{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:N26I4AOVUPCB2ZTXKSTKNNTBIO","short_pith_number":"pith:N26I4AOV","schema_version":"1.0","canonical_sha256":"6ebc8e01d5a3c41d667754a6a6b66143bb22577ba72dd73266565f3ffbcf976e","source":{"kind":"arxiv","id":"2510.17759","version":2},"attestation_state":"computed","paper":{"title":"VERA-V: Variational Inference Framework for Jailbreaking Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.LG","stat.ML"],"primary_cat":"cs.CR","authors_text":"Anamika Lochab, Qilin Liao, Ruqi Zhang","submitted_at":"2025-10-20T17:12:10Z","abstract_excerpt":"Vision-Language Models (VLMs) extend large language models with visual reasoning, but their multimodal design also introduces new, underexplored vulnerabilities. Existing multimodal red-teaming methods largely rely on brittle templates, focus on single-attack settings, and expose only a narrow subset of vulnerabilities. To address these limitations, we introduce VERA-V, a variational inference framework that recasts multimodal jailbreak discovery as learning a joint posterior distribution over paired text-image prompts. This probabilistic view enables the generation of stealthy, coupled advers"},"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":"2510.17759","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-10-20T17:12:10Z","cross_cats_sorted":["cs.CL","cs.CV","cs.LG","stat.ML"],"title_canon_sha256":"3b528bcc210840894ba552e1d244c418b06290fb16a92fefe7c49bc8f414fa08","abstract_canon_sha256":"1ba1355a64be9efad3ef9cc5450512efcfe5c5a5ac174f2c566d4f55d4ee35f6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:05:39.152750Z","signature_b64":"8NHPL7BEJ/+9JpL03vTPqpYLkhz1w6J1AhGdKpy2uX0bpAvYo4/UkOfOW8g3kz/m/v4IsxYZOsV0RHw8gB0LBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ebc8e01d5a3c41d667754a6a6b66143bb22577ba72dd73266565f3ffbcf976e","last_reissued_at":"2026-05-27T01:05:39.152248Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:05:39.152248Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VERA-V: Variational Inference Framework for Jailbreaking Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.LG","stat.ML"],"primary_cat":"cs.CR","authors_text":"Anamika Lochab, Qilin Liao, Ruqi Zhang","submitted_at":"2025-10-20T17:12:10Z","abstract_excerpt":"Vision-Language Models (VLMs) extend large language models with visual reasoning, but their multimodal design also introduces new, underexplored vulnerabilities. Existing multimodal red-teaming methods largely rely on brittle templates, focus on single-attack settings, and expose only a narrow subset of vulnerabilities. To address these limitations, we introduce VERA-V, a variational inference framework that recasts multimodal jailbreak discovery as learning a joint posterior distribution over paired text-image prompts. This probabilistic view enables the generation of stealthy, coupled advers"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.17759","kind":"arxiv","version":2},"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/2510.17759/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":"2510.17759","created_at":"2026-05-27T01:05:39.152300+00:00"},{"alias_kind":"arxiv_version","alias_value":"2510.17759v2","created_at":"2026-05-27T01:05:39.152300+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.17759","created_at":"2026-05-27T01:05:39.152300+00:00"},{"alias_kind":"pith_short_12","alias_value":"N26I4AOVUPCB","created_at":"2026-05-27T01:05:39.152300+00:00"},{"alias_kind":"pith_short_16","alias_value":"N26I4AOVUPCB2ZTX","created_at":"2026-05-27T01:05:39.152300+00:00"},{"alias_kind":"pith_short_8","alias_value":"N26I4AOV","created_at":"2026-05-27T01:05:39.152300+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/N26I4AOVUPCB2ZTXKSTKNNTBIO","json":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO.json","graph_json":"https://pith.science/api/pith-number/N26I4AOVUPCB2ZTXKSTKNNTBIO/graph.json","events_json":"https://pith.science/api/pith-number/N26I4AOVUPCB2ZTXKSTKNNTBIO/events.json","paper":"https://pith.science/paper/N26I4AOV"},"agent_actions":{"view_html":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO","download_json":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO.json","view_paper":"https://pith.science/paper/N26I4AOV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2510.17759&json=true","fetch_graph":"https://pith.science/api/pith-number/N26I4AOVUPCB2ZTXKSTKNNTBIO/graph.json","fetch_events":"https://pith.science/api/pith-number/N26I4AOVUPCB2ZTXKSTKNNTBIO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO/action/storage_attestation","attest_author":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO/action/author_attestation","sign_citation":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO/action/citation_signature","submit_replication":"https://pith.science/pith/N26I4AOVUPCB2ZTXKSTKNNTBIO/action/replication_record"}},"created_at":"2026-05-27T01:05:39.152300+00:00","updated_at":"2026-05-27T01:05:39.152300+00:00"}