{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:APJ3VEKJJOO7DKANKC62QK2IJT","short_pith_number":"pith:APJ3VEKJ","schema_version":"1.0","canonical_sha256":"03d3ba91494b9df1a80d50bda82b484ce215914151d2decce8f6fd7105c81fae","source":{"kind":"arxiv","id":"2605.18845","version":1},"attestation_state":"computed","paper":{"title":"First-Passage Prediction of Grokking Delay: ACalibrated Law under AdamW with Causal Validation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Luu Duc Trung, Phan Thanh Duc, Truong Quynh Hoa, Truong Xuan Khanh","submitted_at":"2026-05-13T06:33:56Z","abstract_excerpt":"We give the first quantitative prediction of grokking delay under AdamW. Treating the delay as a first-passage time, we derive a closed-form law T_grok - T_mem = (1 / 2 kappa_LL eta lambda) log(V_mem / V_star), where V_t = ||theta_t||^2 is the squared parameter norm, V_star is an architecture-dependent threshold, and kappa_LL absorbs the AdamW correction to the clean-SGD contraction rate 2 eta lambda. Calibrating (kappa_LL, V_star) on a single hyperparameter cell predicts grokking delays on 26 held-out runs with MAPE 17.7% over a 41x delay range; the law generalises to MLPs (MAPE 18.0%, N=34) "},"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":"2605.18845","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T06:33:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"536326b2f97ffdca850c8888688ffc3052b9fe9cb8ecf463cceaf2d156fdd0af","abstract_canon_sha256":"e347e542a2a4f4d290caea9069b2d208a49ac7f146c1197509aefdd1bfa85a4d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:06:25.415146Z","signature_b64":"PXnSEgAM10Zk699W5YyhGI3kfiBLImiuKMfhfkE1f91pLyiPsy5w71QnrWF6NVPUcFRvY+CstFaGgLDWnSc0DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03d3ba91494b9df1a80d50bda82b484ce215914151d2decce8f6fd7105c81fae","last_reissued_at":"2026-05-20T00:06:25.414363Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:06:25.414363Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"First-Passage Prediction of Grokking Delay: ACalibrated Law under AdamW with Causal Validation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Luu Duc Trung, Phan Thanh Duc, Truong Quynh Hoa, Truong Xuan Khanh","submitted_at":"2026-05-13T06:33:56Z","abstract_excerpt":"We give the first quantitative prediction of grokking delay under AdamW. Treating the delay as a first-passage time, we derive a closed-form law T_grok - T_mem = (1 / 2 kappa_LL eta lambda) log(V_mem / V_star), where V_t = ||theta_t||^2 is the squared parameter norm, V_star is an architecture-dependent threshold, and kappa_LL absorbs the AdamW correction to the clean-SGD contraction rate 2 eta lambda. Calibrating (kappa_LL, V_star) on a single hyperparameter cell predicts grokking delays on 26 held-out runs with MAPE 17.7% over a 41x delay range; the law generalises to MLPs (MAPE 18.0%, N=34) "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18845","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/2605.18845/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":"2605.18845","created_at":"2026-05-20T00:06:25.414491+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18845v1","created_at":"2026-05-20T00:06:25.414491+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18845","created_at":"2026-05-20T00:06:25.414491+00:00"},{"alias_kind":"pith_short_12","alias_value":"APJ3VEKJJOO7","created_at":"2026-05-20T00:06:25.414491+00:00"},{"alias_kind":"pith_short_16","alias_value":"APJ3VEKJJOO7DKAN","created_at":"2026-05-20T00:06:25.414491+00:00"},{"alias_kind":"pith_short_8","alias_value":"APJ3VEKJ","created_at":"2026-05-20T00:06:25.414491+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/APJ3VEKJJOO7DKANKC62QK2IJT","json":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT.json","graph_json":"https://pith.science/api/pith-number/APJ3VEKJJOO7DKANKC62QK2IJT/graph.json","events_json":"https://pith.science/api/pith-number/APJ3VEKJJOO7DKANKC62QK2IJT/events.json","paper":"https://pith.science/paper/APJ3VEKJ"},"agent_actions":{"view_html":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT","download_json":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT.json","view_paper":"https://pith.science/paper/APJ3VEKJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18845&json=true","fetch_graph":"https://pith.science/api/pith-number/APJ3VEKJJOO7DKANKC62QK2IJT/graph.json","fetch_events":"https://pith.science/api/pith-number/APJ3VEKJJOO7DKANKC62QK2IJT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT/action/storage_attestation","attest_author":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT/action/author_attestation","sign_citation":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT/action/citation_signature","submit_replication":"https://pith.science/pith/APJ3VEKJJOO7DKANKC62QK2IJT/action/replication_record"}},"created_at":"2026-05-20T00:06:25.414491+00:00","updated_at":"2026-05-20T00:06:25.414491+00:00"}