{"as_of":"2026-08-20T06:56:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:803c765c1a10e72af6c809f06054792788bdbbbc2a40775bd7fafa3be0943805","coverage":[{"denominator":0,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":51,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":51,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-20T06:33:59.587034+00:00","state":"measured"},{"denominator":51,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":51,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-16T11:34:05.033901Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":1,"source":"pith","source_observed_at":"2026-08-05T02:28:24.338817Z","state":"measured"}],"external_citation_measurements":[{"count":11,"observed_at":"2026-08-05T02:28:24.338817Z","source":"pith"}],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2404.08144","last_updated":"2024-04-17T04:34:39Z","snapshot_observed_at":"2026-08-15T19:39:28.669101Z","submitted_at":"2024-04-11T22:07:19Z","title":"LLM Agents can Autonomously Exploit One-day Vulnerabilities","version":2},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-05-18T04:18:27.597704Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2404.08144"},"observation_digest":"sha256:1b2d45f74dacba98ba103eac402ec771b9bcd73a838d4a106e3050ba87c88371","observation_id":"5397a145-bf4e-4251-a510-aa73e680b479","resolution":{"observed_at":"2026-05-18T04:18:27.721335Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2408.00118","last_updated":"2024-10-02T15:22:49Z","snapshot_observed_at":"2026-08-02T16:20:09.773989Z","submitted_at":"2024-07-31T19:13:07Z","title":"Gemma 2: Improving Open Language Models at a Practical Size","version":3},"reference_index":37,"source":"arxiv_source","source_observed_at":"2026-05-10T12:11:16.326752Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2408.00118"},"observation_digest":"sha256:3d86cd9bd2951a090eae7edcfbc2ae2f8aa57ba3c075dffc6d3b4e509f226156","observation_id":"5e4744a9-9afb-4475-8606-520bfe93d33a","resolution":{"observed_at":"2026-05-10T12:11:16.476742Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-12T22:01:11.507302Z","title":"Phuong, M","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.08088","last_updated":"2024-11-12T18:45:08Z","snapshot_observed_at":"2026-08-17T17:30:03.420101Z","submitted_at":"2024-11-12T18:45:08Z","title":"Safety case template for frontier AI: A cyber inability argument","version":1},"reference_index":60,"source":"pdf_text","source_observed_at":"2026-08-12T22:01:11.507302Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2411.08088"},"observation_digest":"sha256:18ddf7b9a7e60d6bb7dca6b88c00a46d982e039774a4e269822a7b85a29fcd45","observation_id":"8ebdf006-f847-4ed3-a907-983e3caeb6ac","resolution":{"observed_at":"2026-08-12T22:01:11.507302Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-12T18:15:02.620872Z","title":"A., Webson, A., Ho, L., Lin, S., Farquhar, S., Hutter, M., Del \\' e tang, G., Ruoss, A., El - Sayed, S., Brown, S., Dragan, A","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.11731","last_updated":"2024-11-18T16:59:59Z","snapshot_observed_at":"2026-08-15T06:42:12.354799Z","submitted_at":"2024-11-18T16:59:59Z","title":"Moral Persuasion in Large Language Models: Evaluating Susceptibility and Ethical Alignment","version":1},"reference_index":17,"source":"arxiv_source","source_observed_at":"2026-08-12T18:15:02.620872Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2411.11731"},"observation_digest":"sha256:d81701cfea8ad041049e8505d721866fafbc68eaab1f096e93c7be0e24b80d78","observation_id":"15f0547b-94c7-4ce5-b869-0556aad8a54d","resolution":{"observed_at":"2026-08-12T18:15:02.620872Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-12T17:15:21.945350Z","title":"Evaluating frontier models for dangerous capabilities, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2411.12709","last_updated":"2024-11-19T18:25:30Z","snapshot_observed_at":"2026-08-16T03:04:10.336968Z","submitted_at":"2024-11-19T18:25:30Z","title":"Dimensions of Generative AI Evaluation Design","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-12T17:15:21.945350Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2411.12709"},"observation_digest":"sha256:f8153b854e3ac7885f8e126cb96d353830b396350a38cf239c0c02a2195659e5","observation_id":"0353eefc-190f-4539-bbd6-3a7d36c013b0","resolution":{"observed_at":"2026-08-12T17:15:21.945350Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-11T22:21:46.938158Z","title":"Daniel Povey, Arnab Ghoshal, Gilles Boulianne, Lukas Burget, Ondrej Glembek, Nagendra Goel, Mirko Hannemann, Petr Motlicek, Yanmin Qian, Petr Schwarz, et al","venue":null,"work_id":null,"year":2011},"citing_paper":{"arxiv_id":"2412.03556","last_updated":"2024-12-19T22:37:45Z","snapshot_observed_at":"2026-08-16T14:32:35.895208Z","submitted_at":"2024-12-04T18:51:32Z","title":"Best-of-N