{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GKEVWVMI5RBL2J2OCYILJGE3KU","short_pith_number":"pith:GKEVWVMI","schema_version":"1.0","canonical_sha256":"32895b5588ec42bd274e1610b4989b55166fb564f3b8c0360f2500020b4e2aee","source":{"kind":"arxiv","id":"2605.23157","version":1},"attestation_state":"computed","paper":{"title":"Same Model, Different Weakness: How Language and Modality Reshape the Jailbreak Attack Surface in Frontier MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Casey Ford, Emily Dix, Madison Van Doren, Sicheng Jin","submitted_at":"2026-05-22T02:12:45Z","abstract_excerpt":"The attack surface of a multimodal large language model (MLLM) is language-dependent in ways that reveal the mechanistic structure of alignment failures. We present the first systematic cross-lingual, multimodal red-teaming study comparing jailbreak vulnerability in US English (en-US) and Mexican Spanish (es-MX) across four frontier MLLMs: Claude Sonnet 4.5, GPT-5, Pixtral Large, and Qwen Omni. Using a fixed adversarial benchmark of 363 diverse prompt scenarios administered in text-only and multimodal conditions, we collected 52,272 harm ratings and binary attack success judgements from matche"},"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":"2605.23157","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-22T02:12:45Z","cross_cats_sorted":[],"title_canon_sha256":"7ee4382206a60140920b62b9cabd61cf073001f155ab3b5a565c151f5e2ee01b","abstract_canon_sha256":"a7967009842a9956790f313bacc12098b7ba4bbdd526066e0d3555787ef6af4d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:40.847110Z","signature_b64":"/NGSqoAiLo/wF0wYKWLhIeFsnW2LM87CMEZ4WlxQ6VsxrL6y0MtkPXN+oBGvRRq7cUveqgSqSEW+vOJ44q3CBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"32895b5588ec42bd274e1610b4989b55166fb564f3b8c0360f2500020b4e2aee","last_reissued_at":"2026-05-25T02:01:40.846547Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:40.846547Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Same Model, Different Weakness: How Language and Modality Reshape the Jailbreak Attack Surface in Frontier MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Casey Ford, Emily Dix, Madison Van Doren, Sicheng Jin","submitted_at":"2026-05-22T02:12:45Z","abstract_excerpt":"The attack surface of a multimodal large language model (MLLM) is language-dependent in ways that reveal the mechanistic structure of alignment failures. We present the first systematic cross-lingual, multimodal red-teaming study comparing jailbreak vulnerability in US English (en-US) and Mexican Spanish (es-MX) across four frontier MLLMs: Claude Sonnet 4.5, GPT-5, Pixtral Large, and Qwen Omni. Using a fixed adversarial benchmark of 363 diverse prompt scenarios administered in text-only and multimodal conditions, we collected 52,272 harm ratings and binary attack success judgements from matche"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23157","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/2605.23157/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":"2605.23157","created_at":"2026-05-25T02:01:40.846604+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23157v1","created_at":"2026-05-25T02:01:40.846604+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23157","created_at":"2026-05-25T02:01:40.846604+00:00"},{"alias_kind":"pith_short_12","alias_value":"GKEVWVMI5RBL","created_at":"2026-05-25T02:01:40.846604+00:00"},{"alias_kind":"pith_short_16","alias_value":"GKEVWVMI5RBL2J2O","created_at":"2026-05-25T02:01:40.846604+00:00"},{"alias_kind":"pith_short_8","alias_value":"GKEVWVMI","created_at":"2026-05-25T02:01:40.846604+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/GKEVWVMI5RBL2J2OCYILJGE3KU","json":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU.json","graph_json":"https://pith.science/api/pith-number/GKEVWVMI5RBL2J2OCYILJGE3KU/graph.json","events_json":"https://pith.science/api/pith-number/GKEVWVMI5RBL2J2OCYILJGE3KU/events.json","paper":"https://pith.science/paper/GKEVWVMI"},"agent_actions":{"view_html":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU","download_json":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU.json","view_paper":"https://pith.science/paper/GKEVWVMI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23157&json=true","fetch_graph":"https://pith.science/api/pith-number/GKEVWVMI5RBL2J2OCYILJGE3KU/graph.json","fetch_events":"https://pith.science/api/pith-number/GKEVWVMI5RBL2J2OCYILJGE3KU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU/action/storage_attestation","attest_author":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU/action/author_attestation","sign_citation":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU/action/citation_signature","submit_replication":"https://pith.science/pith/GKEVWVMI5RBL2J2OCYILJGE3KU/action/replication_record"}},"created_at":"2026-05-25T02:01:40.846604+00:00","updated_at":"2026-05-25T02:01:40.846604+00:00"}