{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VWGNNMXIHD67KFTST2ILLRSHZO","short_pith_number":"pith:VWGNNMXI","schema_version":"1.0","canonical_sha256":"ad8cd6b2e838fdf516729e90b5c647cb8f533453db867036856f5b5b4578b734","source":{"kind":"arxiv","id":"2503.16529","version":2},"attestation_state":"computed","paper":{"title":"Safety Evaluation and Enhancement of DeepSeek Models in Chinese Contexts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.CL","authors_text":"Beibei Huang, Jiaojiao Zhao, Junting Guo, Kai Wang, Limin Han, Meijuan An, Ning Wang, Rongjia Du, Shiguo Lian, Shu Yang, Wenjing Zhang, Xuejiao Lei, Zhaoxiang Liu, Zhenhong Long","submitted_at":"2025-03-18T08:38:10Z","abstract_excerpt":"DeepSeek-R1, renowned for its exceptional reasoning capabilities and open-source strategy, is significantly influencing the global artificial intelligence landscape. However, it exhibits notable safety shortcomings. Recent research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 achieves a 100\\% attack success rate when processing harmful prompts. Furthermore, multiple security firms and research institutions have identified critical security vulnerabilities within the model. Although China Unicom has unco"},"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":"2503.16529","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-18T08:38:10Z","cross_cats_sorted":["cs.AI","cs.CY"],"title_canon_sha256":"5c871835b2d9652f920498f47e7236dd3b15af7448622134874d5fdf247b54eb","abstract_canon_sha256":"5ad2c52ce772328332b20eb6deb8c547dbc826407d3faa413efff4a945b3e313"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:03:58.456977Z","signature_b64":"KI45yQrkjuojErdkYnrX0kwqMb3wv45OgjY3xi1Ed7tXZsBp9tDVFvob42R9yQ1kTTLqA5l+ajmYHdQeu4NtCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ad8cd6b2e838fdf516729e90b5c647cb8f533453db867036856f5b5b4578b734","last_reissued_at":"2026-07-05T11:03:58.456495Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:03:58.456495Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Safety Evaluation and Enhancement of DeepSeek Models in Chinese Contexts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.CL","authors_text":"Beibei Huang, Jiaojiao Zhao, Junting Guo, Kai Wang, Limin Han, Meijuan An, Ning Wang, Rongjia Du, Shiguo Lian, Shu Yang, Wenjing Zhang, Xuejiao Lei, Zhaoxiang Liu, Zhenhong Long","submitted_at":"2025-03-18T08:38:10Z","abstract_excerpt":"DeepSeek-R1, renowned for its exceptional reasoning capabilities and open-source strategy, is significantly influencing the global artificial intelligence landscape. However, it exhibits notable safety shortcomings. Recent research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 achieves a 100\\% attack success rate when processing harmful prompts. Furthermore, multiple security firms and research institutions have identified critical security vulnerabilities within the model. Although China Unicom has unco"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.16529","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/2503.16529/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":"2503.16529","created_at":"2026-07-05T11:03:58.456554+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.16529v2","created_at":"2026-07-05T11:03:58.456554+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.16529","created_at":"2026-07-05T11:03:58.456554+00:00"},{"alias_kind":"pith_short_12","alias_value":"VWGNNMXIHD67","created_at":"2026-07-05T11:03:58.456554+00:00"},{"alias_kind":"pith_short_16","alias_value":"VWGNNMXIHD67KFTS","created_at":"2026-07-05T11:03:58.456554+00:00"},{"alias_kind":"pith_short_8","alias_value":"VWGNNMXI","created_at":"2026-07-05T11:03:58.456554+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.12287","citing_title":"The Tower of Babel Revisited: Multilingual Jailbreak Prompts on Closed-Source Large Language Models","ref_index":61,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO","json":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO.json","graph_json":"https://pith.science/api/pith-number/VWGNNMXIHD67KFTST2ILLRSHZO/graph.json","events_json":"https://pith.science/api/pith-number/VWGNNMXIHD67KFTST2ILLRSHZO/events.json","paper":"https://pith.science/paper/VWGNNMXI"},"agent_actions":{"view_html":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO","download_json":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO.json","view_paper":"https://pith.science/paper/VWGNNMXI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.16529&json=true","fetch_graph":"https://pith.science/api/pith-number/VWGNNMXIHD67KFTST2ILLRSHZO/graph.json","fetch_events":"https://pith.science/api/pith-number/VWGNNMXIHD67KFTST2ILLRSHZO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO/action/storage_attestation","attest_author":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO/action/author_attestation","sign_citation":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO/action/citation_signature","submit_replication":"https://pith.science/pith/VWGNNMXIHD67KFTST2ILLRSHZO/action/replication_record"}},"created_at":"2026-07-05T11:03:58.456554+00:00","updated_at":"2026-07-05T11:03:58.456554+00:00"}