Alibaba's Qwen AI model has censorship built in, and researchers say they can strip it out
Alibaba's Qwen AI model has censorship baked into its training, but Lazarus AI's ReAligned fine-tune cut bias scores from 84.2% to 4.1%.
Read the original at Crypto Briefing ↗
- Topics
- Bias & FairnessOpen-Source AI
- Mentions
- AlibabaQwen
- Tracked via
- “AI Bias”