Jailbreaking","version":2},"reference_index":2024,"source":"pdf_text","source_observed_at":"2026-08-11T22:21:46.938158Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.03556"},"observation_digest":"sha256:6c4877ca405bf1ce172489d77f61a3656fe0d426e0d0d16773505dff8849a472","observation_id":"2f64c4e1-828a-4706-9b42-47ed2dace3a8","resolution":{"observed_at":"2026-08-11T22:21:46.938158Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-11T22:06:54.204725Z","title":"Evaluating Frontier Models for Dangerous Capabilities","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2412.03824","last_updated":"2025-06-01T22:36:28Z","snapshot_observed_at":"2026-08-14T08:21:56.307711Z","submitted_at":"2024-12-05T02:37:51Z","title":"Towards Data Governance of Frontier AI Models","version":2},"reference_index":81,"source":"pdf_text","source_observed_at":"2026-08-11T22:06:54.204725Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.03824"},"observation_digest":"sha256:567b15c72107c2507b9c9f40c6dddef6fe29c4bae5d990c2a91ab38854bf75f8","observation_id":"84930135-d3d2-41d7-a0ea-024504b98d5c","resolution":{"observed_at":"2026-08-11T22:06:54.204725Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2412.04984","last_updated":"2025-01-14T20:16:01Z","snapshot_observed_at":"2026-08-15T04:02:05.429942Z","submitted_at":"2024-12-06T12:09:50Z","title":"Frontier Models are Capable of In-context Scheming","version":2},"reference_index":31,"source":"arxiv_source","source_observed_at":"2026-05-16T14:22:01.616448Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.04984"},"observation_digest":"sha256:1521e1dfcd46acd3bc395e299253727029042d112cf19df21cfe802721f421de","observation_id":"fa14d73f-3b98-467b-98df-ed70fa9be0ee","resolution":{"observed_at":"2026-05-16T14:22:01.671094Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-12T11:56:20.991463Z","title":"Evaluating Frontier Models for Dangerous Capabilities, April 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2412.08653","last_updated":"2024-11-26T18:00:36Z","snapshot_observed_at":"2026-08-18T10:33:59.727047Z","submitted_at":"2024-11-26T18:00:36Z","title":"What AI evaluations for preventing catastrophic risks can and cannot do","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-12T11:56:20.991463Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.08653"},"observation_digest":"sha256:96bdd9f5b168c169f764af6d793df04cf8f12aaf0f021dc99145ce7d43fe0828","observation_id":"d7d3fc15-ed60-4b0e-984f-dca0c53661e7","resolution":{"observed_at":"2026-08-12T11:56:20.991463Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-11T19:37:44.732532Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2412.12140","last_updated":"2024-12-09T15:01:37Z","snapshot_observed_at":"2026-08-18T10:33:26.926267Z","submitted_at":"2024-12-09T15:01:37Z","title":"Frontier AI systems have surpassed the self-replicating red line","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-11T19:37:44.732532Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.12140"},"observation_digest":"sha256:d840f0ac60dc2f5d756236c676c75fa45ab071f8c490535717edab40e8c5fa26","observation_id":"d4832d0c-7de9-4495-92f3-9564d4a34d9b","resolution":{"observed_at":"2026-08-11T19:37:44.732532Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-11T11:30:43.131236Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2412.15433","last_updated":"2024-12-19T22:31:34Z","snapshot_observed_at":"2026-08-18T21:45:03.682655Z","submitted_at":"2024-12-19T22:31:34Z","title":"Quantifying detection rates for dangerous capabilities: a theoretical model of dangerous capability evaluations","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-11T11:30:43.131236Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.15433"},"observation_digest":"sha256:57e549ff314e17d4f3c36d42a02533c280b474f2c5d9d8133853b4920fd15af0","observation_id":"b08d6fd2-6db4-4753-ab21-4885354ae96e","resolution":{"observed_at":"2026-08-11T11:30:43.131236Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2412.16720","last_updated":"2026-04-30T02:46:40Z","snapshot_observed_at":"2026-08-19T11:46:55.171293Z","submitted_at":"2024-12-21T18:04:31Z","title":"OpenAI o1 System Card","version":2},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-05-23T06:39:44.542350Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2412.16720"},"observation_digest":"sha256:ddf507cdd0c9c87955f498dfdf57ac9e43fbcdcd7ecd035194f763321a53948a","observation_id":"f85bd960-a8cd-4ffc-8e96-fbaf221f5797","resolution":{"observed_at":"2026-05-23T06:42:39.642455Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-10T19:50:42.581504Z","title":"A., Webson, A., Ho, L., Lin, S., Farquhar, S., Hutter, M., Deletang, G., Ruoss, A., El-Sayed, S., Brown, S., Dragan, A., Shah, R., Dafoe, A., and