{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5BWU3JJDLVR77WZFCIBD6YFWDK","short_pith_number":"pith:5BWU3JJD","schema_version":"1.0","canonical_sha256":"e86d4da5235d63ffdb2512023f60b61aa386c6547e1593273b349348ad141aab","source":{"kind":"arxiv","id":"2312.11523","version":1},"attestation_state":"computed","paper":{"title":"ToViLaG: Your Visual-Language Generative Model is Also An Evildoer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Han Jiang, Shanlin Zhou, Xiaoyuan Yi, Xing Xie, Xinpeng Wang, Zhihua Wei","submitted_at":"2023-12-13T08:25:07Z","abstract_excerpt":"Warning: this paper includes model outputs showing offensive content. Recent large-scale Visual-Language Generative Models (VLGMs) have achieved unprecedented improvement in multimodal image/text generation. However, these models might also generate toxic content, e.g., offensive text and pornography images, raising significant ethical risks. Despite exhaustive studies on toxic degeneration of language models, this problem remains largely unexplored within the context of visual-language generation. This work delves into the propensity for toxicity generation and susceptibility to toxic data ac"},"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":"2312.11523","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-12-13T08:25:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ab52d3fa3a6fcff2caa92ef23ba8224b537f06d62f1626fa605ef236bb1866a7","abstract_canon_sha256":"9c8a59f98d19dd141360a4d13d5036a42c9d48d2b5a07f67d5398329ff0b80a4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:25:50.787773Z","signature_b64":"PCnRBOQz1SbCMJjphho1q2DxMm229T9vomSGQPJZkLHuZqnaRywYYYnEuRduePGxYkeNP1VKg+dW5yP2f9C8CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e86d4da5235d63ffdb2512023f60b61aa386c6547e1593273b349348ad141aab","last_reissued_at":"2026-07-05T07:25:50.787338Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:25:50.787338Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ToViLaG: Your Visual-Language Generative Model is Also An Evildoer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Han Jiang, Shanlin Zhou, Xiaoyuan Yi, Xing Xie, Xinpeng Wang, Zhihua Wei","submitted_at":"2023-12-13T08:25:07Z","abstract_excerpt":"Warning: this paper includes model outputs showing offensive content. Recent large-scale Visual-Language Generative Models (VLGMs) have achieved unprecedented improvement in multimodal image/text generation. However, these models might also generate toxic content, e.g., offensive text and pornography images, raising significant ethical risks. Despite exhaustive studies on toxic degeneration of language models, this problem remains largely unexplored within the context of visual-language generation. This work delves into the propensity for toxicity generation and susceptibility to toxic data ac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.11523","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/2312.11523/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":"2312.11523","created_at":"2026-07-05T07:25:50.787390+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.11523v1","created_at":"2026-07-05T07:25:50.787390+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.11523","created_at":"2026-07-05T07:25:50.787390+00:00"},{"alias_kind":"pith_short_12","alias_value":"5BWU3JJDLVR7","created_at":"2026-07-05T07:25:50.787390+00:00"},{"alias_kind":"pith_short_16","alias_value":"5BWU3JJDLVR77WZF","created_at":"2026-07-05T07:25:50.787390+00:00"},{"alias_kind":"pith_short_8","alias_value":"5BWU3JJD","created_at":"2026-07-05T07:25:50.787390+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.15838","citing_title":"Align Anything: Training All-Modality Models to Follow Instructions with Language Feedback","ref_index":105,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK","json":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK.json","graph_json":"https://pith.science/api/pith-number/5BWU3JJDLVR77WZFCIBD6YFWDK/graph.json","events_json":"https://pith.science/api/pith-number/5BWU3JJDLVR77WZFCIBD6YFWDK/events.json","paper":"https://pith.science/paper/5BWU3JJD"},"agent_actions":{"view_html":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK","download_json":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK.json","view_paper":"https://pith.science/paper/5BWU3JJD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.11523&json=true","fetch_graph":"https://pith.science/api/pith-number/5BWU3JJDLVR77WZFCIBD6YFWDK/graph.json","fetch_events":"https://pith.science/api/pith-number/5BWU3JJDLVR77WZFCIBD6YFWDK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK/action/storage_attestation","attest_author":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK/action/author_attestation","sign_citation":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK/action/citation_signature","submit_replication":"https://pith.science/pith/5BWU3JJDLVR77WZFCIBD6YFWDK/action/replication_record"}},"created_at":"2026-07-05T07:25:50.787390+00:00","updated_at":"2026-07-05T07:25:50.787390+00:00"}