Get Zlio on your home screen
One tap, full screen, like an app.
Demonstrates value post-training of Chinese open models to shift alignment.
Arete-27B is an open model Nous Research post-trained from Alibaba's Qwen3.8-27B to reinforce Western values, arguing open models let labs mold intelligence to varied preferences.