Shevlane, T","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2501.09674","last_updated":"2025-01-16T17:11:21Z","snapshot_observed_at":"2026-08-20T04:56:48.360693Z","submitted_at":"2025-01-16T17:11:21Z","title":"Authenticated Delegation and Authorized AI Agents","version":1},"reference_index":65,"source":"pdf_text","source_observed_at":"2026-08-10T19:50:42.581504Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2501.09674"},"observation_digest":"sha256:c55c3e633d60a221a8b9d8b8c8b0ad27bfa35afd5bf43a2c0619dcd18fb8abd9","observation_id":"28512f1c-7477-4b92-a651-cde7deb30b94","resolution":{"observed_at":"2026-08-10T19:50:42.581504Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-09T14:48:35.492465Z","title":"Evaluating frontier models for dangerous capabilities","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2502.01635","last_updated":"2025-02-03T18:59:13Z","snapshot_observed_at":"2026-08-17T18:29:46.642427Z","submitted_at":"2025-02-03T18:59:13Z","title":"The AI Agent Index","version":1},"reference_index":60,"source":"arxiv_source","source_observed_at":"2026-08-09T14:48:35.492465Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2502.01635"},"observation_digest":"sha256:eff09ff3ada6cb258d6884e75ecc394f9bb9dfd1ac3ef63db1f754bd5fa319a8","observation_id":"54ab1573-da71-4eb2-a70f-7d5bdffbc599","resolution":{"observed_at":"2026-08-09T14:48:35.492465Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2502.18864","last_updated":"2025-02-26T06:17:13Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2025-02-26T06:17:13Z","title":"Towards an AI co-scientist","version":1},"reference_index":38,"source":"arxiv_source","source_observed_at":"2026-05-11T13:02:43.571234Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2502.18864"},"observation_digest":"sha256:82ecd7233f872857156dffd98f9395ad5cefac8f1d07fff40962b8213223c70d","observation_id":"aa0b066f-1fcc-4b5a-812a-f607318d8510","resolution":{"observed_at":"2026-05-11T13:02:44.576679Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-16T11:34:05.033901Z","title":"Evaluating Frontier Models for Dangerous Capabilities","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2504.15181","last_updated":"2025-07-22T19:27:15Z","snapshot_observed_at":"2026-08-19T14:18:31.472529Z","submitted_at":"2025-04-21T15:44:01Z","title":"Mapping Industry Practices to the EU AI Act's GPAI Code of Practice Safety and Security Measures","version":2},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-16T11:34:05.033901Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2504.15181"},"observation_digest":"sha256:124736295101faa21efd3fb07295fbc2554354cbc8ef850208adb49c09bcdb64","observation_id":"c7c49b1b-fc83-4f17-bc1f-82192cf068a2","resolution":{"observed_at":"2026-08-16T11:34:05.033901Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-16T05:57:58.287237Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2504.19440","last_updated":"2025-04-28T03:01:51Z","snapshot_observed_at":"2026-08-19T09:22:52.214724Z","submitted_at":"2025-04-28T03:01:51Z","title":"JailbreaksOverTime: Detecting Jailbreak Attacks Under Distribution Shift","version":1},"reference_index":36,"source":"pdf_text","source_observed_at":"2026-08-16T05:57:58.287237Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2504.19440"},"observation_digest":"sha256:f58bdee6df2f11d573efb14e25b8c52506ad215d66afee299073f90ab88fc01f","observation_id":"60966de0-cb01-47a8-9c2a-e3f296d50abb","resolution":{"observed_at":"2026-08-16T05:57:58.287237Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-16T04:53:11.969029Z","title":"Evaluating frontier models for dangerous capabilities.arXiv preprint arXiv:2403.13793, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.00174","last_updated":"2025-05-05T21:12:46Z","snapshot_observed_at":"2026-08-18T19:06:16.018746Z","submitted_at":"2025-04-30T20:44:42Z","title":"Real-World Gaps in AI Governance Research","version":2},"reference_index":64,"source":"pdf_text","source_observed_at":"2026-08-16T04:53:11.969029Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2505.00174"},"observation_digest":"sha256:9231b1dd9a16269640a941b78dcc1853c37ae37bef104072a74fc85fcefcf772","observation_id":"49cc80c4-77d1-44db-b5ae-09685e1e801c","resolution":{"observed_at":"2026-08-16T04:53:11.969029Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-16T04:22:42.676750Z","title":"Phuong, M","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.01420","last_updated":"2025-07-03T17:52:47Z","snapshot_observed_at":"2026-08-20T00:14:58.043944Z","submitted_at":"2025-05-02T17:57:14Z","title":"Evaluating Frontier Models for Stealth and Situational