{"_canonicalization":{"envelope_id":"axm_ + sha256(envelope minus {signature, axiom_id, anchors})","envelope_signature":"ed25519(envelope minus {signature, axiom_id})","json":"sort_keys=True, separators=(',',':'), ensure_ascii=False, allow_nan=False, utf-8","leaf_hash":"sha256(0x00 || canonical_json(envelope_full))","seal_signature":"ed25519(seal minus {signature, sig_algorithm})"},"axiom_id":"axm_ff5306de89189279f7f774195e50327fc2e485cbba7da7f8113f2783bd6e5db2","bitcoin_anchor":{"bitcoin_attestations":[],"calendar_attestations":[],"ots_url":"","stamped_at":"","status":"pending_next_stamp"},"envelope":{"anchors":[{"chain":"crovia.axiom_graph","height":0,"merkle_proof":"spider_vendor_press_v1","root_at_anchor":"spider_vendor_press_v1"}],"axiom_id":"axm_ff5306de89189279f7f774195e50327fc2e485cbba7da7f8113f2783bd6e5db2","axiom_type":"AX.OBS","body":{"axiom_subtype":"news.vendor_press.v1","category":"news","fingerprint":"68f798fa24c267518bc98ee1ab2fa7ff54b44572acb23e566374249bbf0a0ada","published":"Wed, 22 Jul 2026 00:00:00 -0400","receipt_hash":"68f798fa24c267518bc98ee1ab2fa7ff54b44572acb23e566374249bbf0a0ada","schema":"spider.news.vendor_press.v1","spider":"vendor_press","spider_record":{"axiom_subtype":"news.vendor_press.v1","category":"news","decision_hint":"POSITIVE","envelope_target":"AX.OBS","fingerprint":"68f798fa24c267518bc98ee1ab2fa7ff54b44572acb23e566374249bbf0a0ada","observed_at":"2026-07-22T04:43:18.261256Z","parent_run_hash":"4765c85b8b4b27ff9a690c1ae11c3b009baa2c60297295f395ad422f5afed68c","published":"Wed, 22 Jul 2026 00:00:00 -0400","runtime_version":"0.1.0","schema":"spider.news.vendor_press.v1","source_status":200,"source_url":"https://export.arxiv.org/rss/cs.AI","spider":"vendor_press","summary_excerpt":"arXiv:2607.18292v1 Announce Type: cross \nAbstract: 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 scaling, the knowledge gap falls $\\approx$$6\\times$ while knowledge degradation grows $11$-$39\\times$; (ii) at a fabrication, felt uncertainty $H(p_M)$ relaxes quickly while oracle-referenced risk persists up to $17\\times$ longer, leaving a confident-but-precarious risk regime that bridges consecutive fabrications ($+69\\%$ at $14$B); (iii) this regi","title":"Reliability Scales Inversely: Bigger Models Compound Mistakes Faster via a Hidden Auto-Regressive Risk Regime","url":"https://arxiv.org/abs/2607.18292","vendor":"arxiv_cs_ai"},"summary":"arXiv:2607.18292v1 Announce Type: cross \nAbstract: As language models scale, answers start truer but degrade faster: scaling buys capability but erodes reliability. The knowledge-gap account - more data, retrieval, or scale - misses an auto-regressive risk residual that scale sharpens: the model commits to a low-probability token, conditions on it as established, and snowballs. We track this through per-position disagreement $\\delta = \\log p_M - \\log p_O$ against a stronger same-family oracle, whose second moment splits exactly into bias$^2$ $\\mathrm{KL}(p_M \\,\\|\\, p_O)^2$ and risk $\\mathrm{Var}[\\delta]$. We present four findings: (i) under scaling, the knowledge gap falls $\\approx$$6\\times$ while knowledge degradation grows $11$-$39\\times$; (ii) at a fabrication, felt uncertainty $H(p_M)$ relaxes quickly while oracle-referenced risk persists up to $17\\times$ longer, leaving a confident-but-precarious risk regime that bridges consecutive fabrications ($+69\\%$ at $14$B); (iii) this regi","title":"Reliability Scales Inversely: Bigger Models Compound Mistakes Faster via a Hidden Auto-Regressive Risk Regime","vendor":"arxiv_cs_ai"},"confidence":{"method":"deterministic"},"decision":"POSITIVE","issued_at":"2026-07-22T04:43:18Z","notes":"Spider vendor_press (news) news.vendor_press.v1","object":{"captured_by":"crovia.spider.vendor_press","primary_source_url":"https://arxiv.org/abs/2607.18292"},"predecessors":[],"schema":"crovia.axiom.v1","signature":"ed25519:4a1b362307ea12200919bb4652091212021ab9072faf386c6c4200ed48abe40461eeaebb4c92773098bbfbf7d537925411b9296a2fc103108a070c7dfc92f500","signer":"crovia.substrate","subject":{"observed_at":"2026-07-22T04:43:18Z","source_collector":"spider:vendor_press","target_id":"https://arxiv.org/abs/2607.18292"},"tsa":{"authority":"crovia.substrate.bootstrap","rfc3161_token":"{\"kind\":\"crovia.bootstrap.tsa\",\"source_jsonl\":\"/opt/crovia/spider/data/news/vendor_press_v1.jsonl\",\"source_seal_merkle_root\":\"spider_vendor_press_v1\",\"upgrade_path\":\"Sessione H \\u2014 OpenTimestamps weekly anchor\"}"},"zk_mode":"clear","zk_proof":null},"ledger":{"leaf_hash":"27b004dccd66e9a307ab8fe2fb6c3be19a20bd8f6806f0a55be45d79d48cfb27","leaf_index":340205,"ledger_path":"/opt/crovia/substrate/axiom_ledger.jsonl"},"merkle_proof":{"hash_alg":"sha256","leaf_prefix":"0x00","node_prefix":"0x01","odd_leaf_rule":"duplicate_last","path":[{"sibling":"e8f39012202ac4da8ae0d659cb5aa7c6da35b9c09f8b321802117aacbcbd8da8","side":"left"},{"sibling":"239da9523098480a995bb86fcca228c68b7fd92affd4ce582961df685df6ab89","side":"right"},{"sibling":"e80f2ea64034cd7739b64e087a0ed83fb0c3c0c354c7f17c459b4a930aa50f80","side":"left"},{"sibling":"131865f26b8d39ac78660bc91ab0cafb14250122de188b5bdf7efb3f80fac352","side":"left"},{"sibling":"7030f57e43a484aa5b758076c98dba4f4dff7d1a1c8c3101ec721a74bf6c929c","side":"right"},{"sibling":"43998a1a57a65bd1de81db5cc2eea6975708ea80e20709bdf1d8a24f45d87a28","side":"left"},{"sibling":"3ef60c9ed8999dfaf391c6bba67940285b69f69ba30251f62e74b5bc37deb1a8","side":"left"},{"sibling":"acd872eabf3ff3deb3760bc63ca3a4ca483306d9b2eb2ddf59ac880a653ae17a","side":"left"},{"sibling":"c0cb3fbdae423fd11084f7ab87b57c412a0ef2e6dcda395ebb9831d480ad609b","side":"right"},{"sibling":"c7fc9d4187cdc36f4c03b4b13daf4b880ea65536b051f71a5cc2543839d02697","side":"right"},{"sibling":"1758ec6ac206ce40e8368cb702195322fe3737d0fb03d8bd9e3b30acc4fa7d81","side":"right"},{"sibling":"0c407f0d553cf3fab8f9bd79205b8180e090cbf29fa0490ebb55155041ad5c86","side":"right"},{"sibling":"2dd9cb2521044ee7c6b74f2315e0a0253b8df0d04a7b810bbbbe7da5a9788769","side":"left"},{"sibling":"21d66dd41003813f710b7617944f1bfba3258658a5d3370c21cad8f9e945bc99","side":"left"},{"sibling":"787ee3744642ff909d610b0514cb100784f0ef1ba0ef4c70a0dc91f0ab2bb192","side":"right"},{"sibling":"9fc8a8ebbc1bff7e62b9f1e1c681c91e7196092ce9551573df6e23096df13e4d","side":"right"},{"sibling":"77025bcb374a7ad74f520643e20a8ae1205a7ee78507b0beb93117760f1c29d3","side":"left"},{"sibling":"ee6f33920899d9bdef2eb706dfff29e26eea61e29ea9824c6c8e6bcd48275d76","side":"right"},{"sibling":"1cecb7f447febd025aac272837c80de218aecc6485d2395a509b2a1f1b9c746e","side":"left"}]},"schema":"crovia.axiom_proof.v1","seal":{"first_collector_run_id":"","first_receipt_hash":"","jsonl_path":"/opt/crovia/substrate/axiom_ledger.jsonl","key_id":"430895f101d38164","last_collector_run_id":"","last_receipt_hash":"","leaf_count":340557,"merkle_root":"7d45d94f20b5bf82263df45b87749e19e06161972f25141dd573cc138d566338","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","run_id":"hourly_json_retrofit_20260722T053701Z","schema":"crovia.seal.v1","seal_family_version":"crovia-seal-family/1","seal_kind":"substrate_batch","sealed_at":"2026-07-22T05:38:39Z","sig_algorithm":"ed25519","signature":"812cb61e90d3582ba508db8515c4168bbf8ee1c6762885609049d012f680f8068f15c98b9accf2042047dcfc6a6fc3885820ee34b0be350846386f64f2591309","signer_version":"1.1.0"},"trust_root":{"key_id":"430895f101d38164","public_key_hex":"cf742e26f75669dc673cb5c0786a1ae23ae8ca19c347317192ce40c28a7ff25c","signature_algorithm":"ed25519","url":"/registry/canon/TRUST_ROOT.md"},"verifier":{"spec":"/registry/canon/AXIOM_RECEIPT_v1.md","url":"/v/axm_ff5306de89189279f7f774195e50327fc2e485cbba7da7f8113f2783bd6e5db2"}}