{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:E43A3EKYPOTXWTXBZITPO4DH63","short_pith_number":"pith:E43A3EKY","schema_version":"1.0","canonical_sha256":"27360d91587ba77b4ee1ca26f77067f6cc1bfcd5d234eb685b456f10372555d7","source":{"kind":"arxiv","id":"2607.18292","version":1},"attestation_state":"computed","paper":{"title":"Reliability Scales Inversely: Bigger Models Compound Mistakes Faster via a Hidden Auto-Regressive Risk Regime","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Kushal Chakrabarti","submitted_at":"2026-06-30T17:51:04Z","abstract_excerpt":"As language models scale, answers start truer but degrade faster: scaling buys capability but erodes reliability. The knowledge-gap account - more data, retrieval, or scale - misses an auto-regressive risk residual that scale sharpens: the model commits to a low-probability token, conditions on it as established, and snowballs. We track this through per-position disagreement $\\delta = \\log p_M - \\log p_O$ against a stronger same-family oracle, whose second moment splits exactly into bias$^2$ $\\mathrm{KL}(p_M \\,\\|\\, p_O)^2$ and risk $\\mathrm{Var}[\\delta]$. We present four findings: (i) under sc"},"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":"2607.18292","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-30T17:51:04Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"4587b42aa4f2f5fbd672fe148eee9b0235092a6a3d7ca91f71f977c385538de1","abstract_canon_sha256":"4ee8e9ebe5c0af7d14cec0d5e0974f1870d4a277039a17de9a514a16fce1a842"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-22T00:22:40.374522Z","signature_b64":"3a4j1SpSsaGuhwWsnTXrqkeNuuiMRw4uwmAE8JW06LYntuVy/8/wa6uCgvbw18q3ypxiI3PZNYK+BZz+FkE/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"27360d91587ba77b4ee1ca26f77067f6cc1bfcd5d234eb685b456f10372555d7","last_reissued_at":"2026-07-22T00:22:40.373713Z","signature_status":"signed_v1","first_computed_at":"2026-07-22T00:22:40.373713Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reliability Scales Inversely: Bigger Models Compound Mistakes Faster via a Hidden Auto-Regressive Risk Regime","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Kushal Chakrabarti","submitted_at":"2026-06-30T17:51:04Z","abstract_excerpt":"As language models scale, answers start truer but degrade faster: scaling buys capability but erodes reliability. The knowledge-gap account - more data, retrieval, or scale - misses an auto-regressive risk residual that scale sharpens: the model commits to a low-probability token, conditions on it as established, and snowballs. We track this through per-position disagreement $\\delta = \\log p_M - \\log p_O$ against a stronger same-family oracle, whose second moment splits exactly into bias$^2$ $\\mathrm{KL}(p_M \\,\\|\\, p_O)^2$ and risk $\\mathrm{Var}[\\delta]$. We present four findings: (i) under sc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.18292","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/2607.18292/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":"2607.18292","created_at":"2026-07-22T00:22:40.374131+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.18292v1","created_at":"2026-07-22T00:22:40.374131+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.18292","created_at":"2026-07-22T00:22:40.374131+00:00"},{"alias_kind":"pith_short_12","alias_value":"E43A3EKYPOTX","created_at":"2026-07-22T00:22:40.374131+00:00"},{"alias_kind":"pith_short_16","alias_value":"E43A3EKYPOTXWTXB","created_at":"2026-07-22T00:22:40.374131+00:00"},{"alias_kind":"pith_short_8","alias_value":"E43A3EKY","created_at":"2026-07-22T00:22:40.374131+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/E43A3EKYPOTXWTXBZITPO4DH63","json":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63.json","graph_json":"https://pith.science/api/pith-number/E43A3EKYPOTXWTXBZITPO4DH63/graph.json","events_json":"https://pith.science/api/pith-number/E43A3EKYPOTXWTXBZITPO4DH63/events.json","paper":"https://pith.science/paper/E43A3EKY"},"agent_actions":{"view_html":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63","download_json":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63.json","view_paper":"https://pith.science/paper/E43A3EKY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.18292&json=true","fetch_graph":"https://pith.science/api/pith-number/E43A3EKYPOTXWTXBZITPO4DH63/graph.json","fetch_events":"https://pith.science/api/pith-number/E43A3EKYPOTXWTXBZITPO4DH63/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63/action/storage_attestation","attest_author":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63/action/author_attestation","sign_citation":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63/action/citation_signature","submit_replication":"https://pith.science/pith/E43A3EKYPOTXWTXBZITPO4DH63/action/replication_record"}},"created_at":"2026-07-22T00:22:40.374131+00:00","updated_at":"2026-07-22T00:22:40.374131+00:00"}