Awareness","version":4},"reference_index":29,"source":"arxiv_source","source_observed_at":"2026-08-16T04:22:42.676750Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2505.01420"},"observation_digest":"sha256:ff691a811db31c7aec09d2dfb5e19baa44cbaf1b11743d3f4d60d35a31076068","observation_id":"fff4a8ad-7dad-4e6c-9731-1915a374fa2c","resolution":{"observed_at":"2026-08-16T04:22:42.676750Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T23:27:30.430189Z","title":"Evaluating Frontier Models for Dangerous Capabilities , April 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.04592","last_updated":"2025-05-07T17:35:36Z","snapshot_observed_at":"2026-08-17T23:42:59.217553Z","submitted_at":"2025-05-07T17:35:36Z","title":"AI Governance to Avoid Extinction: The Strategic Landscape and Actionable Research Questions","version":1},"reference_index":156,"source":"arxiv_source","source_observed_at":"2026-08-15T23:27:30.430189Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2505.04592"},"observation_digest":"sha256:21d43f07a11394c88017eff411951956e5f82026238ef689b8c4b0e1af3a3cd0","observation_id":"4badcd9d-26c5-4958-8f5e-052cd73a5167","resolution":{"observed_at":"2026-08-15T23:27:30.430189Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T21:03:00.711731Z","title":"arXiv preprint arXiv:2403.13793 (2024)","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.11565","last_updated":"2025-05-24T03:07:23Z","snapshot_observed_at":"2026-08-19T00:42:03.420156Z","submitted_at":"2025-05-16T08:40:09Z","title":"ACSE-Eval: Can LLMs threat model real-world cloud infrastructure?","version":2},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-15T21:03:00.711731Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2505.11565"},"observation_digest":"sha256:52559f0fcf10b95365051dfc8bce5f8e859c107104ec5433a29b61d0910b9646","observation_id":"1f46d63e-ab30-4b48-b0fb-6d5d6d486985","resolution":{"observed_at":"2026-08-15T21:03:00.711731Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T20:28:59.431537Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.12786","last_updated":"2025-05-27T05:00:23Z","snapshot_observed_at":"2026-08-20T02:04:57.449283Z","submitted_at":"2025-05-19T07:19:06Z","title":"Forewarned is Forearmed: A Survey on Large Language Model-based Agents in Autonomous Cyberattacks","version":2},"reference_index":154,"source":"pdf_text","source_observed_at":"2026-08-15T20:28:59.431537Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2505.12786"},"observation_digest":"sha256:e0583357cc25b46998f003c445819b808f5bcf0721cbfcaa677411d2193161e6","observation_id":"291bbf65-44a7-4f41-afc8-e085cebc48d4","resolution":{"observed_at":"2026-08-15T20:28:59.431537Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2506.06414","last_updated":"2026-04-20T20:58:30Z","snapshot_observed_at":"2026-08-10T17:13:42.931667Z","submitted_at":"2025-06-06T17:33:33Z","title":"Benchmarking Misuse Mitigation Against Covert Adversaries","version":2},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-05-19T10:29:05.104520Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2506.06414"},"observation_digest":"sha256:b7cad6ac5dce9cb934700cf356a47048257f96214f29343b75186238c3990853","observation_id":"316b0a8f-dd34-417e-8046-86a1e7ec13d6","resolution":{"observed_at":"2026-05-19T10:32:14.642218Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-07T04:34:08.884266Z","title":"arXiv preprint arXiv:2403.13793","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2506.10217","last_updated":"2025-07-01T01:49:13Z","snapshot_observed_at":"2026-08-15T12:25:24.343032Z","submitted_at":"2025-06-11T22:26:50Z","title":"Data-Centric Safety and Ethical Measures for Data and AI Governance","version":3},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-07T04:34:08.884266Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2506.10217"},"observation_digest":"sha256:1368ef1882a20fde654c058afbbed59278d2dee14dbca4a3db4da9ff13038a75","observation_id":"40c2ff1d-0346-4269-9b7a-7b608aa67487","resolution":{"observed_at":"2026-08-07T04:34:08.884266Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-07T04:17:20.085006Z","title":"Open AI 2024c (September 2024)","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2506.13798","last_updated":"2025-06-12T19:53:38Z","snapshot_observed_at":"2026-08-14T14:53:24.357334Z","submitted_at":"2025-06-12T19:53:38Z","title":"Contemporary AI foundation models increase biological weapons risk","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-07T04:17:20.085006Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2506.13798"},"observation_digest":"sha256:2ae46853f910e51f50fc1a4e7176eab49b6026ec8dfb7d1494ebbac73d684c33","observation_id":"64cb3526-2168-4d5e-96d6-e947c741a752","resolution":{"observed_at":"2026-08-07T04:17:20.085006Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T19:47:59.062215Z","title":"Evaluating Frontier Models for Dangerous