Can We Locate and Prevent Stereotypes in LLMs?
Do Hallucination Neurons Generalize? Evidence from Cross-Domain Transfer in LLMs