{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YHTKAEPX4YVAWDGRQNO2TE25AC","short_pith_number":"pith:YHTKAEPX","schema_version":"1.0","canonical_sha256":"c1e6a011f7e62a0b0cd1835da9935d008a0173ea87a44df964a8cff750409c27","source":{"kind":"arxiv","id":"2506.03409","version":3},"attestation_state":"computed","paper":{"title":"Technical Options for Flexible Hardware-Enabled Guarantees","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"James Petrie, Onni Aarne","submitted_at":"2025-06-03T21:37:43Z","abstract_excerpt":"Frontier AI models pose increasing risks to public safety and international security, creating a pressing need for AI developers to provide credible guarantees about their development activities without compromising proprietary information. We propose Flexible Hardware-Enabled Guarantees (flexHEG), a system integrated with AI accelerator hardware to enable verifiable claims about compute usage in AI development. The flexHEG system consists of two primary components: an auditable Guarantee Processor that monitors accelerator usage and verifies compliance with specified rules, and a Secure Enclo"},"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":"2506.03409","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-03T21:37:43Z","cross_cats_sorted":[],"title_canon_sha256":"800bb8e3e0ca4ec992c3c423ee6059c1a652f9b1f843f6e4267ca5922f7f3f4a","abstract_canon_sha256":"ca6cb938fedbac1a067ed565a503499774478b7aef5e8d70d36a4aae3ca4f487"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:29.243821Z","signature_b64":"7BFHACQZ3x6qm/gYC185LN7GPyu1apZjMBA+iBrSQ+Ba3jGiy/Gq5uUBRg5dQ8ZzPdrwyAGhdsjENx5WGwuJBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1e6a011f7e62a0b0cd1835da9935d008a0173ea87a44df964a8cff750409c27","last_reissued_at":"2026-07-05T11:23:29.243335Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:29.243335Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Technical Options for Flexible Hardware-Enabled Guarantees","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"James Petrie, Onni Aarne","submitted_at":"2025-06-03T21:37:43Z","abstract_excerpt":"Frontier AI models pose increasing risks to public safety and international security, creating a pressing need for AI developers to provide credible guarantees about their development activities without compromising proprietary information. We propose Flexible Hardware-Enabled Guarantees (flexHEG), a system integrated with AI accelerator hardware to enable verifiable claims about compute usage in AI development. The flexHEG system consists of two primary components: an auditable Guarantee Processor that monitors accelerator usage and verifies compliance with specified rules, and a Secure Enclo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.03409","kind":"arxiv","version":3},"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/2506.03409/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":"2506.03409","created_at":"2026-07-05T11:23:29.243393+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.03409v3","created_at":"2026-07-05T11:23:29.243393+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.03409","created_at":"2026-07-05T11:23:29.243393+00:00"},{"alias_kind":"pith_short_12","alias_value":"YHTKAEPX4YVA","created_at":"2026-07-05T11:23:29.243393+00:00"},{"alias_kind":"pith_short_16","alias_value":"YHTKAEPX4YVAWDGR","created_at":"2026-07-05T11:23:29.243393+00:00"},{"alias_kind":"pith_short_8","alias_value":"YHTKAEPX","created_at":"2026-07-05T11:23:29.243393+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19262","citing_title":"Detecting Hidden ML Training With Zero-Overhead Telemetry","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05433","citing_title":"Zero knowledge verification for frontier AI training is possible","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC","json":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC.json","graph_json":"https://pith.science/api/pith-number/YHTKAEPX4YVAWDGRQNO2TE25AC/graph.json","events_json":"https://pith.science/api/pith-number/YHTKAEPX4YVAWDGRQNO2TE25AC/events.json","paper":"https://pith.science/paper/YHTKAEPX"},"agent_actions":{"view_html":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC","download_json":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC.json","view_paper":"https://pith.science/paper/YHTKAEPX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.03409&json=true","fetch_graph":"https://pith.science/api/pith-number/YHTKAEPX4YVAWDGRQNO2TE25AC/graph.json","fetch_events":"https://pith.science/api/pith-number/YHTKAEPX4YVAWDGRQNO2TE25AC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC/action/storage_attestation","attest_author":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC/action/author_attestation","sign_citation":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC/action/citation_signature","submit_replication":"https://pith.science/pith/YHTKAEPX4YVAWDGRQNO2TE25AC/action/replication_record"}},"created_at":"2026-07-05T11:23:29.243393+00:00","updated_at":"2026-07-05T11:23:29.243393+00:00"}