AI RESEARCH
Do Hallucination Neurons Generalize? Evidence from Cross-Domain Transfer in LLMs
arXiv CS.CL
•
ArXi:2604.19765v1 Announce Type: new Recent work identifies a sparse set of "hallucination neurons" (H-neurons), less than 0.1% of feed-forward network neurons, that reliably predict when large language models will hallucinate. These neurons are identified on general-knowledge question answering and shown to generalize to new evaluation instances.