Capabilities,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2506.15100","last_updated":"2025-06-18T03:10:49Z","snapshot_observed_at":"2026-08-18T08:15:24.074006Z","submitted_at":"2025-06-18T03:10:49Z","title":"International Security Applications of Flexible Hardware-Enabled Guarantees","version":1},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-08-15T19:47:59.062215Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2506.15100"},"observation_digest":"sha256:f73db28c02339ebc927813a3ad45ff15d5514899f445f129951fec343b2fe386","observation_id":"ff62de0c-4161-493e-8c1b-09bf990d1e6f","resolution":{"observed_at":"2026-08-15T19:47:59.062215Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T19:54:52.954857Z","title":"Evaluating frontier models for dangerous capabilities.arXiv preprint arXiv:2403.13793, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2506.15740","last_updated":"2025-07-08T21:23:22Z","snapshot_observed_at":"2026-08-17T23:40:36.693819Z","submitted_at":"2025-06-17T15:46:15Z","title":"SHADE-Arena: Evaluating Sabotage and Monitoring in LLM Agents","version":2},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-15T19:54:52.954857Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2506.15740"},"observation_digest":"sha256:5aa43caaafcc0b60ab9777ce68d13c946f562534d5b1b5ffa04ee08cd007cb9f","observation_id":"4a56befd-c9be-42e5-9209-0b30a053139b","resolution":{"observed_at":"2026-08-15T19:54:52.954857Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2507.06261","last_updated":"2025-12-19T14:25:46Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2025-07-07T17:36:04Z","title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","version":6},"reference_index":64,"source":"arxiv_source","source_observed_at":"2026-05-19T05:48:02.828938Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2507.06261"},"observation_digest":"sha256:d225da8dc52599ff3de34ce44486fdc9aa0e8e0f6f16efbbb22a1dcc9dd12aaf","observation_id":"be9f9fa3-1e59-4bc6-b345-e655913e4d13","resolution":{"observed_at":"2026-05-19T05:52:07.931876Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-06T16:39:37.129859Z","title":"arXiv:2403.13793","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2507.12872","last_updated":"2025-07-17T07:45:53Z","snapshot_observed_at":"2026-08-19T16:29:33.572763Z","submitted_at":"2025-07-17T07:45:53Z","title":"Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework","version":1},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-06T16:39:37.129859Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2507.12872"},"observation_digest":"sha256:9b15d2e5a1bf6aa89f6ffc95c673f196de073fa9a68c4232f18a389e77c35551","observation_id":"e107757b-5e07-42fd-a604-df71d28c0cf0","resolution":{"observed_at":"2026-08-06T16:39:37.129859Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-06T15:14:22.926909Z","title":"OpenAI Re- search Publication","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2507.16534","last_updated":"2025-07-26T12:33:42Z","snapshot_observed_at":"2026-08-15T01:11:58.606591Z","submitted_at":"2025-07-22T12:44:38Z","title":"Frontier AI Risk Management Framework in Practice: A Risk Analysis Technical Report","version":2},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-08-06T15:14:22.926909Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2507.16534"},"observation_digest":"sha256:d481e10d5faa3c06ed7a3b1e3e3f375b5d2a71cd04af787fbd89d2f0ea4accea","observation_id":"ded1746d-a757-43f2-ae1c-421e4d68f658","resolution":{"observed_at":"2026-08-06T15:14:22.926909Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T15:53:53.446112Z","title":"Phuong, M","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2508.19461","last_updated":"2025-08-26T22:29:31Z","snapshot_observed_at":"2026-08-13T22:40:03.202069Z","submitted_at":"2025-08-26T22:29:31Z","title":"Reliable Weak-to-Strong Monitoring of LLM Agents","version":1},"reference_index":45,"source":"pdf_text","source_observed_at":"2026-08-05T15:53:53.446112Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2508.19461"},"observation_digest":"sha256:707932b57a04ad68525a5fa0429ec10c1a6c4894f970f53825084a483f412f78","observation_id":"cdd23794-a56e-411d-9767-1858edc9d8c5","resolution":{"observed_at":"2026-08-05T15:53:53.446112Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T15:43:59.371101Z","title":"Evaluating frontier models for dangerous capabilities, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2509.00085","last_updated":"2025-08-27T07:08:44Z","snapshot_observed_at":"2026-08-13T09:09:43.124306Z","submitted_at":"2025-08-27T07:08:44Z","title":"Private, Verifiable, and Auditable AI