{"paper":{"title":"RTP-LX: Can LLMs Evaluate Toxicity in Multilingual Scenarios?","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CY","cs.LG"],"primary_cat":"cs.CL","authors_text":"Adrian de Wynter, Akiko Kyuba, Andrey Zaikin, Anna Kaminska, Anna Pavlenko, Anna Vickers, Bruno Pereira Vidal, Can G\\\"oren, Davide Turcato, Herdyan Widarmanto, Ishaan Watts, Ivana Milovanovi\\'c, Jongho Lee, Judit Velcsov, Kartik Mathur, Lena Baur, Luciano Strika, Minghui Zhang, Nani Paananen, Nektar Ege Alt{\\i}ntoprak, Noura Farra, Oleksandr Vakhno, Pavel Gajdusek, Petter Merok, Qilong Gu, Ruby Kuo, Samantha Claudet, Si-Qing Chen, St\\'ephanie Visser, Tomasz Kaminski, Tua Wongsangaroonsri, Vesa-Matti Paananen, Yueh Tsao","submitted_at":"2024-04-22T17:56:26Z","abstract_excerpt":"Large language models (LLMs) and small language models (SLMs) are being adopted at remarkable speed, although their safety still remains a serious concern. With the advent of multilingual S/LLMs, the question now becomes a matter of scale: can we expand multilingual safety evaluations of these models with the same velocity at which they are deployed? To this end, we introduce RTP-LX, a human-transcreated and human-annotated corpus of toxic prompts and outputs in 28 languages. RTP-LX follows participatory design practices, and a portion of the corpus is especially designed to detect culturally-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.14397","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/2404.14397/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"}