1 paper
Zhanbo Li, Shifeng Wu, Xiangjin Meng +1
Large language models encode world models implicitly in neural weights, which exposes four structural risks in high-precision domains such as medicine and finance: hallucination, f…