Systems","version":1},"reference_index":225,"source":"pdf_text","source_observed_at":"2026-08-05T15:43:59.371101Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2509.00085"},"observation_digest":"sha256:94c9e28d4705a36950fbbfe7f24799d82eeada19c114062af69741e11ce97412","observation_id":"7c944593-4e77-41ff-8120-20da83a075f2","resolution":{"observed_at":"2026-08-05T15:43:59.371101Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-04T14:53:17.688392Z","title":"Evaluating frontier models for dangerous capabilities","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2509.22504","last_updated":"2026-05-28T15:39:50Z","snapshot_observed_at":"2026-08-18T21:20:26.454560Z","submitted_at":"2025-09-26T15:46:14Z","title":"Estimating the Empowerment of Language Model Agents","version":3},"reference_index":31,"source":"arxiv_source","source_observed_at":"2026-08-04T14:53:17.688392Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2509.22504"},"observation_digest":"sha256:248cd55b86ed5d8f948a90e3349349e9615afed6d9dea1861625d32af9ad4d18","observation_id":"93071eb1-00da-4785-8cae-d7d45886470a","resolution":{"observed_at":"2026-08-04T14:53:17.688392Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2512.10687","last_updated":"2026-04-20T11:46:57Z","snapshot_observed_at":"2026-08-11T08:27:08.305717Z","submitted_at":"2025-12-11T14:34:40Z","title":"Safe for Whom? Rethinking How We Evaluate the Safety of LLMs for Real Users","version":2},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-05-16T23:22:42.431997Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2512.10687"},"observation_digest":"sha256:8318fe56843627e14fcd3c6edd7fd36c24527a505bec32e2f4ce0fae670ceab0","observation_id":"98004cc5-5adf-4f1f-b56a-8b3576daf855","resolution":{"observed_at":"2026-05-16T23:23:39.955248Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2602.17753","last_updated":"2026-05-06T13:43:46Z","snapshot_observed_at":"2026-08-16T03:02:11.597234Z","submitted_at":"2026-02-19T18:57:43Z","title":"The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI Systems","version":2},"reference_index":103,"source":"pdf_text","source_observed_at":"2026-05-15T20:41:49.137745Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2602.17753"},"observation_digest":"sha256:3ae03123ba4960d0764c693f6332900820c7bdcdbe9e08d49642c2096def5f51","observation_id":"9fa259d1-9099-4208-a1de-04ac5e511324","resolution":{"observed_at":"2026-05-15T20:46:36.241926Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2604.14070","last_updated":"2026-04-15T16:39:28Z","snapshot_observed_at":"2026-08-17T11:27:14.703954Z","submitted_at":"2026-04-15T16:39:28Z","title":"From Disclosure to Self-Referential Opacity: Six Dimensions of Strain in Current AI Governance","version":1},"reference_index":69,"source":"pdf_text","source_observed_at":"2026-05-10T11:55:50.822764Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2604.14070"},"observation_digest":"sha256:63ae94ff7c6b036cfea2210971f7dc5a9b119583710e7b60c55188c64eef2c67","observation_id":"bf4cfde7-10f5-433a-9099-a2b0e77d19be","resolution":{"observed_at":"2026-05-10T12:00:22.250927Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2604.20389","last_updated":"2026-04-22T09:44:38Z","snapshot_observed_at":"2026-07-06T23:06:49.210284Z","submitted_at":"2026-04-22T09:44:38Z","title":"CyberCertBench: Evaluating LLMs in Cybersecurity Certification Knowledge","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-05-10T00:32:59.120345Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2604.20389"},"observation_digest":"sha256:4f34f4c79bf85a0e514d3f53338582347228fa096025ad51c494e4471ee2ebef","observation_id":"3fb4d26d-f3b0-4c3e-ad66-36d994d3cc4a","resolution":{"observed_at":"2026-05-10T00:34:47.049775Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2605.11086","last_updated":"2026-05-11T18:00:14Z","snapshot_observed_at":"2026-08-16T23:33:20.820875Z","submitted_at":"2026-05-11T18:00:14Z","title":"ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?","version":1},"reference_index":44,"source":"pdf_text","source_observed_at":"2026-05-13T03:04:14.130717Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2605.11086"},"observation_digest":"sha256:49a333da3d273920da379e62de80e3b56492b99e75c48476eb784086f0d613aa","observation_id":"8a653bfd-6ffc-4603-9442-4775de29256c","resolution":{"observed_at":"2026-05-13T03:07:09.009924Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2605.12131","last_updated":"2026-05-12T13:54:31Z","snapshot_observed_at":"2026-08-13T22:02:24.826201Z","submitted_at":"2026-05-12T13:54:31Z","title":"Rollout Cards: A Reproducibility Standard for Agent Research","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-05-13T05:57:48.614881Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2605.12131"},"observation_digest":"sha256:f847209140f1504e1af047b1dc6a26fae5a23eec89a2bc0b57e5296e352358da","observation_id":"d378f5ed-b3e3-4af3-ab17-32288ec8738b","resolution":{"observed_at":"2026-05-13T06:02:23.587743Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2605.19722","last_updated":"2026-05-19T11:55:54Z","snapshot_observed_at":"2026-08-16T02:40:10.272246Z","submitted_at":"2026-05-19T11:55:54Z","title":"Measuring