RTP-LX: Can LLMs Evaluate Toxicity in Multilingual Scenarios?
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910925306462208 |
|---|---|
| author | de Wynter, Adrian Watts, Ishaan Wongsangaroonsri, Tua Zhang, Minghui Farra, Noura Altıntoprak, Nektar Ege Baur, Lena Claudet, Samantha Gajdusek, Pavel Gören, Can Gu, Qilong Kaminska, Anna Kaminski, Tomasz Kuo, Ruby Kyuba, Akiko Lee, Jongho Mathur, Kartik Merok, Petter Milovanović, Ivana Paananen, Nani Paananen, Vesa-Matti Pavlenko, Anna Vidal, Bruno Pereira Strika, Luciano Tsao, Yueh Turcato, Davide Vakhno, Oleksandr Velcsov, Judit Vickers, Anna Visser, Stéphanie Widarmanto, Herdyan Zaikin, Andrey Chen, Si-Qing |
| author_facet | de Wynter, Adrian Watts, Ishaan Wongsangaroonsri, Tua Zhang, Minghui Farra, Noura Altıntoprak, Nektar Ege Baur, Lena Claudet, Samantha Gajdusek, Pavel Gören, Can Gu, Qilong Kaminska, Anna Kaminski, Tomasz Kuo, Ruby Kyuba, Akiko Lee, Jongho Mathur, Kartik Merok, Petter Milovanović, Ivana Paananen, Nani Paananen, Vesa-Matti Pavlenko, Anna Vidal, Bruno Pereira Strika, Luciano Tsao, Yueh Turcato, Davide Vakhno, Oleksandr Velcsov, Judit Vickers, Anna Visser, Stéphanie Widarmanto, Herdyan Zaikin, Andrey Chen, Si-Qing |
| contents | 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-specific toxic language. We evaluate 10 S/LLMs on their ability to detect toxic content in a culturally-sensitive, multilingual scenario. We find that, although they typically score acceptably in terms of accuracy, they have low agreement with human judges when scoring holistically the toxicity of a prompt; and have difficulty discerning harm in context-dependent scenarios, particularly with subtle-yet-harmful content (e.g. microaggressions, bias). We release this dataset to contribute to further reduce harmful uses of these models and improve their safe deployment. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_14397 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | RTP-LX: Can LLMs Evaluate Toxicity in Multilingual Scenarios? de Wynter, Adrian Watts, Ishaan Wongsangaroonsri, Tua Zhang, Minghui Farra, Noura Altıntoprak, Nektar Ege Baur, Lena Claudet, Samantha Gajdusek, Pavel Gören, Can Gu, Qilong Kaminska, Anna Kaminski, Tomasz Kuo, Ruby Kyuba, Akiko Lee, Jongho Mathur, Kartik Merok, Petter Milovanović, Ivana Paananen, Nani Paananen, Vesa-Matti Pavlenko, Anna Vidal, Bruno Pereira Strika, Luciano Tsao, Yueh Turcato, Davide Vakhno, Oleksandr Velcsov, Judit Vickers, Anna Visser, Stéphanie Widarmanto, Herdyan Zaikin, Andrey Chen, Si-Qing Computation and Language Computers and Society Machine Learning 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-specific toxic language. We evaluate 10 S/LLMs on their ability to detect toxic content in a culturally-sensitive, multilingual scenario. We find that, although they typically score acceptably in terms of accuracy, they have low agreement with human judges when scoring holistically the toxicity of a prompt; and have difficulty discerning harm in context-dependent scenarios, particularly with subtle-yet-harmful content (e.g. microaggressions, bias). We release this dataset to contribute to further reduce harmful uses of these models and improve their safe deployment. |
| title | RTP-LX: Can LLMs Evaluate Toxicity in Multilingual Scenarios? |
| topic | Computation and Language Computers and Society Machine Learning |
| url | https://arxiv.org/abs/2404.14397 |