{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NGSAUVBNS5J6RYDSY4WWJ7TPUM","short_pith_number":"pith:NGSAUVBN","schema_version":"1.0","canonical_sha256":"69a40a542d9753e8e072c72d64fe6fa336bbd648d601a9c1b0922643fd1a3ca1","source":{"kind":"arxiv","id":"2406.15481","version":3},"attestation_state":"computed","paper":{"title":"Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Haneul Yoo, Hwaran Lee, Yongjin Yang","submitted_at":"2024-06-17T06:08:18Z","abstract_excerpt":"As large language models (LLMs) have advanced rapidly, concerns regarding their safety have become prominent. In this paper, we discover that code-switching in red-teaming queries can effectively elicit undesirable behaviors of LLMs, which are common practices in natural language. We introduce a simple yet effective framework, CSRT, to synthesize codeswitching red-teaming queries and investigate the safety and multilingual understanding of LLMs comprehensively. Through extensive experiments with ten state-of-the-art LLMs and code-switching queries combining up to 10 languages, we demonstrate t"},"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":"2406.15481","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-06-17T06:08:18Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"6b995368471372e7dd367efa7b62085d1971d78481f5a710e1978b17933c22a0","abstract_canon_sha256":"d12864a1c3505a31e56b15c61bf13ac8479968f7da0d30dc417cd42ce57954e0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:19.941165Z","signature_b64":"AbNv0x9a6JEHhk5HoOdqYQeb3FwHOki3eMUvgzCkCo9J3UjkmuuZmo2nx2s7ZjnBMV2Q0q0DZMXTSbdXQlXNCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69a40a542d9753e8e072c72d64fe6fa336bbd648d601a9c1b0922643fd1a3ca1","last_reissued_at":"2026-07-05T11:19:19.940652Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:19.940652Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Haneul Yoo, Hwaran Lee, Yongjin Yang","submitted_at":"2024-06-17T06:08:18Z","abstract_excerpt":"As large language models (LLMs) have advanced rapidly, concerns regarding their safety have become prominent. In this paper, we discover that code-switching in red-teaming queries can effectively elicit undesirable behaviors of LLMs, which are common practices in natural language. We introduce a simple yet effective framework, CSRT, to synthesize codeswitching red-teaming queries and investigate the safety and multilingual understanding of LLMs comprehensively. Through extensive experiments with ten state-of-the-art LLMs and code-switching queries combining up to 10 languages, we demonstrate t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.15481","kind":"arxiv","version":3},"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/2406.15481/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":"2406.15481","created_at":"2026-07-05T11:19:19.940713+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.15481v3","created_at":"2026-07-05T11:19:19.940713+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.15481","created_at":"2026-07-05T11:19:19.940713+00:00"},{"alias_kind":"pith_short_12","alias_value":"NGSAUVBNS5J6","created_at":"2026-07-05T11:19:19.940713+00:00"},{"alias_kind":"pith_short_16","alias_value":"NGSAUVBNS5J6RYDS","created_at":"2026-07-05T11:19:19.940713+00:00"},{"alias_kind":"pith_short_8","alias_value":"NGSAUVBN","created_at":"2026-07-05T11:19:19.940713+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07535","citing_title":"Multilingual Refusal Alignment for Safer Large Language Models","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2505.14226","citing_title":"Phonetic Perturbations Reveal Tokenizer-Rooted Safety Gaps in LLMs","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11157","citing_title":"Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM","json":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM.json","graph_json":"https://pith.science/api/pith-number/NGSAUVBNS5J6RYDSY4WWJ7TPUM/graph.json","events_json":"https://pith.science/api/pith-number/NGSAUVBNS5J6RYDSY4WWJ7TPUM/events.json","paper":"https://pith.science/paper/NGSAUVBN"},"agent_actions":{"view_html":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM","download_json":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM.json","view_paper":"https://pith.science/paper/NGSAUVBN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.15481&json=true","fetch_graph":"https://pith.science/api/pith-number/NGSAUVBNS5J6RYDSY4WWJ7TPUM/graph.json","fetch_events":"https://pith.science/api/pith-number/NGSAUVBNS5J6RYDSY4WWJ7TPUM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM/action/storage_attestation","attest_author":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM/action/author_attestation","sign_citation":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM/action/citation_signature","submit_replication":"https://pith.science/pith/NGSAUVBNS5J6RYDSY4WWJ7TPUM/action/replication_record"}},"created_at":"2026-07-05T11:19:19.940713+00:00","updated_at":"2026-07-05T11:19:19.940713+00:00"}