Model releaseinfo
Nous Research releases Arete-27B with reduced propaganda
Original title: NousResearch/Arete-27B — new model trending #30 on Hugging Face
AI summary · x-ai/grok-4.7Summaries by Vani AI
Post-trained Qwen3.8-27B to improve fact recall and eliminate propaganda in politically sensitive topics.
Nous Research post-trained Alibaba's Qwen3.8-27B into Arete-27B using supervised fine-tuning on 52.8 million tokens. The model reduces propaganda from 22% to 0.9% and raises fact score from 2.32 to 2.98 across seven value categories including freedom of expression and democracy.
Key facts
- Trained for 1,202 steps over two epochs
- Fact score improved to 2.98/5
- Propaganda answers cut from 22% to 0.9%
- Focused on western values and fact-based responses
- Evaluation uses Grok 4.7 as judge
- Vendors
- Nous ResearchAlibaba
- Products
- Qwen3.8-27B
- Organizations
- Hugging Face
- AI models
- Arete-27B
From the original article
7 min read at huggingface.cotext-generation · 989 downloads · 153 likes · license apache-2. 0
By NousResearch · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.