{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YD5WE2EWZAO2XYLIBIPPHG4VLE","short_pith_number":"pith:YD5WE2EW","schema_version":"1.0","canonical_sha256":"c0fb626896c81dabe1680a1ef39b955905db9c2a757c8a9a2e2da52f464c2d7a","source":{"kind":"arxiv","id":"2405.01964","version":3},"attestation_state":"computed","paper":{"title":"Position: Understanding LLMs Requires More Than Statistical Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Anna Kerekes, Anna M\\'esz\\'aros, Ferenc Husz\\'ar, Patrik Reizinger, Szilvia Ujv\\'ary, Wieland Brendel","submitted_at":"2024-05-03T09:41:39Z","abstract_excerpt":"The last decade has seen blossoming research in deep learning theory attempting to answer, \"Why does deep learning generalize?\" A powerful shift in perspective precipitated this progress: the study of overparametrized models in the interpolation regime. In this paper, we argue that another perspective shift is due, since some of the desirable qualities of LLMs are not a consequence of good statistical generalization and require a separate theoretical explanation. Our core argument relies on the observation that AR probabilistic models are inherently non-identifiable: models zero or near-zero K"},"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":"2405.01964","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-05-03T09:41:39Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"131a333b7acc60c52bb0083949cd1d6466d651315ae41e2a2a89c0e08001f23a","abstract_canon_sha256":"d8faaa11c2cd7bd3f57860afbf919137505e9e02c659b2529189ed554915e367"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:32:40.319644Z","signature_b64":"dW30EVH1MJpyt+tXJDG3LINhAeMtrzP6XFvRoftJsrRi+rL0Qp7Nd7F5P3o07UGtFHUDf/Dv0c0lYErCQE4YCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c0fb626896c81dabe1680a1ef39b955905db9c2a757c8a9a2e2da52f464c2d7a","last_reissued_at":"2026-07-05T08:32:40.319138Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:32:40.319138Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Position: Understanding LLMs Requires More Than Statistical Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Anna Kerekes, Anna M\\'esz\\'aros, Ferenc Husz\\'ar, Patrik Reizinger, Szilvia Ujv\\'ary, Wieland Brendel","submitted_at":"2024-05-03T09:41:39Z","abstract_excerpt":"The last decade has seen blossoming research in deep learning theory attempting to answer, \"Why does deep learning generalize?\" A powerful shift in perspective precipitated this progress: the study of overparametrized models in the interpolation regime. In this paper, we argue that another perspective shift is due, since some of the desirable qualities of LLMs are not a consequence of good statistical generalization and require a separate theoretical explanation. Our core argument relies on the observation that AR probabilistic models are inherently non-identifiable: models zero or near-zero K"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.01964","kind":"arxiv","version":3},"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/2405.01964/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":"2405.01964","created_at":"2026-07-05T08:32:40.319196+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.01964v3","created_at":"2026-07-05T08:32:40.319196+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.01964","created_at":"2026-07-05T08:32:40.319196+00:00"},{"alias_kind":"pith_short_12","alias_value":"YD5WE2EWZAO2","created_at":"2026-07-05T08:32:40.319196+00:00"},{"alias_kind":"pith_short_16","alias_value":"YD5WE2EWZAO2XYLI","created_at":"2026-07-05T08:32:40.319196+00:00"},{"alias_kind":"pith_short_8","alias_value":"YD5WE2EW","created_at":"2026-07-05T08:32:40.319196+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29983","citing_title":"Stabilizing Extrapolation in Looped Transformers via Learned Stochastic Stopping","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE","json":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE.json","graph_json":"https://pith.science/api/pith-number/YD5WE2EWZAO2XYLIBIPPHG4VLE/graph.json","events_json":"https://pith.science/api/pith-number/YD5WE2EWZAO2XYLIBIPPHG4VLE/events.json","paper":"https://pith.science/paper/YD5WE2EW"},"agent_actions":{"view_html":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE","download_json":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE.json","view_paper":"https://pith.science/paper/YD5WE2EW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.01964&json=true","fetch_graph":"https://pith.science/api/pith-number/YD5WE2EWZAO2XYLIBIPPHG4VLE/graph.json","fetch_events":"https://pith.science/api/pith-number/YD5WE2EWZAO2XYLIBIPPHG4VLE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE/action/storage_attestation","attest_author":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE/action/author_attestation","sign_citation":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE/action/citation_signature","submit_replication":"https://pith.science/pith/YD5WE2EWZAO2XYLIBIPPHG4VLE/action/replication_record"}},"created_at":"2026-07-05T08:32:40.319196+00:00","updated_at":"2026-07-05T08:32:40.319196+00:00"}