Safety Alignment Effects in Autonomous Security Agents","version":1},"reference_index":49,"source":"pdf_text","source_observed_at":"2026-05-20T04:24:46.357505Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2605.19722"},"observation_digest":"sha256:cec94823f6dca6356b7f8633bf3aa1d4f482c6c49523e75a0abd00930902316a","observation_id":"0672b9c5-6fb0-439c-9e09-65e98af93209","resolution":{"observed_at":"2026-05-20T04:28:05.916500Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.00047","last_updated":"2026-05-01T08:30:27Z","snapshot_observed_at":"2026-08-15T13:07:40.446930Z","submitted_at":"2026-05-01T08:30:27Z","title":"Comprehensive AI governance requires addressing non-model gains","version":1},"reference_index":76,"source":"arxiv_source","source_observed_at":"2026-07-01T08:11:23.860723Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.00047"},"observation_digest":"sha256:bb558e27a3b3f3f09f782b56ed52dcc21765537dc0fcc266822d67183de947df","observation_id":"2e85b7ac-258d-4977-8ac5-73376046bb6f","resolution":{"observed_at":"2026-07-01T08:15:31.663670Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.03237","last_updated":"2026-06-02T06:54:55Z","snapshot_observed_at":"2026-08-15T00:58:19.037726Z","submitted_at":"2026-06-02T06:54:55Z","title":"Solipsistic Superintelligence is Unlikely to be Cooperative","version":1},"reference_index":199,"source":"arxiv_source","source_observed_at":"2026-06-28T10:00:41.374121Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.03237"},"observation_digest":"sha256:2c0504876f572feeb65e19ad902bcc7927bbb1c2948f4db0e8a22749590f41d9","observation_id":"b787e572-707f-449d-9223-02a0fb9e8fcb","resolution":{"observed_at":"2026-07-02T03:26:29.746920Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.05330","last_updated":"2026-06-03T18:17:20Z","snapshot_observed_at":"2026-08-14T12:16:24.588905Z","submitted_at":"2026-06-03T18:17:20Z","title":"A Model of Multi-turn Human Persuadability Using Probabilistic Belief Tracing","version":1},"reference_index":103,"source":"pdf_text","source_observed_at":"2026-06-28T06:17:01.173495Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.05330"},"observation_digest":"sha256:656900519493dc07abbf5794cacb5322d8b9a8cdfb47f0dc0c3cc2349edd33db","observation_id":"80ecbb19-e52f-46ae-9021-c419ae918ce2","resolution":{"observed_at":"2026-07-02T08:16:47.572672Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.13079","last_updated":"2026-06-29T13:15:12Z","snapshot_observed_at":"2026-08-13T01:36:54.597641Z","submitted_at":"2026-06-11T09:02:14Z","title":"The Emergence of Autonomous Penetration Capabilities in Large Language Model-Powered AI Systems","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-06-27T06:25:17.298114Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.13079"},"observation_digest":"sha256:5f2e11b01f10b4bb8d0a1437400a5922da5c07ccf7435d4084b030d39adbbd83","observation_id":"238c4677-edad-4e28-805c-67e7e959149c","resolution":{"observed_at":"2026-07-03T15:38:33.531225Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.13079","last_updated":"2026-06-29T13:15:12Z","snapshot_observed_at":"2026-08-13T01:36:54.597641Z","submitted_at":"2026-06-11T09:02:14Z","title":"The Emergence of Autonomous Penetration Capabilities in Large Language Model-Powered AI Systems","version":2},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-06-30T11:27:22.436981Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.13079"},"observation_digest":"sha256:fced67b45578848500a5b66fa0653ad08f35b142618db76012ea373ebcfddfaf","observation_id":"498984e8-a9ec-4251-80d7-a6089f9abafb","resolution":{"observed_at":"2026-06-30T11:34:38.126424Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.18532","last_updated":"2026-06-16T22:57:24Z","snapshot_observed_at":"2026-08-17T01:13:56.343799Z","submitted_at":"2026-06-16T22:57:24Z","title":"AI Sandboxes: A Threat Model, Taxonomy, and Measurement Framework","version":1},"reference_index":122,"source":"pdf_text","source_observed_at":"2026-06-26T23:42:20.304205Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.18532"},"observation_digest":"sha256:6abe03a5b9fdf320d2802a0eff2aa55f850f6d071f7d86d837a3e3e0977a1a12","observation_id":"f4441bf2-4b62-4cd4-9a17-81337fcf7b1b","resolution":{"observed_at":"2026-07-03T22:08:59.996827Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2606.31916","last_updated":"2026-06-30T16:22:12Z","snapshot_observed_at":"2026-08-12T20:49:26.366092Z","submitted_at":"2026-06-30T16:22:12Z","title":"Theory of Mind and Persuasion Beyond Conversation: Assessing the Capacity of LLMs to Induce Belief States via Planning and