{"as_of":"2026-08-06T10:54:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:f03705866a0537e4214635e0332a67f01254dddac875b2431359aa5057b82bb8","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":8,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":8,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-06T06:34:29.942622+00:00","state":"measured"},{"denominator":8,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":8,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-05T20:31:33.221051Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":1,"source":"arxiv_reference","source_observed_at":"2026-08-05T02:28:24.338817Z","state":"measured"}],"external_citation_measurements":[{"count":2,"observed_at":"2026-08-05T02:28:24.338817Z","source":"arxiv_reference"}],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"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":14,"source":"arxiv_source","source_observed_at":"2026-05-11T13:02:43.571234Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2502.18864"},"observation_digest":"sha256:e20f031e5ca44adeddadb8ca154c5e1508c692578a3c3648e395714f981cfec6","observation_id":"14973b43-962e-4911-aedc-d49bf99aeffc","resolution":{"observed_at":"2026-05-11T13:02:44.339484Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T20:31:33.221051Z","title":"Hijacking large language models via adversarial in-context learning","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2508.10404","last_updated":"2025-08-14T07:12:44Z","snapshot_observed_at":"2026-08-05T20:31:28.198027Z","submitted_at":"2025-08-14T07:12:44Z","title":"Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-05T20:31:33.221051Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2508.10404"},"observation_digest":"sha256:65c581972b2fd9914650f8e8205ebf439c0145b0bdbd387b29b3033c821b77b6","observation_id":"c46d0d5a-1d65-40fa-b1f0-f571846445f8","resolution":{"observed_at":"2026-08-05T20:31:33.221051Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-04T13:29:31.557755Z","title":"Qiang, X","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2510.00399","last_updated":"2026-07-07T00:49:52Z","snapshot_observed_at":"2026-08-04T13:29:26.004833Z","submitted_at":"2025-10-01T01:25:01Z","title":"How Can Mamba Learn In Context with Outliers and Generalize Provably?","version":2},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-04T13:29:31.557755Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2510.00399"},"observation_digest":"sha256:cd2bf35f9bb22f97bcd51ec2de06fd950bd0da33cd12a90561021d67f0c71f0e","observation_id":"ef79adea-ac4a-450d-b779-5d9e179f914d","resolution":{"observed_at":"2026-08-04T13:29:31.557755Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"citing_paper":{"arxiv_id":"2601.17887","last_updated":"2026-05-17T15:59:35Z","snapshot_observed_at":"2026-08-02T19:18:47.080046Z","submitted_at":"2026-01-25T15:42:01Z","title":"When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents","version":2},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-05-21T15:11:04.394636Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2601.17887"},"observation_digest":"sha256:afd6db53d0381ea652bf41c845ac63090e69139ea97a6f5f4d8354b959257d3a","observation_id":"2f084c20-4669-487d-9c81-b122f500b841","resolution":{"observed_at":"2026-05-21T15:14:13.376775Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"citing_paper":{"arxiv_id":"2604.23593","last_updated":"2026-04-26T08:03:32Z","snapshot_observed_at":"2026-08-01T01:43:18.294936Z","submitted_at":"2026-04-26T08:03:32Z","title":"When AI reviews science: Can we trust the referee?","version":1},"reference_index":108,"source":"pdf_text","source_observed_at":"2026-05-08T06:19:54.727724Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2604.23593"},"observation_digest":"sha256:dfef5cfbbd2fd9c56b41d97d02d7a54e706981ecb99e7b1d6f75f621ff20ec27","observation_id":"7f40aed0-e9e2-42d2-a0a5-88bce6623002","resolution":{"observed_at":"2026-05-08T23:19:29.999422Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"citing_paper":{"arxiv_id":"2605.05116","last_updated":"2026-05-06T16:47:07Z","snapshot_observed_at":"2026-07-06T23:17:48.161262Z","submitted_at":"2026-05-06T16:47:07Z","title":"On the Hardness of Junking LLMs","version":1},"reference_index":64,"source":"pdf_text","source_observed_at":"2026-05-08T17:38:32.028947Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2605.05116"},"observation_digest":"sha256:6e604b33313c05d988cbce68da347fa14d553a1fde1bb40a711142157358c391","observation_id":"047b4222-96f8-4a2d-9720-97995c22b443","resolution":{"observed_at":"2026-05-11T17:21:10.127581Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"citing_paper":{"arxiv_id":"2605.12813","last_updated":"2026-05-31T17:51:51Z","snapshot_observed_at":"2026-07-06T23:24:27.821980Z","submitted_at":"2026-05-12T23:13:50Z","title":"REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations","version":1},"reference_index":93,"source":"arxiv_source","source_observed_at":"2026-05-14T20:13:10.814899Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2605.12813"},"observation_digest":"sha256:4f83878270769b84a1f86cb16aacd1a0961ab2f22ec0e479cf5959b68d2dbcc3","observation_id":"48ed8345-50d9-4ab7-8d1a-e741a4e5a888","resolution":{"observed_at":"2026-05-14T20:19:26.707805Z","resolver_source":"arxiv_id","status":"verified_exact"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning","version":3},"cited_work":{"arxiv_id":"2311.09948","doi":"10.48550/arxiv.2311.09948","metadata_source":"arxiv_reference","pith_arxiv_id":"2311.09948","snapshot_observed_at":"2026-08-05T02:28:24.338817Z","title":"blue\" data (harmful questions with safe/refusal responses) and","venue":"arXiv (Cornell University)","work_id":"c7c63a4a-7756-4e78-b7c5-d97db73e7bf0","year":2018},"citing_paper":{"arxiv_id":"2605.26350","last_updated":"2026-05-25T21:52:49Z","snapshot_observed_at":"2026-07-06T23:36:11.031436Z","submitted_at":"2026-05-25T21:52:49Z","title":"When Correct Demonstrations Hurt: Rethinking the Role of Exemplars in In-Context Learning","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-06-29T22:23:59.570806Z"},"links":{"cited_paper":"/paper/2311.09948","citing_paper":"/paper/2605.26350"},"observation_digest":"sha256:40cc6e6a62cbc813093dd39cec5a77c3aa3f6489af1f2b4125c2455ab534ec0f","observation_id":"1274c455-a1b4-45f2-bb65-7df0fedbe888","resolution":{"observed_at":"2026-06-29T22:34:02.714544Z","resolver_source":"arxiv_id","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"state":"measured"}}],"links":{"evidence":"/evidence","html":"/paper/2311.09948/citation-record","integrity":"/paper/2311.09948/integrity","json":"/paper/2311.09948/citation-record.json","paper":"/paper/2311.09948"},"outbound":[],"paper":{"arxiv_id":"2311.09948","last_updated":"2025-05-29T14:49:44Z","latest_version":3,"primary_category":"cs.LG","snapshot_observed_at":"2026-07-06T16:48:37.243305Z","submitted_at":"2023-11-16T15:01:48Z","title":"Hijacking Large Language Models via Adversarial In-Context Learning"},"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":"2026-08-06T06:34:29.942622+00:00","source":"crossref"},{"observed_at":"2026-08-06T06:34:23.284952+00:00","source":"retraction_watch"}],"thesis":"As of 6 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 8 inbound Pith citation observations for arXiv:2311.09948."}