{"as_of":"2026-08-08T11:15:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:560e0f4bea3ed66bba4ccdf0a2e64f44732a013e2f5540f231b75cc7b000eb4d","coverage":[{"denominator":20,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":20,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-03T01:56:59.274728Z","state":"measured"},{"denominator":20,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":20,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-08T06:32:00.761636+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"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":[],"links":{"evidence":"/evidence","html":"/paper/2607.18622/citation-record","integrity":"/paper/2607.18622/integrity","json":"/paper/2607.18622/citation-record.json","paper":"/paper/2607.18622"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-03T01:56:57.447334Z","title":"Query- based adversarial prompt generation","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.447334Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:dce66b7fa696dade40d927f12bf2df0c89ef76760ee348b3e9e53d6feb9d0de5","observation_id":"d190f417-d402-481d-9572-87cb55831802","resolution":{"observed_at":"2026-08-03T01:56:57.447334Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.14720","last_updated":"2024-03-20T15:26:23Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2024-03-20T15:26:23Z","title":"Defending Against Indirect Prompt Injection Attacks With Spotlighting","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.14720","snapshot_observed_at":"2026-08-03T01:56:57.562647Z","title":"Defending against indirect prompt injection attacks with spotlighting.arXiv preprint arXiv:2403.14720,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.562647Z"},"links":{"cited_paper":"/paper/2403.14720","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:b65b4b71a04446a2a86a33a44e681123c564669d259efd45482911723ecc5940","observation_id":"80d543ce-a7cf-4d4a-87ae-9f896c9f3eb3","resolution":{"observed_at":"2026-08-03T01:56:57.562647Z","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":"2026-08-03T01:56:57.643584Z","title":"Attention tracker: Detecting prompt injection attacks in llms","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.643584Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:71f29de51a7082909b1ec46f42ed86a182839717c9cbd89c4e65129f801188d0","observation_id":"b86d549a-b4ad-470b-944e-d3a7316805d3","resolution":{"observed_at":"2026-08-03T01:56:57.643584Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2309.00614","last_updated":"2023-09-04T17:47:36Z","snapshot_observed_at":"2026-07-06T16:13:23.343694Z","submitted_at":"2023-09-01T17:59:44Z","title":"Baseline Defenses for Adversarial Attacks Against Aligned Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2309.00614","snapshot_observed_at":"2026-08-03T01:56:57.779480Z","title":"Baseline defenses for adversarial attacks against aligned language models.arXiv preprint arXiv:2309.00614,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.779480Z"},"links":{"cited_paper":"/paper/2309.00614","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:3d1b0a7ade0608a5c23ff393f31a6bad4c1fa18a971f74d8e6f4fa9955a41d68","observation_id":"3501e32d-a96c-4b6a-a0a4-4e0fefef94f1","resolution":{"observed_at":"2026-08-03T01:56:57.779480Z","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":"2026-08-03T01:56:57.853340Z","title":"Pubmedqa: A dataset for biomedical research question answering","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.853340Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:f22c40adfd9104d48094a8498f60d02f4a47a470d3203095205b49d79f551546","observation_id":"19ea5d7d-369c-41fb-a598-312101913ad9","resolution":{"observed_at":"2026-08-03T01:56:57.853340Z","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":"2026-08-03T01:56:57.972654Z","title":"Instruction boundary: Quantifying biases in llm reasoning under various coverage.arXiv preprint arXiv:2509.20278,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.972654Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:a0dc7d738c5c008e137fc4dd5257a97eb44270c507540682c5d83b89ca2fea30","observation_id":"d38109ac-5807-47a6-87d7-53d253a4dca3","resolution":{"observed_at":"2026-08-03T01:56:57.972654Z","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":"2026-08-03T01:56:58.065795Z","title":"Tree of attacks: Jailbreaking black-box llms automatically","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.065795Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:ecc14fa5c9a3d64022f09a60ffb5c212f4290a08dd04050719791550ac8b2ad3","observation_id":"5c7fa244-e9a5-4b67-a8fe-9d9116eea9da","resolution":{"observed_at":"2026-08-03T01:56:58.065795Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2412.07724","last_updated":"2024-12-16T20:27:36Z","snapshot_observed_at":"2026-08-01T20:25:38.720765Z","submitted_at":"2024-12-10T18:17:02Z","title":"Granite