Action","version":1},"reference_index":75,"source":"arxiv_source","source_observed_at":"2026-07-01T05:40:54.002702Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2606.31916"},"observation_digest":"sha256:70abc6ed0b7a151d745ac51eb005d4d2a61489108293febfb0f2205bfbe6ef9c","observation_id":"2943cd42-a5e2-4f2c-b94b-4cd3e36984e8","resolution":{"observed_at":"2026-07-01T10:15:44.646309Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-07-12T01:32:59.751620Z","title":"arXiv preprint arXiv:2403.13793 (2024)","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.05442","last_updated":"2026-07-03T19:09:01Z","snapshot_observed_at":"2026-08-19T23:04:30.160733Z","submitted_at":"2026-07-03T19:09:01Z","title":"The Oracle's Gambit: A Game-Theoretic Framework for Responsible AI Release","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-07-12T01:32:59.751620Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2607.05442"},"observation_digest":"sha256:fffccd44bf68de6b2e42fe1151611aa4daf37758de813d2a4413aed2e702dcb9","observation_id":"c0a59a1d-43da-4459-8037-1aed539ea542","resolution":{"observed_at":"2026-07-12T01:32:59.751620Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":"2403.13793","doi":"10.48550/arxiv.2403.13793","metadata_source":"pith","pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"Evaluating frontier models for dangerous capabilities","venue":"cs.LG","work_id":"2f7df923-9116-4ade-9fca-e78987af4e53","year":2024},"citing_paper":{"arxiv_id":"2607.08066","last_updated":"2026-07-09T02:48:36Z","snapshot_observed_at":"2026-08-18T08:26:26.366694Z","submitted_at":"2026-07-09T02:48:36Z","title":"Persuasion Attacks Can Decrease Effectiveness of CoT Monitoring","version":1},"reference_index":62,"source":"arxiv_source","source_observed_at":"2026-07-10T00:52:47.537142Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2607.08066"},"observation_digest":"sha256:95dac0f5d35029fa8f995c28b53cf068acd0fbed51a112664bd676cf88150138","observation_id":"0746f56a-9ea6-4c5c-9ca1-9d2e0f48bd6a","resolution":{"observed_at":"2026-07-10T00:56:40.994995Z","resolver_source":"local_arxiv","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-01T04:12:58.764877Z","title":"Evaluating frontier models for dangerous capabilities","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.22926","last_updated":"2026-07-24T21:36:28Z","snapshot_observed_at":"2026-08-17T16:17:48.697721Z","submitted_at":"2026-07-24T21:36:28Z","title":"SAGE: Safety-First Defense-in-Depth Guardrails for Verified Lifecycle Control of High-Impact Generative AI","version":1},"reference_index":62,"source":"arxiv_source","source_observed_at":"2026-08-01T04:12:58.764877Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2607.22926"},"observation_digest":"sha256:6794623b2773ebade48350707dbed5e4aa4b7d8f0e39b962a68cd2eb64318737","observation_id":"1cc1b510-a081-4d37-a465-c8fa80c98ce4","resolution":{"observed_at":"2026-08-01T04:12:58.764877Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.13793","snapshot_observed_at":"2026-08-15T14:51:31.797330Z","title":"Evaluating frontier models for dangerous capabilities.arXiv preprint arXiv:2403.13793, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2608.03670","last_updated":"2026-08-04T13:47:54Z","snapshot_observed_at":"2026-08-20T00:36:23.547270Z","submitted_at":"2026-08-04T13:47:54Z","title":"Accountability Asymmetry and Structural Trust in Autonomous AI Systems","version":1},"reference_index":40,"source":"pdf_text","source_observed_at":"2026-08-15T14:51:31.797330Z"},"links":{"cited_paper":"/paper/2403.13793","citing_paper":"/paper/2608.03670"},"observation_digest":"sha256:0ce4fa6981ddfad47a60ed8c5fc211bdbf56f1e232ad93b28524908d51ad2bb9","observation_id":"4b1eb21c-0605-4abb-a4cf-76023772b512","resolution":{"observed_at":"2026-08-15T14:51:31.797330Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"links":{"evidence":"/evidence","html":"/paper/2403.13793/citation-record","integrity":"/paper/2403.13793/integrity","json":"/paper/2403.13793/citation-record.json","paper":"/paper/2403.13793"},"outbound":[],"paper":{"arxiv_id":"2403.13793","last_updated":"2024-04-05T12:26:11Z","latest_version":2,"primary_category":"cs.LG","snapshot_observed_at":"2026-08-16T14:07:56.490852Z","submitted_at":"2024-03-20T17:54:26Z","title":"Evaluating Frontier Models for Dangerous Capabilities"},"reference_resolution":{"displayed":0,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":0,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":0},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-20T06:33:59.587034+00:00","source":"crossref"},{"observed_at":"2026-08-20T06:33:54.927442+00:00","source":"retraction_watch"}],"thesis":"As of 20 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 51 inbound Pith citation observations for arXiv:2403.13793."}