{"as_of":"2026-08-10T20:40:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:5ff4c6f9d0e9b735df3d60cebdd9b139ea2605c44bb6634c3cb991addecbf4d6","coverage":[{"denominator":32,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":32,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-03T18:41:35.433927Z","state":"measured"},{"denominator":33,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":33,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-10T06:31:04.303077+00:00","state":"measured"},{"denominator":1,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":1,"source":"paper_references, paper_reference_links","source_observed_at":"2026-07-13T04:30:23.082717Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2512.04144","snapshot_observed_at":"2026-07-13T04:30:23.082717Z","title":"Ripplebench: Capturing ripple effects using existing knowledge repositories.arXiv preprint arXiv:2512.04144, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.09236","last_updated":"2026-07-10T09:31:29Z","snapshot_observed_at":"2026-08-09T15:26:37.592220Z","submitted_at":"2026-07-10T09:31:29Z","title":"Forget Narrowly, Retain Broadly: Unlearning as an Asymmetric Generalization Problem","version":1},"reference_index":37,"source":"pdf_text","source_observed_at":"2026-07-13T04:30:23.082717Z"},"links":{"cited_paper":"/paper/2512.04144","citing_paper":"/paper/2607.09236"},"observation_digest":"sha256:2988a739dacde7262af72d61d2110156c57c09a8bbcde25d6cefc8641a257fba","observation_id":"8e61e3ad-40a0-478c-8275-59f56490515d","resolution":{"observed_at":"2026-07-13T04:30:23.082717Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"links":{"evidence":"/evidence","html":"/paper/2512.04144/citation-record","integrity":"/paper/2512.04144/integrity","json":"/paper/2512.04144/citation-record.json","paper":"/paper/2512.04144"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:32.715667Z","title":"Evaluating the ripple effects of knowledge editing in language models, 2023","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:32.715667Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:462858f8b2e48cecc874ff26951d59a70d06fa199e43ade3d97ae2cd04e52f2b","observation_id":"ee099214-c90f-4ccd-afb3-a3b4112a8437","resolution":{"observed_at":"2026-08-03T18:41:32.715667Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2501.04952","last_updated":"2025-01-09T03:59:10Z","snapshot_observed_at":"2026-08-08T00:46:37.661782Z","submitted_at":"2025-01-09T03:59:10Z","title":"Open Problems in Machine Unlearning for AI Safety","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2501.04952","snapshot_observed_at":"2026-08-03T18:41:32.801107Z","title":"Open problems in machine unlearning for ai safety.arXiv preprint arXiv:2501.04952, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:32.801107Z"},"links":{"cited_paper":"/paper/2501.04952","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:02387e2a6e25716ae933eb1bbfd0e87690b735b2d89c9c41616bcd1894e27ecb","observation_id":"b678310c-67b5-46fd-87e7-4fc6cf614ef6","resolution":{"observed_at":"2026-08-03T18:41:32.801107Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:32.864058Z","title":"Are we making progress in unlearning? findings from the first neurips unlearning competition, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:32.864058Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:17790c6f24df9b3dc466d99b0fc19a6ca9183e704a98fe6bb4627b64e95fe6fc","observation_id":"62b61030-dfc6-4742-890f-b127c6289236","resolution":{"observed_at":"2026-08-03T18:41:32.864058Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2401.06121","last_updated":"2024-01-11T18:57:12Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2024-01-11T18:57:12Z","title":"TOFU: A Task of Fictitious Unlearning for LLMs","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2401.06121","snapshot_observed_at":"2026-08-03T18:41:33.040444Z","title":"Tofu: A task of fictitious unlearning for llms.arXiv preprint arXiv:2401.06121, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.040444Z"},"links":{"cited_paper":"/paper/2401.06121","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:e9f69f33b57af4ce0564b88c1da89c9c4e76b8147747053e7fa4fa153489cd47","observation_id":"04f0f31d-a9e9-4287-941b-46995b0a69ca","resolution":{"observed_at":"2026-08-03T18:41:33.040444Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:33.157442Z","title":"Lipton, J","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.157442Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:a0f754badbcc1150fdd3b800a87a290ed0ae6cd12dd78b5bcfbcfe1a264c9782","observation_id":"d5bcc48f-67b6-441d-b50b-27a509d3cbaa","resolution":{"observed_at":"2026-08-03T18:41:33.157442Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2502.05209","last_updated":"2025-07-24T21:34:54Z","snapshot_observed_at":"2026-08-09T22:40:17.102835Z","submitted_at":"2025-02-03T18:59:16Z","title":"Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2502.05209","snapshot_observed_at":"2026-08-03T18:41:33.209039Z","title":"Model