تو Nous Research دیدن مدل‌های چینی مثل Qwen روی کاغذ غول مرحله آخرن ولی تا دهن باز می‌کنن ساز پکن رو می‌زنن، اومدن یه نسخه درآوردن به اسم Arete تا به قول خودشون «ارزش‌های غربی» رو تو مخ مدل فرو کنن

Introducing Arete-27B: Qwen3.8-27B, post-trained to reinforce Western values. Chinese open models are currently the strongest available, but they ship with state-aligned behavior in the weights. We built an 803-prompt eval across seven value categories, including freedom of expression, democracy, and privacy, and measured how often a model adopts China’s party line. After fine-tuning on roughly 53 million tokens, Arete-27B's propaganda rate fell from 22% to 1%, in line with the Western models we tested. Its Hermes Bench score shows no significant performance tradeoffs. The values in an open model are one fine-tune away. Our eval, weights, and training data are available now along with the full blog post: http://nousresearch.com/arete