Huawei Launches DeepSeek-R1-Safe to Block Politically Sensitive AI Content: How It Works

Huawei launches DeepSeek-R1-Safe to block politically sensitive AI content: How it works

Huawei has launched DeepSeek-R1-Safe, a safety-focused AI model designed to block politically sensitive content. Developed with Zhejiang University and trained on 1,000 Ascend chips, it achieves 83% security defence, outperforming rivals, while maintaining near-original performance.

Huawei launches Deepseek-R1-Safe to block politically sensitive AI content: how it works

Huawei has launched a safety-oriented version of the Deepseek-artificial intelligence model and claims that it is almost entirely effective in preventing politically sensitive discussions. Huawei unveils on safety -oriented Deepseek AI The model, dubbed Deepseek-R1-Safe, was developed in collaboration with Zhejiang University, a prominent institution and the Alma Mater by the founder of Deepseek, Liang […]

Solving AI Safety and Performance: Huawei ZJU Launches DeepSeek-R1-Safe Model

Huawei and Zhejiang University have jointly announced the launch of their latest large-scale AI model, DeepSeek-R1-Safe, marking a significant step forward in addressing the critical challenges of AI safety and performance. This innovative development aims to strike a better balance between robust security measures and high computational efficiency, a long-standing issue in the AI research community. DeepSeek-R1-Safe is designed to enhance the r…