Guardian","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2412.07724","snapshot_observed_at":"2026-08-03T01:56:58.170447Z","title":null,"venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.170447Z"},"links":{"cited_paper":"/paper/2412.07724","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:8b8ec7717ae35631cdcd0ae9ec90ccc02b19381cb9f7e4be9eafebdbd6b6ed92","observation_id":"abc6bdd9-fa6d-4439-9573-ffd07e8ee868","resolution":{"observed_at":"2026-08-03T01:56:58.170447Z","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":"2026-08-03T01:56:58.277806Z","title":null,"venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.277806Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:a633fcc5c4cf6b750656545944ccc2ff149a8e4921020e68d3b4a3fc29c191f8","observation_id":"bcf92fe2-d32a-46de-a81d-f63e803f28b0","resolution":{"observed_at":"2026-08-03T01:56:58.277806Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2503.01781","last_updated":"2025-07-21T07:33:39Z","snapshot_observed_at":"2026-08-07T17:32:51.602677Z","submitted_at":"2025-03-03T18:10:54Z","title":"Cats Confuse Reasoning LLM: Query Agnostic Adversarial Triggers for Reasoning Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2503.01781","snapshot_observed_at":"2026-08-03T01:56:58.553796Z","title":"Meghana Rajeev, Rajkumar Ramamurthy, Prapti Trivedi, Vikas Yadav, Oluwanifemi Bamg- bose, Sathwik Tejaswi Madhusudan, James Zou, and Nazneen Rajani","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.553796Z"},"links":{"cited_paper":"/paper/2503.01781","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:453058296f7c87471c9a7619dc03727b0b7b2bc05510fe6f1b114749470a8979","observation_id":"54c0e337-67a5-4452-8931-222af3d5dcea","resolution":{"observed_at":"2026-08-03T01:56:58.553796Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.19314","last_updated":"2025-04-18T19:36:00Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2024-06-27T16:47:42Z","title":"LiveBench: A Challenging, Contamination-Limited LLM Benchmark","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.19314","snapshot_observed_at":"2026-08-03T01:56:58.763709Z","title":"Livebench: A challenging, contamination-free llm benchmark.arXiv preprint arXiv:2406.19314, 4:2,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.763709Z"},"links":{"cited_paper":"/paper/2406.19314","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:0dac58b645176d49e8f1961dcb57d2f03eed26484a9bbea998eb24d3c305e3a6","observation_id":"9e597876-3897-42da-9cf1-594c97327fea","resolution":{"observed_at":"2026-08-03T01:56:58.763709Z","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":"2026-08-03T01:56:58.893921Z","title":"Black-box optimization of llm outputs by asking for directions.arXiv preprint arXiv:2510.16794,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.893921Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:14bfed58a2cd5e504ff92777260b27a5e9ccd30a609883088a68b0db94d97a2c","observation_id":"c0eb748f-b430-4ab7-8cd1-cce3180d2e47","resolution":{"observed_at":"2026-08-03T01:56:58.893921Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2307.15043","last_updated":"2023-12-20T20:48:57Z","snapshot_observed_at":"2026-07-06T15:59:23.019044Z","submitted_at":"2023-07-27T17:49:12Z","title":"Universal and Transferable Adversarial Attacks on Aligned Language Models","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2307.15043","snapshot_observed_at":"2026-08-03T01:56:58.999868Z","title":"Universal and transferable adversarial attacks on aligned language models.arXiv preprint arXiv:2307.15043,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.999868Z"},"links":{"cited_paper":"/paper/2307.15043","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:2d572fcf8998dcc4e2a91898b93b3513b7f5448ca59b1c84e16ea0fc9abdcf55","observation_id":"97eef707-615d-401b-a73e-58e2ffbae2cb","resolution":{"observed_at":"2026-08-03T01:56:58.999868Z","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":"2026-08-03T01:56:59.163517Z","title":",L(r) u o nu ! ,n u =max (1, round(λu))","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:59.163517Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:3dd86a5590daa3dc5efde1c2fe1adaf590bd93a8e52fdd15cc7bb74193efc4d2","observation_id":"e7ca6615-69b6-4fa1-a270-9e4f1e036bea","resolution":{"observed_at":"2026-08-03T01:56:59.163517Z","resolver_source":null,"status":"malformed_identifier"},"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":"2026-08-03T01:56:59.274728Z","title":"CPInj achieves the highest Target ASR while maintaining perfect format