{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LGDZCWNCH32BEXTWX3KISXOWNZ","short_pith_number":"pith:LGDZCWNC","schema_version":"1.0","canonical_sha256":"59879159a23ef4125e76bed4895dd66e78d0e168567a6ad66e7c2a8755c8a9d6","source":{"kind":"arxiv","id":"2607.27917","version":1},"attestation_state":"computed","paper":{"title":"One Anchor for All: Unified Multilingual and Multimodal Safety Alignment for LVLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chuancheng Shi, Enyi Shi, Fei Shen, Jinhui Tang, Linxia Zhu, Shuyi Miao, Tat-Seng Chua","submitted_at":"2026-07-30T09:30:03Z","abstract_excerpt":"As large vision-language models (LVLMs) are deployed globally, the combination of multilingual instructions and visual information makes malicious attacks more covert and sophisticated than ever before. However, existing methods isolate language and modality defenses, which, coupled with the scarcity of safety data and high fine-tuning costs, makes it difficult for models to defend against compound attacks. To address this severe challenge, we propose a neuron-level cross-dimensional safety alignment framework driven by modality- and language-shared safety neurons (MLS-Neurons). First, we iden"},"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":"2607.27917","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-30T09:30:03Z","cross_cats_sorted":[],"title_canon_sha256":"1b683c8181aad69eff1bc5dd4d5dd2b1963cc42328acc371cc7933f819c055f8","abstract_canon_sha256":"2e8c7f9986c1240f51fcab409ee782abd7eb854fdf550c288f0f75cce0888d1a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"59879159a23ef4125e76bed4895dd66e78d0e168567a6ad66e7c2a8755c8a9d6","last_reissued_at":"2026-07-31T01:34:47.753347Z","signature_status":"unsigned_v0","first_computed_at":"2026-07-31T01:34:47.753347Z"},"graph_snapshot":{"paper":{"title":"One Anchor for All: Unified Multilingual and Multimodal Safety Alignment for LVLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chuancheng Shi, Enyi Shi, Fei Shen, Jinhui Tang, Linxia Zhu, Shuyi Miao, Tat-Seng Chua","submitted_at":"2026-07-30T09:30:03Z","abstract_excerpt":"As large vision-language models (LVLMs) are deployed globally, the combination of multilingual instructions and visual information makes malicious attacks more covert and sophisticated than ever before. However, existing methods isolate language and modality defenses, which, coupled with the scarcity of safety data and high fine-tuning costs, makes it difficult for models to defend against compound attacks. To address this severe challenge, we propose a neuron-level cross-dimensional safety alignment framework driven by modality- and language-shared safety neurons (MLS-Neurons). First, we iden"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.27917","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/2607.27917/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":"2607.27917","created_at":"2026-07-31T01:34:47.756604+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.27917v1","created_at":"2026-07-31T01:34:47.756604+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.27917","created_at":"2026-07-31T01:34:47.756604+00:00"},{"alias_kind":"pith_short_12","alias_value":"LGDZCWNCH32B","created_at":"2026-07-31T01:34:47.756604+00:00"},{"alias_kind":"pith_short_16","alias_value":"LGDZCWNCH32BEXTW","created_at":"2026-07-31T01:34:47.756604+00:00"},{"alias_kind":"pith_short_8","alias_value":"LGDZCWNC","created_at":"2026-07-31T01:34:47.756604+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/LGDZCWNCH32BEXTWX3KISXOWNZ","json":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ.json","graph_json":"https://pith.science/api/pith-number/LGDZCWNCH32BEXTWX3KISXOWNZ/graph.json","events_json":"https://pith.science/api/pith-number/LGDZCWNCH32BEXTWX3KISXOWNZ/events.json","paper":"https://pith.science/paper/LGDZCWNC"},"agent_actions":{"view_html":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ","download_json":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ.json","view_paper":"https://pith.science/paper/LGDZCWNC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.27917&json=true","fetch_graph":"https://pith.science/api/pith-number/LGDZCWNCH32BEXTWX3KISXOWNZ/graph.json","fetch_events":"https://pith.science/api/pith-number/LGDZCWNCH32BEXTWX3KISXOWNZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ/action/storage_attestation","attest_author":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ/action/author_attestation","sign_citation":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ/action/citation_signature","submit_replication":"https://pith.science/pith/LGDZCWNCH32BEXTWX3KISXOWNZ/action/replication_record"}},"created_at":"2026-07-31T01:34:47.756604+00:00","updated_at":"2026-07-31T01:34:47.756604+00:00"}