{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HSYI2ULGVPPYOLSKGFFDWYIFPB","short_pith_number":"pith:HSYI2ULG","schema_version":"1.0","canonical_sha256":"3cb08d5166abdf872e4a314a3b61057848e7dd3bf5b11238182a2b6fc65df2bc","source":{"kind":"arxiv","id":"2606.00230","version":1},"attestation_state":"computed","paper":{"title":"A Pre-Training Analogue of Grokking in Language Models: Tracing Delayed Grammatical Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anna Rumshisky, Namrata Shivagunde, Sherin Muckatira, Vijeta Deshpande","submitted_at":"2026-05-29T18:04:52Z","abstract_excerpt":"Grokking, the phenomenon in which neural networks generalize long after fitting their training data, has been studied in supervised settings on many epochs. LLM pre-training instead involves next-token prediction over an unlabeled corpus, with limited data repetition and no explicit train/validation split. To address this, we propose an exposure-based framework that enables the study of grokking-like dynamics during LLM pre-training. We ground our evaluation in BLiMP minimal pairs, which provide controlled grammatical contrasts. For every BLiMP minimal pair, we identify a critical phrase, the "},"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":"2606.00230","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-29T18:04:52Z","cross_cats_sorted":[],"title_canon_sha256":"2660ac98a070ffa772de9be464f20f5e4e90d813394c7c320a9dc0476f71c3ab","abstract_canon_sha256":"3ad19bd65075f393620aa73412184226abc1f2efa4dfa6d5181b84996189e460"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:22.303434Z","signature_b64":"8kwdNGLf1SGz2BNo/2HuLeKY5iTlitVhFzRPz0dEdH+FFN1/VUb8+Yi7201odPXi+ly3hblLMnd3eUcoAJSeDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3cb08d5166abdf872e4a314a3b61057848e7dd3bf5b11238182a2b6fc65df2bc","last_reissued_at":"2026-06-02T01:03:22.303064Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:22.303064Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Pre-Training Analogue of Grokking in Language Models: Tracing Delayed Grammatical Generalization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anna Rumshisky, Namrata Shivagunde, Sherin Muckatira, Vijeta Deshpande","submitted_at":"2026-05-29T18:04:52Z","abstract_excerpt":"Grokking, the phenomenon in which neural networks generalize long after fitting their training data, has been studied in supervised settings on many epochs. LLM pre-training instead involves next-token prediction over an unlabeled corpus, with limited data repetition and no explicit train/validation split. To address this, we propose an exposure-based framework that enables the study of grokking-like dynamics during LLM pre-training. We ground our evaluation in BLiMP minimal pairs, which provide controlled grammatical contrasts. For every BLiMP minimal pair, we identify a critical phrase, the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00230","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/2606.00230/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":"2606.00230","created_at":"2026-06-02T01:03:22.303111+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00230v1","created_at":"2026-06-02T01:03:22.303111+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00230","created_at":"2026-06-02T01:03:22.303111+00:00"},{"alias_kind":"pith_short_12","alias_value":"HSYI2ULGVPPY","created_at":"2026-06-02T01:03:22.303111+00:00"},{"alias_kind":"pith_short_16","alias_value":"HSYI2ULGVPPYOLSK","created_at":"2026-06-02T01:03:22.303111+00:00"},{"alias_kind":"pith_short_8","alias_value":"HSYI2ULG","created_at":"2026-06-02T01:03:22.303111+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/HSYI2ULGVPPYOLSKGFFDWYIFPB","json":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB.json","graph_json":"https://pith.science/api/pith-number/HSYI2ULGVPPYOLSKGFFDWYIFPB/graph.json","events_json":"https://pith.science/api/pith-number/HSYI2ULGVPPYOLSKGFFDWYIFPB/events.json","paper":"https://pith.science/paper/HSYI2ULG"},"agent_actions":{"view_html":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB","download_json":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB.json","view_paper":"https://pith.science/paper/HSYI2ULG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00230&json=true","fetch_graph":"https://pith.science/api/pith-number/HSYI2ULGVPPYOLSKGFFDWYIFPB/graph.json","fetch_events":"https://pith.science/api/pith-number/HSYI2ULGVPPYOLSKGFFDWYIFPB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB/action/storage_attestation","attest_author":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB/action/author_attestation","sign_citation":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB/action/citation_signature","submit_replication":"https://pith.science/pith/HSYI2ULGVPPYOLSKGFFDWYIFPB/action/replication_record"}},"created_at":"2026-06-02T01:03:22.303111+00:00","updated_at":"2026-06-02T01:03:22.303111+00:00"}