{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NMAQXCITDFB223FHPLYBEAQYVS","short_pith_number":"pith:NMAQXCIT","schema_version":"1.0","canonical_sha256":"6b010b89131943ad6ca77af0120218ac8829ee60f12d8397545c61ca30fdc4f7","source":{"kind":"arxiv","id":"2607.05069","version":1},"attestation_state":"computed","paper":{"title":"MIRAGE: Defending Long-Form RAG Against Misinformation Pollution","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aleksandr Rubashevskii, Maxim Panov, Preslav Nakov, Ruihong Zeng, Saadeldine Eletter, Yuxia Wang","submitted_at":"2026-07-06T13:36:12Z","abstract_excerpt":"Retrieval-Augmented Generation (RAG) improves factuality by grounding LLMs in external evidence, but real-world retrieval is often polluted: semantically relevant passages may contain subtle misinformation, misleading framings, or fabrications. We introduce MIRAGE, a training-free, model-agnostic defense for long-form RAG. MIRAGE builds an NLI-based cross-document claim graph and applies a Defended-Claims Gate to either condition generation on a consistent, multi-source supported subset or to block retrieval and answer parametrically. We also release a minimal-edit pollution protocol spanning "},"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":"2607.05069","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-06T13:36:12Z","cross_cats_sorted":[],"title_canon_sha256":"66fe3c351f509e72c8b16d43670aa0d94481166ed73b8f05b81d0fe8072ae4b6","abstract_canon_sha256":"f241a122cb0d48b502c74b0eab4d89d5292b63791dd249bde3b38b35a2039ac2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T03:19:15.110336Z","signature_b64":"WQ9DmPcQ6QR/mzDyme4tWJkZdmBU/akzYbew8rMXXPxTlLjzOxSMo+SIQR8/ZaDAunjCE6Prg/a1nQ7yAKNHDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b010b89131943ad6ca77af0120218ac8829ee60f12d8397545c61ca30fdc4f7","last_reissued_at":"2026-07-07T03:19:15.109777Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T03:19:15.109777Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MIRAGE: Defending Long-Form RAG Against Misinformation Pollution","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aleksandr Rubashevskii, Maxim Panov, Preslav Nakov, Ruihong Zeng, Saadeldine Eletter, Yuxia Wang","submitted_at":"2026-07-06T13:36:12Z","abstract_excerpt":"Retrieval-Augmented Generation (RAG) improves factuality by grounding LLMs in external evidence, but real-world retrieval is often polluted: semantically relevant passages may contain subtle misinformation, misleading framings, or fabrications. We introduce MIRAGE, a training-free, model-agnostic defense for long-form RAG. MIRAGE builds an NLI-based cross-document claim graph and applies a Defended-Claims Gate to either condition generation on a consistent, multi-source supported subset or to block retrieval and answer parametrically. We also release a minimal-edit pollution protocol spanning "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.05069","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/2607.05069/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":"2607.05069","created_at":"2026-07-07T03:19:15.109854+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.05069v1","created_at":"2026-07-07T03:19:15.109854+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.05069","created_at":"2026-07-07T03:19:15.109854+00:00"},{"alias_kind":"pith_short_12","alias_value":"NMAQXCITDFB2","created_at":"2026-07-07T03:19:15.109854+00:00"},{"alias_kind":"pith_short_16","alias_value":"NMAQXCITDFB223FH","created_at":"2026-07-07T03:19:15.109854+00:00"},{"alias_kind":"pith_short_8","alias_value":"NMAQXCIT","created_at":"2026-07-07T03:19:15.109854+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS","json":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS.json","graph_json":"https://pith.science/api/pith-number/NMAQXCITDFB223FHPLYBEAQYVS/graph.json","events_json":"https://pith.science/api/pith-number/NMAQXCITDFB223FHPLYBEAQYVS/events.json","paper":"https://pith.science/paper/NMAQXCIT"},"agent_actions":{"view_html":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS","download_json":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS.json","view_paper":"https://pith.science/paper/NMAQXCIT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.05069&json=true","fetch_graph":"https://pith.science/api/pith-number/NMAQXCITDFB223FHPLYBEAQYVS/graph.json","fetch_events":"https://pith.science/api/pith-number/NMAQXCITDFB223FHPLYBEAQYVS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS/action/storage_attestation","attest_author":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS/action/author_attestation","sign_citation":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS/action/citation_signature","submit_replication":"https://pith.science/pith/NMAQXCITDFB223FHPLYBEAQYVS/action/replication_record"}},"created_at":"2026-07-07T03:19:15.109854+00:00","updated_at":"2026-07-07T03:19:15.109854+00:00"}