Nous Research 微调 Qwen3.8-27B，宣传倾向占比从 22% 降至 1% 以下
State propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away
Nous Research 把阿里 Qwen3.8-27B 改造成 Arete-27B，5300 万 token 微调就让宣传内容占比从 22% 掉到不到 1%，能力几乎没掉，方法本身也能借鉴。
Nous Research 用阿里巴巴的 Qwen3.8-27B 重训练出 Arete-27B。团队仅用 5300 万 token 做监督微调，模型亲华倾向的回答比例从 22% 降到不足 1%。通用能力在测试中保持不变。团队认为这类政治偏向是训练后期才加入的浅层调整，微调即可移除。
State propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away
Nous Research retrained Alibaba's Qwen3.8-27B to produce Arete-27B, largely stripping out its pro-Chinese bias. The propaganda rate dropped from 22 percent to less than one percent, while general capabilities remained intact in tests. The method relied on 53 million tokens of supervised fine-tuning. The political bias is "applied late, it is applied thin," the team writes. The article State propaganda in some Chinese AI models is only "a thin layer" that fine-tuning can strip away appeared first on The Decoder .
