{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MDRNDRM4QTLE7CRYCDHZSWDEPY","short_pith_number":"pith:MDRNDRM4","schema_version":"1.0","canonical_sha256":"60e2d1c59c84d64f8a3810cf9958647e312f1bf658e206ee4440424d968d4717","source":{"kind":"arxiv","id":"2506.14922","version":2},"attestation_state":"computed","paper":{"title":"FORTRESS: Frontier Risk Evaluation for National Security and Public Safety","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CY","authors_text":"Christina Q. Knight, Julian Michael, Kaustubh Deshpande, Meher Mankikar, Scale Red Team, SEAL Research Team, Ved Sirdeshmukh","submitted_at":"2025-06-17T19:08:02Z","abstract_excerpt":"The rapid advancement of large language models (LLMs) introduces dual-use capabilities that could both threaten and bolster national security and public safety (NSPS). Models implement safeguards to protect against potential misuse relevant to NSPS and allow for benign users to receive helpful information. However, current benchmarks often fail to test safeguard robustness to potential NSPS risks in an objective, robust way. We introduce FORTRESS: 500 expert-crafted adversarial prompts with instance-based rubrics of 4-7 binary questions for automated evaluation across 3 domains (unclassified i"},"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":"2506.14922","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CY","submitted_at":"2025-06-17T19:08:02Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b334ed808ecb02bbf1eea4198b7d632e5ccb35e41355b1e6e25b615882b1b989","abstract_canon_sha256":"b49c347ec7701b0cd22d7eb9c6c1fd3522a9ee70c8fdd92b7d4b7d5d279430a7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:26:52.943222Z","signature_b64":"wzgg1S1s1ShCh7erDaAU3O46Gwo28EetvJD310u4j9qbg1oja7NcJHxjPLo0Hifzj2YCVaBk/2wr3OCzWka1AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60e2d1c59c84d64f8a3810cf9958647e312f1bf658e206ee4440424d968d4717","last_reissued_at":"2026-07-05T11:26:52.942751Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:26:52.942751Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FORTRESS: Frontier Risk Evaluation for National Security and Public Safety","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CY","authors_text":"Christina Q. Knight, Julian Michael, Kaustubh Deshpande, Meher Mankikar, Scale Red Team, SEAL Research Team, Ved Sirdeshmukh","submitted_at":"2025-06-17T19:08:02Z","abstract_excerpt":"The rapid advancement of large language models (LLMs) introduces dual-use capabilities that could both threaten and bolster national security and public safety (NSPS). Models implement safeguards to protect against potential misuse relevant to NSPS and allow for benign users to receive helpful information. However, current benchmarks often fail to test safeguard robustness to potential NSPS risks in an objective, robust way. We introduce FORTRESS: 500 expert-crafted adversarial prompts with instance-based rubrics of 4-7 binary questions for automated evaluation across 3 domains (unclassified i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.14922","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/2506.14922/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":"2506.14922","created_at":"2026-07-05T11:26:52.942802+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.14922v2","created_at":"2026-07-05T11:26:52.942802+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.14922","created_at":"2026-07-05T11:26:52.942802+00:00"},{"alias_kind":"pith_short_12","alias_value":"MDRNDRM4QTLE","created_at":"2026-07-05T11:26:52.942802+00:00"},{"alias_kind":"pith_short_16","alias_value":"MDRNDRM4QTLE7CRY","created_at":"2026-07-05T11:26:52.942802+00:00"},{"alias_kind":"pith_short_8","alias_value":"MDRNDRM4","created_at":"2026-07-05T11:26:52.942802+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05842","citing_title":"Beyond Refusal: A Same-Lineage Study of Aligned and Abliterated LLMs for Vulnerability Analysis","ref_index":49,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31748","citing_title":"Addressing Over-Refusal in LLMs with Competing Rewards","ref_index":114,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14152","citing_title":"ROK-FORTRESS: Measuring the Effect of Geopolitical Transcreation for National Security and Public Safety","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03121","citing_title":"An Independent Safety Evaluation of Kimi K2.5","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08936","citing_title":"Self-ReSET: Learning to Self-Recover from Unsafe Reasoning Trajectories","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08930","citing_title":"Internalizing Safety Understanding in Large Reasoning Models via Verification","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY","json":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY.json","graph_json":"https://pith.science/api/pith-number/MDRNDRM4QTLE7CRYCDHZSWDEPY/graph.json","events_json":"https://pith.science/api/pith-number/MDRNDRM4QTLE7CRYCDHZSWDEPY/events.json","paper":"https://pith.science/paper/MDRNDRM4"},"agent_actions":{"view_html":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY","download_json":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY.json","view_paper":"https://pith.science/paper/MDRNDRM4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.14922&json=true","fetch_graph":"https://pith.science/api/pith-number/MDRNDRM4QTLE7CRYCDHZSWDEPY/graph.json","fetch_events":"https://pith.science/api/pith-number/MDRNDRM4QTLE7CRYCDHZSWDEPY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY/action/storage_attestation","attest_author":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY/action/author_attestation","sign_citation":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY/action/citation_signature","submit_replication":"https://pith.science/pith/MDRNDRM4QTLE7CRYCDHZSWDEPY/action/replication_record"}},"created_at":"2026-07-05T11:26:52.942802+00:00","updated_at":"2026-07-05T11:26:52.942802+00:00"}