{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZFYYO6SH322AZISCQBWU7GP7EV","short_pith_number":"pith:ZFYYO6SH","schema_version":"1.0","canonical_sha256":"c971877a47deb40ca242806d4f99ff255d9786e8eedc1492bd261f57614b702c","source":{"kind":"arxiv","id":"2603.25450","version":2},"attestation_state":"computed","paper":{"title":"Cross-Model Disagreement as a Label-Free Correctness Signal","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Matt Gorbett, Suman Jana","submitted_at":"2026-03-26T13:46:22Z","abstract_excerpt":"Detecting when a language model is wrong without ground truth labels is a fundamental challenge for safe deployment. Existing approaches rely on a model's own uncertainty -- such as token entropy or confidence scores -- but these signals fail critically on the most dangerous failure mode: confident errors, where a model is wrong but certain. In this work we introduce cross-model disagreement as a correctness indicator -- a simple, training-free signal that can be dropped into existing production systems, routing pipelines, and deployment monitoring infrastructure without modification. Given a "},"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":"2603.25450","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-03-26T13:46:22Z","cross_cats_sorted":[],"title_canon_sha256":"48acb70f3e4df6e5f4c5591d4f087aea58d4c9458cb8382ac32df4031b5bc088","abstract_canon_sha256":"aa4211b9f5851b1b597a39b46788d41ebe7bec0136849580a29fc77f6d33b682"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-12T01:08:24.017957Z","signature_b64":"qIMfxCDl/hqqxLiSmuYfUa5RrdvnvjgJOcfQFrTCTv2+HqS+ZOt21ApCCCyu4keDq5JqTLZSkix9+TniUJzjBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c971877a47deb40ca242806d4f99ff255d9786e8eedc1492bd261f57614b702c","last_reissued_at":"2026-06-12T01:08:24.016876Z","signature_status":"signed_v1","first_computed_at":"2026-06-12T01:08:24.016876Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Cross-Model Disagreement as a Label-Free Correctness Signal","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Matt Gorbett, Suman Jana","submitted_at":"2026-03-26T13:46:22Z","abstract_excerpt":"Detecting when a language model is wrong without ground truth labels is a fundamental challenge for safe deployment. Existing approaches rely on a model's own uncertainty -- such as token entropy or confidence scores -- but these signals fail critically on the most dangerous failure mode: confident errors, where a model is wrong but certain. In this work we introduce cross-model disagreement as a correctness indicator -- a simple, training-free signal that can be dropped into existing production systems, routing pipelines, and deployment monitoring infrastructure without modification. Given a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.25450","kind":"arxiv","version":2},"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/2603.25450/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":"2603.25450","created_at":"2026-06-12T01:08:24.017031+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.25450v2","created_at":"2026-06-12T01:08:24.017031+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.25450","created_at":"2026-06-12T01:08:24.017031+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZFYYO6SH322A","created_at":"2026-06-12T01:08:24.017031+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZFYYO6SH322AZISC","created_at":"2026-06-12T01:08:24.017031+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZFYYO6SH","created_at":"2026-06-12T01:08:24.017031+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/ZFYYO6SH322AZISCQBWU7GP7EV","json":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV.json","graph_json":"https://pith.science/api/pith-number/ZFYYO6SH322AZISCQBWU7GP7EV/graph.json","events_json":"https://pith.science/api/pith-number/ZFYYO6SH322AZISCQBWU7GP7EV/events.json","paper":"https://pith.science/paper/ZFYYO6SH"},"agent_actions":{"view_html":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV","download_json":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV.json","view_paper":"https://pith.science/paper/ZFYYO6SH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.25450&json=true","fetch_graph":"https://pith.science/api/pith-number/ZFYYO6SH322AZISCQBWU7GP7EV/graph.json","fetch_events":"https://pith.science/api/pith-number/ZFYYO6SH322AZISCQBWU7GP7EV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV/action/storage_attestation","attest_author":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV/action/author_attestation","sign_citation":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV/action/citation_signature","submit_replication":"https://pith.science/pith/ZFYYO6SH322AZISCQBWU7GP7EV/action/replication_record"}},"created_at":"2026-06-12T01:08:24.017031+00:00","updated_at":"2026-06-12T01:08:24.017031+00:00"}