{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TED2UFKAEMQN3PSYBE3JGDMTC6","short_pith_number":"pith:TED2UFKA","schema_version":"1.0","canonical_sha256":"9907aa15402320ddbe580936930d9317ac965d3db9fe5b135d5dbb50389b44f9","source":{"kind":"arxiv","id":"2502.13794","version":1},"attestation_state":"computed","paper":{"title":"LESA: Learnable LLM Layer Scaling-Up","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hai Zhao, Libo Qin, Xinbei Ma, Yao Yao, Yifei Yang, Zhi Chen, Zouying Cao","submitted_at":"2025-02-19T14:58:48Z","abstract_excerpt":"Training Large Language Models (LLMs) from scratch requires immense computational resources, making it prohibitively expensive. Model scaling-up offers a promising solution by leveraging the parameters of smaller models to create larger ones. However, existing depth scaling-up methods rely on empirical heuristic rules for layer duplication, which result in poorer initialization and slower convergence during continual pre-training. We propose \\textbf{LESA}, a novel learnable method for depth scaling-up. By concatenating parameters from each layer and applying Singular Value Decomposition, we un"},"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":"2502.13794","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-02-19T14:58:48Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"03c2a3b5dc0e87b522a8632aaa2f3e07fe78a28dfcc2eb6c0e32c3094bed92f9","abstract_canon_sha256":"b26115dad84992896d4c5cb2dbeec7e117e26f4bdff672687d59dda31d6c0563"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:17:00.696324Z","signature_b64":"f0/Rg0m5gxQ6OWcwhnm4PFTsiaKePXZfdPx6MoYj32FMBxZRmfUBqzZRA5rmJnT2ZNDpYuVCIQ/vZo/dBxuhBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9907aa15402320ddbe580936930d9317ac965d3db9fe5b135d5dbb50389b44f9","last_reissued_at":"2026-07-05T10:17:00.695866Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:17:00.695866Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LESA: Learnable LLM Layer Scaling-Up","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hai Zhao, Libo Qin, Xinbei Ma, Yao Yao, Yifei Yang, Zhi Chen, Zouying Cao","submitted_at":"2025-02-19T14:58:48Z","abstract_excerpt":"Training Large Language Models (LLMs) from scratch requires immense computational resources, making it prohibitively expensive. Model scaling-up offers a promising solution by leveraging the parameters of smaller models to create larger ones. However, existing depth scaling-up methods rely on empirical heuristic rules for layer duplication, which result in poorer initialization and slower convergence during continual pre-training. We propose \\textbf{LESA}, a novel learnable method for depth scaling-up. By concatenating parameters from each layer and applying Singular Value Decomposition, we un"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13794","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/2502.13794/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":"2502.13794","created_at":"2026-07-05T10:17:00.695922+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13794v1","created_at":"2026-07-05T10:17:00.695922+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13794","created_at":"2026-07-05T10:17:00.695922+00:00"},{"alias_kind":"pith_short_12","alias_value":"TED2UFKAEMQN","created_at":"2026-07-05T10:17:00.695922+00:00"},{"alias_kind":"pith_short_16","alias_value":"TED2UFKAEMQN3PSY","created_at":"2026-07-05T10:17:00.695922+00:00"},{"alias_kind":"pith_short_8","alias_value":"TED2UFKA","created_at":"2026-07-05T10:17:00.695922+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2512.13751","citing_title":"MIDUS: Memory-Infused Depth Up-Scaling","ref_index":28,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6","json":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6.json","graph_json":"https://pith.science/api/pith-number/TED2UFKAEMQN3PSYBE3JGDMTC6/graph.json","events_json":"https://pith.science/api/pith-number/TED2UFKAEMQN3PSYBE3JGDMTC6/events.json","paper":"https://pith.science/paper/TED2UFKA"},"agent_actions":{"view_html":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6","download_json":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6.json","view_paper":"https://pith.science/paper/TED2UFKA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13794&json=true","fetch_graph":"https://pith.science/api/pith-number/TED2UFKAEMQN3PSYBE3JGDMTC6/graph.json","fetch_events":"https://pith.science/api/pith-number/TED2UFKAEMQN3PSYBE3JGDMTC6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6/action/storage_attestation","attest_author":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6/action/author_attestation","sign_citation":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6/action/citation_signature","submit_replication":"https://pith.science/pith/TED2UFKAEMQN3PSYBE3JGDMTC6/action/replication_record"}},"created_at":"2026-07-05T10:17:00.695922+00:00","updated_at":"2026-07-05T10:17:00.695922+00:00"}