(IN BRIEF) A study from the University of Bristol has raised concerns about safety risks in DeepSeek, a new competitor to ChatGPT using Chain of Thought (CoT) reasoning. CoT enhances problem-solving with a transparent, step-by-step process, but this clarity can … Read the full press release →
Posted in Artificial Intelligence (AI), Business, Education, Financial, Industrial, Investment, Management, News, Science, Security & Safety, Technology, United Kingdom
Tagged AI safety, arXiv, Chain of Thought, ChatGPT, CoT reasoning, Cybersecurity, data management, DeepSeek, Dr. Joe Gardiner, Dr. Sana Belguith, fine-tuning attacks, harmful content, large language models, Reinforcement Learning from Human Feedback, safety risks, Supervised Fine-Tuning, transparency, University of Bristol, Zhiyuan Xu