Nous Research has open-sourced Arete-27B, a Qwen3.8-27B fine-tune trained with plain supervised fine-tuning on 52.8M tokens using Inkling as teacher, with no RL or architecture changes. On its self-built 803-prompt eval, propaganda-style answers dropped from 22% to 0.9%, while factuality mean rose only from 2.32 to 2.98 out of 5 and still trails Muse Glimmer 30B; 216 of 803 answers still score 0, and both eval and training data are self-constructed and self-judged.
Termsslug: preference-optimization
Preference Optimization
偏好优化
Training method aligning model outputs with user preferences using human feedback
Curated Intelligence
Marktechpost (@Marktechpost)T2·Open Source
72 pts
Hugging Face BlogT1·Models
25 pts
The input provides only the title regarding preference optimization for Vision Language Models (VLMs), lacking specific methods, experimental data, or conclusions. Without body text details, technical breakthroughs or practical impacts cannot be assessed.