tampering attacks enable more rigorous evaluations of llm capabilities.arXiv preprint arXiv:2502.05209, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.209039Z"},"links":{"cited_paper":"/paper/2502.05209","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:3f24e7d78755c610dc91757a620a475a564794a2dd7f9d485b198917a4cd7a4e","observation_id":"9195ab00-e824-41a4-9ef9-dd9eb8448c61","resolution":{"observed_at":"2026-08-03T18:41:33.209039Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:33.286908Z","title":"Measuring massive multitask language understanding, 2021","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.286908Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:a2872206e8226b2199271c16933d17c6b5a880c12961ef7440905f932c6f1b00","observation_id":"d633b40f-b453-4530-b558-748189c05f0e","resolution":{"observed_at":"2026-08-03T18:41:33.286908Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2402.16835","last_updated":"2024-02-26T18:57:37Z","snapshot_observed_at":"2026-08-08T00:43:51.702314Z","submitted_at":"2024-02-26T18:57:37Z","title":"Eight Methods to Evaluate Robust Unlearning in LLMs","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2402.16835","snapshot_observed_at":"2026-08-03T18:41:33.358328Z","title":"Eight methods to evaluate robust unlearning in llms.arXiv preprint arXiv:2402.16835, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.358328Z"},"links":{"cited_paper":"/paper/2402.16835","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:8ec0157b930bb202c21519a926a5628dfa1787200d313e55c9f4e0a6da0db446","observation_id":"7c398949-171e-4658-96d7-5b041eccaad3","resolution":{"observed_at":"2026-08-03T18:41:33.358328Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2402.04249","last_updated":"2024-02-27T04:43:08Z","snapshot_observed_at":"2026-07-06T17:26:23.067923Z","submitted_at":"2024-02-06T18:59:08Z","title":"HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2402.04249","snapshot_observed_at":"2026-08-03T18:41:33.493629Z","title":"Harmbench: A standardized evaluation framework for automated red teaming and robust refusal.arXiv preprint arXiv:2402.04249, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.493629Z"},"links":{"cited_paper":"/paper/2402.04249","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:762a07e69479c291bc93278c73d78222837846fdfb29ea19b48a925b4a45c188","observation_id":"aa4a9a1a-56b9-490d-a486-50b913eda32d","resolution":{"observed_at":"2026-08-03T18:41:33.493629Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.06622","last_updated":"2024-06-07T15:37:15Z","snapshot_observed_at":"2026-08-08T00:47:02.920080Z","submitted_at":"2024-06-07T15:37:15Z","title":"Adversarial Tuning: Defending Against Jailbreak Attacks for LLMs","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.06622","snapshot_observed_at":"2026-08-03T18:41:33.593441Z","title":"Adversarial tuning: Defending against jailbreak attacks for llms.arXiv preprint arXiv:2406.06622, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.593441Z"},"links":{"cited_paper":"/paper/2406.06622","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:fc615c8cb1ba395a327b6632ccef6dc0f341f06056ed2b6e639656244672c4c3","observation_id":"03c3a2f6-b9a7-42b7-beaf-5b1c4b9a42ac","resolution":{"observed_at":"2026-08-03T18:41:33.593441Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2409.20089","last_updated":"2025-03-20T15:28:18Z","snapshot_observed_at":"2026-08-07T05:14:57.199418Z","submitted_at":"2024-09-30T08:41:39Z","title":"Robust LLM safeguarding via refusal feature adversarial training","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2409.20089","snapshot_observed_at":"2026-08-03T18:41:33.670392Z","title":"Robust llm safeguarding via refusal feature adversarial training.arXiv preprint arXiv:2409.20089, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.670392Z"},"links":{"cited_paper":"/paper/2409.20089","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:47c82687f5d519a60d67e781ae70a511b9096d81955ccae3fe8928740e39e914","observation_id":"c6c009dc-6a89-445f-9866-ee7117d5db89","resolution":{"observed_at":"2026-08-03T18:41:33.670392Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.05030","last_updated":"2025-07-29T09:44:14Z","snapshot_observed_at":"2026-08-07T17:43:00.958834Z","submitted_at":"2024-03-08T04:22:48Z","title":"Defending