{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HUQEYUCZSAHTFQBSUNZADP5BB7","short_pith_number":"pith:HUQEYUCZ","schema_version":"1.0","canonical_sha256":"3d204c5059900f32c032a37201bfa10ff6bdc1f795249380f8094ee9fb4a4591","source":{"kind":"arxiv","id":"2605.20369","version":1},"attestation_state":"computed","paper":{"title":"DEL: Digit Entropy Loss for Numerical Learning of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chenhang He, Lei Zhang, Ming-Ming Cheng, Shihao Wang, Yuxuan Li, Zhaohui Zheng","submitted_at":"2026-05-19T18:18:59Z","abstract_excerpt":"Number prediction stands as a fundamental capability of large language models (LLMs) in mathematical problem-solving and code generation. The widely adopted maximum likelihood estimation (MLE) for LLM training is not tailored to number prediction. Recently, penalty-driven approaches, e.g., Number Token Loss and Discretized Distance Loss, introduce an inductive bias of numerical distance but induce over-sharpened and over-flattened digit distributions, respectively. In this paper, we make an in-depth analysis on LLM numerical learning, and show that existing numerical learning methods conceptua"},"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":"2605.20369","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T18:18:59Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"352b33351711903d4bfb9e1d226e0df230f2f02aa79c2041a8db323298302af4","abstract_canon_sha256":"9c1d940b8a967f6cd951b1e16da33190838eb3df538ff84afbf95631e25ad9da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:35.305568Z","signature_b64":"Pxv5lTsLXN1dzezgCb98tNylQ6Xt5pnZpTXZ3tlgL/qiLAUdEy5uUFgjllgRBioYnW5yMcRsl3rEuzWJQJ+1Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3d204c5059900f32c032a37201bfa10ff6bdc1f795249380f8094ee9fb4a4591","last_reissued_at":"2026-05-21T01:04:35.304867Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:35.304867Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DEL: Digit Entropy Loss for Numerical Learning of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chenhang He, Lei Zhang, Ming-Ming Cheng, Shihao Wang, Yuxuan Li, Zhaohui Zheng","submitted_at":"2026-05-19T18:18:59Z","abstract_excerpt":"Number prediction stands as a fundamental capability of large language models (LLMs) in mathematical problem-solving and code generation. The widely adopted maximum likelihood estimation (MLE) for LLM training is not tailored to number prediction. Recently, penalty-driven approaches, e.g., Number Token Loss and Discretized Distance Loss, introduce an inductive bias of numerical distance but induce over-sharpened and over-flattened digit distributions, respectively. In this paper, we make an in-depth analysis on LLM numerical learning, and show that existing numerical learning methods conceptua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20369","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/2605.20369/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":"2605.20369","created_at":"2026-05-21T01:04:35.304974+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20369v1","created_at":"2026-05-21T01:04:35.304974+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20369","created_at":"2026-05-21T01:04:35.304974+00:00"},{"alias_kind":"pith_short_12","alias_value":"HUQEYUCZSAHT","created_at":"2026-05-21T01:04:35.304974+00:00"},{"alias_kind":"pith_short_16","alias_value":"HUQEYUCZSAHTFQBS","created_at":"2026-05-21T01:04:35.304974+00:00"},{"alias_kind":"pith_short_8","alias_value":"HUQEYUCZ","created_at":"2026-05-21T01:04:35.304974+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/HUQEYUCZSAHTFQBSUNZADP5BB7","json":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7.json","graph_json":"https://pith.science/api/pith-number/HUQEYUCZSAHTFQBSUNZADP5BB7/graph.json","events_json":"https://pith.science/api/pith-number/HUQEYUCZSAHTFQBSUNZADP5BB7/events.json","paper":"https://pith.science/paper/HUQEYUCZ"},"agent_actions":{"view_html":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7","download_json":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7.json","view_paper":"https://pith.science/paper/HUQEYUCZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20369&json=true","fetch_graph":"https://pith.science/api/pith-number/HUQEYUCZSAHTFQBSUNZADP5BB7/graph.json","fetch_events":"https://pith.science/api/pith-number/HUQEYUCZSAHTFQBSUNZADP5BB7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7/action/storage_attestation","attest_author":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7/action/author_attestation","sign_citation":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7/action/citation_signature","submit_replication":"https://pith.science/pith/HUQEYUCZSAHTFQBSUNZADP5BB7/action/replication_record"}},"created_at":"2026-05-21T01:04:35.304974+00:00","updated_at":"2026-05-21T01:04:35.304974+00:00"}