Get Zlio on your home screenOne tap, full screen, like an app.
Open Models · first seen 06 November 2026, 19:58 UTC
Nous Research releases Arete-27B, a Qwen post-train
Shows fine-tuning can strip embedded state bias from open Chinese models.
Arete-27B is an open-weights model built by post-training Alibaba's Qwen3.8-27B to reinforce Western values, Nous Research says. It reports cutting pro-Beijing answers from 22% to under 1% with fine-tuning alone.