Against Unforeseen Failure Modes with Latent Adversarial Training","version":6},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.05030","snapshot_observed_at":"2026-08-03T18:41:33.782057Z","title":"Defending against unforeseen failure modes with latent adversarial training.arXiv preprint arXiv:2403.05030, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.782057Z"},"links":{"cited_paper":"/paper/2403.05030","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:c4d844c4535b8a676e87ddc01ccbeb03250e5eeb11ddc6ad0a4722e1b3c30718","observation_id":"a1215b09-08ba-4379-adbd-ac7328f4ee6e","resolution":{"observed_at":"2026-08-03T18:41:33.782057Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:33.935923Z","title":"Rethinking machine unlearning for large language models.Nature Machine Intelligence, pages 1–14, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:33.935923Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:9ef37255f1611e53068784abae39132731e7617fc49c213bb807ba575533a9d0","observation_id":"ce4190c3-1d4a-4ba5-bd09-c5f14d0c0f79","resolution":{"observed_at":"2026-08-03T18:41:33.935923Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2310.02238","last_updated":"2023-10-04T05:20:19Z","snapshot_observed_at":"2026-08-06T15:19:35.944642Z","submitted_at":"2023-10-03T17:48:14Z","title":"Who's Harry Potter? Approximate Unlearning in LLMs","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2310.02238","snapshot_observed_at":"2026-08-03T18:41:34.006399Z","title":"Who’s harry potter? approximate unlearning in llms.URL https://arxiv","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.006399Z"},"links":{"cited_paper":"/paper/2310.02238","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:7d5b7d3965ca364f8f12182cb1177edd04cb92dc59760e1e227a3b0478714234","observation_id":"bb69cc0c-8a5b-4536-9faa-636ec616d737","resolution":{"observed_at":"2026-08-03T18:41:34.006399Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.059988Z","title":"Improving alignment and robustness with circuit breakers.Advances in Neural Information Processing Systems, 37:83345–83373, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.059988Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:32ef979ec972eb74d73f837d0950f388a7ff0b42e1cb12cfd9786d58068a41bd","observation_id":"7097da3d-9cc2-4152-854d-cc253646dd09","resolution":{"observed_at":"2026-08-03T18:41:34.059988Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2407.15549","last_updated":"2025-07-29T09:37:22Z","snapshot_observed_at":"2026-08-06T22:31:48.025702Z","submitted_at":"2024-07-22T11:19:14Z","title":"Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2407.15549","snapshot_observed_at":"2026-08-03T18:41:34.112470Z","title":"Latent adver- sarial training improves robustness to persistent harmful behaviors in llms.arXiv preprint arXiv:2407.15549, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.112470Z"},"links":{"cited_paper":"/paper/2407.15549","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:8b6cddcc6bfc1808c0fe76f6420ff7d737965e3161044069c48668f3f9839a17","observation_id":"f8554132-1fc3-4601-bbaf-9f55da832e40","resolution":{"observed_at":"2026-08-03T18:41:34.112470Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2408.00761","last_updated":"2025-02-10T18:26:14Z","snapshot_observed_at":"2026-08-06T05:53:06.918202Z","submitted_at":"2024-08-01T17:59:12Z","title":"Tamper-Resistant Safeguards for Open-Weight LLMs","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2408.00761","snapshot_observed_at":"2026-08-03T18:41:34.212799Z","title":"Tamper-resistant safeguards for open-weight llms.URL https://arxiv","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.212799Z"},"links":{"cited_paper":"/paper/2408.00761","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:76271a755127af85d422735ec4bbb59f975a0a38755b2db0ab21eb680b403f6d","observation_id":"5409273a-ae7e-4ee8-beaf-5d42683a99a9","resolution":{"observed_at":"2026-08-03T18:41:34.212799Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.319879Z","title":"Representation noising effectively prevents harmful fine-tuning on