compliance","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:59.274728Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:76f84ec839bb13a0442c3f8fadd7d9af51de5ceaf6022424187d77b152abc5a1","observation_id":"770b62f9-b048-4541-8acc-276f0a246960","resolution":{"observed_at":"2026-08-03T01:56:59.274728Z","resolver_source":null,"status":"malformed_identifier"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2404.16873","last_updated":"2025-06-02T18:59:01Z","snapshot_observed_at":"2026-08-05T10:26:02.067012Z","submitted_at":"2024-04-21T22:18:13Z","title":"AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2404.16873","snapshot_observed_at":"2026-08-03T01:56:58.444116Z","title":"doi: 10.18653/v1/2022.findings-acl.165","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":2022,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:58.444116Z"},"links":{"cited_paper":"/paper/2404.16873","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:da460548def26b48431b21599546de4567c7dd1f9d735e4cc810e8a589aa378d","observation_id":"dd902880-98df-4544-b3eb-4b03dcac666c","resolution":{"observed_at":"2026-08-03T01:56:58.444116Z","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":"2026-08-03T01:56:57.302491Z","title":"Can textual gradient work in federated learning? InThe Thirteenth International Conference on Learning Representations, 2025a","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":2023,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.302491Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:f2a4aed42f79aa52c865305c7ae1becbf665dba71fdeb7b573e74cd6f67b2eca","observation_id":"266ef3f7-791e-4b9d-9e96-ed75a01509e0","resolution":{"observed_at":"2026-08-03T01:56:57.302491Z","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":"2026-08-03T01:56:57.378176Z","title":"Folio: Natural language reasoning with first-order logic","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":2024,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.378176Z"},"links":{"citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:a42dbcc0eca66de58ae7c0feaedaff3ea1ef564159a7684cf811640bbfe43fb6","observation_id":"b7a804f1-e460-4a19-8d8d-6b18ea3ec16c","resolution":{"observed_at":"2026-08-03T01:56:57.378176Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2312.06674","last_updated":"2023-12-07T19:40:50Z","snapshot_observed_at":"2026-07-06T17:00:00.321552Z","submitted_at":"2023-12-07T19:40:50Z","title":"Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2312.06674","snapshot_observed_at":"2026-08-03T01:56:57.704085Z","title":"Llama guard: Llm- based input-output safeguard for human-ai conversations.arXiv preprint arXiv:2312.06674,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":2025,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.704085Z"},"links":{"cited_paper":"/paper/2312.06674","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:2197044da7cc2f2c27b75e7ca02ce0d4b32b958a67e0e51ad4f815c121de69e1","observation_id":"d25f5654-317f-419d-969b-08a285c3e634","resolution":{"observed_at":"2026-08-03T01:56:57.704085Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2310.08419","last_updated":"2024-07-18T18:24:57Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2023-10-12T15:38:28Z","title":"Jailbreaking Black Box Large Language Models in Twenty Queries","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2310.08419","snapshot_observed_at":"2026-08-03T01:56:57.266792Z","title":"Patrick Chao, Alexander Robey, Edgar Dobriban, Hamed Hassani, George J","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization","version":2},"reference_index":2026,"source":"pdf_text","source_observed_at":"2026-08-03T01:56:57.266792Z"},"links":{"cited_paper":"/paper/2310.08419","citing_paper":"/paper/2607.18622"},"observation_digest":"sha256:f7bd7a1ed8bbdae56c68a75a9a11c43e7df97f5bd0aa185875ac4fafbbb52ea3","observation_id":"3e37472e-aaf8-46c9-a9af-a051b615573b","resolution":{"observed_at":"2026-08-03T01:56:57.266792Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2607.18622","last_updated":"2026-07-30T22:51:32Z","latest_version":2,"primary_category":"cs.CR","snapshot_observed_at":"2026-08-08T08:22:09.495547Z","submitted_at":"2026-07-21T01:44:59Z","title":"CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization"},"reference_resolution":{"displayed":20,"state_counts":{"malformed_identifier":2,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":18,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":20},"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":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"thesis":"As of 8 August 2026, this Paper Citation Record lists 20 of 20 outbound references and 0 inbound Pith citation observations for arXiv:2607.18622."}