Hirundo releases Westernized Qwen model with reduced CCP alignment
🤖 AI-generated content — The title and summary were produced automatically by artificial intelligence, without human editorial review.
AI safety lab Hirundo released Westernized versions of Alibaba's Qwen open-weight models, reducing Chinese Communist Party aligned censorship and political bias from 89.8 percent to 2.8 percent across a 500-prompts benchmark. The original Qwen3.6-35B-A3B model exhibited biased framing across 15 sensitive topics, including Tiananmen, Xinjiang, Taiwan, Hong Kong, Tibet, and COVID origins. Independent testing by CBS News confirmed the findings, while Hirundo reported that the model's reasoning, coding, and instruction-following performance remained essentially unchanged.
Continue in the app — vote & join in ➔Source: VentureBeat · via ahirlevel.hu