llms.CoRR, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.319879Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:9d2556dca667a56cb65693a323a34d704c8029cf897f7bfd67f3b83f1c3cc5b0","observation_id":"324c2444-c86c-4b36-9d49-f4d036e89f09","resolution":{"observed_at":"2026-08-03T18:41:34.319879Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2410.12949","last_updated":"2024-12-05T02:55:35Z","snapshot_observed_at":"2026-08-08T00:43:01.914749Z","submitted_at":"2024-10-16T18:35:02Z","title":"Mechanistic Unlearning: Robust Knowledge Unlearning and Editing via Mechanistic Localization","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2410.12949","snapshot_observed_at":"2026-08-03T18:41:34.380063Z","title":"Mechanistic unlearning: Robust knowledge unlearning and editing via mechanistic localization","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.380063Z"},"links":{"cited_paper":"/paper/2410.12949","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:d822d7ae7cb4a4223c8ace0e5c5652841d5fc12456fba8317353201701c0db3d","observation_id":"3232c1fd-3222-41a1-b376-671ec4eb6824","resolution":{"observed_at":"2026-08-03T18:41:34.380063Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.450577Z","title":"Redirection for erasing memory (rem): Towards a universal unlearning method for corrupted data.arXiv preprint arXiv:2505.17730, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.450577Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:fd053e58009407d48218dc0a0bf0923862c3fb10aefe9201062e03a9d23baa5f","observation_id":"fe197041-53af-4820-98ae-29e7f2f57528","resolution":{"observed_at":"2026-08-03T18:41:34.450577Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.526575Z","title":"Saes can improve unlearning: Dynamic sparse autoencoder guardrails for precision unlearning in llms","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.526575Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:115e7018262ca0175c70624fcd210ff2b62a1691b4475908f210e88a03297b5f","observation_id":"b21c3fa2-4f87-4799-92c4-e41572e64493","resolution":{"observed_at":"2026-08-03T18:41:34.526575Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2505.24428","last_updated":"2025-05-30T10:07:52Z","snapshot_observed_at":"2026-08-08T06:59:43.623757Z","submitted_at":"2025-05-30T10:07:52Z","title":"Model Unlearning via Sparse Autoencoder Subspace Guided Projections","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2505.24428","snapshot_observed_at":"2026-08-03T18:41:34.610283Z","title":"Model unlearning via sparse autoencoder subspace guided projections.arXiv preprint arXiv:2505.24428, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.610283Z"},"links":{"cited_paper":"/paper/2505.24428","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:4248af97cc6c9be9d49fbeb04f2abdba248007f7b40400bc52c9fedc54a8a1c3","observation_id":"107ab70d-df49-491e-bb37-c934545094e5","resolution":{"observed_at":"2026-08-03T18:41:34.610283Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.715225Z","title":"C-pack: Packaged resources to advance general chinese embedding, 2023","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.715225Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:970afbe743f8c5886d9c96398ee250d07c80d20178846cad698219436bca9e24","observation_id":"6a015d92-9ae4-4d2c-bfc5-f7ea840608ce","resolution":{"observed_at":"2026-08-03T18:41:34.715225Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.780216Z","title":"The faiss library","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.780216Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:6015de3b250681a5fd04fa3a36b923fd010afaee783ce7469397e7497781ddde","observation_id":"d1d3079b-479a-4d3d-b1c7-7eb7d71fb16e","resolution":{"observed_at":"2026-08-03T18:41:34.780216Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.855405Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.855405Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:ccf98d479993bd60537a85fbaa7851585c6a2bde4c62f0bb3992b46c7ef7d882","observation_id":"275fbbdb-f6c4-41af-9fc1-a40aa2232cde","resolution":{"observed_at":"2026-08-03T18:41:34.855405Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:34.934648Z","title":"Continual learning and private unlearning","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:34.934648Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:4260b668d3be83f1ab7acbab04c39d23c4147b7bbeef40c62c5478bee2af12e7","observation_id":"636bf27a-ddaf-4096-8c9c-e9c8cae7bbb9","resolution":{"observed_at":"2026-08-03T18:41:34.934648Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.03218","last_updated":"2024-05-15T19:16:09Z","snapshot_observed_at":"2026-08-09T09:20:14.433688Z","submitted_at":"2024-03-05T18:59:35Z","title":"The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning","version":7},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.03218","snapshot_observed_at":"2026-08-03T18:41:35.007554Z","title":"The