{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MKYIQFJSET5TH4VNNAKETDIW36","short_pith_number":"pith:MKYIQFJS","schema_version":"1.0","canonical_sha256":"62b088153224fb33f2ad6814498d16df9c5ecf1212f1ef29e4891bb40563d1b2","source":{"kind":"arxiv","id":"2505.13348","version":1},"attestation_state":"computed","paper":{"title":"Investigating the Vulnerability of LLM-as-a-Judge Architectures to Prompt-Injection Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bislan Ashinov, Dmitry Namiot, Narek Maloyan","submitted_at":"2025-05-19T16:51:12Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly employed as evaluators (LLM-as-a-Judge) for assessing the quality of machine-generated text. This paradigm offers scalability and cost-effectiveness compared to human annotation. However, the reliability and security of such systems, particularly their robustness against adversarial manipulations, remain critical concerns. This paper investigates the vulnerability of LLM-as-a-Judge architectures to prompt-injection attacks, where malicious inputs are designed to compromise the judge's decision-making process. We formalize two primary attack strateg"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2505.13348","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-19T16:51:12Z","cross_cats_sorted":[],"title_canon_sha256":"7684ec2a2d00e52105f0aa11fe04c0fa93b1acbc0a31cc978444524986b55131","abstract_canon_sha256":"995439aa78849394a189aca68017acc78b9a407b07ef14eaaba126a8da095579"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:05:28.936309Z","signature_b64":"jIlZfOJ4XxRtoi/6/lGouowrBn7I0ONIQTr+Zo2ZQgCk1kA5KTLEyD3MfHKo2bz8RI2lswjm2Bzm1OInpFAsBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"62b088153224fb33f2ad6814498d16df9c5ecf1212f1ef29e4891bb40563d1b2","last_reissued_at":"2026-07-05T11:05:28.935836Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:05:28.935836Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Investigating the Vulnerability of LLM-as-a-Judge Architectures to Prompt-Injection Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bislan Ashinov, Dmitry Namiot, Narek Maloyan","submitted_at":"2025-05-19T16:51:12Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly employed as evaluators (LLM-as-a-Judge) for assessing the quality of machine-generated text. This paradigm offers scalability and cost-effectiveness compared to human annotation. However, the reliability and security of such systems, particularly their robustness against adversarial manipulations, remain critical concerns. This paper investigates the vulnerability of LLM-as-a-Judge architectures to prompt-injection attacks, where malicious inputs are designed to compromise the judge's decision-making process. We formalize two primary attack strateg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.13348","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2505.13348/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2505.13348","created_at":"2026-07-05T11:05:28.935895+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.13348v1","created_at":"2026-07-05T11:05:28.935895+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.13348","created_at":"2026-07-05T11:05:28.935895+00:00"},{"alias_kind":"pith_short_12","alias_value":"MKYIQFJSET5T","created_at":"2026-07-05T11:05:28.935895+00:00"},{"alias_kind":"pith_short_16","alias_value":"MKYIQFJSET5TH4VN","created_at":"2026-07-05T11:05:28.935895+00:00"},{"alias_kind":"pith_short_8","alias_value":"MKYIQFJS","created_at":"2026-07-05T11:05:28.935895+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.20312","citing_title":"Pramana: A Protocol-Layer Treatment of Claim Verification in Autonomous Agent Networks","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2512.20405","citing_title":"ChatGPT: Excellent Paper! Accept It. Editor: Imposter Found! Review Rejected","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36","json":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36.json","graph_json":"https://pith.science/api/pith-number/MKYIQFJSET5TH4VNNAKETDIW36/graph.json","events_json":"https://pith.science/api/pith-number/MKYIQFJSET5TH4VNNAKETDIW36/events.json","paper":"https://pith.science/paper/MKYIQFJS"},"agent_actions":{"view_html":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36","download_json":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36.json","view_paper":"https://pith.science/paper/MKYIQFJS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.13348&json=true","fetch_graph":"https://pith.science/api/pith-number/MKYIQFJSET5TH4VNNAKETDIW36/graph.json","fetch_events":"https://pith.science/api/pith-number/MKYIQFJSET5TH4VNNAKETDIW36/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36/action/storage_attestation","attest_author":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36/action/author_attestation","sign_citation":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36/action/citation_signature","submit_replication":"https://pith.science/pith/MKYIQFJSET5TH4VNNAKETDIW36/action/replication_record"}},"created_at":"2026-07-05T11:05:28.935895+00:00","updated_at":"2026-07-05T11:05:28.935895+00:00"}