{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:7RASXS3VVIFPBBCVDB7HT3EBS6","short_pith_number":"pith:7RASXS3V","schema_version":"1.0","canonical_sha256":"fc412bcb75aa0af08455187e79ec8197b10429f3adc89458295383e1322b5e94","source":{"kind":"arxiv","id":"2409.11052","version":1},"attestation_state":"computed","paper":{"title":"A logical alarm for misaligned binary classifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andr\\'es Corrada-Emmanuel, Ilya Parker, Ramesh Bharadwaj","submitted_at":"2024-09-17T10:19:22Z","abstract_excerpt":"If two agents disagree in their decisions, we may suspect they are not both correct. This intuition is formalized for evaluating agents that have carried out a binary classification task. Their agreements and disagreements on a joint test allow us to establish the only group evaluations logically consistent with their responses. This is done by establishing a set of axioms (algebraic relations) that must be universally obeyed by all evaluations of binary responders. A complete set of such axioms are possible for each ensemble of size N. The axioms for $N = 1, 2$ are used to construct a fully l"},"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":"2409.11052","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-17T10:19:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"62d8ea19d0e248a97f780e92ace47ff79e93d5fa81e46b19c04469da17d2361c","abstract_canon_sha256":"61b5cf9e136546259b2f0365b78c58c8e0dba355d699563d9363ed8363e5bf69"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:08:01.140082Z","signature_b64":"Kzytu+XzI0h9UdgbKl5VmtT5srB0UzwoLd/IIX+2Kuc4km5CzJ2bIazselrG9/Tz5g0/+OOAdIS64UutfUNQBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc412bcb75aa0af08455187e79ec8197b10429f3adc89458295383e1322b5e94","last_reissued_at":"2026-07-05T09:08:01.139556Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:08:01.139556Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A logical alarm for misaligned binary classifiers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andr\\'es Corrada-Emmanuel, Ilya Parker, Ramesh Bharadwaj","submitted_at":"2024-09-17T10:19:22Z","abstract_excerpt":"If two agents disagree in their decisions, we may suspect they are not both correct. This intuition is formalized for evaluating agents that have carried out a binary classification task. Their agreements and disagreements on a joint test allow us to establish the only group evaluations logically consistent with their responses. This is done by establishing a set of axioms (algebraic relations) that must be universally obeyed by all evaluations of binary responders. A complete set of such axioms are possible for each ensemble of size N. The axioms for $N = 1, 2$ are used to construct a fully l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.11052","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/2409.11052/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":"2409.11052","created_at":"2026-07-05T09:08:01.139629+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.11052v1","created_at":"2026-07-05T09:08:01.139629+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.11052","created_at":"2026-07-05T09:08:01.139629+00:00"},{"alias_kind":"pith_short_12","alias_value":"7RASXS3VVIFP","created_at":"2026-07-05T09:08:01.139629+00:00"},{"alias_kind":"pith_short_16","alias_value":"7RASXS3VVIFPBBCV","created_at":"2026-07-05T09:08:01.139629+00:00"},{"alias_kind":"pith_short_8","alias_value":"7RASXS3V","created_at":"2026-07-05T09:08:01.139629+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.16238","citing_title":"Algebraic Evaluation Theorems","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6","json":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6.json","graph_json":"https://pith.science/api/pith-number/7RASXS3VVIFPBBCVDB7HT3EBS6/graph.json","events_json":"https://pith.science/api/pith-number/7RASXS3VVIFPBBCVDB7HT3EBS6/events.json","paper":"https://pith.science/paper/7RASXS3V"},"agent_actions":{"view_html":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6","download_json":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6.json","view_paper":"https://pith.science/paper/7RASXS3V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.11052&json=true","fetch_graph":"https://pith.science/api/pith-number/7RASXS3VVIFPBBCVDB7HT3EBS6/graph.json","fetch_events":"https://pith.science/api/pith-number/7RASXS3VVIFPBBCVDB7HT3EBS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6/action/storage_attestation","attest_author":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6/action/author_attestation","sign_citation":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6/action/citation_signature","submit_replication":"https://pith.science/pith/7RASXS3VVIFPBBCVDB7HT3EBS6/action/replication_record"}},"created_at":"2026-07-05T09:08:01.139629+00:00","updated_at":"2026-07-05T09:08:01.139629+00:00"}