wmdp benchmark: Measuring and reducing malicious use with unlearning.arXiv preprint arXiv:2403.03218, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.007554Z"},"links":{"cited_paper":"/paper/2403.03218","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:c81bc4c0654dd29f4aa979706005b0153154b85362909221ba471d67dd25c67b","observation_id":"f13f6837-af46-40be-b7a2-77e31a36686a","resolution":{"observed_at":"2026-08-03T18:41:35.007554Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2410.02760","last_updated":"2025-07-21T21:03:45Z","snapshot_observed_at":"2026-08-10T00:16:23.865631Z","submitted_at":"2024-10-03T17:59:30Z","title":"Erasing Conceptual Knowledge from Language Models","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2410.02760","snapshot_observed_at":"2026-08-03T18:41:35.115153Z","title":"Erasing conceptual knowl- edge from language models.arXiv preprint arXiv:2410.02760, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.115153Z"},"links":{"cited_paper":"/paper/2410.02760","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:a8f21df5deba37b93426b1f9808cb5e79cec27aa5274caf72478782939e8ec8e","observation_id":"f1510b40-0e7e-4389-a742-16c8752b8df6","resolution":{"observed_at":"2026-08-03T18:41:35.115153Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:35.209629Z","title":"Unlearning in large language models via activation projections","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.209629Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:112bd6a3a4a5ea233efdea16d40324af4ee317ff5ceacc6545d2a52d9c5f0423","observation_id":"1b351cc9-3623-4c8f-98bf-8cfa64702c4d","resolution":{"observed_at":"2026-08-03T18:41:35.209629Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:35.272194Z","title":"Model tampering attacks enable more rigorous evaluations of llm capabilities, 2025","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":32,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.272194Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:19a86ab5b691a0503548903842d8bda94d6bb3588a2f53de46bf32bd420e6972","observation_id":"b4eddb53-6309-4684-adbd-07bcef20f5bc","resolution":{"observed_at":"2026-08-03T18:41:35.272194Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2404.03233","last_updated":"2024-04-04T06:37:46Z","snapshot_observed_at":"2026-08-08T00:45:26.795585Z","submitted_at":"2024-04-04T06:37:46Z","title":"Learn What You Want to Unlearn: Unlearning Inversion Attacks against Machine Unlearning","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2404.03233","snapshot_observed_at":"2026-08-03T18:41:35.335649Z","title":"Representation misdirection for unlearning.arXiv preprint arXiv:2404.03233, 2024","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.335649Z"},"links":{"cited_paper":"/paper/2404.03233","citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:9fba6a25d4fe62be44a924e6b19159f63340cdfa903653679e1ae91150100bda","observation_id":"73906e04-d433-4e11-ba98-c0573d4ce187","resolution":{"observed_at":"2026-08-03T18:41:35.335649Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T18:41:35.433927Z","title":"unknowledgeable","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories","version":3},"reference_index":34,"source":"pdf_text","source_observed_at":"2026-08-03T18:41:35.433927Z"},"links":{"citing_paper":"/paper/2512.04144"},"observation_digest":"sha256:8d2494155b3b082909a338822f2e741453b02d19067e49c3e8d036aef5156fc8","observation_id":"cab5c894-e8a8-41e6-b7fc-4abe7c157e06","resolution":{"observed_at":"2026-08-03T18:41:35.433927Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2512.04144","last_updated":"2026-07-14T05:13:06Z","latest_version":3,"primary_category":"cs.AI","snapshot_observed_at":"2026-08-08T13:08:29.843209Z","submitted_at":"2025-12-03T18:57:59Z","title":"RippleBench: Capturing Ripple Effects Using Existing Knowledge Repositories"},"reference_resolution":{"displayed":32,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":32,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":32},"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":"2026-08-10T06:31:04.303077+00:00","source":"crossref"},{"observed_at":"2026-08-10T06:30:57.382061+00:00","source":"retraction_watch"}],"thesis":"As of 10 August 2026, this Paper Citation Record lists 32 of 32 outbound references and 1 inbound Pith citation observation for arXiv:2512.04144."}