{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FW6BKMKLLQDRFORAJTAWH3F3GV","short_pith_number":"pith:FW6BKMKL","schema_version":"1.0","canonical_sha256":"2dbc15314b5c0712ba204cc163ecbb35726ac7c3d791f9f3b3f9ae8b5509d830","source":{"kind":"arxiv","id":"2608.13840","version":1},"attestation_state":"computed","paper":{"title":"ASSERT: A Measurement Pipeline for GenAI Audits","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.CL","authors_text":"Abhinav Palia, Chad Atalla, Dan Vann, Emily Sheng, Hannah Washington, Hanna Wallach, Heba Elfardy, Jean Garcia-Gathright, Nicholas Pangakis, P. Alex Dow, Riccardo Fogliato, Sandeep Atluri, Sharman Tan, Xiawei Wang","submitted_at":"2026-08-14T00:07:32Z","abstract_excerpt":"Audits of generative AI (GenAI) systems often summarize behavior as a reported rate: how often the audited system complies with policy. Researchers and stakeholders use that rate to compare systems, track regressions, and gate deployment. A reported rate reflects both the system under audit and the measurement choices behind it, so a change in the rate can leave it unclear whether the system or those choices moved. We introduce ASSERT, a specification-driven measurement pipeline for GenAI audits that ties each reported rate to a written specification of the measurement choices used to produce "},"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":"2608.13840","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-14T00:07:32Z","cross_cats_sorted":["cs.AI","cs.CY"],"title_canon_sha256":"cac8d23fdebbfb4a8199d0ba29a7aaeda661cd30fae581e2cebd63f4ddbfe4e4","abstract_canon_sha256":"d29f17f8c50dee07db4e78394e2c02b09906c7ee99a891ba7bb121d3a144c37b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-17T01:11:16.274125Z","signature_b64":"L5o4tD0/41m1cbZoCAdShWxBVRJD/kCMBH6JPYJG8UUSXNf965rKGtC+Ua3vEWqL2Sv8mdSue0J1s8zIjXi6BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2dbc15314b5c0712ba204cc163ecbb35726ac7c3d791f9f3b3f9ae8b5509d830","last_reissued_at":"2026-08-17T01:11:16.271824Z","signature_status":"signed_v1","first_computed_at":"2026-08-17T01:11:16.271824Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ASSERT: A Measurement Pipeline for GenAI Audits","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.CL","authors_text":"Abhinav Palia, Chad Atalla, Dan Vann, Emily Sheng, Hannah Washington, Hanna Wallach, Heba Elfardy, Jean Garcia-Gathright, Nicholas Pangakis, P. Alex Dow, Riccardo Fogliato, Sandeep Atluri, Sharman Tan, Xiawei Wang","submitted_at":"2026-08-14T00:07:32Z","abstract_excerpt":"Audits of generative AI (GenAI) systems often summarize behavior as a reported rate: how often the audited system complies with policy. Researchers and stakeholders use that rate to compare systems, track regressions, and gate deployment. A reported rate reflects both the system under audit and the measurement choices behind it, so a change in the rate can leave it unclear whether the system or those choices moved. We introduce ASSERT, a specification-driven measurement pipeline for GenAI audits that ties each reported rate to a written specification of the measurement choices used to produce "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.13840","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/2608.13840/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":"2608.13840","created_at":"2026-08-17T01:11:16.272379+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.13840v1","created_at":"2026-08-17T01:11:16.272379+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.13840","created_at":"2026-08-17T01:11:16.272379+00:00"},{"alias_kind":"pith_short_12","alias_value":"FW6BKMKLLQDR","created_at":"2026-08-17T01:11:16.272379+00:00"},{"alias_kind":"pith_short_16","alias_value":"FW6BKMKLLQDRFORA","created_at":"2026-08-17T01:11:16.272379+00:00"},{"alias_kind":"pith_short_8","alias_value":"FW6BKMKL","created_at":"2026-08-17T01:11:16.272379+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/FW6BKMKLLQDRFORAJTAWH3F3GV","json":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV.json","graph_json":"https://pith.science/api/pith-number/FW6BKMKLLQDRFORAJTAWH3F3GV/graph.json","events_json":"https://pith.science/api/pith-number/FW6BKMKLLQDRFORAJTAWH3F3GV/events.json","paper":"https://pith.science/paper/FW6BKMKL"},"agent_actions":{"view_html":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV","download_json":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV.json","view_paper":"https://pith.science/paper/FW6BKMKL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.13840&json=true","fetch_graph":"https://pith.science/api/pith-number/FW6BKMKLLQDRFORAJTAWH3F3GV/graph.json","fetch_events":"https://pith.science/api/pith-number/FW6BKMKLLQDRFORAJTAWH3F3GV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV/action/storage_attestation","attest_author":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV/action/author_attestation","sign_citation":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV/action/citation_signature","submit_replication":"https://pith.science/pith/FW6BKMKLLQDRFORAJTAWH3F3GV/action/replication_record"}},"created_at":"2026-08-17T01:11:16.272379+00:00","updated_at":"2026-08-17T01:11:16.272379+00:00"}