כתבה
arXiv cs.CL ·
WinoQueer-NL: בדיקת גזענות בדגמי שפה הולנדית כלפי זהויות LGBTQ+
WinoQueer-NL: Assessing Bias in Dutch Language Models toward LGBTQ+ Identities
בדיקת גזענות בדגמי שפה הולנדית כלפי זהויות LGBTQ+, כולל דגמי Gemini ו-GPT-5. המחקר חשף גזענות חריפה כלפי זהויות טרנסג'נדריות ולא-בינאריות.
תקציר מקורי באנגליתarXiv:2609.02651v2 Announce Type: replace Abstract: While English language models have been widely examined for anti-queer bias, Dutch models remain understudied. To address this gap, we developed a culturally and linguistically adapted Dutch dataset based on the English WinoQueer benchmark, containing pairs of stereotypical and counter-stereotypical sentences. To validate and expand it, we conducted an online survey with 43 Dutch queer participants, confirming 145 of 171 stereotypes as culturally relevant and identifying 22 new biases through free-text responses. The final released dataset, comprising 42,906 sentences, was evaluated using a range of Dutch-specific and multilingual models, including both masked language models (MLMs) and autoregressive language models (ARLMs), with bias me
קרא במקור המקורי
arxiv.org
פתח כתבה מקורית