":{"title":"An Information Theory of Compute-Optimal Size Scaling, Emergence, and Plateaus in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.IT"],"primary_cat":"cs.IT","authors_text":"Anuj K. Nayak, Lav R. Varshney","submitted_at":"2024-10-02T05:12:13Z","abstract_excerpt":"Recent empirical studies show three phenomena with increasing size of language models: compute-optimal size scaling, emergent capabilities, and performance plateauing. We present a simple unified mathematical framework to explain all of these language model scaling phenomena, building on recent skill-text bipartite graph frameworks for semantic learning. Modeling the learning of concepts from texts as an iterative process yields an analogy to iterative decoding of low-density parity check (LDPC) codes in information theory. Thence, drawing on finite-size scaling characterizations of LDPC decod"},"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":"2410.01243","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IT","submitted_at":"2024-10-02T05:12:13Z","cross_cats_sorted":["math.IT"],"title_canon_sha256":"19742c102441f5a04b41ac574fee074f23c1aeafba60a55de333caf19c2cb7f9","abstract_canon_sha256":"6aae15ea136a11383fa8d1e08b547af9113ce8ccdebcec90392ed7d9192befe1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:21:07.756885Z","signature_b64":"oCJfzFzfZ6mxPbsY7fZ1seKImwy6aLBAYFy0g/Fhnow5OOPtKs1X1Sfi5AQBbo+EzszRzCnoGDuZGa6QJHygBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1441d535be393ededaacf6e1d664bc14df763bf41363c335a5eeae2da9688c37","last_reissued_at":"2026-07-05T09:21:07.756423Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:21:07.756423Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"An Information Theory of Compute-Optimal Size Scaling, Emergence, and Plateaus in Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.IT"],"primary_cat":"cs.IT","authors_text":"Anuj K. Nayak, Lav R. Varshney","submitted_at":"2024-10-02T05:12:13Z","abstract_excerpt":"Recent empirical studies show three phenomena with increasing size of language models: compute-optimal size scaling, emergent capabilities, and performance plateauing. We present a simple unified mathematical framework to explain all of these language model scaling phenomena, building on recent skill-text bipartite graph frameworks for semantic learning. Modeling the learning of concepts from texts as an iterative process yields an analogy to iterative decoding of low-density parity check (LDPC) codes in information theory. Thence, drawing on finite-size scaling characterizations of LDPC decod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.01243","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/2410.01243/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":"2410.01243","created_at":"2026-07-05T09:21:07.756481+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.01243v2","created_at":"2026-07-05T09:21:07.756481+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.01243","created_at":"2026-07-05T09:21:07.756481+00:00"},{"alias_kind":"pith_short_12","alias_value":"CRA5KNN6HE7N","created_at":"2026-07-05T09:21:07.756481+00:00"},{"alias_kind":"pith_short_16","alias_value":"CRA5KNN6HE7N5WVM","created_at":"2026-07-05T09:21:07.756481+00:00"},{"alias_kind":"pith_short_8","alias_value":"CRA5KNN6","created_at":"2026-07-05T09:21:07.756481+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17767","citing_title":"Feature Learning in Linear-Width Two-Layer Networks: Two vs. One Step of Gradient Descent","ref_index":219,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17767","citing_title":"Feature Learning in Linear-Width Two-Layer Networks: Two vs. One Step of Gradient Descent","ref_index":219,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT","json":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT.json","graph_json":"https://pith.science/api/pith-number/CRA5KNN6HE7N5WVM63Q5MZF4CT/graph.json","events_json":"https://pith.science/api/pith-number/CRA5KNN6HE7N5WVM63Q5MZF4CT/events.json","paper":"https://pith.science/paper/CRA5KNN6"},"agent_actions":{"view_html":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT","download_json":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT.json","view_paper":"https://pith.science/paper/CRA5KNN6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.01243&json=true","fetch_graph":"https://pith.science/api/pith-number/CRA5KNN6HE7N5WVM63Q5MZF4CT/graph.json","fetch_events":"https://pith.science/api/pith-number/CRA5KNN6HE7N5WVM63Q5MZF4CT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT/action/storage_attestation","attest_author":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT/action/author_attestation","sign_citation":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT/action/citation_signature","submit_replication":"https://pith.science/pith/CRA5KNN6HE7N5WVM63Q5MZF4CT/action/replication_record"}},"created_at":"2026-07-05T09:21:07.756481+00:00","updated_at":"2026-07-05T09:21:07.756481+00:00"}